Mastering Public Health

A postgraduate guide to examinations and revalidation

Public [Link] 1 3/6/08 [Link]

Public [Link] 2 3/6/08 [Link]
 Mastering Public Health
A postgraduate guide to examinations and revalidation

Geraint H Lewis MA MB BChir MSc MRCP MFPH

Specialist Registrar in Public Health, London Deanery
Jessica Sheringham MA MSc MFPH
Specialist Trainee in Public Health, London Deanery
Kanwal Kalim MB ChB MSc
Specialist Registrar in Public Health, London Deanery
Tim JB Crayford MB BS MSc FFPH
Director of Public Health, Croydon PCT;
President, UK Association of Directors of Public Health

Public [Link] 3 3/6/08 [Link]

 © 2008 Royal Society of Medicine Press Ltd

Published by the Royal Society of Medicine Press Ltd

1 Wimpole Street, London W1G 0AE, UK
Tel: +44 (0)20 7290 2921
Fax: +44 (0)20 7290 2929
Email: publishing@[Link]
Website: [Link]

Apart from any fair dealing for the purposes of research or private study, criticism or review, as permitted under the UK Copyright,
Designs and Patents Act, 1988, no part of this publication may be reproduced, stored or transmitted, in any form or by any means,
without the prior permission in writing of the publishers or in the case of reprographic reproduction in accordance with the
terms of licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of licences issued by the
appropriate Reproduction Rights Organization outside the UK. Enquiries concerning reproduction outside the terms stated here
should be sent to the publishers at the UK address printed on this page.

The rights of Geraint H Lewis, Jessica Sheringham, Kanwal Kalim and Tim JB Crayford to be identified as authors of this work have
been asserted by them in accordance with the Copyright, Designs and Patents Act, 1988.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN-13 978-1-85315-781-3

Distribution in Europe and Rest of World:

Marston Book Services Ltd

PO Box 269
Abingdon
Oxon OX14 4YN, UK
Tel: +44 (0)1235 465500
Fax: +44 (0)1235 465555
Email: [Link]@[Link]

Distribution in the USA and Canada:

Royal Society of Medicine Press Ltd

c/o BookMasters Inc
30 Amberwood Parkway
Ashland, OH 44805, USA
Tel: +1 800 247 6553/+1 800 266 5564
Fax: +1 419 281 6883
Email: order@[Link]

Distribution in Australia and New Zealand:

Elsevier Australia
30–52 Smidmore Street
Marrickville NSW 2204, Australia
Tel: +61 2 9517 8999
Fax: +61 2 9517 2249
Email: service@[Link]

Typeset by Phoenix Photosetting, Chatham, Kent

A Eurasia production in China

Public [Link] 4 3/6/08 [Link]

 CONTENTS

Foreword vii
Preface viii
Acknowledgements ix
Abbreviations xiii

Section 1 Research Methods 1

1A Epidemiology 3
1B Statistics 69
1C Assessment and Evaluation 107
1D Principles of Qualitative Methods 139

Section 2 Disease Causation and Prevention; Health Promotion 145

2A Epidemiological Paradigms 147
2B Epidemiology of Specific Diseases 149
2C Diagnosis and Screening 163
2D Genetics 183
2E Health and Social Behaviour 199
2F Environment 211
2G Communicable Diseases 235
2H Principles and Practice of Health Promotion 273
2I Disease Prevention and Models of Behaviour Change 309

Section 3 Health Information 323

3A Populations 325
3B Sickness and Health 343
3C Applications of Health Services Information 359

Public [Link] 5 3/6/08 [Link]

 C o nt e nt s

Section 4 Medical Sociology, Social Policy and Health Economics 373

4A Health and Illness 375
4B Health Care 389
4C Equality, Equity and Policy 399
4D Health Economics 421

Section 5 Organisation and Management of Health Care and Health-Care Programmes 449
5A Individuals, Teams and Groups 451
5B Understanding Organisations, Their Function and Structure 471
5C Management and Change 477
5D Understanding the Theory and Process of Strategy Development 485
5E Finance, Management Accounting and Relevant Theoretical Approaches 501

Section 6 Skills Tested at Part A 509

6A Research Design and Critical Appraisal 511
6B Drawing Conclusions from Data 515
6C Written Presentation Skills 523
6D Formulae Required to Pass Part A 533

Appendix A: Revision Tips 541

Appendix B: Answer Frameworks 549
Appendix C: Top Fives 559
Further Reading 565
References 581
Index 593

vi

Public [Link] 6 3/6/08 [Link]

 Foreword

FOREWORD

‘Health for All’ is much more than just a slogan. It is a reminder of the enormous scope of public health. Its practitioners
must be able to get to grips with the explosion of knowledge on the bewildering array of health determinants, drawing
on insights from, among others, genetics, biology, economics and political science. As if this were not enough, they
must be experts in the critical appraisal of evidence, so that they can judge what is likely to work in a given set of
circumstances and what not, and must then deploy the managerial and organisational skills to engage with a wide
range of actors so as to turn their vision into reality. The scale of the canvas on which they must work is so great, and
the environment in which they work so dynamic, that the acquisition of the skills and knowledge that they require
can be achieved only through a process of life-long learning. It is not possible for anyone to be an expert in every
aspect of public health; some degree of specialisation is required but this must be built on a common foundation of
knowledge and skills that will equip the next generation of public health specialists to set out on this journey.
In the United Kingdom this foundation has been defined by the curriculum for the examinations of the Faculty of
Public Health. Recently revised, it comes in two parts. The first tests candidates’ understanding of the scientific
principles of public health; the second, the Objective Structured Public Health Examination, tests their ability to apply
this understanding.
Whatever one’s disciplinary background, there is an enormous amount of new information for the candidate to absorb.
Public health requires an understanding of quantitative and qualitative methods, and of natural and social sciences,
and high levels of numeracy and literacy. There are, of course, many books that those training in public health can
look to for guidance. However, it has long been apparent that there is a need for a concise single volume in which the
candidate can find the key elements of knowledge that he or she will need to pass the exams. The authors of this book
are to be congratulated for meeting this need. They have assembled an excellent overview of the essential knowledge
required by tomorrow’s public health practitioners in a clear and highly readable manner. I am confident that this will
rapidly become required reading for all those taking the Faculty’s exams, as well as for those undertaking training in
public health in many other countries. Of course knowledge is not enough: the real challenge is to develop the skills
to apply it. But that only comes with practice.

Martin McKee CBE MD DSc MSc FRCP FRCPI FFPH FMedSci

Professor of European Public Health
London School of Hygiene and Tropical Medicine

vii

Public [Link] 7 3/6/08 [Link]

 P re fa c e

PREFACE
This is the type of book that we wish had existed when we were revising for the membership examination of the Faculty
of Public Health.
It began life as a mass of revision notes based on dozens of books and lectures. Over the ensuing 18 months it has
developed, thanks to the contributions of a wide range of people including our colleagues and our international
editors, into this text. It is a revision guide that we hope will be of use to candidates of postgraduate public health
examinations across many countries. However, it will also be of use for medical students and for people who are about
to embark on a course of study such as an MPH or MSc in public health, for continual professional development and as
a quick, practical reference text for practitioners of the specialty.
For better or worse, we decided that the book should follow strictly the structure of the UK Faculty’s Part A examination
syllabus (which is also used by the Australian Faculty of Public Health). This consists of five areas of public health
knowledge plus a sixth section relating to public health skills. In order to avoid excessive duplication, we have at
times cross-referenced the reader to other parts of the book where the syllabus covers the same material. Occasionally
(as in Section 6 − public health skills), the overlapping content has been kept in order to reflect different emphases
drawn out by different parts of the syllabus.
Our aim throughout the book has been to strike a balance somewhere between brief lecture notes and a full-prose
textbook. We hope that this compromise will be suited both to revision and to quick reference needs. Being a
postgraduate revision book, we have assumed a certain degree of prior knowledge, and have prioritised breadth of
content over depth. The appendices offer the candidate of the Part A examination some revision strategies, essay
frameworks and lists of key public health names and concepts that are specifically designed to support exam preparation.
Text given in italics under headings indicate the exact syllabus item.
We acknowledge the assistance and guidance of our international editors, who generously offered their time and
expertise to provide us with international perspectives and comments on our draft chapters. We have included the
examples that they provided (indicated by national symbols) but recognise that the book remains Anglo-centric. For
this reason we would be delighted to receive further international details from readers for any future edition of the
book. Particular thanks go to Drs David Pencheon and Paul Crook for their thoughtful comments.
We and the publishers have attempted to seek permission from, and to acknowledge, all those whose work we used
in compiling this revision guide. If, however, if we have unwittingly omitted any acknowledgements then do please
contact the RSM Press.
GH Lewis
J Sheringham
K Kalim
TJB Crayford

viii

Public [Link] 8 3/6/08 [Link]

 Acknowledgements

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the contributions of the international editorial panel:
AUSTRALIA: Professor Kerry Kirke, Clinical Associate Professor, University of Adelaide, The Australasian Faculty of
Public Health Medicine, Royal Australasian College of Physicians
ENGLAND: Professor David Strachan, Chair, MFPH Part A examiners, Professor of Epidemiology, Division of
Community Health Sciences, St George’s Hospital, University of London
HONG KONG: Professor Sian Griffiths, Professor of Public Health, Director, School of Public Health, Faculty of
Medicine, Chairman, Department of Community and Family Medicine, The Chinese University of Hong Kong
HONG KONG: Professor Emily YY Chan, Assistant Professor, School of Public Health, Faculty of Medicine, The
Chinese University of Hong Kong
IRELAND: Dr Emer Shelley, Population Health Directorate Health Services Executive, Stewarts Sports Centre, Dublin
NEW ZEALAND: Dr Michael Baker, Senior Lecturer, Department of Public Health, Wellington School of Medicine and
Health Sciences
NORTHERN IRELAND: Dr John WG Yarnell, Reader in Cardiovascular Epidemiology, Epidemiology and Public Health,
Queen’s University Belfast
SCOTLAND: Dr Anne Maree Wallace, Scottish Director of Training in Public Health
SOUTH AFRICA: Dr Stephen Knight, Lecturer/Public Health Medicine Specialist, School of Family and Public Health
Medicine, Nelson R Mandela School of Medicine, University of KwaZulu-Natal
WALES: Dr Nigel Monaghan, Consultant in Public Health NPHS Wales/Public Health Director, Rhondda Cynon Taff
Local Health Board, National Public Health Service, Cardiff

ix

Public [Link] 9 3/6/08 [Link]

 Ac k no w l e dge ments

In the preparations for their Part A exams and in writing this book, the authors are indebted to the teaching
provided by:
London School of Hygiene and Tropical Medicine, Masters in Public Health, 2004−06, in particular:
Judy Green, Kara Hansen, Helen Hogan, Valerie Iles, Stephen Jan, Karen Lock, Martin McKee, Barbara McPake, Ellen
Nolte, Justin Parkhurst, Carine Ronsmans, Colin Sanderson and Rosalind Raine
Edmund Jessop, Part A Examination Course and Manual, Winchester, December 2005
North Central London Strategic Health Authority, Part A Revision Course, September−December 2005, in
particular:
Marion Deacon, Meic Goodyear, Andy Hall and Oliver Morgan
Gwyn Bevan, Public Policy Course, Commission for Health Improvement, 2003−04
We would also like to thank the following for their valuable advice and detailed comments on sections of the book:
Jamila Aboulhab, Abdul Azad, Glen Brice, Araceli Busby, Helen Crabbe, Daragh Fahey, Jennie Mussard, Geoff Ridgway,
Abdul Roudsari, Murad Ruf, South West London Health Intelligence Team, Louisa Shillito and Chris Weston

ACKNOWLEDGEMENTS BY CHAPTER
1A
Ellen Nolte and Martin McKee, London School of Hygiene and Tropical Medicine, Issues in Public Health Course,
2004–05
Carine Ronsmans and colleagues, London School of Hygiene and Tropical Medicine, Basic Epidemiology Course,
2004–05
Colin Sanderson, London School of Hygiene and Tropical Medicine, Healthcare Evaluation Course, 2004–05
Justin Parkhurst, London School of Hygiene and Tropical Medicine, Health Power, Process and Policy Course, 2004–05
Edmund Jessop, Part A Examination Course, Winchester, December 2005
Andy Hall, North Central London Part A Examination Course, September 2005

1B
Jamila Aboulhab, Specialist Registrar in Public Health
Murad Ruf, Specialist Registrar in Public Health
Edmund Jessop, Part A Examination Course and lecture notes, Winchester, December 2005

1C
Karen Lock, London School of Hygiene and Tropical Medicine, Issues in Public Health Unit, 2004–05
Colin Sanderson, London School of Hygiene and Tropical Medicine, Public Health: Integrating Unit, 2004–05
London School of Hygiene and Tropical Medicine, Economic Evaluation Unit, 2004–05
London School of Hygiene and Tropical Medicine, Introduction to Health Economics Unit, 2004–05
Ellen Nolte, London School of Hygiene and Tropical Medicine, Health Services Linear Unit, 2004–05
Andy Hall, North Central London Part A Examination Course, September 2005
Jamila Aboulhab, Specialist Registrar in Public Health, London Training Scheme

Public [Link] 10 3/6/08 [Link]

 Acknowledgements

1D
London School of Hygiene and Tropical Medicine, Introduction to Social Research Unit, 2004–05

2A
Ellen Nolte, London School of Hygiene and Tropical Medicine, Issues in Public Health Course, Lecture 2, 2004–05

2G
Geoff Ridgway, Microbiologist, Senior Medical Advisor, Health Protection Division, DH
Bharat Patel, Health Protection Agency, North Central London Strategic Health Authority Part 1 revision slides
Andy Hall, London School of Hygiene and Tropical Medicine, North Central London Strategic Health Authority Part 1
revision

2H
Abdul Azad, Health Improvement and Sure Start Co-ordinator, Redbridge PCT
Carine Ronsmans and colleagues, London School of Hygiene and Tropical Medicine, Basic Epidemiology Course
2004–05

2I
London School of Hygiene and Tropical Medicine Integrating Unit in Public Health notes, 2005

3B
Meic Goodyear, North Central London course notes, December 2005

3C
Jennie Mussard, Health Intelligence Lead, Croydon PCT
South West London Health Intelligence Team
Abdul Roudsari, Head of Centre for Health Informatics, School of Informatics, City University

4A
Judy Green, London School of Hygiene and Tropical Medicine lecture notes, 2004–05

4C
Rosalind Raine, Healthcare Evaluation lecture notes, London School of Hygiene and Tropical Medicine, 2005
Stephen Jan, Economic Analysis for Management and Policy: Equity and Priority Setting lecture notes, London School
of Hygiene and Tropical Medicine, 2005
Gwyn Bevan, Public Policy: Priority Setting in Oregon lecture notes, January 2004
Oliver Morgan, North Central London Part A Revision Course, 2005

4D
Daragh Fahey, North Central London Part 1 economics revision notes

xi

Public [Link] 11 3/6/08 [Link]

 Ac k no w l e dge ments

5A
Valerie Iles, London School of Hygiene and Tropical Medicine, Organisational Management, 2005

5B
Oliver Morgan, North Central London Part A Revision Course, Organisational Management lecture notes, 2005

5C
Oliver Morgan, North Central London Part A Revision Course, Organisational Management presentation, 2005

6A
Edmund Jessop, Part A Examination Course, Winchester, December 2005

6C
North Central London Part A Revision Course

xii

Public [Link] 12 3/6/08 [Link]

 Abbreviations

ABBREVIATIONS

ADL activities of daily living

AFP a-fetoprotein
ANOVA analysis of variance
BMI body mass index
COPD chronic obstructive pulmonary disease
COSHH Control of Substances Hazardous to Health
DALY disability-adjusted life-year
DDD defined daily dose
DEFRA UK Department for Environment, Farming and Rural Affairs
DFID UK Department for International Development
DH Department of Health
DNA deoxyribonucleic acid, did not attend
DRG diagnosis-related group
DVT deep vein thrombosis
DWI Drinking Water Inspectorate (England & Wales)
EBM evidence-based medicine
EIA environmental impact assessment
FEV1 forced expiratory volume over 1 second
FRR familial relative risk
GDP gross domestic product
GHS General Household Survey
GHQ general health questionnaire
GMC General Medical Council
GUM genitourinary medicine
HALE health-adjusted life expectancy
HES hospital episode statistics
HIA health impact assessment
HNA health needs assessment
HPA UK Health Protection Agency
HPV human papillomavirus
HRG healthcare resource group
HRQoL health-related quality of life
ICD International Classification of Diseases

xiii

Public [Link] 13 3/6/08 [Link]

 A b b re v ia t io ns

ICP integrated care pathway

LAA local area agreement
LR likelihood ratio
LSP local strategic partnership
MANOVA multivariate analysis of variance
MHRA Medicines and Healthcare products Regulatory Authority
MRSA meticillin-resistant Staphylococcus aureus
NCEPOD National Confidential Enquiry into Patient Outcome and Death
NICE National Institute for Health and Clinical Excellence
NMC Nursing and Midwifery Council
NNT number needed to treat
NSC UK National Screening Committee
NSSEC National Statistics Socio-economic Classification
ONS Office for National Statistics
PACT prescribing analysis and cost
PALS Patient Advice and Liaison Service
PCR polymerase chain reaction
PCT primary care trust
PDP personal development plan
PEM protein–energy malnutrition
PSA prostate-specific antigen
QALY quality-adjusted life-year
QMAS quality management and analysis system
QOF quality and outcomes framework
RCT randomised controlled trial
RDA recommended daily amount
ROC receiver operating characteristic
RRR relative risk ratio
SARS severe acute respiratory syndrome
SD standard deviation
SE standard error
SIDS sudden infant death syndrome
SMR standardised mortality ratio
WARNER weekly analysis report of notifications above expected rates
WHO World Health Organization
YLL years of life lost

xiv

Public [Link] 14 3/6/08 [Link]

 Section 1
RESEARCH METHODS

Public health practitioners need to understand how health knowledge is generated not only to be able to select and
use appropriate research methods for their own work, but also so that they can appraise the quality of published
research and provide credible professional advice.
Section 1 outlines the research methods that underpin public health. Epidemiology is the key to public health
practice: the discipline involves scrutinising data to generate meaningful inferences that can be used as the basis of
health policy. The validity of public health research relies on the appropriate use of statistics, from the design and
instigation of data collection, through to their analysis and interpretation.
Qualitative research methods are necessary for understanding the reasons why things happen the way that they do.
Section 1 ends with the field of health-care assessment that draws on all of these disciplines in order to evaluate
the structure, function and performance of health systems and services.

Public [Link] 1 3/6/08 [Link]

Public [Link] 2 3/6/08 [Link]
 1A Epidem iology

1A
Epidemiology

1A.1 Health statistics 4 1A.22 Numbers needed to treat 42

1A.2 Numerators, denominators and populations 1A.23 Time trend analysis, time series designs 43
at risk 7 1A.24 Nested case–control studies 44
1A.3 Time at risk 8 1A.25 Methods of sampling from a population 45
1A.4 Methods for summarising data 9 1A.26 Methods of allocation in intervention
1A.5 Incidence, prevalence and standardisation 12 studies 48
1A.6 Years of life lost 16 1A.27 Recording survey data 49
1A.7 Measures of disease burden 17 1A.28 Construction of valid questionnaires 50
1A.8 Variation 18 1A.29 Validating observational techniques 50
1A.9 Errors in epidemiological measurement 19 1A.30 Studies of disease prognosis 51
1A.10 Concepts and measures of risk 21 1A.31 Statistical analysis of epidemiological
1A.11 The odds ratio 25 studies 52
1A.12 Rate ratio and risk ratio (relative risk) 25 1A.32 Epidemic theory 52
1A.13 Association and causation 26 1A.33 Combining studies 56
1A.14 Bias 28 1A.34 Electronic bibliographical databases 59
1A.15 Confounding and interaction 31 1A.35 Grey literature 60
1A.16 Control of confounding 33 1A.36 Evidence-based medicine and policy 60
1A.17 Descriptive and ecological studies 34 1A.37 Hierarchy of research evidence 61
1A.18 Small area analysis 35 1A.38 Publication bias 62
1A.19 Study design 35 1A.39 The Cochrane Collaboration 62
1A.20 Intention-to-treat analysis 41 1A.40 Epidemiological research ethics 63
1A.21 Clustered data 41 1A.41 Genetics 67

Epidemiology is the study of the distribution and determinants of health and disease in populations. It is a
collection of techniques for studying the characteristics of many different people. Epidemiology is the science of the
‘who’, the ‘what’, the ‘where’ and the ‘when’, and it is the tool for exploring the underlying questions of ‘how’ and
‘why’ as they relate to health.

Public [Link] 3 3/6/08 [Link]

 1A E p ide m io l o gy

It is therefore fundamental to public health practice, for it uses data from large populations to generate meaningful
information that can be used as the basis of health policy.
The techniques covered in this chapter are needed for the interpretation of the scientific literature, and for
providing credible professional advice. Public health practitioners need to learn how to apply these techniques well
enough to design and analyse with confidence their own basic studies.

1A.1 HEALTH STATISTICS

Use of routine vital and health statistics to describe the distribution of disease in time, place and person
Information about a defined population that is collected in a consistent manner for administrative reasons is often
called routine data. These data may be used to describe the needs of and services provided to different population
groups. The sources of routine health data, the information collected and the frequency of collection vary between
different countries. However, almost all countries process their data into vital statistics. These concern the
important events in human life, such as births, deaths and migrations.
Two standard and important vital statistics used across the globe for assessing a population’s health are life-
expectancy and infant mortality.
• Life-expectancy is the expectation of life at birth. It is defined as the period after which half of all persons born
have died.
• Infant mortality is the number of children per 1000 live births who die in their first year of life.
Strengths and weaknesses of vital statistics are shown in Box 1A.1.1.

Box 1A.1.1
Strengths of vital statistics Weaknesses of vital statistics
• Cheap and readily available • Incomplete
• Almost complete data recording • Potential for biases (e.g. postmortem inflation of
• Contemporary socioeconomic status; diseases with stigma under-
• Can be used for ecological studies to develop reported)
hypotheses • Currency − can become out of date (e.g. census
• Recorded at regular intervals − can be used for data only recorded every 10 years)
following trends
Improving the reliability, validity and completeness of routine data is important to avoid waste and to maximise the
use of resources. There should therefore be a good reason to begin or to stop collecting each item of data.
The quality of data can be improved as shown in Box 1A.1.2.

Box 1A.1.2
Computerised Improves the accuracy and timeliness of the preparation and dissemination of information
data collation and
analysis
Feedback Improving feedback of collated data to providers is essential if their interest is to be
maintained and their attention to providing quality data sustained
Presentation Data should be presented in a variety of ways which are meaningful to policy makers, the
media, professionals and the lay public
Training Training the coders and those responsible for data entry in the use of standard definitions,
terminology, etc

Public [Link] 4 3/6/08 [Link]

 1A Epidem iology

ROUTINE STATISTICS IN THE UK

Eng The key sources of routine health data in England are listed in Table 1A.1.1. More details are available in
Section 3B.

Table 1A.1.1 Routine statistics collected in England

Type Examples
Demography Census
General Household Survey†
Mortality Mortality statistics (Office for National Statistics)
10-yearly supplement on occupational mortality
Morbidity Primary care GP disease registries, e.g. Quality and Outcomes Framework (QOF)
GP research database
General Household Survey
Hospital NHS Secondary Uses Service (SUS)
Laboratory results
A&E attendances
Registers Regional cancer registries
National childhood cancer register
Confidential Enquiry into Maternal and Child Health
Congenital abnormalities
Prostheses
Transplants
Confidential enquiries
Other Notifiable disease statistics
Health Survey for England*
*Collects information from about 13 000 people living in private households in Great Britain. Started in 1971 and has
been carried out approximately annually since then (ONS 2003).
†
Uses multistage stratified sampling with the postcode address files as the sampling frame. Comprises: questionnaire,
height, weight and blood pressure measurement, plus a blood sample.

MORTALITY INDICES
UK Data on mortality are highly accurate in the UK, owing to the legal processes to which registrations of deaths
are subject. Mortality indices are shown in Table 1A.1.2.

FERTILITY INDICES
A number of measures exist to describe different features of the reproductive behaviour of a population. These are
commonly known as fertility indices. See Table 1A.1.3.

Public [Link] 5 3/6/08 [Link]

 1A E p ide m io l o gy

Table 1A.1.2 Mortality indices

Index* Typical Numerator Denominator
reference
period
Crude mortality rate 1 year Number of deaths Mid-year population
Age-specific mortality rate 1 year Number of deaths aged X Mid-year population aged X
Child mortality rate 1 year Number of deaths in children aged Mid-year number of children
1−4 years aged 1−4 years
Infant mortality rate 1 year Number of deaths under 1 year old Number of live births
Postnatal mortality rate 1 year Number of deaths in infants aged Number of live births
4−52 weeks
Neonatal mortality rate 1 year Number of deaths in the first 28 Number of live births
days
Perinatal mortality rate 1 year Number of stillbirths + deaths <7 Number of live births +
days stillbirths
Standardised mortality ratio See Section 1A.5
*Commonly expressed per 1000 or per 100 000.

Table 1A.1.3 Fertility indices

Rate Typical Numerator Denominator Advantages/limitations
reference
period
Crude birth rate 1 year Number of live 1000 total population Poor indicator of fertility:
births denominator includes men,
children and postmenopausal
women
General fertility rate 1 year Number of live 1000 women aged 15−44 Denominator includes
births years only women, typically of
childbearing age
Age-specific fertility 1 year Number of live 1000 women within a More precise – takes into
rate births in 1 particular age band, e.g. account differences in
year 20−24 years fertility at different ages
Total period fertility 1 year Sum of the age-specific rates across an Enables comparisons
rate average woman’s reproductive lifetime. UK between countries over time
= 1.66 in 2005

DESCRIPTIVE EPIDEMIOLOGY
Epidemiology is the study of the patterns, causes and control of disease in groups of people. In descriptive
epidemiology, the three major dimensions used to describe the occurrence of disease are time, place and person.

TIME
Considers when the disease occurs and how it changes/has changed over time, described by:

Public [Link] 6 3/6/08 [Link]

 1A Epidem iology

• Epidemic curves: acute increases in disease frequency

• Seasonal variation: cyclical patterns in disease frequency, e.g. seasonal influenza
• Secular trends: trends over decades and centuries
• Point events: sudden emergence of disease at a particular time.

PLACE
Describes where the incidence is high/low, where incidence is changing/has changed, considering:
• International: ecological comparisons may suggest hypotheses regarding causation
• National: highlights urban–rural patterns and patterns related to deprivation
• Small area: compare census data/index of multiple deprivation/town-centre data.

PERSON
Describes who is affected, taking into account characteristics such as:
• Age
• Sex
• Occupation/social class
• Ethnicity
• Behaviour/lifestyle.

1A.2 NUMERATORS, DENOMINATORS AND POPULATIONS AT RISK

In epidemiology, a numerator is a feature that has been counted (e.g. number of deaths); it forms the upper part
of a fraction. A denominator, for epidemiological purposes, is usually the population from which the numerator was
drawn. The denominator is the lower part of a fraction and is often restricted to a particular time period.
A ‘population at risk’ is a subgroup of a wider population that has been subjected to the exposure of interest and
which therefore has a propensity to have been affected by that exposure. For example, the population at risk of
prostate cancer is all males who have not previously had a prostatectomy.
The numerator and denominator can be combined into a ratio, a proportion or a rate, as detailed below. Some of
the most frequent mistakes in applying epidemiological principles to real-world problems come from a failure to
define and to count numerators and denominators accurately.

RATIO
n1
Ratio = = n1:n2
n2
where n1 and n2 are numbers.
For example,
Number of males in a population
Ratio of males to females in a population =
Number of females in a population
A ratio is often expressed as odds, which is a single number. For example, if a bag contains 2 white balls (n1), 3
black balls (n2) and 5 grey balls (n3), then the ratio of black balls to white balls is 3:2 (three to two). This can be
simplified by dividing both the numerator and the denominator by the denominator to give 1.5:1 (one and a half to
one). Expressed in odds, it is simply 1½.

PROPORTION
n
Proportion N1
where n1 is a subpopulation of the whole study population, N.

Public [Link] 7 3/6/08 [Link]

 1A E p ide m io l o gy

If a bag contains 2 white balls (n1), 3 black balls (n2) and 5 grey balls (n3), then the proportion of black balls is
3/10 or 0.3.
For example,
Number of males in a population
Proportion of men in a population* =
Number of males + Number of females in a population
*Assumes that all people in the population are either male or female

RATE
N
Rate =
P¥T
where N is the numerator, P is the number of people in the population and T is a period of time.
This is a measure of frequency of the occurrence of a phenomenon. The denominator is constituted from both
population and time.
For example,
Number of new cases of epilepsy
Incidence of epilepsy =
Population ¥ Reporting period
Denominators should include only those at risk, and not those who cannot possibly develop the disease. For
example, people who have been immunised against a disease should not be included in studies looking at the rate
at which people acquire that disease, since they cannot be considered in the same way to be at risk as those who
have not been immunised.

1A.3 TIME AT RISK

Time at risk describes the total amount of time that individuals within a study spend at risk of developing the
disease of interest. The concept is particularly important in the analysis of cohort studies. One of the problems with
conducting cohort studies is that some subjects will join the study after data collection has started. Others will
leave the at-risk population early because they have:
• Died
• Moved away
• Become a case
• Been lost to follow-up for another reason (e.g. withdrawn from the study)
• Been censored (see below).
To correct for this variation, the ‘total person-time at risk’ is used as the denominator in calculations of morbidity or
mortality relating to cohort studies. Person-time represents the sum of all the individual times at risk.
Number of new cases in a specified period
Incidence rate =
Total person-time at risk
Box 1A.3.1 represents a small cohort study, and illustrates the concept of time at risk. As time progresses, the
number of individuals at risk will fall as people die, become a case or are lost to follow-up. Censoring occurs when
the value of an observation is only partially known. At the end of this study, patient 3 was still alive and therefore
spent at least 10 years at risk. This is known as right-censoring. Left-censoring (where a data point is below a
certain value but it is unknown by how much) and interval censoring (where a data point is somewhere in the
interval between two values) are also possible.
Sometimes, participants are replaced by new people recruited into the study. For studies of short duration and where
few individuals leave or join the at-risk population, it is reasonably accurate to use the number of individuals at the
start of the study as the denominator in incidence calculations.

Public [Link] 8 3/6/08 [Link]

 1A Epidem iology

Box 1A.3.1 Cohort study illustrating the concept of time at risk

Patient 1 Patient 2 Patient 3 Patient 4
Date of entry to study 1998 2000 1995 1997
Event Death Loss Censored Death
Date of event 2004 2003 2005 1997
Person-years in study 6 3 10 2

1995 96 97 98 99 2000 01 02 03 04 2005

Patient lost
to follow-up/
1
left the study
Patient

2
Patient died
3

CALCULATION OF MORTALITY RATE USING PERSON-TIME AT RISK

Where the study duration is longer or the likelihood of individuals leaving the at-risk population is greater, it is
preferable to use a more accurate calculation of the incidence rate using person-time at risk. See Box 1A.3.2.

Box 1A.3.2
Calendar time Person-years Events Death rate (number/year)
1995−1999 8 1 0.125
2000−2004 12.5 1 0.080
(2005−2009) (0.5) (0) (0)

1A.4 METHODS FOR SUMMARISING DATA

See also Section 1B.7.
Data are considered as either qualitative (non-numeric) or quantitative (numeric). Subtypes of each type of data
are listed below.

QUALITATIVE DATA
Qualitative data are non-numeric. Some qualitative data are categorical: they describe different categories or states
that a subject may fall into. These may be further categorised into nominal data and ordinal data.

Public [Link] 9 3/6/08 [Link]

 1A E p ide m io l o gy

NOMINAL DATA
Nominal data have no order and thus only give names or labels to various categories. Examples include the ABO
blood groups (A, B, AB, O), the names of colours (red, yellow, orange, green, violet, purple) and types of hospital
ward (rehabilitation, surgical, medical). See Box 1A.4.1.

ORDINAL DATA
Ordinal data have order, but the interval between measurements is not meaningful.
Some qualitative data are ordinal data: there is a natural order to the states, but no clear numerical relationship
between them. Examples are [poor, fair, good, better, best] and [very quiet, quiet, normal volume, loud, loudest].
Although ordinal data should not be used for calculations, it is not uncommon to find averages calculated from
data collected of the type strongly disagree, disagree, neither agree nor disagree, agree, strongly agree. See Box
1A.4.1.

QUANTITATIVE DATA
Quantitative data are numeric, and they are further classified as either discrete or continuous. Data may come from
one or more of the following categories.

DISCRETE DATA
Discrete data have a finite number of possible numerical values. Examples include the number of children with brown
eyes in a class of 30 children, or the number of times someone has been admitted to hospital in their lifetime.

CONTINUOUS DATA
Continuous data include measurable quantities of length, volume, time, mass, etc. They frequently have an upper or
lower limit, e.g. height cannot be <0.

INTERvAL DATA
Interval data have meaningful intervals between measurements, e.g. the age groups 0−4, 5−9, 10−14 ... 90+. They
are typically displayed as a table or histogram.

RATIO DATA
Ratio data are the most flexible data to work with, since they contain the most information of any data type. It
becomes meaningful to say not only that A scored 1 and B scored 2, but that B is twice as good as A. Ratio data are
ideal for use as outcome variables in regression. See Box 1A.4.1.

BINARY DATA
Binary data are a special type of data that have just two values. Depending on how they are analysed, they
may be considered to have properties of ordinal data, interval data or ratio data. They are very common data in
epidemiology, since they accurately describe many of the states that are of interest. For example, did a patient
improve following a particular treatment – or not? Is the case alive or dead? Was the case in the treatment group or
the control group?
Binary data often take values such as True/False or Male/Female. For analysis, they are usually transformed into
values of 0 and 1. Data of this type are often used in logistic regression (see Section 1B.14), where they might be
the outcome variable itself, or a binary coefficient reported as an odds ratio.

10

Public [Link] 10 3/6/08 [Link]

 1A Epidem iology

Box 1A.4.1
Example: Colours
To most people, black, brown, red, orange, yellow, green, blue, violet, grey and white are just names of colours
– nominal data. To an electronics student familiar with colour-coded resistors, these data are in ascending order
and thus represent ordinal data. To a physicist, red, orange, yellow, green, blue and violet correspond to specific
wavelengths of light and would be considered ratio data.
Example: Temperatures
Celsius Fahrenheit Kelvin
−273.15°C −459.67°F 0K
0°C 32°F 273.15 K
100°C 212°F 373.15 K
−17.8°C 0°F 255.4 K
Only the Kelvin scale has a true zero and can thus be used as a ratio scale. A temperature of 200 K is twice
as hot as a temperature of 100 K. This is not true for the Celsius and Fahrenheit scales.

Table 1A.4.1 Prevalence of

SUMMARISING DATA different ABO blood groups in
Discrete variables (e.g. blood groups) are typically summarised as proportions. the UK population
For example, the distribution of the common ABO blood groups in England is Blood group Proportion
shown in Table 1A.4.1.
A 0.41
Continuous variables (e.g. blood pressure) are described using both:
B 0.08
• A measure of central tendency (mean/median/mode) AB 0.03
• A measure of spread (range/variance/standard deviation).
O 0.48
If a population is normally distributed, then it can be described by both its
Reproduced from [Link].
mean and its standard deviation alone. For an example see Box 1A.4.2.
[Link].

Box 1A.4.2
Ages of children in a childhood leukaemia
cluster
Define n as the set of individual observations
1 5
Define ni as a particular observation
2 5
2 6 Define N as the total number of observations
2 7 Measures of central tendency of observations:
2 7
Mean age =
∑n i
= 100/18 = 5.6 years
3 9
N
3 10
4 13 Modal age (most frequently occurring) = 2 years
4 15 Median age = 4.5 years. This is the value midway between the
9th and 10th observations of the 18 ranked observations in the
N = 18 children data

11

Public [Link] 11 3/6/08 [Link]

 1A E p ide m io l o gy

MEASURES OF THE SPREAD OF OBSERvATIONS

Some commonly used terms are listed in Table 1A.4.2 with further properties described in Section 1B.7.

Table 1A.4.1 Measures of spread

Measure Description
Range The difference between the largest and smallest value
Percentiles The value below which p% of the observations in a population fall is called the pth
percentile
Interquartile range The difference between the value at the 25th centile and the value at the 75th centile
(i.e. between the 1st and 3rd quartiles)
Standard deviation Measure of the spread of observations about the mean of the sample. Takes the same units
as the data; 1.96 standard deviations either side of the mean covers 95% of the population
Used to describe the data ‘Deviation  Description’
Square root of the variance of the sample, i.e. standard deviation s = variance
∑( x − n )
2

where variance = s2 = i

n−1
Standard error of The error associated with measuring a sample mean is determined by both the standard
the mean deviation and the number of observations in the sample. The more observations, the more
precise is the estimate of the mean
Measure of variability of the mean of the sample
Used to make estimations about the true mean of the population
‘Error  Estimation’
s
SE =
n
Central limit If repeated samples are taken from any population, then the means of these samples will
theorem tend towards a normal distribution, even if the population is not normally distributed

1A.5 INCIDENCE, PREvALENCE AND STANDARDISATION

Incidence and prevalence (direct and indirect standardisation)
Incidence and prevalence are measures of occurrence.

INCIDENCE
Incidence relates to new occurrences. There are three related measures of incidence as shown in Table 1A.5.1.

PREVALENCE
Prevalence relates to existing occurrences.
Prevalence is also called the ‘point prevalence’, i.e. the proportion of a population with a disease. It is approximately
equal to incidence x duration provided that the incidence and the death/recovery rate have been stable for the

12

Public [Link] 12 3/6/08 [Link]

 1A Epidem iology

disease over some preceding time. Period prevalence relates to the proportion of the population with a disease
during a specified period. These two measures of prevalence are shown in Table 1A.5.2.

Table 1A.5.1 Measures of incidence

Concept Alternative name Definition
Incidence Incidence rate* Number of new events during a specified time period
Incidence rate Force of morbidity Number of new events divided by the total person-time at risk
Incidence density Used to describe studies where there have been varying periods of
follow-up (e.g. 3 cases during 25 person-years ∫ 12 cases per 100 person-
years)
Cumulative Risk Proportion of a population who become diseased in a defined time period.
incidence This is a measure of the risk that an individual will become diseased
during a defined time period, e.g. the attack rate during an epidemic
Note:
1. Incidence (unlike prevalence) is not affected by disease survival
2. The denominator should include only those ‘at risk’

*Rather confusingly, the term incidence rate is sometimes used to mean incidence. The incidence and incidence rate
can generally be considered as the same concept, with the number of person-years being more or less accurately
measured.

Table 1A.5.2 Measures of prevalence

Concept Definition
Point prevalence Proportion of a population with a disease at a particular point in time
Period prevalence Proportion of a population who had the disease during a specified period
Note: since period prevalence considers both cases at the start and new cases arising during the
time period, it includes features of both prevalence and incidence.

STANDARDISATION
Also known as adjustment, this technique is required to make comparisons between populations of differing
demographic structures, where crude mortality rates would be misleading. For example, the technique is typically the
first step in addressing questions such as, ‘Does town A have a higher mortality rate than town B?’ or, ‘Are 10 incident
cancers in a workforce of 100 people over a 10-year period more or less than we would have expected had national
rates applied?’ The technique is mostly used to compare populations that differ in their age structure – but it may
also be applied to correct for differences in gender or social class (or combinations of these and other variables).
To standardise data, populations are divided into strata based on differences that might potentially influence the
comparison that is being made. For example, to compare the numbers of deaths in a retirement community with a
town having a high proportion of young mothers, the mortality rates would need to be age standardised.
Two forms of standardisation that are commonly used are direct and indirect.

DIRECT STANDARDISATION
Direct standardisation may be used to compare two populations in different regions, or the same population at
two different periods of time. In direct standardisation, the age- or stratum-specific mortality rates of the observed

13

Public [Link] 13 3/6/08 [Link]

 1A E p ide m io l o gy

population are known (e.g. rate of lung cancer in 20- to 24-year-old men). These Table 1A.5.3 European
rates are then applied to a standard or reference population. Data typically need Standard Population
to be available concerning a large number of subjects in order to have adequate
numbers in each stratum to be confident about the estimate. Age group Number in
age group
• Start by choosing a single population (e.g. one of those being compared, 0 1600
their average or an outside population) 1–4 6400
• Break this single population down into individual age bands (‘standard age
5–9 7000
structure’)
• For the populations being compared, take the age-specific mortality rate for 10–14 7000
each age band and multiply it by the size-weighting of that age band from 15–19 7000
the standard age structure 20–24 7000
• Sum all of these to obtain the age-standardised mortality rate. The actual 25–29 7000
value of this adjusted rate is meaningless but it shows the true difference in 30–34 7000
rates between the populations being compared. 35–39 7000
40–44 7000
THE EUROPEAN STANDARD POPULATION
45–49 7000
The European Standard Population (see Table 1A.5.3) is used to compute 50–54 7000
directly age-standardised rates. The same population is used for males, females 55–59 6000
and all persons. 60–64 5000
The example in Box 1A.5.1 shows how mortality data from 2 years can be 65–69 4000
compared using direct standardisation. 70–74 3000
75–79 2000
INDIRECT STANDARDISATION 80–84 1000
With indirect standardisation, the population under study is usually small. The 85+ 1000
technique hinges on the calculation of the rate of death that would have been TOTAL 100 000
expected in the study population had a comparison rate applied instead. This
permits the determination of the standardised mortality ratio (SMR). Reproduced from the World
Health Annual of Statistics
• Start with the stratum-specific death rates of a standard population (e.g. (1991), based on Waterhouse
European Standard Population if age is the effect to be standardised) et al (1976).
• Use these to calculate expected number of deaths in each stratum of the
study population
• Add up the expected number of deaths for each age band:
Define E as the expected number of deaths
Define k as the number of strata
Define ni as the number of people in the ith stratum
Define Ri as the rate of death in the ith stratum
k
E = ∑ ni Ri
i =1

• Standardised mortality ratio (SMR) is then calculated as (Observed deaths/Expected deaths).

The SMR indicates whether mortality in the study group (after correction for age, sex, etc.) is unusual when
compared with a standard population. In occupational mortality studies, comparisons are often made against two
standard populations: (i) an unexposed population from the same occupation and (ii) the general population.
When used to estimate the association between an occupational exposure and a disease, the SMR underestimates
the true magnitude of association because the general population contains both exposed and unexposed individuals.

14

Public [Link] 14 3/6/08 [Link]

 1A Epidem iology

Two SMRs should never be compared with each other: the number of expected cases in each group depends on the
group’s actual age/sex/race composition. Instead, direct standardisation can be used to compare the two groups.
See Box 1A.5.2 for an example of this calculation.

Box 1A.5.1
Example: Comparison of 2 years’ mortality data using directly age-standardised rates
Deaths from all malignant neoplasms (ICD-10 C00–C97) in persons aged under 75 in a defined population†

Stage 1: calculation of age-specific death rates per 100 000 population in a defined population
(a) Number of deaths by age group in the defined population
Year <1 1−4 5−9 10−14 … 65−69 70−74
2001 0 0 1 1 … 53 83
2002 0 0 0 0 … 63 68

(b) Population by age group in the defined population

Year <1 1−4 5−9 10−14 … 65−69 70−74
2001 1900 8200 11 100 11 800 … 9300 8300
2002 1900 8200 11 100 11 800 … 9300 8300

(c) Age-specific death rates per 100 000 in each age group/year = (number of deaths ∏ population in each age
group) ¥ 100 000
Year <1 1−4 5−9 10−14 … 65−69 70−74
2001 0 0 9.01 8.47 … 569.89 1000.00
2002 0 0 0 0 … 677.42 819.28

Stage 2: calculation of European age-standardised annual rates

Method: find the expected number of cases in the European Standard Population had the death rate observed
applied. Multiply the death rate in a given age group/year by the number of people from the correct cell of the
European Standard Population. Sum these to derive the ‘expected’ numbers of deaths in the population as a
whole.

Year <1 1−4 5−9 10−14 … 65−69 70−74 Sum Standard

(<75 s) rates*
2001 0 0 0.63 0.59 … 22.80 30.00 112.60 117.29
2002 0 0 0 0 … 27.10 24.58 109.84 114.42
*Obtained by dividing the sum of expected numbers at ages under 75 by the European Standard Population at ages
under 75 (96 000/100 000).
†
Adapted from: NHS performance indicators: standardisation. Available online at: [Link]/Ratings/Downloads/
direct_standardn_meth.pdf.

15

Public [Link] 15 3/6/08 [Link]

 1A E p ide m io l o gy

Box 1A.5.2
Example: Indirect standardisation of deaths from cirrhosis of the liver in British doctors (after Bland 2000)

Age group Mortality per million Number of male doctors Expected deaths
men per year in the UK in the UK
15−24 5.859 1080 0.0063
25−34 13.050 12 860 0.1678
35−44 46.937 11 510 0.5402
45−54 161.503 10 330 1.6683
55−64 271.358 7790 2.1139
TOTAL 4.4966
If 14 deaths were observed among British doctors, then:
Indirectly, SMR compared with the background population = O/E = 14/4.4966 = 3.11
SMR (*100 as integer) = 311
Exact 95% confidence interval = 1.70–5.22 (SMR 170–522)

1A.6 YEARS OF LIFE LOST

One way of considering the impact of a particular disease or risk factor is to consider how many years people might
have expected to have lived had their lives not been curtailed by the disease. For example, deaths from road traffic
collisions affect mostly young males who otherwise may have been expected to live into their 70s or 80s.
‘Years of life lost (YLL)’ is a measure of premature mortality and it explicitly places more importance upon deaths
that occur in the young than upon those in the elderly. It is therefore a value-laden statistic, which reflects
the wish of society to prevent avoidable causes of death in younger people, and to avoid the loss to society of
investment in raising children and young adults.
To calculate YLL in its most simple form, an upper age limit (e.g. 75) is chosen. A person dying at the age of
60 contributes 15 YLL (75 – 60 = 15) to the calculation. A person dying at age 15 contributes 60 years. Deaths
occurring in people aged over 75 are not included in the calculation. Infant deaths may or may not be included.
In more complex calculations, an actuarial assessment of the expectation of life for each person will be undertaken.
Life-expectancy is calculated using current life-tables: age-specific mortality rates are applied to a hypothetical
population

∑N × L
where:
N = number of deaths in a particular age–sex group
L = life-expectancy of this age–sex group in this population.
A specific type of YLL can be calculated for smoking:
Smoking-attributable YLL = (Age- and sex-specific mortality attributable to smoking) ¥ (Remaining life-expectancy
in this age and sex group).

16

Public [Link] 16 3/6/08 [Link]

 1A Epidem iology

Note that YLL will underestimate the burden of disease due to chronic conditions, which have a low mortality at a
young age. To take this factor into account, health-adjusted life-expectancy (HALE) can be used instead of crude
life-expectancy.
HALE = S (Number of life years lived in each age group) ¥ (Mean health state score for that age group)

1A.7 MEASURES OF DISEASE BURDEN

Measures of disease burden (event and time based) and population attributable risks, including identification of
comparison groups appropriate to public health
See Section 1A.10 for risks.
A number of measures exist to assess the weight of disease affecting a community, taking into account both the
number of events and the impact of these events on the population. They each depend on the availability of data.

MEASURES OF DISEASE BURDEN BASED ON EVENTS

These studies of incidence require data from routine sources:
• Death certificates (if the disease burden is in mortality rather than morbidity)
• Hospital episode data (if the disease results in hospital admissions)
• Disease register (if a dedicated register for that disease exists)
• Statutory notifications: infectious disease, procedures, e.g. abortion.

MEASURES OF DISEASE BURDEN BASED ON TIME

Where there are no routine event-based data collected, a prevalence (‘cross-sectional’) survey may be needed. Figure
1A.7.1 is a graph that illustrates the relative contributions of various conditions to the sum of ‘disability-adjusted
life-years’ (DALYs) for the UK.

Ischaemic heart disease

Unipolar major depression

Stroke

Alcohol use

Road traffic injuries

Lung cancer

Dementia

Osteoarthritis

Diabetes

Self-inflicted injuries Figure 1A.7.1 Disability

burden of a range of
0 2000 4000 6000 8000 10 000 conditions in the UK.
DALYs Reproduced from the WHO

17

Public [Link] 17 3/6/08 [Link]

 1A E p ide m io l o gy

1A.8 vARIATION
Sources of variation, its measurement and control
Variation arises from differences between populations or between individuals within a population. It may be due to
random or non-random factors.
Measurements can differ for many reasons. Some differences will be due to genuine dissimilarities between subjects.
This sort of variation is often random, and follows a normal distribution. Some people are taller than average, some
are shorter than average. This is everyday, random variation.
Other differences will be due to discrepancies in the way that measurements were made or recorded. A ruler may
measure only to the nearest 5 cm. It judges a person’s height as 175 cm and another’s as 170 cm. A more precise
ruler, accurate to 0.1 cm, would confirm that the first subject is in fact 176.1 cm tall and the second is 172.4 cm
tall. The aim for any investigator is to minimise the error associated with making the measurements, thereby
focusing attention on the real differences between individuals.
Measurements can be biased or unbiased, precise or imprecise, or some combination of these.
By way of analogy, consider a markswoman firing shots at a target. Her sights might be on-target or off-target,
and she might be a good shot or a bad shot. The diagrams in Figure 1A.8.1 illustrate the problems that she might
encounter when aiming at a target. Making measurements in epidemiology is much the same.

Sights precise and unbiased Sights precise but biased

0.4 0.4

0.3 0.3

0.2 0.2
f(x)
f(x)

0.1 0.1

0 0
–6 –4 –2 0 2 4 6 –6 –4 –2 0 2 4 6
x x

Distribution of shots Distribution of shots

18

Public [Link] 18 3/6/08 [Link]

 1A Epidem iology

Sights imprecise yet unbiased Sights imprecise and biased

0.2 0.2

0.15 0.15

0.1 0.1
f(x)

f(x)

0.05 0.05

0 0
–6 –4 –2 0 2 4 6 –6 –4 –2 0 2 4 6
x x

Distribution of shots Distribution of shots

Figure 1A.8.1 Imprecision and bias for a gun and a target

Sometimes a group of people differs systematically from another. One subject might belong to an African pygmy
tribe and another subject might be from a tall population, e.g. the Netherlands. The observed differences in
height between these two individuals are not due to random variation but are attributable to systematic genetic
differences.
It is this sort of variation that is of interest to epidemiologists. Most epidemiological techniques aim to control
random variation in order to establish whether the systematic variation is significant.

1A.9 ERRORS IN EPIDEMIOLOGICAL MEASUREMENT

Common errors in epidemiological measurement, their effect on numerator and denominator data, and their avoidance
See also Section 1B.10.
Measurement error can be defined as any mistake that occurs during the process of applying a standard set of values
(i.e. a measurement scale) to a set of observations. In epidemiology, such errors can lead to the misclassification of
cases and controls. Acknowledgement of measurement error is important in research, and leads to more robust and
defensible scientific results.
Measurement errors are either random (i.e. occurring due to chance) or systematic (i.e. persistent, non-random,
differences between the observed measurement and its true value). These errors will lead respectively to non-

19

Public [Link] 19 3/6/08 [Link]

 1A E p ide m io l o gy

differential errors (affecting all groups equally) and differential errors (affecting one group more than another).
Table 1A.9.1 compares random and systematic errors in epidemiological measurement.

Table 1A.9.1 Random and systematic errors in epidemiological measurement

Random error Systematic error
Occurs due to chance Occurs due to non-random factors
Affects all groups equally (non-differential) Affects some groups more than others
(differential)
Exposure: misclassification of exposures is equal for cases and Exposure: misclassification of exposures
controls (e.g. 20% of cases and 20% of controls are misclassified as differs between cases and controls (e.g.
being exposed) 20% of cases misclassified as having
been exposed but only 5% of controls
misclassified)
Outcome: misclassification of outcomes is equal for exposed and Outcome: misclassification of outcomes
non-exposed (e.g. 30% of exposed and 30% of non-exposed are differs between exposed and non-exposed
misclassified as having the disease) (e.g. 19% of the exposed are misclassified
as having the disease but only 7% of non-
exposed are misclassified as having the
disease)
Bias is towards the null-hypothesis (i.e. dilution of the study’s Bias can be in any direction
findings)
Less threatening for a study than systematic error More threatening for a study than random
error
Example: a study of lung cancer in relation to proximity of Example: a study assessing the association
residence to a coke oven classifies subjects (cases and populations) between visual display unit (VDU) usage
by distance of residence from the oven at the time of follow-up. and spontaneous abortion. Here, cases are
Here there is misclassification due to migration (not all people more likely to recall VDU usage compared
living near the oven at the time of follow-up will have lived there with controls, particularly if there has been
at the aetiologically relevant time) but this error occurs randomly media interest in this hypothesis. Therefore
the association measured is likely to be
greater than the true association

EFFECT OF MEASUREMENT ERRORS

Measurement errors can affect:
• Dependent variables (outcomes such as disease)
• Independent variables (risk factors such as exposures)
• Confounder variables: see Section 1A.15
• Effect modifiers: see Section 1A.15.
With regard to the measurement of confounder variables, an apparent association between the exposure and the
disease may persist after statistical adjustment. Such residual confounding is particularly problematic where the
variable being measured is difficult to quantify (e.g. measurement of socioeconomic status). For this reason, residual
confounding should always be considered where an association persists following statistical adjustment for a known
confounder.

20

Public [Link] 20 3/6/08 [Link]

 1A Epidem iology

AVOIDANCE OF MEASUREMENT ERROR

Measurement error can be avoided or accounted for at various stages of the study. See Table 1A.9.2.

Table 1A.9.2 Avoiding and accounting for measurement error

Stage of study Strategy for dealing with measurement error
Design Set out to measure reliability: correlation coefficients (continuous variables) or Cohen’s kappa
(categorical variables)
Blinding
Standardised measurement instruments
Use multiple sources of information (questionnaires, direct measurements, registries, case
records)
Use multiple controls
Data collection Administer instruments equally to:
cases/controls
exposed/unexposed
Data analysis Perform a sensitivity analysis to test the robustness of the findings (see Section 4D.5)
Reporting Consider and mention potential random and systematic errors

1A.10 CONCEPTS AND MEASURES OF RISK

See also Sections 1A.5, 1A.7 and 2F.2.

RISK
Also called ‘cumulative incidence’.
Risk, in an epidemiological sense, is the likelihood of some future event. For example, the risk that a 30-year-old
male will die at some point in the future is 100%. The risk that this 30-year-old male will die in the next year is
about 1 in a 1000.

ATTRIBUTABLE RISK
Also called ‘absolute risk’ or ‘excess risk’, the attributable risk is the difference in the risk or rate of disease
between the exposed group, Ie (pronounced I sub e), and the unexposed group, I0 (I sub zero), assuming causality
(see Section 1A.13).
Ie – I0

The attributable risk per cent, also called proportional attributable risk or aetiological fraction when expressed as
a proportion, is the percentage of disease in the exposed group, which is attributable to the exposure.
Ie − I0 RR – 1
× 100% or × 100% where data on disease incidence is not available RR = relative risk (see page 23).
Ie RR

21

Public [Link] 21 3/6/08 [Link]

 1A E p ide m io l o gy

This is a useful epidemiological statistic that is readily understood by non-epidemiologists. For example, in people
who smoke, about 90% of lung cancers arise because they smoke. See Box 1A.10.1.
Population attributable risk is the excess rate of disease in the whole population that is attributable to the
exposure. Define IT as the rate of disease in the total study population (both exposed and unexposed groups). Define
I0 as the rate in the unexposed population.
Population attributable risk = IT – I0

Box 1A.10.1
Example: Population attributable risk of smoking in lung cancer
Mortality in whole population = 55 per 100 000
Mortality in non-smokers = 16 per 100 000
Population attributable risk = Rate in population – Rate in non-smokers
= 55 – 16
= 39 deaths per 100 000/year
If generalising for a broader population, the true prevalence of exposure in that broader population must be known
from an outside study; otherwise it will have to be assumed that the broader population has the same exposure as
the study population.

POPULATION ATTRIBUTABLE (RISK) FRACTION

This statistic measures the proportion of disease in the study population that is attributable to the exposure. This is
the preventable fraction. It is one of the most important indices in prioritising population interventions and may
also be called the ‘population attributable risk per cent’. It can be calculated in two ways. See Box 1A.10.4 for an
example of its calculation.
Population attributable risk
Population attributable fraction =
Rate of disease in population
Prevalence in exposed population ¥ (RR–1)
=
[1 + Prevalence in exposed population (RR–1)]

COMPARISON GROUPS USED IN PUBLIC HEALTH

Comparison groups may be used in the control arm of a study (see Section 1A.19). At the population level,
comparators are used for needs assessment studies and for assessing the quality of health care. Factors to consider
when choosing comparison groups include:
• Size of population
• Area characteristics (rural/urban)
• Age structure
• Ethnicity
• Socioeconomic characteristics
• Disease burden (mortality and morbidity)
• Health service usage

22

Public [Link] 22 3/6/08 [Link]

 1A Epidem iology

• Health service provision, funding and organisation (relevant to international comparisons)

• Provision of other relevant services, e.g. social care.

RELATIVE RISK (RATIO)

Relative risks are used to measure the strength of association between an exposure and an outcome. They are used
to assess aetiological strength and can be calculated as the risk ratio, rate ratio, odds ratio or standardised
mortality ratio.

RISK RATIO
Risk ratio = (Risk of disease in exposed) ∏ (Risk of disease in non-exposed).
See Box 1A.10.2 for an example of how to calculate a risk ratio.

Box 1A.10.2
Example: Calculating the risk ratio for hypertension among cardiac patients
A consultant believes that hypertension is unusually common in patients in her
clinic who have had a coronary bypass operation.

Hypertension
Yes No
Coronary bypass Yes 15 467 482
No 70 6364 6434
85 6831

Risk ratio = (Risk of disease in exposed) ∏ (Risk of disease in unexposed)

= (15 ∏ 482) ∏ (70 ∏ 6434)
= 0.0311 ∏ 0.0109
= 2.86
In this clinic, patients who had had a coronary bypass operation were 2.86
times as likely to have hypertension as those who had not had a bypass.

RATE RATIO
Rate ratio = (Incidence rate in exposed) ∏ (Incidence rate in non-exposed).

23

Public [Link] 23 3/6/08 [Link]

 1A E p ide m io l o gy

ODDS RATIO*
Odds ratio = (Odds of exposure in cases) ∏ (Odds of exposure in controls).
*Mainly used for case–control studies only
Note: in cohort studies the time period must be stated (because the risk of dying will always be 100% in the long
run for both groups).
See Box 1A.10.3 for an example of how to calculate an odds ratio.

STANDARDISED MORTALITY RATIO (SMR)

Standardised mortality ratio = (Number of cases observed) ∏ (Number of cases expected) ¥ 100%.
The SMR is often used in occupational settings.

Box 1A.10.3
Example: calculating an odds ratio for risk of stroke among rheumatology patients
A consultant believes that, among the patients attending her rheumatology clinic, stroke is
more common among those who have recently begun immunotherapy.

Stroke
Yes No
Immunotherapy Yes 8 168
started in last
year No 184 9813

Odds ratio = (Odds of disease in exposed group) ∏ (Odds of disease in non-exposed group)
= (8 ∏ 168) ∏ (184 ∏ 9813)
= 0.048 ∏ 0.019
= 2.54
So, the consultant is proved right: stroke is 2.54 times more common in patients attending
the rheumatology clinic who have had immunotherapy in the previous year compared with
those attending the rheumatology clinic who have not had immunotherapy.

ATTACK RATE AND SECONDARY ATTACK RATE

These measures are used in outbreaks or epidemics of infectious diseases. Both are risks.
Number of new cases
secondary Attack rate =
Population at risk in a defined time period

Number of cases from exposure to primary case

Secondary attack rate =
Total number exposed

24

Public [Link] 24 3/6/08 [Link]

 1A Epidem iology

1A.11 THE ODDS RATIO

See Section 1A.10.

Box 1A.10.4
Example: Attributable risk fraction of lung cancer associated with smoking
34% of the population smoke. The incidence of lung cancer is 376.8/100 000 per year among
smokers and is 13.4/100 000 per year among non-smokers.
The key point to remember is that smokers suffer from lung cancer because of the effect of
smoking and because of the background causes which non-smokers also experience.
So, in 100 000 of the population there will be:
34 000 smokers (who will have the risk from smoking plus the background risk); and
66 000 non-smokers (who will just have the background causes)
(a) Number of cases of lung cancer among smokers that are due to their smoking habit:
= [(376.8 – 13.4) ∏ 100 000] ¥ 34 000
= 123.6 cases (this is the population attributable risk)
(b) Number of cases of lung cancer among smokers due to background risk:
= (13.4 ∏ 100 000) ¥ 34 000
= 4.6 cases
(c) Number of cases of lung cancer among non-smokers due to background risk:
= (13.4 ∏ 100 000) ¥ 66 000
= 8.8 cases
Number of cases of lung cancer in population:
= (a) + (b) + (c)
= 123.6 + 4.6 + 8.8
= 137.0
Attributable risk fraction:
Population attributable risk
= ¥ 100%
Rate of disease in population
(a)
= ¥ 100%
(a) + (b) + (c)
= (123.6 ∏ 137.0) x 100%
= 90.2%

1A.12 RATE RATIO AND RISK RATIO (RELATIvE RISK)

See Section 1A.10.

25

Public [Link] 25 3/6/08 [Link]

 1A E p ide m io l o gy

1A.13 ASSOCIATION AND CAUSATION

The concepts of association and causation are fundamental to the science of epidemiology. An association is a
statistical link between two variables. It is assessed using a statistical test that calculates how unlikely it is that
the finding occurred due to chance alone. An association does not imply that one factor brought about the other.
For example, there is an association between eating ice-cream and hot weather, but eating ice-cream does not cause
hot weather.
Association = statistical link
Association π cause and effect.
Causality implies that one factor caused the other. Once an association has been established, then determining
causality involves a subsequent judgement of evidence. Causality is determined according to a number of set criteria
(see below). An epidemiologist must never confuse association with causality.
It is important to be cognisant of three phenomena that can erroneously make an association appear to be present
when no such association exists in reality. These three phenomena are chance, bias and confounding.

CHANCE
Epidemiological studies make inferences regarding the wider population based on observations from a sample.
However, an association might be observed in the sample due to luck of the draw, i.e. the sample does not truly
reflect the wider population. The chance of observing such an unrepresentative association falls as the sample size
increases.
The null hypothesis states that the association observed in the sample was due to chance alone. So the probability
that the null hypothesis is true (called the p value) is the probability that an association at least as extreme as that
observed could have occurred due to chance alone. If the p value is sufficiently low, i.e. sufficiently unlikely, then
the null hypothesis is rejected in favour of the alternative hypothesis (which states that the association observed in
the sample truly exists in the wider population).

Box 1A.13.1
Type I error (a) Mistaken rejection of the null hypothesis when it was in fact true
Experiment says that the treatments are different, when the truth is that they are the
same (false positive)
Value of a is the significance level (p value) of the test; it is decided before data are
collected
Type II error (b) Mistaken rejection of the alternative hypothesis when it was in fact true
Experiment says that the treatments are not different when the truth is that they are
different (false negative)
Occurs when the sample size is too small
Power (1 - b) Probability of correctly rejecting the null hypothesis, usually set at 80%

BIAS
For details of different types of bias see Section 1A.14. Estimates of rates and risks rely on obtaining an accurate
denominator figure for the population at risk. Where the denominator is not available (or is systematically inflated
or reduced), then errors in epidemiological measurement occur. For example, GP practice lists often include people

26

Public [Link] 26 3/6/08 [Link]

 1A Epidem iology

who have in fact moved out of the area or registered with another practice, commonly known as ghosts. When such
a list is used as a denominator, results may become biased.

MINIMISING BIAS
See also Section 1A.14.
Bias can be reduced in a number of ways:
• Randomisation with true concealment ensures that investigators are unable to predict or affect group
allocation. Random allocation (in intervention studies) effectively removes selection bias.
• Blindness: double blindness (investigator and subject both unaware of allocation) if possible, but if no
placebo is possible then ensure that the person assessing outcome is blinded to the exposure status. Blinding
of participants avoids respondent bias; blinding of practitioners who are treating study participants avoids
instrument bias; and blinding of observers who collect measurements or analyse results avoids observer bias.
• Collect irrelevant factors to check for bias between groups and to mask the hypothesis under investigation.
• Enhance the reliability/reproducibility of measurements, e.g. repeating measurements, and assure inter-
observer agreement (instrument bias).
• Study personnel should receive standardised training and have their performance monitored (e.g. time the
interviews to assess the degree of probing by interviewers).
• Data collection should be by detailed written protocol using closed questions and direct observations.
• Choose hospitalised controls to increase comparability (they will have similar recall of events before
admission). However, be aware of the hospital having different catchment areas for different specialties.
• Choose cohorts that can be easily followed up long term, e.g. alumni groups.
• Choose cohorts with above-average risk of disease to shorten follow-up period, and therefore reduce losses to
follow-up.
• Use multiple sources of data to assess exposure/disease status.
Bias in questionnaires can be reduced by:
• Checking for known associations
• Seeking the same information in different ways
• Checking characteristics of data collection (time taken).
Bias in intervention studies can be assessed by:
• Self-reports
• Pill counts
• Measuring biochemical parameters
• Incorporating a safe biochemical marker in the placebo that can be detected in urine.

CONFOUNDING
A confounded association is one in which the effect of the variable of interest is inextricably combined with that of
another uncontrolled variable. For example, a study that showed an association between increased level of foreign
travel and decreased risk of stroke might be confounded by age (because, on average, older people travel abroad less
frequently and are at higher risk of stroke than younger people). This is illustrated in Figure 1A.13.1.
Confounding can be controlled either at the measurement stage or the analysis stage (see Section 1A.16).

CAUSALITY
Once an association has been established, the issue of causality may be considered. Causality is a judgement of
cause and effect. It is based on valid study data plus a background of other evidence. To be valid, all alternative

27

Public [Link] 27 3/6/08 [Link]

 1A E p ide m io l o gy

Exposure
Disease (stroke)
(foreign travel)

Confounder (age)

Figure 1A.13.1 Age is a confounder in the association between foreign travel and stroke
explanations for the findings (i.e. chance, bias and confounding) must be dismissed. Judgement of causality is then
based on the criteria listed in Table 1A.13.1, which were first described by Bradford Hill (1965).

Table 1A.13.1 Bradford Hill criteria for determining causality

Strength of association The greater the magnitude of the risk observed (positive or negative), the less
likely it is to be due to confounding. The opposite is not true
Biological credibility/ A finding is more credible if there is a known or postulated biological mechanism
plausibility to explain it. However, an apparently implausible association might turn out to be
a scientific discovery
Consistency of findings Most persuasive evidence comes from several studies with different methodologies
showing similar results
Temporal sequence The exposure of interest should precede the outcome by a biologically plausible
period
Dose–response There may be a gradient of risk according to the amount of exposure. Note,
however, that many biological phenomena demonstrate a threshold (rather than a
linear) relationship. The relationship between alcohol and cardiovascular disease
can show a ‘J’-shaped curve, implying greater cardiovascular disease risk for those
who drink no alcohol compared with those who drink small amounts
Specificity If the exposure is associated with one outcome or range of outcomes, this is
evidence in favour of a causal effect relationship. However, there are few such
one-to-one relationships, and many exposures cause several outcomes, e.g.
smoking causes a plethora of diseases
Coherence The relationship should not conflict with the natural history of the disease
Reversibility If the risk of the disease decreases when the exposure is removed, then this is
strong evidence of causality

1A.14 BIAS
A bias is a systematic error that leads to a difference between the comparison groups with regard to how they are
chosen, treated, measured or interpreted. This error leads to an incorrect estimate of the association between the

28

Public [Link] 28 3/6/08 [Link]

 1A Epidem iology

exposure and the risk of disease. Unlike confounding and the role of chance, the magnitude of a bias cannot be
quantified.
There are five main groups of bias:
• Misclassification
• Selection
• Measurement
• Biases relating to intervention studies
• Types of bias that are associated with screening programmes (e.g. lead-time bias, length-time bias and
volunteer bias – see Section 2C.2).
Different epidemiological study designs are particularly susceptible to the types of bias shown in Box 1A.14.1.

Box 1A.14.1
Type of study Type of bias
All studies Misclassification bias
Case–control Recall bias and selection bias
Retrospective cohort Follow-up bias and selection bias
Prospective cohort Follow-up bias
Intervention study Bias with comparison group, placebo, ascertainment

MISCLASSIFICATION BIAS
Misclassification will occur to some extent in all studies. It involves allocating subjects to the wrong group –
either at random or in a non-random fashion.

RANDOM MISCLASSIFICATION BIAS

Also known as ‘non-differential’ misclassification bias, this causes the result of the study to be diluted towards the
null.

NON-RANDOM MISCLASSIFICATION BIAS

If the errors in classification are non-random (i.e. ‘differential’), then the result of the study may be biased in any
direction or in none. For example, see Box 1A.14.2.

Box 1A.14.2
Example: Non-random misclassification bias among smokers
When asked, approximately 10% of smokers deny their habit. This
fact can be validated through analysis of urinary cotinine. These
untruthful responses can introduce a non-random misclassification
bias into any study where the analysis considers the effect of
smoking.

Some people have a tendency to ‘yea-say’, i.e. to agree with a statement rather than to disagree with
it. This phenomenon may result in the misclassification of these subjects.

29

Public [Link] 29 3/6/08 [Link]

 1A E p ide m io l o gy

SELECTION BIAS
This bias occurs when the presence or absence of exposure influences either allocation to particular study
groups or participation in the study. Examples of selection bias include the following.

HEALTHY WORKER EFFECT

This effect occurs when a sample drawn from a workplace is selected as representative of the general population.
The general population includes those too ill to work so the workplace sample is healthier than the general
population.

vOLUNTEER BIAS
This occurs because those who agree to participate in studies tend to be both healthier and more compliant than
those who do not.

FOLLOW-UP BIAS
This occurs when participants who are lost to follow-up differ systematically from those who remain contactable.

MEASUREMENT BIAS
There are three types of measurement bias.

INSTRUMENT BIAS
This occurs where there are systematic inaccuracies in a test or instrument.

RECALL BIAS
It is a feature of human psychology that experiences that precede an adverse event or outcome are recalled
particularly well. For this reason, participants who develop a disease will recall antecedent exposures differently
from those without disease.

OBSERvER BIAS
This occurs if systematically there are differences in the ways that the exposure or outcome data are collected
between the two groups. There may be inaccuracies or incompleteness in data collection that affects the groups to
differing degrees. Interviewer bias is a type of observer bias in which there is a systematic difference in the way
that the investigator collects, probes for or interprets data between groups.

BIASES IN INTERVENTION STUDIES

Intervention study designs are prone to the following types of bias.

COMPARISON GROUP
This occurs where there are systematic differences between the control group and the intervention group.
Comparison group biases can be eliminated by rigorous randomisation and intention-to-treat analysis (see Section
1A.20). Differences between the intervention group and the control group are typically logged in the first table of a
published study.

PLACEBO
It is a feature of human psychology that people tend to report favourable outcomes to any intervention that they
receive, regardless of its physiological effect. Studies that do not employ a placebo control will therefore be biased.

30

Public [Link] 30 3/6/08 [Link]

 1A Epidem iology

ASCERTAINMENT
If the collection of data from one group is more accurate or complete than that from the other, then a bias will
result. The longer the follow-up time, the higher the propensity to ascertainment bias. Ascertainment bias is a
particular problem where the outcome being measured is subjective (e.g. degree of pain).

MINIMISING BIAS
Several measures can be taken to avoid bias (see Section 1A.9).

EVALUATING BIAS
Although bias cannot be quantified (compare confounding and chance), attempts should always be made to:
• Contemplate the direction of the alleged bias
• Use internal validation. For example, if some participants were incorrectly diagnosed initially, then, if there
were no bias, these participants should have had the same exposures as those of the controls.

1A.15 CONFOUNDING AND INTERACTION

Confounding, interactions and methods for ascertainment of effect modification
See Sections 1A.13, 1A.16 and 1B.14.

CONFOUNDING
This occurs when an association arises because of differences between groups other than the exposure being
studied. The observation is distorted because of the presence of a confounder, i.e. a variable that is both associated
with the exposure and also independently associated with the outcome. For example, see Box 1A.15.1.

CONFOUNDERS
• A confounder must be associated with the disease and be independently associated with the exposure under
study
• A confounder must predict disease independently of the exposure under study (including in non-exposed
individuals)
• A confounder cannot simply be an intermediate step in the causal chain.

MEDIATING FACTORS
Not every factor that is associated with both the exposure and the disease is a confounding variable. A variable
that is associated with both the independent and dependent variables may be an intermediate step along the causal
chain. This is called a mediating factor and it is not a confounder. See Box 1A.15.2 for an example of a mediating
factor.

31

Public [Link] 31 3/6/08 [Link]

 1A E p ide m io l o gy

Box 1A.15.1
Example: Smoking as a confounder
Populations with high alcohol consumption tend to have high lung cancer rates, but this is simply because
people who misuse alcohol are more likely to smoke than people who do not. Smoking, in turn, is causally
associated with lung cancer. Smoking therefore confounds the apparent relationship between alcohol
consumption and lung cancer.

Association
Alcohol Lung cancer

Association Causal association

Smoking
Confounder

Box 1A.15.2
Example: Cholesterol as a mediating factor
There is a relationship between a poor diet and the incidence of coronary heart disease (CHD). There is also a
relationship between a poor diet and a high serum cholesterol level. We also know that a high serum cholesterol
level is causally associated with a higher risk of a CHD event. However, serum cholesterol is not a confounding
factor here. Because a high cholesterol level may be caused by a poor diet, it is in fact on the causal pathway of
the relationship between diet and CHD. It is therefore a mediating factor.

Association Association
Coronary heart
Poor diet High cholesterol
disease
Mediating
factor

INTERACTION
An interaction occurs when the effect on the outcome of one causal factor differs according to the level of a
third variable. This third variable is termed an effect modifier. For example, the effect of smoking on the risk of
myocardial infarction is stronger in the young: age is an effect modifier in this case.

32

Public [Link] 32 3/6/08 [Link]

 1A Epidem iology

DIRECTION OF CONFOUNDING
Positive confounding makes the association appear more pronounced (this may be in a positive or negative
direction).
Negative confounding makes the association appear less pronounced (i.e. diluted towards the null).

RESIDUAL CONFOUNDING
This occurs as a result of unknown confounders (i.e. those with effects that remain after the known confounders
have been taken into account) or where confounders are inaccurately measured. The process of randomisation,
which distributes both known and unknown confounders equally between groups, effectively eliminates all residual
confounding.

1A.16 CONTROL OF CONFOUNDING

Strategies to allow/adjust for confounding in design and analysis
Confounding can be countered at the design stage or it can be adjusted for at the analysis stage.

DESIGN
At the design stage, confounding can be addressed through restriction, matching and randomisation. See Table
1A.16.1.

Table 1A.16.1 Techniques for dealing with confounding at the design stage
Technique Method Advantages Disadvantages
Randomisation Participants allocated to If the sample is large Not always possible
groups at random enough then randomisation
removes known and unknown
confounders
Restriction If sex and race are considered Cheap Smaller pool of potential
to be potential confounders, recruits; residual confounders
then consider only one if restriction is insufficiently
particular group from across narrow; cannot assess varying
both factors (e.g. just black levels of a factor
men)
Matching Nowadays only used for case– Intuitive appeal; unique Difficult and expensive,
control studies: match for age, benefits of twin studies (see especially where there are
sex, race, smoking history, etc Section 1A.41); useful in small multiple controls for each
case series case; cannot explore factors
that have been matched; no
control for factors that have
not been matched

ANALYSIS
At the analysis stage, confounding can be corrected for by means of stratification or multivariate analysis. See Table
1A.16.2.

33

Public [Link] 33 3/6/08 [Link]

 1A E p ide m io l o gy

Table 1A.16.2 Techniques for dealing with confounding at the analysis stage
Method Details Disadvantages
Stratification Divide confounding variables into strata, and Unable to control simultaneously for more
provide stratum-specific relative estimates (with than a few confounders because number of
confidence intervals) plus a weighted-average strata increases exponentially, so number of
overall single estimate of the confounding individuals in each stratum falls
effect (e.g. Mantel-Haenszel method)
Multivariate Mathematical models (multiple regression, Black box, therefore transparency is lost (a
analysis logistical regression, etc) minor problem: this is usually the preferred
method)

1A.17 DESCRIPTIvE AND ECOLOGICAL STUDIES

Design, applications, strengths and weaknesses of descriptive studies and ecological studies
Descriptive studies describe patterns of disease with regard to time, person and place. They include case reports
(a description of a one-off unusual finding) and case series (a description of several unusual findings that are
linked in some way).
Ecological studies are characterised by the unit of observation being a group (e.g. a population or community)
rather than an individual.
Both types of study can use routinely collected data, and as a result they are relatively cheap and rapid to conduct.
Both are useful for the formulation of research questions. See Table 1A.17.1.

Table 1A.17.1 Features of descriptive studies

Type Descriptive studies (case reports/case Ecological studies
series)
Design Astute clinician notices unusual Describe disease pattern for an entire population with
occurrence regard to another parameter
Uses group level data: aggregate (summaries of
individual data), environmental (measurable at
individual level but easier at group level) and global
(attributes of a place not applicable at individual
level, e.g. district, legislation)
Correlation coefficient (r) can range from –1 to +1
Application Case series are particularly useful in International comparisons
identifying the beginning of an epidemic
Study of group level effects, e.g. legislation
Hypothesis formulation
Hypothesis formulation
Strengths Rapid reporting Rapid results
Low cost Low cost

34

Public [Link] 34 3/6/08 [Link]

 1A Epidem iology

Type Descriptive studies (case reports/case Ecological studies

series)
Weaknesses Case report cannot be used to test valid Unable to control for unknown confounders
statistical association (may just be a
Only considers average exposure so would be unable
coincidence)
to detect a J-shaped curve
Case series difficult to test as no
Spatial autocorrelation – analysis assumes that all
appropriate comparison group
areas are independent but may not be
Leakage of exposures through migration
No information on individuals
Risk of ecological fallacy (see below)

ECOLOGICAL FALLACY
An ecological fallacy is an error of logic that occurs when inferences are made regarding individuals, based on
aggregate data from the population to which the individuals belong. See Box 1A.17.1.

Box 1A.17.1
The ‘original’ ecological fallacy
The term ecological fallacy was first used in relation to an analysis of the 1930 US census. This found that the
higher the proportion of immigrants in a state, the higher its average literacy. Analysis at the individual level
showed, however, that immigrants were less literate than native citizens but that they tended to settle in states
where the native population was more literate. It would have been an ecological fallacy to conclude from the
original analysis that immigrants were more literate people.

1A.18 SMALL AREA ANALYSIS

Analysis of health and disease in small areas
The prevalence rates for a particular disease at a small area level may be markedly different from the prevalence
rates at the regional or national levels. Analysing data at these larger geographical areas would therefore mask
pockets of high risk.
Similarly, it may not be valid to extrapolate health survey data (such as the Health Survey for England) down to a
local level. There is, therefore, a case for high-quality research (using valid survey methods and instruments) to be
commissioned at a local level, although the cost implications may be prohibitive.
See Section 3A for details of small area boundaries in the UK.

1A.19 STUDY DESIGN

Design applications, strengths and weaknesses of cross-sectional, analytical studies and intervention studies (including
randomised controlled trials)
See Section 1B.9.
Study design is key to epidemiological research. An epidemiological question can often be answered using several
designs: practical issues (cost, timescale, etc.) then determine which should be used.
35

Public [Link] 35 3/6/08 [Link]

 1A E p ide m io l o gy

For all analytical studies, the first step is to state the research hypothesis (for the statistical testing of
hypotheses). This involves stating the exposure(s) of interest, the outcome(s) of interest and any possible
confounders.
The main types of study design are:
• Cross-sectional
• Case–control
• Cohort
• Intervention.

CROSS-SECTIONAL STUDIES
In this type of study, all of the variables (exposures and outcomes) are measured at the same time. Cross-sectional
studies can be descriptive, analytical or ecological. See Box 1A.19.1.

Box 1A.19.1
Descriptive Description of the point prevalence of a disease in a population linked to health service usage
data
Analytical Comparison of several exposures with the outcome of interest
Ecological* Neither exposure nor outcome is measured at the individual level
*Note that ecological studies need not necessarily be cross-sectional; they can also be longitudinal.

See Table 1A.19.1 for the features of cross-sectional studies. An example of a cross-sectional ecological study is one
that compared HIV prevalence and rates of male circumcision in several African countries. The study found that HIV
prevalence was lower in countries where male circumcision was commoner. Interpretation of this study needs to take
account of ecological fallacy (see Section 1A.17) and the potential for confounding (e.g. religious practice may be
associated with higher circumcision rates and with lower sexual promiscuity).

Table 1A.19.1 Features of cross-sectional studies

Design Determine the simultaneous prevalence of exposure and disease
Measure prevalence of various diseases, characteristics and health-care usage and make
comparisons
Sampling Sampling needs to be representative of the population under study. A random sample is preferable
and the sample needs to be sufficiently large
Application Hypothesis formulation
If the exposure is fixed (e.g. variables present at birth), then a cross-sectional survey is analytical
and can be used to test hypotheses
Analysis Disease frequency: odds or prevalence
Measure of effect: prevalence ratio, prevalence difference, odds ratio
Strengths Can study multiple exposures and outcomes
Rapid and cheap
Useful for rare diseases
Weaknesses Because they are measuring prevalence, not incidence, findings cannot differentiate the
determinants of aetiology and survival

36

Public [Link] 36 3/6/08 [Link]

 1A Epidem iology

CASE–CONTROL STUDIES
In this type of study, individuals with the outcome of interest (cases) are matched with individuals who do not have
the outcome of interest (controls). See Table 1A.19.2. For example, in the study of sudden infant death syndrome
(SIDS), the characteristics (e.g. sleeping position, type of mattress) of a group of children who had died from SIDS
were compared with a matched group of children who had not.

Table 1A.19.2 Features of case–control studies

Design Hypothesis: state exposure(s) and outcome
Choose subjects who do and do not have the disease of interest and compare with respect to the
exposure of interest
Selection of controls is the most critical issue (see Section 1A.7)
Ideal ratio of cases to controls is 1:1 but, if cases are limited, then can increase number of
controls disproportionately up to 1:4 (beyond this there is little gain in power)
Sampling Selection of cases should be representative of all cases in the population. Controls should be
selected from the same population as the controls and may be matched for certain characteristics
not being tested. In all cases, the selection of cases and controls must be independent of their
exposure to the putative risk factors of interest
Application Case–control studies can be retrospective (all cases of disease have been diagnosed when study
starts) or prospective (new cases identified during lifetime of the study)
Analysis First compare cases and controls for other baseline differences
Disease frequency: not possible to determine from this design; cases chosen precisely because
they have disease
Measure of effect: odds ratio
Strengths Rapid and cheap
Ideal for rare diseases/outcomes
Useful for diseases with long latent periods
Can simultaneously examine a large number of potential exposures
Weaknesses Selection bias (exposure and disease have already occurred)
Temporal relationships may be difficult to establish
Recall bias (of information on exposure and disease)
Poor for rare exposures (unless high attributable risk percentage)
Cannot compare incidence rates (unless population based)
Misclassification of exposure/disease status (random ➔ null; non-random ➔ any direction)
Temptation of data fishing: if multiple hypotheses are tested, then on average 1 in 20 will be
significant at the p <0.05 level just because of chance (see Bonferroni correction, Section 1B.11)

37

Public [Link] 37 3/6/08 [Link]

 1A E p ide m io l o gy

COHORT STUDIES
In a cohort study, a group of individuals is selected who do not initially have the outcome of interest. A range
of exposures is quantified for cohort members and at the end of the study those people who have developed
the outcome of interest are compared (according to the exposure of interest) with those who have not. In
retrospective cohort studies, both the beginning and end of the time period of interest are in the past. For example,
the Framingham cohort studies followed residents of a town in Massachusetts. The outcomes of interest were
cardiovascular endpoints (e.g. myocardial infarction) and the exposures of interest included serum cholesterol and
blood pressure. See Table 1A.19.3 for the features of cohort studies.

Table 1A.19.3 Features of cohort studies

Design Hypothesis: state exposure and outcome(s)
Choose subjects on basis of exposure (all subjects should be disease free at the start)
Follow cohort to study temporal relationships, so consider ease of follow-up when selecting
cohort (e.g. alumni)
Sampling Cohort members should be representative of the population of interest. If a workplace cohort is
chosen, then cohort members are likely to be healthier than the general population (the ‘healthy
worker effect’, see Section 1A.14)
Application Able to measure incidence directly in both groups
Analysis Disease frequency: rate, risk, odds, mean or median
Measure of effect: rate/risk/odds ratio (relative), rate/risk/odds difference (absolute). Other –
vaccine efficacy, difference in mean/median
Must also compare groups to ensure similarity of potential confounders
Can estimate effect of loss to follow-up by comparing the two extreme situations (i.e. all those
lost to follow-up did develop the disease, and then that none of them did)
Strengths Able to follow temporal relationships
Well suited to rare exposures
Multiple effects of a single exposure
Minimise selection bias (prospective cohort studies)
Useful for diseases with long latency periods (retrospective cohort studies)
Weaknesses Time-consuming
Expensive
Risk of loss to follow-up (➔ poor validity [power and bias], especially if loss to follow-up >30%
or disproportionate between two groups)
Inefficient for rare diseases
Records may be inadequate for ascertainment (retrospective cohort studies)
Healthy worker effect (see Section 1A.14)

38

Public [Link] 38 3/6/08 [Link]

 1A Epidem iology

INTERVENTION STUDIES (INCLUDING RANDOMISED CONTROLLED TRIALS)

In an intervention study, the epidemiologist is able to allocate the exposure of interest across the study population.
See Table 1A.19.4. For example, in the 4S Study (1994), a group of Scandinavian volunteers was randomised to
receive either 20 mg of simvastatin daily or placebo. The intervention and control groups were compared according
to numbers of deaths and numbers of major cardiovascular events.

Table 1A.19.4 Features of intervention studies

Design These are epidemiological experiments in which the investigators allocate the exposure. The
intervention is generally allocated at random (see Section 1A.26 for methods), but an alternative
is to use systematic allocation
Factorial design
This allows simultaneous testing of two or more hypotheses at marginal increase in cost
Assign to A or B and independently to X or Y
There should be no known interaction between the two hypotheses, nor any effect on participant
eligibility, loss to follow-up or compliance
If there is actually an interaction then:
• Estimate of the treatment effects will tend towards the null
• Such interaction can be detected during analysis
Stopping rules
An independent group must monitor interim results to ensure the welfare of participants
If there is extreme benefit or extreme harm the trial should be stopped early, but only if this is
definitely not a temporary, random fluctuation in results
Procedures are also needed for immediate unblinding in the event of an isolated serious
complication
Extremely high significance (extremely low p value) is needed to justify early termination of trial
It is controversial whether different stringencies in significance are needed to stop a trial early
because of beneficial and harmful effects
Sample size
Must be considered during planning stage so that the sample size has sufficient power to detect
differences between the groups
Power depends on:
• Sample size
• Total number of endpoints
• Difference in compliance between two groups
• Increased number of endpoints by selecting a high-risk population (notes ‘healthy
volunteer effect’ so number of endpoints will be lower than expected) or increased duration
of follow-up
Remember possibility of secular changes, i.e. whole population becoming healthier over time so
fewer endpoints
Table contd overleaf

39

Public [Link] 39 3/6/08 [Link]

 1A E p ide m io l o gy

Table 1A.19.4 contd

Design Compliance
(contd)
Non-compliance ➔ Two groups more similar ➔ Result tends towards the null ➔ Trial less able to
detect a true effect
Can improve compliance by using run-in period pre-randomisation during which all potential
participants receive treatment/placebo to determine acceptability to patient
Sampling Study population needs to be representative of the reference population (i.e. the population to
which the results of the trial are to be extrapolated)
Application Can investigate therapeutic or preventive exposures, at the level of either the individual or the
population (e.g. fluoridation)
Analysis Disease frequency: rate, risk, odds, mean or median
Measure of effect: rate/risk/odds ratio (relative), rate/risk/odds difference (absolute). Other –
vaccine efficacy, difference in mean/median
Start with a table comparing the two groups for baseline characteristics
Consider the placebo effect
Analyse by intention to treat (see Section 1A.20)
Strengths Evidence provided can be of extremely high quality
If sample large enough, then validity can be guaranteed
Blinding minimises observation bias (where blinding impossible then independent examiners
should assess endpoints)
Weaknesses Generalisability:
Is the study population similar enough to the reference population for the results to be
applicable? If inclusion/exclusion criteria too strict, then it may not be
Is the intervention comparable to the treatment that would be received outside a trial (with
regard to clinicians’ interest and attentiveness)? Efficacy is the effect under ideal conditions;
effectiveness is the real-life effect
There must be sufficient doubt (equipoise) about the alternatives
Study will be not be feasible if the treatment under investigation is already too widely accepted
Study will be unethical if existing evidence suggests harm from giving or withholding the
treatment
Potential for enormous cost
Potential biases from:
• Loss to follow-up of many subjects (more likely with long follow-up periods)
• Unequal follow-up (accuracy or completeness) between two groups
• Observation bias (unlikely with mortality but more likely with cause of death – minimised
by double-blinding)
• Placebo effect (response to any therapy regardless of physiological effect: true effect of
intervention is Percentage effect in treatment group – Percentage effect due to placebo)

40

Public [Link] 40 3/6/08 [Link]

 1A Epidem iology

1A.20 INTENTION-TO-TREAT ANALYSIS

While it may be intuitive to compare the groups in an intervention study according to the actual interventions
that they received, such analysis-by-treatment-received is fundamentally flawed for the three reasons given
below. Instead the analysis should always be performed based on the a priori group to which each participant
was originally assigned. This is known as intention-to-treat analysis and it ignores the reality of whether the
participant actually received the intervention or placebo assigned to that group.

ADVANTAGES
The advantages of intention-to-treat analysis over analysis by treatment received are:
• The actual research question being asked concerns the effect of offering this treatment to patients in the real
world, as opposed to the actual physiological effect of their receiving it.
• The balance of unknown confounders achieved at randomisation is impossible to recreate once non-compliers
are removed from the two groups (hence the maxim, ‘once randomised, always analysed’).
• Those who switched from their allotted groups may differ systematically from those who remained (e.g. patients
who were too sick to cope with side effects may have tended to transfer to the placebo group).

1A.21 CLUSTERED DATA

Clustered data – effect on sample size and approaches to analysis
Clustered data are not fully independent of each other. Rather, they are linked according to a particular
characteristic (e.g. time, place or person). Standard statistical techniques depend on the assumption that
observations are independent of each other. This is not the case when data are clustered: observations in the same
cluster will be more similar to each other than if they were independent. See Box 1A.21.1 for an example of a
clustered study.

CLUSTERED DATA
Common types of clustered data include:
• Multiple observations on the same subject at the same time
• Repeated observations on the same subject over time
• Cluster randomised trials
• Clustered sampling surveys.

Box 1A.21.1
Example: Ten Towns Study
This project, which started in 1990, compares 10 towns in England and Wales. Five of the towns had high levels
of cardiovascular disease and five had low levels. In one study within the project, researchers surveyed school
children to determine, ‘Whether markers of nutrition, cardiovascular health and type 2 diabetes differ between
school pupils who eat school dinners and those whose school day meal is provided from home’.
In order to take account of clustering, the researchers assigned the town of residence as a fixed effect and
assigned the school attended as a random effect (see below). Because the analysis used both fixed and random
effects, it is termed a ‘mixed model’.
Adapted from Ten Towns Heart Health Study: [Link]; Whincup et al (2005).

41

Public [Link] 41 3/6/08 [Link]

 1A E p ide m io l o gy

In each of these circumstances, the effect of clustering must be taken into account at the statistical analysis stage,
otherwise associations that do not truly exist are more likely to be detected.

SAMPLE SIZE
Clustered trials and surveys require an increased sample size to compensate for the tendency of individuals within
each cluster to be more similar to each other than to individuals in other clusters.
The amount by which the sample size needs to be increased is called the design effect and this depends on the
number of individuals per cluster and the intraclass correlation coefficient, i.e. the ratio of intercluster variance to
total variance.

ANALYSIS
At the analysis stage there are several ways of accounting for the effect of clustering:
• Calculate summary statistics for each cluster and then analyse these using standard techniques
• Calculate robust standard errors that account for clustering
• Use more sophisticated techniques (such as generalised estimating equations or random effects models).

RANDOM AND FIxED EFFECTS

These are statistical models that can be employed within ANOVA* or regression analysis of clustered data.
• Random effects models explicitly model the similarities between individuals in the same cluster, i.e. a random
sample of all possible values of that variable
• In contrast, a fixed effects model refers to assumptions about the independent variable (e.g. that it can only
take the values high, medium or low). With a fixed effects model, the conclusions will be restricted to these
three values.
*ANOVA (analysis of variance) is a collection of statistical procedures that compares means by dividing the overall
observed variance into different parts

1A.22 NUMBERS NEEDED TO TREAT

Numbers needed to treat – calculation, interpretation, advantages and disadvantages
The numbers needed to treat (NNT) is a useful statistic that indicates how many patients would need to be treated
with a particular intervention to reduce by one the expected number of a defined outcome. It is the reciprocal of
the absolute risk reduction, and is particularly useful since it takes into account the frequency of the outcome – and
thus reflects the public health impact of the intervention (e.g. impact of offering low-dose aspirin in preventing
cardiovascular disease).
Analogous measures can be used to compare screening programmes* and harmful exposures (see Box 1A.22.1)†.

Box 1A.22.1
Calculation 1
NNT =
Absolute risk reduction
Interpretation The NNT answers the question, ‘How many patients would I need to treat in order for
one extra patient to benefit?’
*An analogous measure in screening studies called the number needed to screen
†
If the treatment or exposure is harmful (i.e. the result is a negative number), then
the minus sign is omitted and the measure is renamed the number needed to harm

42

Public [Link] 42 3/6/08 [Link]

 1A Epidem iology

Advantages More intuitive expression of absolute risk (itself expressed as a percentage)

Disadvantages It is tempting to compare NNTs for different therapies but this is justified only when
the baseline risks of the disease are similar. A drug that reduced mortality rate from
90% to 80% has the same NNT as one that reduces mortality rate from 40% to 30%
but the former would have greater impact
Caution is needed when calculating the NNT from meta-analysis

See Box 1A.22.2 for an example of how NNT is calculated.

Box 1A.22.2
Example: The 4S Study
In the 4S Study, during a 5.4-year period, 256 of 2223 patients
(11.5%) in the placebo group died, compared with 182 of 2221
patients (8.2%) in the group who were treated with simvastatin.
Absolute risk reduction = 0.115 – 0.082 = 0.033
1
NNT = = 30 patients
0.033

1A.23 TIME TREND ANALYSIS, TIME SERIES DESIGNS

Time series analysis considers the situation where both the exposure and the outcome are measured over time.
If there are secular increases or decreases (or seasonal variations) in both variables, then the variables will be
correlated with time. This is known as serial correlation, and time will therefore be a confounding factor in the
analysis. For example, both rainfall and hospital admissions are higher in winter. In answering the question whether
high rainfall is associated with increased hospital admissions, this seasonal correlation would first need to be
accounted for.
Difficulties in attributing effect to the exposure can be due to the causes shown in Box 1A.23.1.

TIME SERIES DESIGNS

Figure 1A.23.1 shows four different time series designs. In each of the four graphs, the horizontal axis represents
time and the vertical axis represents the magnitude of the effect observed.

Box 1A.23.1
Secular changes For example, changes in age structure, classifications of disease, diagnostic
techniques
Concurrent interventions/ For example, a decrease in deaths from heart disease might be observed when
exposures smoking prevalence falls and more effective therapies (e.g. statins) are used
Latency period Long time for exposure to manifest its effect (e.g. smoking and cancer)
Diffuse exposure Exposure spread out over months or years

ANALYSIS TO ESTIMATE EFFECT SIZE

1. Smooth out ‘noise’ (peaks and troughs) in curves using moving averages
2. Segmented regression – analysis of each separate part of an interrupted time series.

43

Public [Link] 43 3/6/08 [Link]

 1A E p ide m io l o gy

1. Simple before/after 2. Repeated measures before/after

Simple and quick, but does not describe the Simple, can see overall trend but takes longer,
overall trend, as illustrated by dotted lines, each needs more data points
representing a possible scenario

Intervention Intervention

3. Cross-over designs: (baseline – exposure – 4. Time series at different locations

baseline)
Different locations should cancel the effect of
If concurrent exposures cause the effect, then concurrent interventions but different locations
removal should reverse it but it is not always can be expensive
possible to remove exposure

Intervention Remove Intervention

Figure 1A.23.1 Time series designs

1A.24 NESTED CASE–CONTROL STUDIES

Sometimes a case–control study is enclosed (‘nested’) within a cohort study. In this instance, both the cases and the
controls are taken from within the population of a cohort study. Nested case–control studies can be useful when it
would be too expensive or otherwise unfeasible to perform laboratory tests on the entire cohort.

NESTED STUDY DESIGN

A nested study might be conducted as follows:

44

Public [Link] 44 3/6/08 [Link]

 1A Epidem iology

1. Take baseline samples for all entrants into cohort but do not analyse (e.g. blood samples are frozen but not
tested)
2. Run cohort study
3. Identify cases as members of the cohort who develop the disease
4. Choose controls as matched disease-free members of the cohort
5. Test frozen blood samples of cases and controls (and discard samples from the rest of the cohort).
Advantages and disadvantages of nested studies are shown in Box 1A.24.1.

Box 1A.24.1
Advantages Disadvantages
Cost-effective Non-diseased cohort members (from whom the controls are
selected) may not be fully representative of the original
Avoid recall bias by using data collected before
cohort, due to death or losses to follow-up
the onset of disease
Avoid selection bias by drawing cases and controls
from the same cohort

1A.25 METHODS OF SAMPLING FROM A POPULATION

Since investigating the entire population (by a census) is costly, sampling is a more cost-effective and convenient
means of collecting information. Sampling methods produce either probability samples (which allow calculation of
sampling error and hence allow inferences to be made regarding the population) or non-probability samples (which
do not permit such inferences).

PROBABILITY SAMPLING
A sampling frame is required for all studies that aim to make an inference about the population (i.e. a complete list
of the population from which the sample is to be drawn). Different methods of probability sampling are shown in
Table 1A.25.1.

Table 1A.25.1 Methods of probability sampling

Type Method Advantages Disadvantages
Random Start with a sampling frame Purest form of probability Relatively inconvenient in
sampling practice
Sampling frames include postcode
address files, electoral register, GP Allows calculation of Inefficient for rare
practice list sampling error outcomes
Draw random sample from sampling Sample frame not always
frame easily available
Systematic Start with sampling frame More convenient than Potential for bias if there
random sampling are underlying patterns to
Calculate sampling interval (n):
the sampling frame
(number in population) Allows calculation of
n= sampling error
(number in sample)
Draw every nth person from the sampling
frame
Table contd overleaf

45

Public [Link] 45 3/6/08 [Link]

 1A E p ide m io l o gy

Table 1A.25.1 contd

Type Method Advantages Disadvantages
Stratified Assign members of the population into Improves accuracy of Requires accurate
relatively homogeneous subgroups estimation information about the
(‘strata’) before sampling population
Efficient
Draw random sample of subjects from Choice of relevant
Allows calculation of
within each stratum stratification variables can
sampling error
be difficult
Unhelpful if there are no
homogeneous subgroups
Cluster Used when there are ‘natural’ clusters in Convenient for fieldwork Increased sampling error
the population, e.g. GP practices within
Cost-efficient
a borough
Allows calculation of
A random sampling technique is used to
sampling error
choose which clusters to include in the
study
In single-stage cluster sampling, all
the elements from each of the selected
clusters are used (e.g. all patients in
selected practices). In two-stage cluster
sampling, elements from each of the
selected clusters are selected at random
(sample of patients from within the
selected practices)

NON-PROBABILITY SAMPLING
In general, for non-probability sampling, the sampling frame is not known. Different methods of non-probability
sampling are shown in Table 1A.25.2.

Table 1A.25.2 Methods of non-probability sampling

Type Method Advantages Disadvantages
Convenience Subjects chosen on basis of being Useful for preliminary Sampling error cannot be
readily available research as it is extremely calculated
efficient
Volunteer bias
Judgement Subjects chosen purposively on Useful for rare characteristics Sampling error cannot be
the basis of having particular calculated
Useful for qualitative research
features, such as controls required
Volunteer bias
for epidemiological studies

46

Public [Link] 46 3/6/08 [Link]

 1A Epidem iology

Type Method Advantages Disadvantages

Quota Select demographic characteristics Representative with regard to Potentially unrepresentative
of interest (e.g. age, sex, known characteristics in terms of other
ethnicity) characteristics
Select subjects to represent Sampling error cannot be
the proportional distribution of calculated
these characteristics within the
Volunteer bias
population
Snowball Ask subjects to recommend Very cost-efficient Sampling error cannot be
acquaintances who meet the calculated
Useful where no sample frame
sample criteria
exists Volunteer bias
Enables researcher to reach
groups that are otherwise
hard to reach

CHOOSING CONTROLS FOR EPIDEMIOLOGICAL STUDIES

The selection of controls is often the most critical issue for an epidemiological study, particularly in case–control
studies which are very vulnerable to selection bias. The choice of this comparison group will vary depending on the
situation. Two ways of choosing controls are shown in Table 1A.25.3. Note that controls are the people who would
have been identified as cases had they developed the disease: they are not the entire non-diseased population.

Table 1A.25.3 Choice of controls in case–control studies

Type of control Sampling strategy Advantages Disadvantages
Controls within Purposive: through registers, Easily identified Ill, therefore different from
the health-care hospital records, referral data healthy population (e.g. more
By being in hospital are (like
system likely to smoke, use oral
the diseased cases) more
contraceptive pill and drink
aware of antecedent events
heavily)
(i.e. more similar to diseased
cases, therefore less recall A given hospital may have
bias) secondary and tertiary
activities, so patient selection
Same hospital, therefore same
differs
influences on choosing that
hospital Different specialties within
the same hospital may have
More cooperative than healthy
different catchment areas
people
Population Random sampling Healthy population General public does not recall
controls antecedent events as well as
or
those who are ill
Purposive: friends/relatives/
Less motivated to participate
neighbours of cases (may be
more motivated to participate) Healthier people less likely to
be at home during the daytime
Costly and time-consuming

47

Public [Link] 47 3/6/08 [Link]

 1A E p ide m io l o gy

1A.26 METHODS OF ALLOCATION IN INTERvENTION STUDIES

Interventional studies differ from observational studies in that the epidemiologist allocates the exposures of
interest. The process of allocation is therefore of pivotal importance. Ideally, subjects will be allocated in such
a way that the two groups are identical, meaning that the only difference between the two groups will be the
intervention under investigation.
So as to minimise allocation bias, subjects should be allocated to their groups only after ensuring that they are
eligible, have consented to participate and have been fully enrolled.
Allocation may be systematic, volunteered or at random.

SYSTEMATIC ALLOCATION
Allocation is determined in advance (e.g. it is decided that recruits will be allocated alternately into groups, or that
all recruits presenting on alternate days will be allocated to the same alternate group). There is much potential for
selection bias because:
• The order is predictable so the recruiter may interfere
• There may be underlying patterns to the order in which recruits present.

VOLUNTEER ALLOCATION
Allocation is determined on the basis of which recruits volunteer to participate. This is highly unsatisfactory
because of extreme selection bias.

RANDOMISATION
The unique advantage of random allocation is that if the groups are large enough then, on average, they will be
similar with respect to all variables. This includes all confounders – both known and unknown.
If randomisation is strict (computer-generated or random table) and concealed, then selection bias is avoided.
Note the difference between allocation concealment and blinding/masking: it is not always possible to mask the
intervention from participants or assessors (e.g. surgery versus drugs), but it should always be possible for the
allocation method to be concealed so that the person recruiting a patient does not know to which arm of the study
they will be allocated.
There are several types of randomisation that may be used for allocating recruits. See Table 1A.26.1.

Table 1A.26.1 Types of randomisation

Type of Description
randomisation
Unrestricted Participants allocated purely on the basis of chance
Potential for unequal group sizes – insignificant in large studies
Blocked This is a technique for obtaining equal group sizes. First the ratio of participants between
allocation groups is set (e.g. 1:1). A block size is then chosen (e.g. 4). All permutations of
assignments that meet the ratio are then listed for that block size (with these examples the
blocks would be AABB, BBAA, ABAB, BABA). Finally, the allocation sequence is generated
by random selection of blocks from this list. The block size can be varied periodically to
make prediction of the allocation sequence even more difficult

48

Public [Link] 48 3/6/08 [Link]

 1A Epidem iology

Type of Description
randomisation
Stratified Recruits are subdivided into strata (e.g. according to ethnicity or gender) and individuals
within each stratum are randomised
Cluster Groups (rather than individuals) are randomised to receive the different interventions. This
needs to be taken into consideration at the analysis stage (see Section 1A.21)
Matched pair Individuals or groups are first matched according to baseline data – matching them on as
many variables as possible. The intervention is then randomly allocated to one member of
the pair, with the other member of the pair receiving the control
Stepped wedge The population is divided into groups and then the intervention is progressively introduced,
in random order, across the groups until every group is receiving it. Used when other
allocation methods would be unfeasible because of widespread belief that the intervention
is beneficial

1A.27 RECORDING SURvEY DATA

Design of documentation for recording survey data
Note that survey/questionnaire data are quantitative. Either the participant or the researcher can record survey
data. In both cases, the design of the survey documentation should ensure that the data recorded is:
• Complete
• Accurate
• Understandable and legible
• Stored in a format suitable for analysis
• Secure to protect respondents’ confidentiality.

SURVEY DOCUMENTATION
Documentation of data can be:
• Hard copy: questionnaires for postal surveys for participants to complete or questionnaires plus prompts for
researchers to complete during face-to-face or telephone surveys
• Online: allows documentation to be tailored to responses, can provide prompts when invalid/incomplete/no
responses received when required.
Those recording survey data need to know:
• Where to record information
• Type of answer required
• Directions for completing the survey (if items are not relevant, respondents need to know which question to
answer next, e.g. ‘NO ➔ move to question 5’).
Researchers recording survey data from participants also need prompts/clarifications with information on:
• Screening: to ensure that appropriate participant is interviewed
• Directions if there is no response to a particular item.
Documentation layout can help by:
• Appropriately sized spaces to indicate where responses are needed and in what form
• Providing clear navigation to guide participants through the survey.

49

Public [Link] 49 3/6/08 [Link]

 1A E p ide m io l o gy

1A.28 CONSTRUCTION OF vALID QUESTIONNAIRES

validity is the extent to which a tool explores what it is purported to measure (i.e. provides the ‘true’ answer and
minimises bias). A questionnaire’s validity will depend on the range of factors listed in Table 1A.28.1.

Table 1A.28.1 Characteristics of valid questionnaires

Characteristic Implications Technique to maximise validity
Sample selected Unrepresentative sample may lead to bias Obtain a sampling frame (exhaustive list of
all possible experimental units)
Random selection of sample
Stratification of sample
Response rate achieved Uneven or very low response rate is Advance warning letters
potentially subject to bias
Incentives for responding
Data method: face to face or telephone
better than postal (though more costly)
Content of Questions should be chosen carefully to Researching content through:
questionnaire ensure that the questionnaire addresses • Existing tools
(content validity and the research question • Literature review
construct validity)* • Qualitative interviews
For psychometric questionnaires:
• Expert opinion
conceptual model needed first to define
dimensions of the construct to be
measured
Quality of the Participants need to understand what the Generate clear and non-judgemental
questions asked questions asked require and be prepared to questions: pilot and test on potential
give this information honestly participants and revise.
Data collection to suit the subject: face
to face better to explain questions; postal
may be better for privacy
*Note the difference between psychometric questionnaires (analysed by a summative score) and survey tools (analysed
by individual question).

Other ways to maximise validity are to:

• Use an existing tool if an appropriate one exists
• Maximise reliability by increasing the sample size
• Triangulate with other sources of evidence (e.g. observation).

1A.29 vALIDATING OBSERvATIONAL TECHNIQUES

Methods for validating observational techniques
Observational studies can be improved by using the techniques described below.

OBSERVATIONAL TECHNIQUES
• Participant observation, e.g. nurse studying her own patients

50

Public [Link] 50 3/6/08 [Link]

 1A Epidem iology

• Non-participant observation, e.g. researcher visiting a ward to observe effects of nurse care on patients (see
also Section 1D).

ENHANCING VALIDITY
• Sampling strategy: appropriate selection of participants and settings for making observations
• Considering reflexivity: the observers’ effect on what they are observing
• Enhancing reliability: recording observations using observers trained comprehensively and systematically, using
a checklist to record observations.

VALIDATE OBSERVATIONAL TECHNIQUES THROUGH CROSS CHECKING

• Repeated observations by the same, or other, observers in the same setting
• Recorded observations (e.g. videotape)
• Triangulation with other research methods.

1A.30 STUDIES OF DISEASE PROGNOSIS

The prognosis is the expected course and outcome of a disease. Studies of prognosis estimate the frequency with
which different outcomes can be expected to occur. Prognostic factors are patient characteristics that guide the
prediction of outcomes. Cohort studies are used for analysing prognostic factors (e.g. the Framingham study*
investigated cardiovascular prognostic factors). Disease registers (e.g. regional cancer skin cancer register) can also
be used to analyse prognosis.
*See the Framingham Heart Study website [Link].

SURVIVAL ANALYSIS
Ideally, all events would be recorded for those patients in the cohort or on the register. In practice the information
available may only be that no event had occurred by a certain date. This is known as censoring (discussed on page
8, Section 1A.3 in more detail) and it includes the components shown in Box 1A.30.1.

Box 1A.30.1
Fixed censoring End of study
Loss to follow-up Moving away or dropping out
Competing event Death by other cause

Censoring should be assumed to be non-informative, i.e. that the distribution of event times is identical
between censored and non-censored subjects. Also, it should be assumed that the survival probability is the same
for those recruited early or late into the study.
Fixed censoring does tend to be non-informative, but loss to follow-up may be informative (e.g. patients may
withdraw due to harmful effects or from being too ill). Informative censoring can bias results.
Survival analysis is generally analysed by survival models:
• Time to death/time to recurrence, etc
• Display by Kaplan–Meier chart
• Log rank test
• Cox’s proportional hazards model for multivariate analysis.

51

Public [Link] 51 3/6/08 [Link]

 1A E p ide m io l o gy

Its uses include:

• Calculating life-expectancy
• Constructing and comparing survival curves
• Comparing survival between groups.

1A.31 STATISTICAL ANALYSIS OF EPIDEMIOLOGICAL STUDIES

Appropriate use of statistical methods in the analysis and interpretation of epidemiological studies, including life-table
analysis

CHOICE OF STATISTICAL METHOD

• What is the hypothesis being tested?
• Are the data independent? Or are they matched, clustered or correlated?
• What are the input variables (e.g. intervention or control)?
• What are the output variables (e.g. measure of morbidity)?

HYPOTHESIS TESTING
The null hypothesis (Ho) states that there is no association between exposure and disease.
Assuming that the null hypothesis is true, the p value is the possibility of obtaining a result at least as extreme as
that observed due to chance alone. Usually a two-sided p value is used (i.e. probability of observing such a large
difference, without specifying the direction of the difference). A one-sided p value is used to increase the precision
of an estimate where there is a strong prior belief. For example, when comparing radical mastectomy against
lumpectomy there is a strong prior belief that the more radical procedure will be at least as curative as the limited
procedure and the question is simply whether lumpectomy is as effective as mastectomy.

CHOICE OF OUTCOME MEASURE

The choice of statistical method and outcome measure will depend on the nature of the study. See Table 1A.31.1

CHOICE OF STATISTICAL TEST

See Section 1B.

1A.32 EPIDEMIC THEORY

Effective and basic reproduction numbers, epidemic thresholds
Techniques for analysing infectious disease data include:
• Construction and use of epidemic curves, generation of numbers
• Exceptional reporting
• Identification of significant clusters.

EPIDEMIC THEORY
An epidemic is the occurrence of a number of cases of a disease that exceeds the number of cases normally expected
for that disease in that area at that time.
Compartmental epidemic models allocate individuals to specific subgroups. See Box 1A.32.1.

52

Public [Link] 52 3/6/08 [Link]

 1A Epidem iology

Table 1A.31.1 Outcome measures

Study type Outcome measure
Correlation study Correlation coefficient (r) See Section 1B.14 (linear regression and correlation
coefficients)
Case–control study Odds ratio Odds (cases)
Odds ratio =
Odds (controls)
Where:
Proportion exposed
Odds =
Proportion unexposed
Confidence intervals can be constructed using the natural
log of standard error of the odds ratio (lnOR). Note that
the confidence interval is not symmetrical around the odds
ratio. If the confidence interval includes 1, then there is no
significant difference between the groups
Cohort study Risk ratio Risk (exposed)
Risk ratio =
Risk (unexposed)
Where:
Number of cases
Risk =
Number at risk to start
Rate ratio Rate (exposed)
Rate ratio =
Rate (unexposed)
Where:
Number of cases
Rate =
Number of person-years
SMR Observed deaths
SMR =
Expected deaths
Intervention study Risk ratio, rate ratio,
attributable risk or
population attributable risk
Life-table analysis Survival probability Cumulative chance of death in a given time period
(see Sections 1B.15
Confidence intervals 95% confidence intervals calculated using a formula by
and 1B.16)
Kalbfleisch and Prentice
Life-expectancy
Time after which half of a given cohort have died
Proportional hazards
= 0.5 + S (Length of survival ¥ Cumulative chance of
survival)
A statistical method for comparing survival rates in different
groups. If there is no difference between groups, then the
ratio of hazards between the groups is constant over time
(even though the underlying hazards will change) and the
logged cumulative hazard curves will be parallel

53

Public [Link] 53 3/6/08 [Link]

 1A E p ide m io l o gy

Box 1A.32.1
Passively immune Susceptible Exposed Infective Recovered
M S E I R

Over time, individuals will move from one

compartment to another, as illustrated by
Births Births
the diagram in Figure 1A.32.1 in which
compartments are represented by boxes
and the transitions between compartments Incidence
are shown as arrows.
M S E I R
Different models are based on the flow
patterns between these compartments,
e.g. MSEIR, MSEIRS, SEIR or SIR. Many
diseases (e.g. measles) can be modelled Deaths Deaths Deaths Deaths Deaths
accurately using SIR. The number of
individuals in any compartment will vary
over time, and hence is expressed as a Figure 1A.32.1 A compartmental epidemic model of a disease.
function of time. For example, the number Reproduced from Hethcote (2000) with permission from the Society for
of susceptible individuals at time t is Industrial and Applied Mathematics
expressed as S(t). The different rates of
transition are applied to the relevant
arrows in the model.
The basic reproduction number (R0) is defined as the average number of secondary infections produced when one
infected individual is introduced into a host population where everyone is susceptible.
For many epidemiology models, an infection will take hold in a fully susceptible population only if R0 > 1. Thus, the
basic reproduction number R0 is often considered as the threshold quantity that determines when an infection can
invade and persist in a new host population.
The epidemic threshold is a critical value for the fraction of the population who are susceptible. Below this value
an epidemic outbreak will not occur. The epidemic threshold is determined by the attack rate during the major
outbreaks.
Epidemic curves are graphic representations of the number of new cases by date of onset. They are used to display
an outbreak’s magnitude and time trend. Features of the curve are listed in Table 1A.32.1.
Uses of an epidemic curve include:
• Determining location within the course of the epidemic, and possibly projecting its future course
• Estimating the probable time of exposure (if the disease and its usual incubation period are known) so that
the investigation can be focused on that specific time period
• Identifying outliers (all of whom are worthy of investigation since their unusual exposure may give clues of the
source)
• Inferences about the epidemic pattern can be drawn, e.g. whether it is an outbreak resulting from a common
source exposure, from person-to-person spread, or both.
Assessment of an outbreak by place provides information on the geographical extent of a problem and may
also show clusters or patterns that provide clues to the identity and origins of the problem. See Table 1A.32.2.
A simple and useful technique for looking at geographical patterns is to plot where the affected people live or
work or may have been exposed onto a ‘spot map’ of the area.

54

Public [Link] 54 3/6/08 [Link]

 1A Epidem iology

Table 1A.32.1 Features of an epidemic curve

Feature on epidemic curve Phenomenon or phenomena
Sudden rise in number of cases Common source
Steep rise then gradual fall Single source (or ‘point source’) epidemic,
with all cases occurring within a single
incubation period
Plateau (rather than peak) Prolonged exposure period (‘continuous
common source epidemic’)
Series of progressively taller peaks Person-to-person spread (a ‘propagated’
one incubation period apart epidemic)
Early outlier Background (unrelated) case
Or a source of the epidemic
Or a person who was exposed earlier than
most of the people affected
Late outlier Unrelated to the outbreak
Or a case with a long incubation period
Or exposure was later than for most of
the people affected
Or secondary cases (i.e. the person may
have become ill after being exposed
to someone who was part of the initial
outbreak)
Reproduced from the National Center for Chronic Disease Prevention and Health Promotion (2006).

EXCEPTION REPORTING
Surveillance involves the mechanistic collection and analysis of data, and the communication of results. Early
warning procedures aim to detect any divergence from usual frequency of disease or symptoms as soon as possible.
UK In the UK this work is coordinated by the Health Protection Agency (HPA). See Figure 1A.32.1.

Table 1A.32.2 Geographical patterns in epidemiology

Setting Clusters
Community Water supply
Wind currents
Proximity to a restaurant or shop
Hospital ward Focal source (person-to-person spread)
Scattered (common source, e.g. catering)
Surgical infection Operating room
Recovery room
Ward

55

Public [Link] 55 3/6/08 [Link]

 1A E p ide m io l o gy

Deaths Disease Laboratory reports GP research unit

notifications

Health Protection Agency

Communicable Weekly report/HPA Euro-surveillance WARNER

disease website report*

*Weekly Analysis Report of Notifications above Expected Rates.

Figure 1A.32.1

SIGNIFICANT CLUSTERS
A cluster is a collection of events in space and/or time that is believed to be greater than would be expected by
chance. If the population density varies between suspect areas, then a spot map will be misleading. To correct for
this, the attack rate in each area should be calculated. The attack rate is the proportion of an exposed population
at risk that becomes infected during a defined time period.

1A.33 COMBINING STUDIES

Systematic reviews, methods for combining data from several studies and meta-analyses
The process of combining the results from several studies offers a number of advantages. These are listed in Box
1A.33.1.

Box 1A.33.1
Increased power and A single trial is always preferable to a combination, but often many underpowered
precision studies will be published, all of which show a similar effect size but lack
significance
Greater generalisability Results taken from several studies may be relevant to a wider patient population
Efficiency and cost It is quicker and cheaper to perform a systematic review than to embark on a new
study

SYSTEMATIC REVIEWS
A systematic review is a summary of the medical literature conducted by using explicit methods. The process
involves three steps: see Box 1A.33.2.

56

Public [Link] 56 3/6/08 [Link]

 1A Epidem iology

Box 1A.33.2
Literature search A thorough, systematic search of the published and grey literature (see Section 1A.35)
Critical appraisal Review of the studies found to determine which are relevant and valid
Amalgamation Merger of the valid studies. When this is done systematically it is termed a meta-analysis

Two organisations that undertake systematic reviews are listed in Box 1A.33.3.

Box 1A.33.3
Examples: Systematic reviews
Cochrane Collaboration: collection of systematic reviews of medical and public health interventions. See Section
1A.39
[Link]
Campbell Collaboration: collection of systematic reviews of social and educational policies, including some
health-related outcomes
[Link]

CRITICAL APPRAISAL
The critical appraisal (see Section 6A.2) of a systematic review addresses the same issues as those for individual
studies, namely:
• Specific question stated
• Study population specified
• Unbiased collection of information (search strategy and data extraction)
• Valid conclusions (using appropriate statistical techniques if a meta-analysis).

META-ANALYSES
Combining trials depends on the quality of the constituent studies and on their being similar with regard to
interventions and outputs. If there are enough studies with sufficiently similar characteristics that their results can
be combined, then a meta-analysis is possible. Since the patients in one trial are likely to differ in a systematic
way from those in another, it is wrong to compare individuals. However, since the individual studies are internally
randomised, their effect sizes can be compared.

METHODS
• Fixed effects: assumes statistical homogeneity (i.e. the size of the treatment effect would be the same in all
studies if there were no differences in the samples used)
• Random effects: allows for statistical heterogeneity. Will result in a more conservative estimate of effect than
the fixed-effects model
• Results: summary measure (with confidence intervals produced)
• Findings presented (usually as a Forest plot or a funnel plot).
Figure 1A.33.1 shows a funnel plot of mortality results from trials of post-myocardial infarction b-blockade. As the
sample size increases so the variability in odds ratios narrows.

57

Public [Link] 57 3/6/08 [Link]

 1A E p ide m io l o gy

4000

3000
No of patients

2000

1000

0
0.25 0.5 0.75 1 2 3
Odds ratio

Figure 1A.33.1 Funnel plot of mortality results from trials of post-myocardial infarction b-blockade

Meta-analysis assesses the grand total of participants in all studies regarding observed and expected numbers.
If the null hypothesis is true then: (Observed – Expected) = 0.
If treatment is beneficial then: (Observed – Expected) <0.

BIAS IN META-ANALYSIS
• Poor trial quality, e.g. inadequate concealment may exaggerate treatment effect
• Publication bias: studies that show an effect are more likely to get published than those that do not. Can
be examined using funnel plots displaying the effect size versus the study size – where the funnel would be
expected to be symmetrical if no publication bias existed (see Sectio 1B.18).

RANDOMISED CONTROLLED TRIALS vERSUS OBSERvATIONAL STUDIES

Meta-analysis is often described with reference to randomised controlled trials (RCTs). It is possible to conduct
meta-analyses of observational studies when it is not possible to collect data from RCTs (e.g. intervention already in
common usage) but these are subject to other types of bias and the effect of confounding factors is hard to control.

FOREST PLOTS
The findings of a meta-analysis can be displayed using a Forest plot. See Figure 1A.33.2 for a Forest plot of four
studies.

58

Public [Link] 58 3/6/08 [Link]

 1A Epidem iology

95% confidence Summary statistic

intervals

Ó
Ô
Ì
Ô
Ï
Study 1
Study 2
Study 3
Study 4
Individual study: size µ
Summary weight of each study

Summary estimate: 0.1 0.25 0.5 1 2 4 8 16

width corresponds to
confidence intervals Summary estimate of relative risk

Figure 1A.33.2 Forest plot showing the estimates of relative risk for four studies

1A.34 ELECTRONIC BIBLIOGRAPHICAL DATABASES

Electronic bibliographical databases and their limitations
The extent of the health literature is colossal: there are currently over 30 000 medical journals, and the content of
some of these extends over 100 years. Bibliographic databases make it possible to search these publications using
Boolean operators such as AND, OR and NOT. Some databases (e.g. PubMed) contain journal material only; others
(e.g. Popline) include other forms of literature such as reports, books, etc.
Databases can be searched either using free text words in the title or abstract (‘keyword’), or by using a thesaurus
term that may be grouped hierarchically. The latter are called ‘exploded’ or ‘MeSH’ terms and they have a number of
advantages and disadvantages. See Box 1A.34.1.

Box 1A.34.1
Advantages Automatically include all synonyms for a particular term, including:
• American and British spellings
• Plural and singular
Disadvantages Prolonged time delays between publication and indexing mean that thesaurus terms may not
keep pace with new areas of research

SEARCH STRATEGY
The following steps should be followed when conducting a search of the literature (Eyers 1998):
• Define the research question
• Choose which databases to search (see below)
• Define limits (e.g. time period)
• List the individual terms that constitute the research question
• Choose either keyword or thesaurus search strategy
• If keyword, then list all similar terms, spellings, etc by using a wildcard term denoted by typing an asterisk
(e.g. diabet* will search for diabetes, diabetology, diabetologist and diabetic)
• If thesaurus, then identify thesaurus terms within the hierarchy
• Search for individual terms

59

Public [Link] 59 3/6/08 [Link]

 1A E p ide m io l o gy

• Combine individual terms (use AND to narrow the search and OR to broaden the search)
• Set further limits (e.g. restrict by language, or just review papers, etc)
• Save or print the papers identified
• Scan the titles for additional related terms to include in the search strategy
• Save the refined research strategy for future use.

LIMITATIONS OF DATABASES
No one database has access to all forms of publication, and no one library will have access to all databases. There
is a tendency in public health to concentrate on health databases, but, being a multidisciplinary specialty, other
databases should also be searched (e.g. economics, anthropology, etc.). There is also often a bias towards English
language publications
Databases have a limited span of years: research that pre-dates databases is often ignored. There is often a time
delay between publication and appearance in database. Using the internet can sometimes overcome this. However,
using the internet as a database is risky as there are no quality controls. It does, however, offer access to the grey
literature and to full-text information.

1A.35 GREY LITERATURE

Grey literature is written material issued by a body with a primary activity that is not publishing. As such, it is not
readily available through traditional publication channels such as books and journals. Advantages and disadvantages
of grey literature are shown in Box 1A.35.1.

Box 1A.35.1
Advantages The internet makes this easier to access
Presents less orthodox views
Gives perspective to published material
Disadvantages Traditionally difficult to access – especially publications in paper form only
No quality control: the onus is on the reader to assess quality and credibility

Box 1A.35.2 gives some examples of grey literature.

Box 1A.35.2
• Technical and scientific reports
• Conference papers
• Internal reports from government and non-governmental organisations
• Government documents
• Theses
• Fact sheets
• Unpublished reports

1A.36 EvIDENCE-BASED MEDICINE AND POLICY

The term ‘evidence-based medicine’ (EBM) first appeared in 1992 and it relates to the explicit use of the current
best evidence for decision-making at the level of the individual patient. In addition to clinical skills, a practitioner
of EBM requires expertise in:

60

Public [Link] 60 3/6/08 [Link]

 1A Epidem iology

• Retrieving, ranking and interpreting the evidence

• Communicating evidence to patients
• Applying evidence to clinical decisions.
Advantages and diasadvantages of EBM are shown in Box 1A.36.1.

Box 1A.36.1
Advantages Explicit use of best evidence
Opinion of ‘medical expert’ demoted to least valid form of evidence
Disadvantages Publication bias (failure to publish negative results)
Retrieval bias (limitations of databases)
Lack of evidence π lack of benefit
Lack of robust evidence for treatments other than drugs
Evidence typically applies to populations, not necessarily to individuals
Diminishes value of clinical nous

EVIDENCE-BASED POLICY
Beginning in the late 1990s, there has been a drive to place more explicit importance on evidence when making
public policy decisions, i.e. a push for engineered policies rather than policies of conviction. The theory of evidence-
based policy is that decisions should be shaped as shown in Box 1A.36.2.

Box 1A.36.2
Problem identified by policy-makers ➔ Problem solved by researchers ➔ Solution adopted as policy
Or
New knowledge ➔ Knowledge adopted into policy

In the real world there are very few examples of evidence-based policy occurring in this way. Instead, decision-
makers tend to absorb the evidence which then appears unexpectedly in the future as policy: the so-called
enlightenment model (Buse et al 2005).

1A.37 HIERARCHY OF RESEARCH EvIDENCE

Hierarchy of research evidence – from well-conducted meta-analyses down to small case series
The Centre for Evidence Based Medicine has ranked the different types of research evidence based on how likely they
are to be true, as listed in Table 1A.37.1 The strongest evidence comes from a systematic review that demonstrates
consistent findings from several high-quality RCTs. This is termed level 1 evidence, and recommendations based on
level 1 results are called grade A. Further down the hierarchy come more heterogeneous findings and evidence from
less robust sources.
A similar system for ranking research evidence is used by the National Institute for Health and Clinical Excellence
(NICE) in the production of guidelines and technology appraisals to indicate its relative level of confidence in the
information used (see Section 1C.5).

61

Public [Link] 61 3/6/08 [Link]

 1A E p ide m io l o gy

Table 1A.37.1 Hierarchy of evidence

Level Study type
1 A Systematic review of RCTs
B Individual RCTs
C All or none
2 A Systematic review of cohort studies
B Individual cohort study
C Ecological studies
3 A Systematic review of case–control studies
B Individual case–control studies
4 Case series
5 Expert opinion
Reproduced from the Centre for Evidence-Based Medicine (2007).

1A.38 PUBLICATION BIAS

Publication bias is a tendency for journals to report positive results (where something was found to happen) rather
than negative results (where something was found not to happen) and neutral results (no conclusive finding).
It exists because researchers are less likely to submit, and publishers less likely to publish, negative and neutral
results. This bias distorts meta-analyses (see Section 1A.33).
Non-reporting of RCTs is beginning to be regarded as scientific and ethical misconduct. In an effort to thwart the
problem, registers such as the ‘Meta Register of Controlled Trials’ have been established, and medical journals have
agreed to publish only registered RCTs. Other options to detect or reduce publication bias include:
• Active discouragement of studies that do not have sufficient power to detect effects
• Examining effects in meta-analyses using funnel plots (see Figure 1A.38.1).

1A.39 THE COCHRANE COLLABORATION

The Cochrane Collaboration was established in 1993 and named after the British epidemiologist Archie Cochrane
(1909–1988). Cochrane was a notable contributor to the development of epidemiology as a science. Between 1960
and 1974 he was Director of the Medical Research Council Epidemiology Research Unit in Cardiff.
The Cochrane Collaboration is an international, non-profit, independent organisation. It produces and disseminates
systematic reviews of health-care interventions, and promotes the search for evidence in the form of clinical trials
and other studies of the effects of interventions. As of 2004, there were over 11 500 people working within the
Cochrane Collaboration in over 90 countries – half of whom were authors of Cochrane Reviews.
A key function of the collaboration is to produce systematic reviews (meta-analyses) of randomised controlled
trials (see Section 1A.33). Cochrane Reviews are systematic assessments of evidence of the effects of health-care
interventions, intended to help people to make informed decisions about health care. Cochrane Reviews ensure
that health-care decisions throughout the world can be informed by high-quality, timely research evidence. These
are published as part of the quarterly Cochrane Library, in the Cochrane Database of Systematic Reviews. Other
components of the library include the Cochrane Methodology Register and the Health Technology Assessment
Database.

62

Public [Link] 62 3/6/08 [Link]

 1A Epidem iology

Precision of In this non-biased funnel plot,

estimate of the circles represent the point
treatment
effect
estimates of the trials. The pattern
of distribution resembles an inverted
funnel. Larger studies tend to be
closer to the pooled estimate (the
interrupted line). In this case, the
effect sizes of the smaller studies
are more or less symmetrically
distributed around the pooled
Favour intervention Favour control estimate
Outcome measure

This biased funnel plot shows

Precision of that the smaller studies are not
estimate of symmetrically distributed around
treatment
effect either the point estimate (dominated
by the larger trials) or the results
of the larger trials themselves. The
trials expected in the bottom right
quadrant are missing. This suggests
Favour intervention Favour control
publication bias and an overestimate
Outcome measure of the true treatment effect

Figure 1A.38.1 Funnel plots for the assessment of publication bias. Reproduced from Montori et
al (2000)

1A.40 EPIDEMIOLOGICAL RESEARCH ETHICS Box 1A.40.1

Ethics and etiquette of epidemiological research Autonomy Human rights, dignity, freedom
Ethics is the philosophical discipline concerned with Non-malfeasance Do no harm
understanding how human beings should act, what is Beneficence Do good
good and what kind of life is best. In 1979 Beauchamp
and Childress (2001) identified four principles associated Justice Equity, fairness
with ethical medical practice: see Box 1A.40.1.
In public health, beneficence implies acting in the best interest of the population or society as a whole (McKeown
and Weed 2002). This should be done in a just way, ensuring the fair distribution across the population of both
benefits and risks. At a population level there is sometimes a tension between these principles, because of the
conflicting aims associated with benefiting an individual and those of providing the optimal conditions for the
wellbeing of the community. See Box 1A.40.2 for examples of such tensions.

Box 1A.40.2
Examples: conflicts between ethical principles in public health
• Fluoridation of the water supply does not permit individual informed consent. It may occur despite the
opposition of people who are opposed to the intervention
• If health-care resources are redistributed with the aim of providing equitable services, it may harm those from
whom resources are removed

63

Public [Link] 63 3/6/08 [Link]

 1A E p ide m io l o gy

In research studies, two ethical issues of particular importance are informed consent and confidentiality.

INFORMED CONSENT
Truly informed consent requires the features listed in Box 1A.40.3.

Box 1A.40.3
Competence Do subjects understand what is involved? Children over a certain age can assent to take part
but may lack full capacity for truly informed consent
voluntariness Are participants free to leave a trial at any point? Have they been put under excessive
pressure to enrol in a research project?
Patients need to be confident that their standard of care will not be affected by their decision
to take part in a study
Understanding Understanding the risks, burdens and benefits of the study
Documentation Written consent is required in trial settings. In cluster-randomised trials, communities may be
asked to give consent but it is not always clear who should give consent for the community

CONFIDENTIALITY
Research may involve the collection of private or sensitive information. Such confidential information should not
be shared with anyone without consent except when there is a clear ethical justification (e.g. approval by a human
subjects research review panel) or a legal requirement (e.g. regulations to protect children).
The use of identifiable data in research without consent requires demonstrating the importance of the research, the
minimal risk to those whose information is used, the promise of benefit to society and an obligation to maintain the
confidentiality of the information.

DATA PROTECTION
UK In the UK, the confidentiality of participants’ information is protected by the Data Protection Act 1998. This
contains eight principles, which state that the data must be:
1. Processed fairly and lawfully
2. Obtained and used only for specified and lawful purposes
3. Adequate, relevant and not excessive
4. Accurate and, where necessary, kept up to date
5. Kept for no longer than necessary
6. Processed in accordance with the individual’s rights (as defined)
7. Kept secure
8. Transferred only to countries that offer adequate data protection.
Ire Similar principles apply in Ireland through the Data Protection Acts of 1988 and 2003. The Data Protection
Commissioner provides information and guidance and ensures compliance with the legislation by those who keep
personal data. Responsibility rests with data controllers, i.e. those who decide what information is to be collected or
stored, to what use it is put and when it should be deleted or altered. Data controllers in public bodies and
organisations specified in the 1988 Act (such as banks, insurance companies and those who keep personal data of a
sensitive nature) are required to register with the Commissioner.

64

Public [Link] 64 3/6/08 [Link]

 1A Epidem iology

Aus Current privacy legislation in Australia provides for sensitive information about an individual to be collected
without that individual’s consent only when the information meets all three criteria below:
• The information is necessary for
# research relevant to public health; or
# compilation or analysis of public health statistics; or
# management or monitoring of a health service
• The purpose cannot be served by collection of non-identified information; and
• It is impractical to seek the consent of individuals.
Australian states and territories have their own legislation that gives protection to people whose work requires them
to deal with identified information, e.g. cancer registry staff, communicable disease control officers.
NZ In New Zealand, the main legislation and codes covering the privacy of personal information are:
• The Privacy Commissioners Act 1991 – established the office of Privacy Commissioner and the legal requirement
for data matching
• The Privacy of Information Act 1993 – which focuses on good personal information handling practices and
applies to almost every person, business and organisation in New Zealand. It includes 12 information privacy
principles covering collection, holding, use and disclosure of personal information and use of unique identifiers.
The Act also gives the Privacy Commissioner the power to issue codes of practice, specifies complaints
mechanisms and rules governing data matching.
• The Health Information Privacy Code 1994 – which specifies a code of practice for the health sector regarding
the collection, use, holding and disclosure of personal health information (i.e. that relates to identifiable
individuals), access to health information and the assignment of unique identifiers. For the health sector,
this code takes the place of the information privacy principles in the Act. It applies to all levels of the health
system from large institutions to sole practitioners.

CALDICOTT GUARDIANSHIP
Eng WalThe Caldicott report, published in 1997, found that compliance with data protection statutes was
variable. As a result, the Department of Health (DH) in England and the Welsh Assembly in Wales established a
register of Caldicott guardians. Every NHS and social services institution must appoint one senior member of staff to
take responsibility for protecting patient-identifiable information.
The other Caldicott recommendations (Walker 1999) are that NHS organisations should:
• Develop protocols to manage the sharing of patient information with other institutions
• Permit access to patient information only to employees who need to know that specific information
• Review and justify all uses of patient information
• Instil a culture of data protection through training, database design, etc.
Similar arrangements for protection of patient confidentiality exist in Scotland.

ELECTRONIC DATA STORAGE

UK The main information security standards that apply across the UK are listed below (Knott 2006).

BS7799
Part 1 of this British standard is known as the ‘Code of Practice for Information Security Management’ and provides
guidance on best practices in information security management. Part 2, the ‘Specification for Information Security
Management Systems’, is the standard against which an organisation’s security management systems are assessed
and certified. It has been revised and now exists in the form of ISO/IEC 27001:2005.

65

Public [Link] 65 3/6/08 [Link]

 1A E p ide m io l o gy

IEC 61508
This standard of the International Electro-technical Commission contains requirements for ensuring that computer
systems are designed, operated and maintained in ways that have sufficient integrity. Section 3 sets out the
requirements on computer suppliers for new equipment.

INFORMATION GOvERNANCE
NHS organisations are subject to controls over the storage of information for both routine purposes and research.
The overarching framework to assure the quality of these processes is called information governance.

ETHICS COMMITTEES
In research studies, ethical principles are protected by requirements of ethics local/national ethics committees.
Proof of ethical approval is often a prerequisite for journals to consider publishing a study.
Eng Scot In England and Scotland, any research conducted in the NHS must receive approval from an ethics
committee before commencing. A research ethics committee will consider issues such as whether the study has the
potential to benefit society/participants, how participants are recruited, how participants’ confidentiality will be
protected, as well as any possible effects of the study on health or wellbeing and processes for obtaining informed
consent.
Wal Any research requires the approval of a local ethics committee. It will subsequently be subject to local research
governance arrangements that seek to ensure that the research is conducted in a manner consistent with the
approval given by the ethics committee. Where the research setting is primary care, the researcher is required to
seek local management approval from the local health board – which may decline approval if it believes that the
research poses a risk either to patients or to delivery of NHS services.
Eng Where research is not located in a single area, the NHS Central Office of Research Ethics Committees
Wal
(COREC) is used. COREC also provides guidance, training and support to local committees.
Ire The Irish Council for Bioethics Guidance (2004) stipulates that all research involving or impacting on human
participants requires ethics review by a research ethics committee (REC). Procedures vary for obtaining permission to
undertake research in the hospital setting. Usually an application must be submitted to the hospital’s REC, but
national committee approval may be accepted as sufficient by some hospitals and regional approval may be accepted
by smaller hospitals in the regional network. For community-based research, protocols may be submitted to the REC
under the joint auspices of the Faculties of Occupational Health and of Public Health Medicine of the Royal College
of Physicians of Ireland.
SA In South Africa there is a national process controlled by the Ethics in Health Research: Principles, Structures and
Processes Research Ethics Guidelines (2004). Biomedical research ethics committees at the various universities now
have to be registered and structured according to the guidelines.
Aus Ethical approval comes from a formally constituted Human Research Ethics Committee (HREC). The National
Health and Medical Research Council has guidelines for the establishment and accreditation of research ethics
committees. Ethics committee approval must be obtained before research on humans can go ahead.
NZ In New Zealand, all health and disability research that involves human subjects must be sent for ethical review.
This includes observational research (e.g. descriptive studies using already collected personal information or
information obtained by questionnaires and interviews) as well as experimental research (e.g. clinical trials). The
only exceptions are certain type of audits and related activities (e.g. quality assurance and programme evaluation),
public health investigations (e.g. outbreak investigations and surveillance) and where a statutory exclusion applies
(e.g. official statistics).

66

Public [Link] 66 3/6/08 [Link]

 1A Epidem iology

The main system for ethical review is the set of health and disability ethics committees (HDECs). There are six
regional committees and a multiregional committee for research carried out nationally or in more than one region.
Research may be reviewed by an accredited institutional ethics committee (IEC) established by universities and
private companies, but in most circumstances will also need to be reviewed by a HDEC. Additional review is required
in some specific areas, notably: clinical trials of a pre-registration medicine, research involving assisted human
reproduction and research involving manipulation of human genetic material (e.g. gene therapy).

1A.41 GENETICS
Understanding of basic issues and terminology in the design, conduct, analysis and interpretation of population-based
genetic association studies, including twin studies, linkage and association studies
See Section 2D for details of public health genetics. The principles of genetic epidemiological studies (Dorak 2007)
are outlined below and are summarised in Table 1A.41.1

FAMILY STUDIES
Family studies are performed in order to determine whether there is a genetic component to a particular disorder.
They aim to detect a higher occurrence rate of disease in siblings or offspring of an affected person. This can be
measured by the relative recurrence risk (RRR) or familial risk ratio (FRR).
RRR = Probability that a particular type of relative (sibling, cousin, etc.) of an affected individual is also affected ∏
Prevalence of the disease in the general population.
A higher RRR or FRR is a necessary (though not sufficient) attribute to decide that there is a genetic component to
a disorder.

Table 1A.41.1 Genetic studies

Study question Appropriate study designs Unit of analysis/results obtained
Is there a genetic component to the Family studies RRR, FRR
disorder?
What is the contribution of genetics as Family – twin, adoption Percentage Concordance,
opposed to environment/other sources to discordance
the trait?
What is the model of transmission of the Family – segregation
genetic trait?
What is the location of the disease Family – linkage Lod score
gene(s)?
Recombination fraction
What is the allele associated with the Population based (and family Linkage disequilibrium (LD)
disease susceptibility? studies) – association

TWIN STUDIES
Twin studies explore the relative contributions of genes and the environment by comparing identical and non-
identical twins to explore the relative contributions of genes and environment to health and disease.
Identical twins can be considered to share both genes and environment. Non-identical twins act as a control: they
share the same environmental factors, but are only as genetically alike as any other sibling.

67

Public [Link] 67 3/6/08 [Link]

 1A E p ide m io l o gy

Where a disease occurs with a higher probability in identical twins than in non-identical twins, this is known
as concordance and indicates a shared genetic basis for the trait. The strength of the genetic predisposition is
expressed as the heritability, e.g. asthma has a heritability of 60%. Note that heritability is a population-based
statistic (i.e. measure of variance within a population), not a measure of the relative contribution of genes in an
individual, nor a measure of their risk. Furthermore, the heritability of a condition is unrelated to the number of
genes that influence it.
• Pairwise concordance is the percentage of concordant pairs (i.e. both twins affected) in a group of twins where
at least one member of each pair is affected.
• Probandwise concordance is the percentage of twins whose twin becomes affected during the study, in a group
of twins where just one member of each pair is affected.

LINKAGE STUDIES
Genetic linkage occurs when two alleles (two forms of a gene) are inherited together. In general, a parent can pass
on either allele of a particular gene to his or her offspring, regardless of which alleles of other genes are passed on.
This is because chromosomes are sorted randomly during meiosis. Sometimes, however, two alleles have a tendency
in the population to be inherited together more often than would be expected by chance. This is called linkage
disequilibrium and it occurs because the alleles are linked, i.e. found close to each other on the same chromosome.
When considering two genetic traits that appear to be linked, the offspring of people who have these two traits are
studied. The lower the percentage of the offspring who do not have both traits, the smaller the physical distance
between the two genes on the chromosome.
A study of the linkages between many genes enables the creation of a genetic map called a linkage map. Linkage
may be quantified using the lod score, which is the logarithm of odds of linkage. Traditionally, a lod score >+3 is
considered to be significant. This corresponds to a one-sided p value of 10–4.

ASSOCIATION STUDIES
These studies measure the relative frequency with which a particular polymorphism occurs together with the disease
of interest in a population, i.e. the extent to which it is associated with the disease. Designs of association studies
include the following.

POPULATION-BASED CASE–CONTROL STUDIES

These are studies in which two groups are chosen from the population: unrelated affected cases and unrelated
healthy controls. The relative frequencies of alleles at a single marker locus are tested statistically (using Pearson’s
c2 test or logistic regression).

FAMILY-BASED ASSOCIATION STUDIES

Families are identified that contain affected and unaffected individuals. The transmission of alleles from parents to
their affected and unaffected offspring is compared statistically.
Linkage disequilibrium is measured in association studies. The closer the physical (genetic) distance between
the disease and marker alleles, the lower the recombination frequency and the stronger the magnitude of linkage
disequilibrium.

68

Public [Link] 68 3/6/08 [Link]

 1B Statistics

1B
Statistics

1B.1 Elementary probability theory 69 1B.10 Type I and type II errors 87

1B.2 Methods for the quantification of 1B.11 Problems of multiple comparisons 88
uncertainty 70 1B.12 Parametric and non-parametric tests 88
1B.3 Estimation of confidence intervals 71 1B.13 Sample size and statistical power 97
1B.4 Conditional probability 74 1B.14 Regression and correlation 98
1B.5 Standard statistical distributions 76 1B.15 Regression techniques 99
1B.6 Principles of making inferences from a 1B.16 Comparison of survival rates 100
sample to a population 78
1B.17 Heterogeneity 101
1B.7 Measures of location and dispersion, and
1B.18 Funnel plots 101
their appropriate uses 79
1B.19 Role of Bayes’ theorem 103
1B.8 Graphical methods in statistics 82
1B.20 Choice of statistical test 104
1B.9 Hypothesis testing 86

An understanding of basic statistical principles is required in many parts of the membership examination. For the
mathematically inclined we have included formulae and their derivations to assist the learning process. However, the
only formulae that must be memorised for the examination are listed in Section 6D.

1B.1 ELEMENTARY PROBABILITY THEORY

Probability is a measure of the likelihood of an event. It is expressed as a positive number between 0 (event never
occurs) and 1 (event is certain to occur).
There are several approaches to calculating the probability of an event: see Table 1B.1.1.

Table 1B.1.1 Approaches to calculating the probability of events

Subjective Personal degree of belief that an event will occur, such as a doctor’s opinion for clinical
decision-making
Frequentist Proportion of times an event would occur in a large number of similar repeated trials, e.g.
number of ‘heads’ in coin tossed 100 times
A priori Requires knowledge of the probability distribution

69

Public [Link] 69 3/6/08 [Link]

 1B S t a t i s t ic s

Rules of probability
Two rules determine the ways in which two or more probabilities may be combined.

Addition rule (OR)

This is used to find the probability P of at least one event occurring out of two or more possible events.

Mutually exclusive events:

A B
P of obtaining A or B = P(A or B) = P(A) + P(B)

Non-mutually exclusive events:

A B
P(A or B) = P(A) + P(B) - P(A and B)

Multiplication rule (AND)

The probability of the joint occurrence of two or more events depends on whether the events are:
• Dependent (i.e. the occurrence of one event affects the probability occurrence of the other event), or
• Independent (the occurrence of each event does not affect the other).
Independent events: P of obtaining A and B = P(A) ¥ P(B)
Dependent events: P(A and B) = P(A) ¥ P(B|A).
Note that P(B|A) is known as conditional probability, i.e. probability of B occurring given that A has already
occurred (see Section 1B.19).

1B.2 Methods for the quantification of uncertainty

Uncertainty, in a statistical sense, refers to the fact that measurements do not always report the exact truth.
A measurement may be defined as a translation of a quantity in the real world, through the use of a measuring
device or instrument, into a number and a unit. Together the number and the unit represent the reality of interest.
For example, a microscope might be used to estimate the size of an average bacterium on a slide (0.5 mm). A
cardiologist might want to assess the amount of myocardial damage sustained by a patient by measuring the blood
troponin. A researcher might want to measure whether someone has a minor psychiatric disorder using the General
Health Questionnaire (GHQ). All of these measurements are associated with a degree of imprecision.
There are uncertainties inherent to measurement and prediction but this can be managed by quantifying and
accounting for this uncertainty.

Probability Theory Table 1B.2.1 Interpretation of probability values

Probability theory, i.e. the mathematical discipline of
Value of P(A) Interpretation
studying probability, offers a framework for dealing with
and quantifying probabilities. 1 A is certain to happen

An event A with probability P(A) takes a numerical value Near 1 A is very likely to happen
between 0 and 1. Table 1B.2.1 provides a summary of the >0.5 and <1 A is more likely to happen than not
interpretation of probability values.
0.5 A is as likely to happen than not
>0 and <0.5 A is less likely to happen than not
Probability distributions
Near 0 A is very unlikely to happen
The probability distribution is the range of values that
a random variable can take, and the relative frequency 0 A cannot happen

70

[Link] 70 6/6/08 [Link]

 1B Statistics

of occurrence of these values. Probability distributions are theoretical distributions, and they are expressed
mathematically with a mean and a variance. If a random variable is discrete or continuous, then its probability
distribution will likewise be discrete or continuous. See Box 1B.2.1.

Box 1B.2.1
Discrete variables Continuous variables
Distribution* Binomial distribution Normal distribution
Poisson distribution Chi-squared (c2)
t-distribution
f-distribution
Example Alive/dead Body weight
Number of hospital admissions p.a. Age
Sum of probabilities Sum of all the probabilities = 1 When drawn as a curve (called the
probability density function) the total area
under the curve = 1
*See Section 1B.5.

STATISTICAL INFERENCE
Statistical inference is a deduction about a population that is based on measurements made on a random sample
drawn from that population. Inference includes the aspects shown in Box 1B.2.2.

Box 1B.2.2
Point estimation Sample mean used to estimate population mean (see Section 1B.6)
Interval estimation Confidence interval (see Section 1B.3)
Hypothesis testing Statistical significance testing (see Section 1B.9)

1B.3 ESTIMATION OF CONFIDENCE INTERvALS

Confidence intervals are a means of making inferences about the accuracy of a sample value that is being used as
an estimate population value.

PRECISION IN SAMPLING
A major cause of apparent differences between observations is the effect of chance, and the magnitude of this
chance is determined by the size of the sample. This effect can be modelled statistically. Statistical tests give the
possibility (p) of chance alone being responsible for obtaining a result at least as extreme as that observed. This is
the underlying principle behind all tests of statistical significance.
Note that the value of this probability (the p value) reflects only the role of chance. It does not take account of any
biases or confounding that may exist, nor does it imply causality. Every p value is a composite measure of:
• Effect size
• Sample size.
By convention, a p value of <0.05 is taken to be suggestive of a statistically significant result, since there is less
than a 1-in-20 chance of the observed result being due to chance. However, it is preferable to think of a spectrum

71

Public [Link] 71 3/6/08 [Link]

 1B S t a t i s t ic s

of evidence rather than a single cut-off: the lower the p value, the stronger the indication that the finding is
significant (i.e. that the observed finding was not due to chance alone).
• 0.05 < p < 0.10 (‘suggestive’ evidence)
• 0.01 < p < 0.05 (‘quite strong’ evidence)
• 0.001 < p < 0.01 (‘strong’ evidence)
• p < 0.001 (‘very strong’ evidence)
Because the p value depends on both the size of the effect and the sample size, given a large enough sample size,
even a minute difference will be statistically significant. For this reason, confidence intervals are more informative
than the p value alone. See Box 1B.3.1

Box 1B.3.1
p values and confidence intervals
• A narrow confidence interval implies a large sample size or a small number of very similar results
• If a p value is non-significant and there is a narrow confidence interval, then it is likely that there truly is no
effect
• If a p value is non-significant and there is a wide confidence interval, then this suggests that the sample size
may be too small
Note that no p value, however small, can exclude chance completely. Likewise, no p value, however large, can
guarantee that an association was definitely due to chance.

Calculation of confidence intervals

Confidence intervals are calculated by taking a sample value, then adding and subtracting a multiple of the standard
error for that value. For a 95% confidence interval (p = 0.05) this multiple is 1.96 (see Box 1B.3.2); for a 99%
confidence interval (p = 0.01) it is 2.58.

Box 1B.3.2
95% confidence interval = sample value ± (1.96 ¥ standard error)
Begin by calculating the standard error for the sample value:
s
For a mean: standard error =
n

For a proportion: standard error = p(1 − p)

n
p1 (1 − p1 ) p2 (1 − p2 )
For two proportions: standard error = +
n1 n2

Here s = standard deviation for the sample; n = number in the sample; p = proportion.

Boxes 1B.3.3–1B.3.5 give three examples.

72

[Link] 72 6/6/08 [Link]

 1B Statistics

Box 1B.3.3
Example 1: calculating the confidence interval of a mean
The mean pO2 arterial blood test for a sample of 56 patients with COPD was 8.9 kPa, with a standard deviation
of 0.8 kPa. What is the 95% confidence interval for the mean pO2 of the population of all COPD patients?
0.8
Standard error =
56
= 0.11

95% confidence interval = 8.9 ± (1.96 × 0.11)

= between 8.7 and 9.1 kPaa

Box 1B.3.4
Example 2: Calculating the confidence interval for two proportions
57% of the 864 patients at a stop-smoking clinic had quit smoking by the end of the programme, compared
with 42% of 795 patients at a neighbouring clinic.

0.57(1 − 0.57) 0.42(1 − 0.42)

Standard error (difference in proportions) = +
864 795
= 0.0243
95% confidence interval for the difference = sample value ± (1.96 × standard error)
= (0.57 − 0..42) ± (1.96 × 0.024)
= between 0.198 and 0.102
= between 10.2% and 19.8% difference

Box 1B.3.5
Example 3: Calculating the confidence interval for a proportion
87% of the 265 patients with a deep vein thrombosis (DVT) at an anticoagulation clinic completed their
6-month course of warfarin. What is the standard error and 95% confidence interval for completion of the
course?

0.87(1 − 0.87)
Standard error =
265
= 0.02
95% confidence interval = 0.87 ± (1.96 × 0.02)
= between 0.83 and 0.91
= bettween 83% and 91% of patients

INTERPRETATION OF CONFIDENCE INTERVALS

A confidence interval is a range of values indicating the precision with which the sample estimate is likely to
represent the population from which the sample was drawn. For a 95% confidence interval, the true value for the

73

Public [Link] 73 3/6/08 [Link]

 1B S t a t i s t ic s

population at large will lie within this range 19 times out of 20. For a 99% confidence interval (which will be a
wider range), the true value will lie within the range expressed 99 times out of 100.
• For measures of absolute risk, where a 95% confidence interval includes zero, there is no evidence at the
p = 0.05 level that there is a true difference.
• For measures of relative risk, where a 95% confidence interval includes one, there is no evidence at the
p = 0.05 level that there is a true difference.
In any diagram showing two or more point estimates and their associated confidence intervals, there are three
possibilities regarding the overlap of the points and intervals. These may be interpreted as shown in Box 1B.3.6.

Box 1B.3.6
Overlap of intervals or parts Interpretation
Confidence intervals do not overlap Significant difference at that significance level
Confidence intervals overlap but the point estimates are Unclear: requires calculation of a significance test
outside the confidence intervals of the other
Point estimate of one sample falls within the confidence No difference at that significance level
intervals of the other

1B.4 CONDITIONAL PROBABILITY

See also Section 1B.1.
Conditional probability represents the chance that one event will occur, given that a second event has already
occurred. Conditional probability allows the evaluation of how different treatments or exposures influence the
probability of outcomes such as disease or mortality. It also provides a useful way to evaluate diagnostic tests. Box
1B.4.1 gives an example.

74

Public [Link] 74 3/6/08 [Link]

 1B Statistics

Box 1B.4.1
Example: Breastfeeding in preterm infants
In a study of breastfeeding in preterm infants, the infants were randomised into two groups. The treatment
group was fed by nasogastric (NG) tube, and the control group was bottle-fed. The purpose of the study was
to test whether using the NG tube for feeding rather than using a bottle would increase the likelihood of
breastfeeding at discharge from hospital and afterwards.
In this table, the rows represent the group (NG tube versus bottle) and the columns represent feeding status at
discharge (exclusive breastfeeding versus partial/no breastfeeding).
Exclusive breastfeeding at discharge
No Yes Row total
Bottle fed 27 20 47
NG tube fed 10 32 42
Column total 37 52 89

To obtain cell probabilities, each entry in the table above is divided by the total sample size.

Exclusive breastfeeding at discharge

No Yes Row total
Bottle fed (27/89 ¥ 100) 30.3% (20/89 ¥ 100) 22.5% 52.8%
NG tube fed (10/89 ¥ 100) 11.2% (32/89 ¥ 100) 36% 47.2%
Column total 41.6% 58.4% 100%
The probabilities in the row and column totals represent unconditional probabilities. The interior probabilities
represent the probabilities of the intersection between two events.
Conditional probabilities represent the probability that an event will occur when attention is restricted to a
specific row (or sometimes a specific column) of a 2 ¥ 2 table.

Exclusive breastfeeding at discharge

No Yes Row total
Bottle fed (30.3/52.8 ¥ 100) 57.4% (22.5/52.8 ¥ 100) 42.6% 100
NG tube fed (11.2/47.2 ¥ 100) 23.8% (36/47.2 ¥ 100) 76.2% 100
Column total (41.6/100 ¥ 100) 41.6% (58.4/100 ¥ 100) 58.4% 100
A vertical bar, ‘|’ denotes conditional probability and is pronounced as the word ‘given’.
For example,
P(exclusive breastfeeding | NG feeding), the probability of exclusive breastfeeding, given that NG tube feeding
occurred, is 76.2%
P(exclusive breastfeeding | bottle-feeding), the probability of exclusive breastfeeding, given that bottle-feeding
occurred, drops to 42.6%
The unconditional probability of breastfeeding is 58.4%, i.e. somewhere between the two conditional
probabilities.
Reproduced with permission from Fleiss (1981).

75

Public [Link] 75 3/6/08 [Link]

 1B S t a t i s t ic s

1B.5 STANDARD STATISTICAL DISTRIBUTIONS

A statistical distribution is an arrangement of the values of a variable that shows how frequently each value is
observed or is expected theoretically to occur.

BINOMIAL DISTRIBUTION
The binomial distribution shows the frequency of events that have two possible outcomes. It is constructed from
two parameters: n (sample size) and p (true probability). When the sample size is large, the binomial distribution
approximates to the normal distribution: see Figure 1B.5.1.
This distribution is used for discrete data with two outcomes (e.g. success or failure) and the sampling distribution
for proportions.

Binomial distribution, n� 10, p �0.5 Binomial distribution, n� 100, p�0.5

0.08

0.20
Probability of x successes 0.06
Probability of x successes

0.15

0.04
0.10

0.02
0.05

0.00 0.00
2 4 6 8 10 30 40 50 60 70
x successes x successes
Figure 1B.5.1 Graph of binomial distribution approximating normal distribution as sample size increases, where n
is the sample size, p is the probability and x represents the number of successes obtained

POISSON DISTRIBUTION
The Poisson distribution shows the frequency of events over time in which the events occur independently of each
other. An example is the discharge of alpha particles during radioactive decay. In the Poisson distribution, the
parameter is the variance which is equal to the mean (and therefore the standard deviation is equal to the square
root of the mean). This means that small samples give asymmetrical distributions, and large samples approximate
the normal distribution: see Figure 1B.5.2. The horizontal axis shows x, the number of occurrences of an event
within a particular time period. The vertical axis shows the probability of obtaining x events during that period.
The distribution assumes that the data are discrete, occurring at random and independent of each other.
The Poisson distribution is used in the analysis of rates (e.g. incident rates of disease). Since it leads to a
prediction of randomly occurring events, it allows determination as to whether observed events are occurring
randomly or not.

76

Public [Link] 76 3/6/08 [Link]

 1B Statistics

Poisson distribution, λ� 5 Poisson distribution, λ�20

0.08
0.15
Probability of x occurrences

0.06

P (x occurrences)
0.10
0.04

0.05
0.02

0.00 0.00
0 5 10 15 5 10 15 20 25 30 35 40
x events or occurrences x events or occurrences

Figure 1B.5.2 Approximation of Poisson distribution to normal distribution as l (the mean or expected value)
increases. Adapted from Zoonekynd (2007)

CENTRAL LIMIT THEOREM

This theorem states that for any population, no matter what its distribution, if a sufficiently large number of
samples is taken, then the distribution of the means of these samples will always be normally distributed, i.e. they
will follow a normal (or ‘gaussian’) distribution.

NORMAL DISTRIBUTION
The normal distribution is a bell-shaped, symmetrical curve that is described by two parameters:
• Mean (m)
• Variance (s2).
The standard normal distribution has a mean of 0 and variance of 1. In the standard normal distribution, 68% of the
area under the curve is within 1 standard deviation of the mean, 95% of the area is within 1.96 standard deviations
and 99% of the area is within 2.58 standard deviations: see Figure 1B.5.3.

Mean = m

Figure 1B.5.3 The standard normal

–2.58 –1.96 –1 µ=0 +1 +1.96 +2.58 distribution
77

Public [Link] 77 3/6/08 [Link]

 1B S t a t i s t ic s

OTHER DISTRIBUTIONS
Table 1B.5.1 lists features and applications of some other distributions of continuous variables.

Table 1B.5.1 Probability distributions of continuous variables

Distribution Features Use and assumptions
t-distribution Parameter: degrees of freedom Uses:
(also called • Estimating the mean of a normally
The shape is similar to the normal
‘Student’s distributed population when the sample
distribution but tails are more spread out
t-distribution’) size is small
As the degrees of freedom increases, the • Testing hypotheses with a single mean
t-distribution approaches the normal (small sample)
distribution • Testing hypotheses with two means (small
samples)
• Confidence intervals
Chi-squared Parameter: degrees of freedom Uses:
• Analysing categorical data, e.g.
The shape is right-skewed, taking positive
significance test for two categorical
values
variables (comparison of observed and
With increasing degrees of freedom, it expected events)
approximates the normal distribution
Assumptions:
• If n is >40 then all x are >1
• If n is >20 but <40 all x should be >5,
otherwise use Fisher’s exact test
f-distribution • It is skewed to the right Uses:
• Values are positive • Useful for comparing two variances and
• It is defined by a ratio of two more than two means using analysis of
parameters: degree of freedom of the variance (ANOVA)
numerator and denominator of the ratio

1B.6 PRINCIPLES OF MAKING INFERENCES FROM A SAMPLE TO A POPULATION

If the population is sampled in such a way that the sample correctly represents the population, then conjectures
can be made about the population from the sample. Using a sample to make such inferences is more efficient
in terms of time and resources than obtaining details on the whole population (which may also be impossible).
However, the process of sampling introduces a degree of error. This error is minimised by taking a sufficiently large,
random sample.
There are two methods of reaching a statistical inference: estimation and hypothesis testing.

ESTIMATION
An estimate is a measurement made from a sample, which has been drawn from a population. Estimates allow
inferences to be drawn about the population, providing both a point estimate and a confidence interval (see
Section 1B.5). The key to estimation is the probability with which particular values will occur during sampling − this
allows the inference about the population to be made.
The values that occur are inevitably based on the sampling distribution of the population. In order to make an
accurate inference about a population, random sampling is required, where each member of the population has the
same probability of being selected for the sample (see Section 1A.25).

78

Public [Link] 78 3/6/08 [Link]

 1B Statistics

HYPOTHESIS TESTING
See also Section 1B.9.
Hypothesis testing involves comparing the observed findings in the sample with the findings that would have
been expected theoretically (the hypothesis). The process differs according to whether a parametric test or non-
parametric distribution is used: see Table 1B.6.1.

Table 1B.6.1 Differences between parametric and non-parametric tests

Parametric Non-parametric
Assumptions Data follow a known distribution (e.g. normal or Makes no assumptions about the
Poisson) underlying distribution
Use Continuous data are sampled from a population When the distribution is not known
with an underlying normal distribution (or the or the distribution is small
distribution of which can be rendered normal by
mathematical transformation, e.g. log or 1/x)
Power Powerful Less powerful
Confidence Straightforward Difficult
interval
calculation
Standard Requires a similar standard deviation to population Do not need similar standard
deviation with which a comparison is being made deviation
Sample size Any Best for <50 data cases
Examples t-test Wilcoxon’s signed rank
ANOVA Mann−Whitney U
MANOVA Rank correlation (e.g. Spearman,
Kendall)
Regression (all types)
Chi-squared
Correlation

1B.7 MEASURES OF LOCATION AND DISPERSION, AND THEIR APPROPRIATE USES

Data are summarised according to two parameters: location and dispersion. See Table 1B.7.1.

Table 1B.7.1 Examples of measures of location and dispersion

Parameter Description Examples
Location Average value for the data Mean
Median
Mode
Dispersion Spread of the data Variance
Standard deviation

79

Public [Link] 79 3/6/08 [Link]

 1B S t a t i s t ic s

MEASURES OF CENTRAL LOCATION

Three measures of central tendency are commonly used in statistics, each with its advantages and disadvantages:
see Table 1B.7.2.

Table 1B.7.2 Advantages and disadvantages of measures of central location

Measure Definition Calculation Advantages Disadvantages
Mean The average value
x=
∑n i
Amenable to Sensitive to outliers
N statistical analysis
For a sample, the mean is denoted Poor for
by x (pronounced ‘x-bar’)
µ=
∑ ni asymmetrical
N distributions

For a population, the mean is

denoted by the Greek letter m
(pronounced ‘mu’)
Median Value at the centre of the If all observations Unaffected by Value determined
distribution are arranged in extreme outliers solely by rank,
ascending order so carries no
Good for skewed
of value, then information about
distributions
the median is the any other values
middle value (half
the items lie above
and half below)
Mode Commonest value or values Value(s) that occurs Not greatly affected Not amenable to
with the highest by extreme values statistical analysis
frequency
Can give extra Sometimes no mode
insights (e.g. exists
suicide is bimodal,
Data containing
affecting young and
more than one
elderly)
mode can be
difficult to interpret

MEASURES OF DISPERSION
The principal measures of dispersion are shown in Table 1B.7.3.

Table 1B.7.3 Principal measures of dispersion

Measure Definition and calculation Advantages Disadvantages
Range Difference between maximum and Intuitive Sensitive to the size of the
minimum sample values (units are the sample
Simple – only depends on
same as data)
two observations

80

Public [Link] 80 3/6/08 [Link]

 1B Statistics

Percentiles Values are ranked and divided into Useful for comparing Comparisons made
100 groups. The 100th percentile is measurements, e.g. BMI in at extreme ends of
the largest value. The nth percentile groups of similar age and distribution less
will have n% values below it and 100 sex informative than at the
− n% values above it centre
Quartiles Values are ranked and then divided More stable than the range Unstable for small samples
into four groups, each containing 25% as sample size increases and not allowing further
of the data mathematical manipulation
Interquartile range: middle 50% of the
sample
Quintiles As for quartiles, but five groups rather
than four
variance Average squared deviation of each Useful for making Units are the squared units
number from its mean inferences about the of the data
Variance in a sample: population

∑( x − n )
2
i
s 2
=
n −1
Population variance:

∑( µ − n )
2

σ= i

N
Standard variance Most commonly used Sensitive to some extent to
deviation Units are the same as the extreme values
data
Not sensitive to a change
in units
Useful for making
inferences about the
population
Coefficient of Ratio of the standard deviation to the Allows comparison of the Mean value is near zero,
variation mean to give an idea of the size of variation of populations the coefficient of variation
the variation relative to the size of that have significantly is sensitive to change in
the observation different mean values the standard deviation
Standard error Measure of how precisely the Used to calculate Depends on sample size,
population mean is estimated by the confidence interval i.e. small sample = large
sample mean standard error
SE = s / n

Confidence Range of values in which the ‘true’, or Indicates the reliability of By chance alone about 1
interval population value, is likely to lie (see an estimate in 20 significant findings
Section 1B.3) at the p=0.05 level will be
spurious. Wide confidence
intervals provide less
information

81

Public [Link] 81 3/6/08 [Link]

 1B S t a t i s t ic s

1B.8 GRAPHICAL METHODS IN STATISTICS

Graphics can be an effective means of conveying statistical information. Different graphics are used for qualitative
and quantitative data.

QUALITATIVE DATA
The type of chart used should be chosen carefully based on the nature of the data, the analysis and the message to
be conveyed.

BAR CHARTS
These use bars to represent the frequencies (or relative frequencies), such that the height of each bar equates to the
frequency or relative frequency of its category. An example is shown in Figure 1B.8.1.
• Frequencies: counts
• Relative frequencies: percentage.

PIE CHARTS
The pie is a circle divided into a number of slices, each representing a different category. The size of each of each
slice is proportional to the relative frequency of that category. An example is shown in Figure 1B.8.2.

5
Count or percentage

A
4

3
C
2

1 B

0
A B C
Category
A B C
Figure 1B.8.1 Schematic of a bar chart
Figure 1B.8.2 Schematic of a pie
chart
QUANTITATIVE DATA
The type of graphical method used to display quantitative data is chosen according to whether the data are
univariate, bivariate or multivariate.

UNIvARIATE DATA
Here there is an observation of a single numerical variable. Graphical methods include stem-and-leaf displays,
boxplots and histograms.

82

Public [Link] 82 3/6/08 [Link]

 1B Statistics

STEM-AND-LEAF DISPLAY
This is a quick technique for displaying numerical data graphically. A vertical stem is drawn, consisting of the first
few significant figures. Any digit after the stem is called the leaf, i.e. the leaf is the last digit of the data value. An
example is shown in Figure 1B.8.3. A back-to-back stem-and-leaf plot can be used to display data from two groups.
Box 1B.8.1 lists advantages and disadvantages of stem-and-leaf display.

Box 1B.8.1 Advantages and disadvantages of stem-and-leaf display

Advantages Disadvantages
Quick and easy to construct Cumbersome for large data sets
Actual values retained

104, 145, 113, 155, 179, 120, 157, 130, 113, 162, 134, 159
Stem Leaf
10 4

11 3

12 0

13 0 4

14 5

15 5 7 9

16 2

17 9

Figure 1B.8.3 Example of stem-and-leaf display constructed from 12 data points

BOXPLOTS
Also known as a box-and-whiskers plot, a boxplot shows a measure of central location (the median), two measures
of dispersion (the range and interquartile range), the skewness (from the orientation of the median relative to the
quartiles) and potential outliers (marked individually). An example is shown in Figure 1B.8.4, and advantages and
disadvantages are listed in Box 1B.8.2.

Box 1B.8.2 Advantages and disadvantages of boxplots

Advantages Disadvantages
The spacing between the different parts of the box can Exact values not
help indicate variance and skew, and identify outliers retained
Boxplots are especially useful when comparing two or
more sets of data

83

Public [Link] 83 3/6/08 [Link]

 1B S t a t i s t ic s

Maximum observation (outlier)

.
Largest non-outlying observation

Inter quartile range

75th percentile

Median (if not in centre of box the data are skewed)

Whiskers

25th percentile

Smallest non-outlying observation

.
Minimum observation (outlier)

Figure 1B.8.4 Schematic of a boxplot

HISTOGRAMS
Histograms divide sample values into many intervals called ‘bins’. The bars represent the number of observations
falling within each bin (i.e. the bin’s frequency). Examples are shown in Figure 1B.8.5. Note that, in contrast to bar
charts, there are no gaps between the bars of a histogram. This is important because it is a reflection of the fact
that the data are continuous. Box 1B.8.3 lists advantages and disadvantages of histograms.

Box 1B.8.3 Advantages and disadvantages of histograms

Advantages Disadvantages
Demonstrates central tendency (mean, mode, medium) Cannot read exact values because data are grouped into
categories
Demonstrates shape of frequency distribution
(symmetrical or skewed, unimodal, bimodal or More difficult to compare two data sets
multimodal)
Use only with continuous data

Mean = Mode = Median Mode Median Mean Mean Median Mode

Figure 1B.8.5 Schematic of histograms, demonstrating shape of the frequency distribution

84

Public [Link] 84 3/6/08 [Link]

 1B Statistics

BIvARIATE DATA
Here there are two numerical variables. Bivariate data are best displayed using scatter plots.

SCATTER PLOT
The data from two variables are plotted one against the other to explore the association between them. A trend line
is drawn to illustrate whether any correlation is:
• Positive, negative or non-existent
• Linear or non-linear
• Strong, moderate or weak.
The strength of any correlation is determined by observing the spread of the points about the line. Examples of
scatter plots are shown in Figure 1B.8.6, and advantages and disadvantages are listed in Box 1B.8.4.

Box 1B.8.4 Advantages and disadvantages of scatter plots

Advantages Disadvantages
Shows a trend in the data relationship Hard to visualise results in large data sets
Retains exact data values and sample size Flat trend line gives inconclusive results
Shows minimum, maximum and outliers Data on both axes should be continuous

Strongly positive correlation Strongly negative correlation Moderately positive correlation

Weakly negative correlation No correlation

Figure 1B.8.6 Scatter plots displaying the correlation between variables

85

Public [Link] 85 3/6/08 [Link]

 1B S t a t i s t ic s

1B.9 HYPOTHESIS TESTING

To answer a statistical question, a hypothesis is first formulated. This is a statement that can be subjected to a
test. The test will generate a probability of how likely it is that the observed outcome could have occurred due to
chance alone. Depending on this probability, the hypothesis will be accepted or rejected.

STEP 1
Define a null hypothesis (H0). This is a statement that there is no difference (or no relationship) between the
variables being tested. For every null hypothesis, there is an alternative hypothesis (HA), which assumes that
a difference or relationship does exist. Both the null hypothesis and the alternative hypothesis are true/false
statements that are answered in view of the significance level chosen (see Section 1B.10). For an example see Box
1B.9.1.

Box 1B.9.1
Example: Randomised controlled trial
H0 = the rates of disease are the same in the intervention group
and the control group
HA = the rates of disease differ between the intervention group
and the control group

STEP 2
Collect the data.

STEP 3
Calculate a test statistic.
A significance test consists of calculating the probability of obtaining a statistic as different or more different from
the null hypothesis (given that the null hypothesis is correct) than the statistic obtained in the sample.
µ
The z test for significance is conducted using the formula: z =
σ N
where m = mean, s = SD and N = number of subjects in the sample.

STEP 4
Derive the p value for the test statistic from a known probability distribution (found in statistical tables – although
note that these are not currently provided to candidates in the UK MFPH Part A examination).

STEP 5
Interpret the p value to accept or reject the null hypothesis, i.e. the outcome of hypothesis testing is always either
‘reject H0’ or ‘do not reject H0’. See Box 1B.9.2.
If the null hypothesis is rejected, then the alternative hypothesis may be true. However, if the H0 is not rejected,
then this does not imply that the HA is correct − only that there is insufficient evidence against the H0.
Note that a very small p value (e.g. 0.001) does not signify a large effect. Rather, it signifies that the observed
effect (whatever its size) is highly improbable given the null hypothesis. A very small p value can arise when an
effect is small but the sample sizes are large. Conversely a larger p value can arise when the effect is large but the
sample size is small.

86

Public [Link] 86 3/6/08 [Link]

 1B Statistics

Box 1B.9.2
Example: interpreting the p value (see also Section 1B.3)
If p = 0.001, there is 0.1% likelihood that, if the null hypothesis were true, this result would be obtained by
chance alone. Therefore, there is strong evidence against the null hypothesis, suggesting that the observed
effect is not due to chance.
If p = 0.1, there is 10% likelihood that, if the null hypothesis were true, we would obtain this result by chance.
There is therefore little evidence to reject the null hypothesis.

1B.10 TYPE I AND TYPE II ERRORS

Findings from a study can lead to the null hypothesis being accepted or rejected. However, this acceptance or
rejection may or may not reflect the ‘true’ situation, as seen in the contingency table shown in Table 1B.10.1, where
the null hypothesis is that A = B.

Table 1B.10.1 Contingency table displaying type I and type II errors

‘True’ finding
A=B AπB
Results from A=B Correct Type II error (b)
study AπB Type I error (a) Correct

The magnitudes of a and b error rates are generally set in advance as the outcome of a power calculation that is
used to determine the appropriate sample size of the study.
Table 1B.10.2 summarizes the differences between type I and type II errors.

Table 1B.10.2 Differences between type I and type II errors

Type I error Type II error
Definition False positive False negative
Description Null hypothesis was wrongly rejected Null hypothesis was not rejected when it should
have been
Study shows an effect which in reality does not
exist Study does not detect an effect that existed in
reality
Symbol a b
Alternative Significance level = a Power = 1 − b
name
Typical value 0.05 or 0.001 0.8
Aide memoire Type I error = p value Type II (two) error = sample size too small

Type II errors are generally considered to be less serious than type I errors: a type II error is only an error in the
sense that an opportunity to reject the null hypothesis was lost.
Note that there is a trade-off between type I and type II errors: the more the study restricts type I errors by setting
a low level of a, the greater the chance that a type II error will occur.

87

Public [Link] 87 3/6/08 [Link]

 1B S t a t i s t ic s

SECULAR TRENDS
Erroneous findings may also result from temporal changes. A secular trend is a long-term change in the burden of a
disease (as opposed to a short-term or cyclical fluctuation). It may be due to:
1. Changes in diagnostic techniques/diagnostic codes
2. Changes in accuracy/completeness of enumerating cases
3. Change in the age distribution of the population
4. Change in survival (which affects prevalence)
5. Change in the true incidence of the disease.*
*This reason should be considered only once items 1–4, and other important possible explanations, have been
accounted for.

1B.11 PROBLEMS OF MULTIPLE COMPARISONS

The hypothesis to be tested should always be stated before the study begins, together with the list of which
variables will be analysed. The reason for this is that after the results have been collected, there may be a
temptation to search for associations with additional variables. This is known as ‘data mining’ or ‘data fishing’. As
more and more tests are performed, the likelihood of a type I error increases, i.e. the chance of falsely concluding
that an association is significant. This is because, at the p = 0.05 level of significance, 1 in 20 tests will appear
significant due to chance alone.
To compensate for multiple comparisons, use a method to adjust the p value for the number of tests performed, such
as Bonferroni’s correction.
If testing n outcomes instead of a single outcome, divide a level by n:
Original α
i.e. Adjusted α =
n
An example is shown in Box 1B.11.1.
Box 1B.11.1
Example: Bonferroni’s correction
Suppose we were investigating the association between salt intake and 15 different types of cancer. We would
test at:
Adjusted a = 0.05 ∏ 15 = 0.0033 level.
This would ensure that the overall chance of making a type I error is kept at <1 in 20.

1B.12 PARAMETRIC AND NON-PARAMETRIC TESTS

Parametric and non-parametric tests for comparing two or more groups
Samples that are normally distributed can be described by two parameters: the mean and the standard deviation. A
difference between samples or between a sample and the population may be measured by examining differences in
their means. These differences can be tested using parametric tests, such as the z-test or the t-test.
Where a population is not normally distributed, its shape cannot be described by two parameters (the mean and
standard deviation) alone, and any difference between two groups must be tested using a non-parametric test.
The Wilcoxon test for paired data, and the Mann–Whitney U-test for unpaired data are rank non-parametric tests
used for small samples taken from a non-normally distributed population. If samples are large, then the lack of
normality is not problematical and the above tests can be used.

88

Public [Link] 88 3/6/08 [Link]

 1B Statistics

The z-test and t-test are parametric tests used for comparing percentages, proportions and means of groups, with
the z-test for large samples and the t-test for samples <60.
The chi-squared test is used for comparing observed versus expected numbers. If the numbers involved are small,
then consider using Fisher’s exact test instead or applying Yates’ correction.

t-Test
The ‘unpaired’ t-test may be used to compare one group to another group. Often one group is the experimental
group and the other is the control. An example of an experiment that might utilise the t-test is a comparison of the
mean blood pressure in 25 patients given a new b-blocker compared with 25 given a standard b-blocker, labetolol.

x1 − x 0
t= , df = n1 + n0 − 2
se
x1 − x 0
t=
1 1
s +
n1 n0

where x = sample mean, n = sample size, s = pooled standard deviation, df = degree of freedom.

(n 1
− 1) s12 + ( n0 − 1) s02
s=
n1 + n0 − 2

The paired t-test typically uses observations in one sample that have been paired with observations in another. This
results in a different test statistic, where x represents the means of the paired differences. The usual hypothesis in
this situation is that the mean of the differences is 0.
x
t= , df = n − 1
s n

One-sample t-test
This tests the likelihood that a sample came from a population.
1. Mean of the population, m, and the sample, x , will both be known
( − x) n
2. Use the formula to calculate a value for the t-test t =
SD
 Difference in means
i.e. t =
Standard error of sample mean

3. Look up the p value in a t-table, using:

Degrees of freedom = n – 1
An example is shown in Box 1B.12.1.

89

[Link] 89 6/6/08 [Link]

 1B S t a t i s t ic s

Box 1B.12.1
Example: One-sample t-test
In a population it is known that the mean finger-tapping time is 100 ms per tap
In a study of eight participants with caffeine addiction, tapping times were found to be:
• Mean = 89.4 ms
• Standard deviation = 20 ms.
Does this prove that caffeine addiction is associated with faster tapping speed? The null hypothesis (H0) is that
tapping speed is not affected by caffeine addiction. We should begin by pre-selecting a significance level at
which we would be convinced of an effect, in this case 0.05.

(µ − x ) n
t=
SD
(100 − 89.4) × 8
t= = 1.5
20
Find the p value from statistical tables, i.e. the area below t(1.5) using seven degrees of freedom, since
df = n - 1.
This value of p is 0.07. Since this value is >0.05, there is little evidence to reject the null hypothesis. We
therefore do not conclude that caffeine addiction is associated with faster tapping speed.
Reproduced with permission from Whiteley (2007).

TWO-SAMPLE (UNPAIRED) t-TEST

This tests the likelihood that two samples came from the same population:
1. Calculate the means for both groups (x1 and x2)
(n1 − 1)s12 + (n2 − 1)s22
2. Calculate the pooled variance
n1 + n2 − 2

3. Calculate the standard error of the difference in means = s 1 + 1

n1 n2

Difference in means ( x − x2 ) − 0 ( x1 − x 2 )
4. t = = 1 =
Standard error of diffeerence in means SE( x1 − x2 ) 1 1
s +
n1 n2

5. Look up t value (using degrees of freedom = n1 + n2 - 2) in a t-table to find the p value.

An example is shown in Box 1B.12.2.

90

Public [Link] 90 3/6/08 [Link]

 1B Statistics

Box 1B.12.2
Example: Unpaired t-test
In an RCT, one group of patients is given an inhaled steroid while the other group is given a placebo. The two
groups are compared at 6 months with regard to their forced expiratory volume (FEV1).
Is there a significant difference in FEV1 at 6 months between the two groups?
Treatment group 1 receive inhaled steroid [n1 = 50; mean FEV1 (x1) = 1.64 l; SD1 = 0.29 l]
Treatment group 2 receive placebo [n2 = 48; mean FEV1 (x2) = 1.54 l; SD2 = 0.25 l]
An unpaired t-test is performed to compare means in the two groups:
• Null hypothesis H0: mean FEV1 is same in the two groups, i.e. no difference
• Alternative hypothesis HA: FEV1 is not the same in the two groups, i.e. there is a difference.

(n1 − 1)s12 + (n2 − 1))s22 (50 − 1)0.292 + (48 − 1)0.252

Pooled SD of the two groups = = = 0.2712
n1 + n2 − 2 50 + 48 − 2
( x1 − x 2 ) − 0 ( x1 − x 2 ) (1.64 − 1.54)
t= = = = 1.8251
SE(x1 − x2 ) 1 1 1 1
s + 0.2712 +
n1 n2 50 48

Degrees of freedom = n1 + n2 – 2
= 50 + 48 – 2
= 96
According to the t-distribution table for 96 degrees of freedom, p >0.05.
H0 is therefore accepted at the 5% level and we conclude that there is no difference in the mean FEV1 between
the two groups at 6 months.
Reproduced from Petrie and Sabin (2005).

PAIRED t-TEST
Used for matched studies (e.g. case−control trials and matched-pair RCTs):
1. Calculate the mean of the differences
2. Find the standard deviation of the differences
SD
3. Calculate the standard error of the mean =
n
4. t = Mean of the differences ∏ Standard error of mean of the differences
5. Look up t value (using degrees of freedom = number of pairs - 1) in a t-table to find the p value.

z-SCORES
An observation from a population may be converted into a standard normal deviate (also called a z-score or normal
score). This is the number of standard deviations that separate the observation from the mean of the population. It
is calculated by subtracting the population mean from an individual (‘raw’) score and then dividing the difference by
the population standard deviation.
The z-score reveals how many units a case is above or below the mean. The z-score allows comparisons to be made
between the results of different normal distributions.

91

Public [Link] 91 3/6/08 [Link]

 1B S t a t i s t ic s

x −µ
z=
σ

where x = raw score to be standardised, m = population mean, s = population standard deviation.

The z-test is the generalisation of the t-test for larger numbers, where n >60. Both t-tests and z-tests are used to
compare two groups. When more than two groups need to be compared, an analysis of variance (ANOvA) is used –
see below.
In the same way that the t-test can be used to examine differences in the means of two smaller populations, the
z-test may be calculated as follows:

Difference in means x − x0
z= + 1
SE of the difference in
n the means s12 s02
+
n1 n0

The test statistic for the paired situation is:

x
z=
s n

Z-TEST FOR THE DIFFERENCE BETWEEN TWO PROPORTIONS

When testing for a difference in proportions, assume the null hypothesis that both groups come from the same
population and have the same proportions (P) of the characteristic of interest.
1. Combine the two groups to find the overall proportion, P
P(1 − P ) P(1 − P )
2. Standard error (difference) = +
n1 n2
3. Calculate the difference in proportions between the two groups (P1 - P2)
4. z = Difference in proportions ∏ Standard error of the difference
5. Look up z value in the normal distribution table to find the p value (note that this is identical to the chi-
squared test).

CHI-SQUARED TEST
This test can be used only for actual numbers (not for proportions, percentages, etc.). Note that the Greek letter
(c2) is used for the test and distribution, but that the Latin letter (X2) is used for the calculated statistic.
1. For each observed number calculate the expected number
2. Subtract the expected number from the observed number (O - E)
3. Square the result and divide this by the expected number (O - E)2 ∏ E
4. X2 = total of these results for all cells, i.e. sum of (3)
5. Look up X2 (using degrees of freedom = [rows - 1] ¥ [columns - 1]) to find the p value.
An example is shown in Box 1B.12.4.

92

Public [Link] 92 3/6/08 [Link]

 1B Statistics

Box 1B.12.4
Example: Chi-squared test
The paté at a restaurant is implicated in an outbreak of listeriosis.
Here are the observed numbers of guests who ate paté, and the numbers who were ill with listeriosis:

Listeriosis
Yes No
Ate paté Yes 12 24 36
No 42 48 90
Total 54 72 126
First calculate the expected number for each cell assuming that paté was not linked to listeriosis, e.g. 36 ¥ 54 ∏
126 = 15.43

Listeriosis
Yes Yes
Ate paté Yes 15.43 20.57 36
No 38.57 51.43 90
Total 54 72 126
(O − E )2
Now calculate for each cell
E

Listeriosis
Yes Yes
Ate paté Yes 0.76 0.57
No 0.31 0.23

(O − E )2
Now calculate Σ E
= 0.76 + 0.57 + 0.31 + 0.23

= 1.87
Note that statistical tables are not provided in the Part A examination. However, for 2 ¥ 2 tables (i.e. 1 degree
of freedom), the cut-off point for the 95% significance level is 3.84 (i.e. 1.962). As 1.87 is <3.84, we can
conclude that eating paté was not significantly associated with listeriosis at the p<0.05 level.

ANOvA
The same principles as the t-test (and z-test) are used in the analysis of variance (ANOVA), which is employed when
more than two groups need to be compared. In this situation, an f-test is used and the null hypothesis is that the
means of all the groups of observations are equal.
ANOVA is called one-way analysis of variance. For example, if three groups of animals are treated separately with
hormone 1, hormone 2 and saline solution (control) to assess whether there are differences in the resultant growth,
then the ‘factor’ concerned here is hormone and this is a ‘one-way’ or ‘one factor’ ANOVA.

93

Public [Link] 93 3/6/08 [Link]

 1B S t a t i s t ic s

Two factors can also be tested with ANOVA. Hormones may affect male animals to a greater or lesser extent than
females. By selecting samples of females and samples of males for each treatment, sex then becomes ‘factor 2’.
Comparisons could then be made between the different hormones and between the different sexes.
A second extension of one-way ANOVA is made when comparing two dependent variables simultaneously across two
or more groups. This extension is called multivariate analysis of variance (MANOvA). For example, the means of
dependent variables (reading, writing, IQ, maths) may be tested across two groups (males, females).
An example is shown in Box 1B.12.5.
Analysis of variance is a special type of regression analysis, and most data sets for which analysis of variance
is appropriate can be analysed by regression with the same results. With two groups one-way ANOVA is exactly
equivalent to the usual two-sample t-test, and we have f = t2.

WILCOxON’S TEST
This is used for paired data with differences, when plotted, that look roughly symmetrical.
1. Find differences between individual pairs
2. Omit zero values
3. Ignoring the signs (+ or -) for the moment, rank the differences in order, placing identical values halfway
between the ranks that they would have occupied if they were unique
4. Reapply the signs (+ or -)
5. Find the sum of the positive ranks and the sum of the negative ranks (these are called the ‘rank totals’)
6. Ignoring the signs again, take the smaller ‘rank total’ from (5)
7. Look up this ‘rank total’ in the Wilcoxon table (either the 1% or the 5% section) according to the number of
pairs in the sample
8. If the ‘rank total’ is larger than the number in the table then the result is insignificant at that p value.

MANN−WHITNEY U-TEST
This is used for unpaired data: where the data do not appear symmetrical.
1. Rank the results from both groups in a single list (writing each in a different colour)
2. Add up the ranks for the two samples separately (by reading off the different colours): these are the ‘rank
totals’
3. Use the smaller of the ‘rank totals’
4. Look up in the Mann–Whitney table (either the 1% or the 5% section) using n1 as the number of observations
in one group and n2 as the number of observations in the other
5. If the number from (3) is larger than the number found in (4), then the result is insignificant at this p value.

INTER-RATER RELIABILITY
A common situation in epidemiology arises when two people measure the same phenomenon or object. The kappa
statistic takes account of the possibility of such chance agreement, and indicates how closely two different
measurements (binary or nominal) are aligned. Table 1B.12.1 is a contingency table for the kappa statistic.

94

Public [Link] 94 3/6/08 [Link]

 1B Statistics

Box 1B.12.5
Example: one-way ANOvA
A researcher predicts that students will learn most effectively with a constant background sound, as opposed
to an unpredictable sound or no sound at all. Twenty-four students are divided at random into three groups of
eight. All students study a passage of text for 30 min. Those in group 1 study with sound at a constant volume
in the background. Those in group 2 study with noise that changes volume periodically. Those in group 3 study
with no sound at all. After studying, all students take a 10-point multiple choice test on the material.
• Null hypothesis H0 = no difference between the groups exposed to constant, random or no sound
• Alternative hypothesis H1 = there is a difference between the groups

Group Test scores

Constant sound 7 4 6 8 6 6 2 9
Random sound 5 5 3 4 4 7 2 2
No sound 2 4 7 1 2 1 5 5
1. Calculate the total sum of squares for all the scores, x; and the participants, n:
(Âxi)2
SStotal = Âxi2 –
n
2. Calculate the sum of squares due to the difference between the observations:
(Âx)2
SSbetween = Ânixi2 +
n
3. Calculate the sum of squares f within groups
SSwithin = SStotal – SSbetween

(48 + 32 + 27)2
SStotal = (322 + 148 + 125) − = 117.96
24
 2304 1024 729  (48 + 32 + 27)2
SSbetween = + + − = 30.8
 8 8 8  24

4. Calculate (a) the mean square for SSbetween, where df = No. of groups –1
SS
MSbetween = between
df
(b) The mean square for SSwithin, where df = No. of participants – No. of groups
SSwithin
MSwithin =
df

Source SS df MS f
Between 30.8 2 15.4 3.68
Within 87.88 21 4.18
SSwithin = 117.96 - 30.8 = 87.88

Table contd overleaf

95

Public [Link] 95 3/6/08 [Link]

 1B S t a t i s t ic s

Box 1B.12.5 contd

5. Calculate the f-test statistic
MSbetween
f=
MSwithin
f = 3.59
With df = (2,21), reference tables tell us that f must be at least 3.4668 to reach p <0.05, so the observed f score
is statistically significant.
Adapted with permission from Hall (2007).

Table1B.12.1 Contingency table for kappa statistic

Observer 1
No Yes Total
Observer 2

No a b a+ b

Yes c d c+d

Total a+c b+d a+b+c+d

(a + c )(a + b) + (b + d )(c + d )
Ie = Predicted agreement =
(a + b + c + d )2
a+d
Io = Observed agreement =
a+b+c +d
I −I
κ= o e
1 − Ie

The agreement due to chance varies according to the proportions of results that are reported as positive or negative.
The kappa value is interpreted as shown in Box 1B.12.6.

Box 1B.12.6
Kappa value Degree of agreement beyond chance
0 None
0−0.2 Slight
0.2−0.4 Fair
0.4−0.6 Moderate
0.6−0.8 Substantial
0.8−1.0 Almost perfect
Kappa has some limitations:
• Using kappa to assess agreement assumes that the raters are independent
• It is useful only when raters are using the same rating scale
• It tells the reader nothing about the reasons for variation.
An example of the use of the kappa statistic is shown in Box 1B.12.7.

96

Public [Link] 96 3/6/08 [Link]

 1B Statistics

1B.13 SAMPLE SIZE AND STATISTICAL POWER

Sample size and power calculations allow the researcher to decide during the design phase of an experiment:
• How large a sample is needed to obtain accurate and reliable results
• How likely it is that the statistical test to be used will detect effects of a given size in a particular situation.

Box 1B.12.7
Example: kappa statistic
Two doctors report results of 29 patients. They agree with each other in
22 cases (75.9%), i.e. 10 + 12. From the table below, we can see that the
resultant kappa statistic is 0.542, representing moderate agreement.
Doctor A
No Yes Total
No 17 (58.6%)
Doctor B

10 (34.5%) 7 (24.1%)

Yes 0 (0.0%) 12 (41.4%) 12 (41.4%)

Total 10 (34.5%) 19 (65.5%) 29

Io = 0.759
Ie= 0.473
kappa = 0.542

SAMPLE SIZE
Sample size calculations are a key part of the study design because they ensure that studies have sufficient
participants to answer the question posed but not so many so that resources are wasted unnecessarily. In general,
there are four factors to consider when calculating the required sample size: see Table 1B.13.1.

Table 1B.13.1 Factors in sample size calculations

Factors Effect on sample size
Size of difference Smaller effect size requires larger
• Effect size needed for result to be clinically/experimentally meaningful sample
Significance level Smaller p requires larger sample
• p value, i.e. type I error
• Usually 0.05 is used
Power Higher power requires larger sample
• Probability that the study will be able to detect a difference if it really
exists: related to type II error (1 - b)
• Usually a power of 80% is used
Exposure in baseline population
• For case−control study: exposure in controls
• For cohort/intervention: disease rate

97

Public [Link] 97 3/6/08 [Link]

 1B S t a t i s t ic s

POWER
The power of a study to detect a true effect is generally set at 80% or more. If the power of an experiment is low,
then there is a high probability that the experiment will be inconclusive. Power can be increased at the expense of
significance. In case−control studies, power can be increased for the same number of cases by increasing the ratio of
controls to cases, although this too will increase the overall sample size.
Different methods are used for estimating power in the various study designs. In all cases, power should be
calculated in advance, and then a sample of appropriate size is recruited. However, for logistical or financial reasons
this is not always possible. In the case of very rare conditions, for example, the sample size may be fixed. It can
then be used to calculate the power of the study (i.e. to assess the likelihood of detecting a statistically significant
effect).
Meta-analyses can be conducted to pool the results from several studies and thereby increase the power of detecting
a finding if one exists (see Section 1A.33).

OTHER FACTORS TO CONSIDER

The sample size should be increased when:
• Loss to follow-up or a low response rate is expected
• Cluster sampling is used (see Section 1A.21)
• Confounding is expected (i.e. several variables will need to be controlled for)
• Interaction is expected.
The sample size should be decreased when:
• Matched case controls are used
Note that sample size can reduce only any errors caused by chance: it cannot compensate for bias.

1B.14 REGRESSION AND CORRELATION

Regression and correlation are related statistical techniques. Both examine the relationships between two or more
variables, but they do so in different ways.

REGRESSION
Regression is the process of deriving an equation for the best fitting line through the points on the scatter diagram.
This line can be found by minimising the squared distances between points and the line (known as the least
squares technique). The regression equation can be used to determine by how much one variable (y) changes when
another variable (x) changes by a certain amount. However, the equation is valid only between the limits of the
scatter diagram: to extrapolate beyond this may be invalid. Regression can be used for explanatory or predictive
purposes.

MULTIPLE REGRESSION
Multiple regression is used for continuous variables:
• Assume that the line can be described by its intercept (c) and slope − called the ‘regression coefficient’ (m)
• If necessary, divide up the line into multiple components: y = c + mx  y = c + m1x1 + m2x2 + m3x3
• The coefficients (mi) are calculated using the least squares technique
• The size of each coefficient (mi) represents the effect size for that variable.

98

Public [Link] 98 3/6/08 [Link]

 1B Statistics

LOGISTIC REGRESSION
Logistic regression is used for binary variables:
• Take natural log of odds (called the ‘logit’) which will always be between 0 and 1
• ln (odds of disease) = (y/(1 - y)) = c + m1x1 + m2x2 + m3x3.
Tests for the significance of the slope give identical results to tests for the significance of the correlation
coefficient.

CORRELATION
Correlation estimates the strength of any linear association between two variables. Unlike regression, correlation
is bidirectional, i.e. when one variable (x) changes by a given amount, so the other variable (y) changes by the
proportional amount.
The strength of the association can be represented by Pearson’s correlation coefficient (r) using a formula based
on the sample size and the values of x and y.
The value of r can vary from -1 through 0 to +1. Box 1B.14.1 explains the interpretation of these values.

Box 1B.14.1
r Implication
+1 Perfect positive correlation
Positive values When one value increases, the other variable increases
0 No correlation
Negative values When one value increases, the other variable decreases
−1 Perfect negative correlation

The association can be tested using a form of the t-test (based on r and n) at degrees of freedom = n - 2:

n−2
t=
1 − r2
Pearson’s correlation coefficient is a parametric statistic. Non-parametric correlation methods, such as Spearman’s r
(rho), are used when distributions are not normal.

1B.15 REGRESSION TECHNIQUES

Appropriate use, objectives and value of multiple linear regression, multiple logistic regression, principles of life-tables
and Cox’s regression
Table 1B.15.1 summarises the usage of the different regression techniques.

REGRESSION LINE EQUATIONS

Simple linear regression: y = a + bx
Multiple linear regression: y = a + b1x1 + b2x2 +, etc
where a = constant, x = explanatory variable, y = intercept, b = gradient (regression coefficient, i.e the increase in y
that corresponds to a unit change in x).

99

Public [Link] 99 3/6/08 [Link]

 1B S t a t i s t ic s

LIFE-TABLES
See Section 3A.6.

Table 1B.15.1 Regression techniques

Univariate regression Appropriate use Objectives
methods
Linear regression Describe relationship Provides a line of best fit through data
between two numerical points.
variables
Calculate relationship between an
Estimated using statistical outcome and an explanatory variable:
technique of fitting straight the correlation coefficient, r. Values
line (line of best fit) to range from r = -1 (perfect negative
scatter plot of variables correlation), through r = 0 (no
using the method of least correlation) to r = 1 (perfect positive
squares correlation)
Multivariate regression Appropriate use Objectives
methods
Multiple linear Analysis of several Adjusts for confounding
regression explanatory variables with
Understand which variables are
one outcome variable
associated with an outcome
Multiple logistic Analysis of several Adjust for confounding
regression explanatory variables with
a binary outcome variable
Cox’s regression (see Analysis of the effect of
Section 1B.16) several variables upon the
time it takes for a specified
event to happen

1B.16 COMPARISON OF SURvIvAL RATES

Two methods can be used to generate survival functions and survival curves: see Box 1B.16.1.

Box 1B.16.1
Life-tables Lead to smooth survival curves (see Section 3A.6)
Kaplan−Meier estimates Produce stepped survival curves
Used when the time of an event and follow-up time
are known (‘censoring’)

Differences in survival rates between two groups can be apparent from survival curves, but quantifying the
differences requires statistical methods.

LOG-RANK TEST
Comparison of life-tables can be made visually by constructing survival curves, or assessed quantitatively by the
log-rank test. This is a special application of the Mantel−Haenszel chi-squared procedure, and is carried out by

100

Public [Link] 100 3/6/08 [Link]

 1B Statistics

constructing a separate 2 ¥ 2 table for each interval of the life tables to compare the proportions dying during each
interval.
The log-rank test is used to test the null hypothesis that there is no difference between the populations in the
probability of an event (here a death) at any time point. The analysis is based on the times of events (such as
deaths). For each such time, the observed number of deaths in each group is calculated, together with the number
that would be expected if there were no difference between the groups.
The log-rank test is most likely to detect a difference between groups when the risk of an event is consistently
greater for one group than another. It is unlikely to detect a difference when survival curves cross, as can happen
when comparing a medical with a surgical intervention. When analysing survival data, the survival curves should
therefore always be plotted.
Because the log-rank test is purely a test of significance, it cannot provide an estimate of the size of the difference
between the groups, nor a confidence interval. For these some assumptions must be made about the data. Common
methods used are the hazard ratio, including the Cox proportional hazards model.

COX’S REGRESSION
This is also known as proportional hazards regression (see also Section 1B.15).
This form of multivariate regression is based on the proportional hazards assumption, i.e. that the ratio of hazards
(the instantaneous risk of dying at time t) in both groups remains the same. It is a non-parametric technique that
makes full use of survival data without making assumptions of survival curve shape. It adjusts for confounders, i.e.
it tests the effects of a number of explanatory variables on the hazard. Cox’s regression can assess only the effects
of one factor on the time to endpoint.
Relative hazard is interpreted in a similar way to a relative risk (i.e. a value of 1 indicates no difference; values >1
indicate a raised hazard; values <1 a decreased hazard).

λi (t ) = λ0 (t )exp {β1 x1 + β2 x2 + … βn x n }
λi (t ) = hazard for individual i at time t

where λ0 (t ) = baseline hazard (not usually interested in), b = coefficients, x = covariates associated with individual
i.

1B.17 HETEROGENEITY
Heterogeneity refers to differences between observations that would lead an investigator to consider that the
observations might have been drawn from different (i.e. heterogeneous) populations. It is often used in a
systematic review to assess whether the pooling of data is reasonable and how meta-analyses should be analysed
(see Section 1A.33). There are three types of heterogeneity: statistical, methodological and clinical. See Table
1B.17.1.

1B.18 FUNNEL PLOTS

The funnel plot (Figure 1B.18.1) is a type of scatter plot that is commonly used to visualise potential publication
bias in meta-analyses (Egger 1997) (see Section 1A.33). It is also increasingly being used in other situations, such
as comparing performance between several organisations.
For a meta-analysis, the estimated treatment effect values from the component studies are plotted against a
measure of precision (study size or standard error). Appropriate confidence intervals are then added. As the sample
sizes of the component studies increase, so the precision of their estimated treatment effect increases. This gives

101

Public [Link] 101 3/6/08 [Link]

 1B S t a t i s t ic s

the graph its funnel shape. So, in the absence of bias, the results from small studies should scatter widely at one
end of the graph, with the spread narrowing among larger studies.

Table 1B.17.1 Types of heterogeneity

Statistical This relates to differences in the reported effects between studies, calculated using the
heterogeneity chi-squared test for heterogeneity. Statistical heterogeneity can be accounted for by
using a random effects model, as opposed to a fixed effects model in the analysis (see
Section 1A.33)
Methodological This is due to differences in study design and presents comparability problems for meta-
heterogeneity analysis
Clinical heterogeneity This relates to differences between studies relating to:
• characteristics of the patient population
• interventions, or
• outcome measures and
• length of follow-up
As with methodological heterogeneity, clinical heterogeneity may preclude the pooling
data from these studies

Potential outlier
Effect

Confidence
intervals

Sample size

Figure 1B.18.1 Funnel plot

INTERPRETING FUNNEL PLOTS

Funnel plots demonstrate any biases graphically by means of the shape of the plot, although formal statistical tests
can also be used.
If the plot of a meta-analysis is symmetrical, then this suggests that there is no publication bias. If the plot is
asymmetrical then publication bias is a possibility. However, asymmetry could also be due to small study effects (the
tendency for the smaller studies in a meta-analysis to show larger treatment effects).
In a plot of performance (e.g. surgical complication rates) those institutions plotted outside the confidence
interval envelopes are revealed as outliers who may warrant further investigation.

102

Public [Link] 102 3/6/08 [Link]

 1B Statistics

1B.19 ROLE OF BAYES’ THEOREM

Bayesian methods incorporate prior beliefs into calculations of probability. In real clinical situations, for example,
existing knowledge about a particular patient will affect how much credence clinicians place on a laboratory test
performed on that patient. Bayesian methods incorporate this prior knowledge into the probability calculations.
According to Bayes’ theorem, the probability of A occurring, given B, depends on three factors. See Table 1B.19.1.

Table 1B.19.1 Bayesian probability

Factor Alternative Symbol Description
name
Probability of A Prior P(A) Probability of A occurring on its own,
irrespective of B
Probability of B Normalising P(B) Probability of B occurring on its own,
constant regardless of A
Probability of B, Likelihood P(B|A) Probability of B occurring given that A
given A occurred
Using these three measures, the probability of A occurring given that B occurred is computed as:

P(B A)P(A)
P(A B) =
P(B)

Advantages and disadvantages of bayesian statistics compared with classical statistics are listed in Box 1B.19.1.

Box 1B.19.1
Advantages More flexible
Makes use of all available knowledge, therefore possibly
more ethical
Mathematics is not controversial
Disadvantages Different users will obtain different conclusions if they
choose different priors

DIAGNOSTIC TEST IN A BAYESIAN FRAMEWORK

Posterior odds of disease = prior odds ¥ likelihood ratio of a positive test result

Prior probability
Prior odds =
1 − Prior probability
Sensitivity
Likelihood ratio =
1 − Specificity
Posterior odds
Posterior probability =
1 + Posterior odds

An example of the use of bayesian statistics is shown in Box 1B.19.2.

103

Public [Link] 103 3/6/08 [Link]

 1B S t a t i s t ic s

Box 1B.19.2
Example: bayesian statistics for a diagnostic test
Likelihood ratio for a positive cytomegalovirus (CMV) test = 13.3
Prevalence of CMV infection after bone marrow transplantation ~ 33%
Prior probability of severe disease = 0.33

Prior probability 0.33

Prior odds = = = 0.493
1 − Prior probability 0.67
Posterior odds = 0.493 × likelihood ratio = 0.493 × 13.3 = 6.557
6.557
Posterior probability = = 0.868
1+6.557

87% chance of developing severe disease with CMV after transplantation, which is greater than the pre-test
probability, indicating usefulness of the test.
Reproduced from Petrie and Sabin (2005).

1B.20 CHOICE OF STATISTICAL TEST

Figures 1B.20.1–1B.20.4 illustrate appropriate tests for particular types of outcome variables. Reproduced with
permission from Kirkwood and Sterne (2003).

104

Public [Link] 104 3/6/08 [Link]

 Two exposure groups
Measurement� mean Two exposure groups Ordered or numerical Multiple exposure
(difference between More than two groups
value (paired measurements) exposure variable variables
means)

Public [Link] 105

Large Large samples
samples �z- test (tests based on Multiple
Paired tests One-way ANOVA Linear regression
(based on normal regression
normal distribution) distribution)

Smaller samples, Correlation

Small samples – equal standard Multiple (Spearman’s rank/ Two-way ANOVA
‘t ’- test deviations regression Kendall’s tau)
(‘t ’-test)

Figure 1B.20.1
Linear regression Standardisation Numerical outcome
variable

105
Ordered or
Single exposure Two exposure More than two Multiple exposure
numerical Survival analysis
group groups exposure groups variables
exposure variable

Poisson’s
Confidence Poisson’s Mantel–Haenszel
Rate difference regression with Cox’s regression
interval for a rate regression method
indicator variables

Survival analysis Cox’s regression

• Life-table Poisson’s
Rate ratio with indicator Cox’s regression
• Kaplan–Meier regression
variables
estimates

Direct
Poisson regression
standardisation
1B

Survival analysis
Indirect
• Hazard ratio Figure 1B.20.2
standardisation
Statistics

• Cox’s regression
Rates and survival times

3/6/08 [Link]
 Ordered or
1B

Single exposure Two exposure More than two Multiple exposure Paired or matched
numerical
group groups exposure groups variables measurements
exposure variable

Public [Link] 106

S t a t i s t ic s

Difference in Chi-squared test Chi-squared test Mantel–Haenszel Comparison of two

Proportion
proportions for r ¥ 2 tables for trend method proportions

Test that the Logistic regression Odds ratios and

proportion = a Risk ratio with indicator Logistic regression Logistic regression McNemar’s chi-
particular value variables squared test

Regression analysis Conditional logistic

Odds Odds ratio
of risk ratios regression

Direct
Chi-squared test
standardisation of
for 2 ¥ 2 tables
proportions

Fisher’s exact test

106
Figure 1B.20.3
Logistic regression Binary outcome
variable

Categorical outcome with more

than two levels

Multinomial
Chi-squared test Ordinal logistic
logistic
for larger tables regression
regression

Figure 1B.20.4 Categorical outcome with more than

two levels

3/6/08 [Link]
 1C Assessment and Evaluation

1C
Assessment and Evaluation

1C.1 Need for health services 107 1C.10 Equity in health care 126
1C.2 Participatory needs assessment 110 1C.11 Clinical audit 128
1C.3 Service utilisation and performance 110 1C.12 Confidential enquiry processes 129
1C.4 Measures of supply and demand 113 1C.13 Delphi methods 130
1C.5 Study design 113 1C.14 Economic evaluation 131
1C.6 Structure, process and outcomes 116 1C.15 Appropriate and acceptable services 131
1C.7 Measuring health 119 1C.16 Epidemiological basis for preventive
1C.8 Deprivation measures 121 strategies 133
1C.9 Evaluation 125 1C.17 Health and environmental impact
assessments 135

Approaches to the assessment of health-care needs, utilisation and outcomes, and the evaluation of health and health
care

Just because a particular health service is provided, it does not necessarily mean that it is effective, or that it
is appropriate for the population served. Public health has an important role in assessing the need for a service,
and the quality and appropriateness of what is currently provided. Practitioners need both knowledge of research
methods and practical skills in order to:
• Measure patients’ health and illness, quality of life and living conditions
• Study the nature of health services
• Involve relevant practitioners, patients and communities in making assessments and effecting improvements.

1C.1 NEED FOR HEALTH SERvICES

Uses of epidemiology and other methods in defining health service needs and in policy development
In public health, the generic term ‘need’ is used to indicate a number of concepts relating either to the illnesses
that people experience (need for health) or to their treatment (need for health care). Bradshaw (1972) described
four specific types of need: see Box 1C.1.1.

107

Public [Link] 107 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

Box 1C.1.1
Type of need Description
Normative Needs as deemed by a clinician
Felt Requirement of patients to feel better
Expressed Demand, e.g. visits to a GP
Comparative Need in one area compared with that in another

Strictly speaking, people are only said to ‘need’ interventions that can benefit them. Therefore, if an intervention
offers no apparent clinical benefit, then there can be no clinical need for it.
A key role of a public health practitioner is to assess the health needs of a given population. When conducted
systematically, this process is known as a health needs assessment (HNA). The findings of an HNA can be used to
guide the allocation of resources in order to improve the health of the population and to reduce health inequalities.
It offers an opportunity to:
• Consult the population
• Cultivate cross-sectoral partnerships
• Develop new interventions
• Ensure that health-care provision is evidence based.

CORPORATE NEEDS ASSESSMENT

In this type of HNA, the practitioner considers the views of interested parties, and aims to tailor local provision of
health care according to these opinions. See Box 1C.1.2.

Box 1C.1.2
Advantages Disadvantages
• Being an incremental process, it allows services • Can be driven by power rather than by need
to be altered a little at a time, guided by feedback
• Can be disproportionately influenced by the
from interested parties
reaction of participants to certain events, e.g.
• Can be done quickly newspaper headlines
• No need to collect many data • An incremental process may be inappropriate where
large-scale or radical reform is required
• Responsive to interested parties

COMPARATIVE NEEDS ASSESSMENT

This process uses data from surveys or hospital activity data. It compares findings observed locally with those
that would be expected from a reference population (e.g. regional or national data). For the comparisons to be
valid, adjustments to the reference population need to be made in order to reflect the characteristics of the local
population (e.g. age, ethnicity, smoking rates, etc.). Note that local prevalence may not be known, e.g. many cases
of chlamydia infection will be undetected.

EVIDENCE-BASED NEEDS ASSESSMENT

Evidence from the literature – ideally in the form of guidelines or consensus statements – should be used to shape
local provision of health care. Provision should be adjusted to take account of local circumstances such as the

108

Public [Link] 108 3/6/08 [Link]

 1C Assessment and Evaluation

prevalence of a disease in the local population and the findings of patient surveys. Note that the literature may not
exist or it may be conflicting.

DATA SOURCES FOR CONDUCTING A NEEDS ASSESSMENT

The data to estimate need may relate to activity or epidemiology (prevalence/incidence): see Table 1C1.1.

Table 1C.1.1 Data for a needs assessment

Type of data Problems with data sources
Activity Data sources such as hospital episode statistics (HES) underestimate need since they do not
include:
• Unmet need (i.e. people who do not seek treatment)
• People who seek private treatment
Epidemiology Where local epidemiological data exist, they may overestimate need if it is assumed that
prevalence equals need. This is because some of the people with the condition:
• Do not need treatment (e.g. hip patients not fit to undergo surgery)
• Do not want treatment
• May have already had treatment
Some conditions that exist on a spectrum require an arbitrary cut-off (e.g. blood pressure)

STEPS INVOLVED IN A NEEDS ASSESSMENT

The UK’s Health Development Agency (now part of the National Institute for Health and Clinical Excellence) outlined
a five-step approach to health needs assessment, as shown in Table 1C.1.2.

Table 1C.1.2 Steps involved in a needs assessment

1. Scope Identify:
• Population to be considered
• Aims of the HNA (health needs assessment)
• Stakeholders that will be involved
• Resources needed to conduct the HNA
• Risks involved
2. Identification of potential priorities Gather data to describe the population (activity data, epidemiology
and opinions from surveys)
3. Selection of a priority Choose a disease according to its health burden
Select interventions that are effective and acceptable
4. Change Specify the aims of the intervention. Use change management
techniques (action planning, monitoring, risk management)
5. Review Learn from the project by measuring impact, then disseminate
findings and choose the next priority
Adapted from Cavanagh and Chadwick (2005).

109

Public [Link] 109 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

1C.2 PARTICIPATORY NEEDS ASSESSMENT

HNAs are traditionally conducted by commissioners of health services or by a third party. As a result, the local
population may sometimes be reluctant to accept its findings and conclusions. In order to improve engagement, the
local community should participate in the HNA process. For this to occur effectively, the participatory HNA requires:
• Clear objectives
• Use of accepted methods and data sources
• Support from appropriate experts
• Effective communication techniques
• Views of groups whose voices are not normally heard to be taken into account during the HNA (e.g. ethnic
minorities, older people, children, young mothers and transient populations)
• Involvement of the community in all steps of the HNA process.
Methods in this type of HNA often produce qualitative rather than quantitative data. Such methods of data
generation may include:
• Key informant interviews
• Group workshops
• Focus groups
• Visual methods (e.g. mapping or ‘transect walks’ where groups are given disposable cameras to take pictures
that illustrate their needs).
If done well, the participatory process will identify potential needs that service providers would not have recognised
without the input of community members.

1C.3 SERvICE UTILISATION AND PERFORMANCE

Formulation and interpretation of measures of utilisation and performance
With such a high proportion of the GDP being spent on health care, it is incumbent on health services to account for
how this money is spent (who is accessing a service), and to demonstrate that it is achieving an impact (the extent
to which a service is meeting its stated aims).
Health service utilisation statistics are usually used to measure utilisation of health services, but other dimensions
of performance include:
• Quality
• Safety
• Effectiveness
• Patient satisfaction
• Waiting times.

WAITING TIMES
Waiting times are indicators of barriers to the use of services. They reflect poor responsiveness of services in
meeting demand, and may be due to:
• Insufficient capacity
• Poorly designed care pathways that fail to manage demand (either referring inappropriately or referring too
early), or are unnecessarily long and complex
• Failure to deploy resources in line with fluctuations in demand.
In many health-care systems, targets have been introduced to reduce these barriers. Box 1C.3.1 gives examples of
waiting time targets in England.

110

Public [Link] 110 3/6/08 [Link]

 1C Assessment and Evaluation

Eng Box 1C.3.1

Service or procedure Target maximum wait
Emergency A&E: maximum waits of 4 h
Ambulances to respond to 75% of category A calls within 8 min
Cancer Urgent GP referral to treatment: 1 month
Urgent GP referral to first outpatient appointment: maximum wait
of 2 weeks
Scheduled operations Maximum wait of 6 months for inpatients
Outpatient appointments Maximum wait of 3 months
Primary care Percentage able to offer GP appointment within 48 h of the
request

Reproduced from Department of Health (2004), National Standards, Local Action: Health and social care standards
and planning framework 2005/06–2007/08, available online at [Link].

ACTIVITY
Where data are not routinely available but need to be collected specifically, tools can be used to ensure consistent
data collection. These tools include those shown in Box 1C.3.2.

Box 1C.3.2
Minimum data A minimum amount of information is collected and collated for each patient using the
sets service (e.g. patient’s postcode, age, sex, ethnicity and reason for seeking care)
System prompts GPs’ clinical systems can automatically prompt them to ask patients about relevant health
services (e.g. mothers whose young children are due for routine immunisations)

Data sources, types of information collected and organisations in England that use the data are shown in Box
1C.3.3. Table 1C.3.1 gives (roughly) equivalent organisations across the different countries of the UK.

Eng Box 1C.3.3

Data source Type of information English organisations that use the data
Hospital episode Number of procedures (operations, Department of Health (DH): to shape strategic
statistics maternity services, psychiatric care) direction of health services and secure funding
taking place in NHS services, broken down
Healthcare Commission: to inspect health
by:
services
• Diagnoses
• Health-care resource groups Regional public health observatories: provide
• NHS trusts local information
• Length of stay
Strategic health authorities: performance
• Time waited
management of NHS trusts
• Admission methods
• Patients’ age, sex and ethnic group Primary care trusts (PCTs) and hospital
trusts: commissioning, monitoring contracts,
performance management
Box contd overleaf

111

Public [Link] 111 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

Box 1C.3.3 contd

Data source Type of information English organisations that use the data
PACT, ePACT Prescription data (cost quantity), mainly PCTs: monitoring and performance
from primary care management of primary care practitioners
Prescribing Support Unit at DH and National
Prescribing Centre (see Section 3B.5)
Stop-smoking Profile and numbers using services: Strategic health authorities: performance
services: numbers • General accessibility of service management
attempting to quit • Groups not accessing services
NHS trusts: shape service delivery
Coverage data: Percentage population offered services DH/strategic health authorities: performance
screening, receiving services management
immunisations
Percentage eligible population receiving NHS trusts: shape service delivery
services

UK Table 1C.3.1 UK equivalence table

England Scotland Wales Northern Ireland
Department of Health Scottish Executive Health Welsh Assembly DH and Social Services
Department Government
Healthcare Commission NHS Quality Improvement Healthcare Inspectorate for Second review of public
Scotland Wales administration (RPA)
Regional Public Health Scottish Public Health Wales Centre for Health All Ireland Public Health
Observatories Observatory and National Public Health Observatory
Service Wales
Strategic health authorities NHS boards (responsible Regional Offices of the New single Health and
for hospitals, community Welsh Assembly Social Services Agency to
health services and healthy replace the existing four
improvement) health boards
Primary care trusts Local health boards Seven new local
Hospital trusts Hospital trusts commissioning groups to
be formed
Health Protection Agency HPA HPA HPA
(HPA)
Local and regional services Health Protection Scotland National Public Health Communicable Disease
of the HPA Service for Wales Surveillance Centre,
Northern Ireland

UK The following central elements of the HPA support all four countries of the UK:
• Centre for Infections (Colindale)
• Centre for Emergency Preparedness and Response (Porton Down)
• Centre for Radiological, Chemical and Environmental Hazards (Chilton).
In England, the HPA’s local and regional services (LaRS) provide surveillance, planning and health protection at

112

Public [Link] 112 3/6/08 [Link]

 1C Assessment and Evaluation

a local level, including local offices called health protection units. In Northern Ireland, Scotland and Wales, the
equivalent functions of LaRS are conducted by the Infection and Communicable Disease Service teams of the
respective national organisations detailed above. See Section 2G for more information about health protection
organisations and roles.

1C.4 MEASURES OF SUPPLY AND DEMAND

See Section 4D.1.

1C.5 STUDY DESIGN

Study design for assessing effectiveness, efficiency and acceptability of services, including measures of structure,
process, service quality and outcome of health care
A range of different methods can be used to evaluate the impact and cost-effectiveness of a health service. The
technique used will depend on the exact question being asked, the time-frame, level of evidence needed and the
budget for the study.
Examples include:
• Health technology assessment (HTA)
• Service delivery and organisation (SDO)
• Experimental study designs (RCTs, qualitative studies, surveys and documentary evidence)
• Comparative studies
• HNA.
Study data can either be routinely collected or else collected especially for the particular study. Examples are
described in Box 1C.5.1.

Box 1C.5.1
Routinely available data Specifically collected
Structure Staffing: human resources records Equipment audit
Finance: accounts, annual reports, receipts
Process Patient care: records, referral letters, treatment plans, Patient diaries
information leaflets
Outputs Activity: hospital episode statistics Clinical audit and clinical
research reports
Outcomes Mortality statistics National confidential enquiries
Health policy Media (TV programmes, newspaper articles) Interviews
Official records – parliamentary debates, bills, meeting
minutes

HEALTH TECHNOLOGY ASSESSMENT

This type of study asks what technology is best for meeting the needs of specific patients or procedures, and what
impact the health technology will have on:
• Outcomes (morbidity, mortality, quality of life, patient satisfaction)
• Structure (resources required to provide service).

113

Public [Link] 113 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

Common study designs include:

• Epidemiological study designs (see Section 1A.19)
• Economic evaluations (see Section 1C.14).

NICE Health Technology Appraisals are summarised in Box 1C.5.2.

Box 1C.5.2
NICE Health Technology Appraisals
Types of technologies that may be appraised include:
• Pharmaceuticals
• Medical devices
• Diagnostic techniques
• Surgical procedures
• Other therapeutic technologies
• Health promotion activities
Focus of NICE reviews:
• Clinical effectiveness
• Cost-effectiveness
• Acceptability (to patients and/or professionals)
• Feasibility
• Equity
Study designs used by NICE:
• Systematic review: evidence is rated according to the hierarchy of evidence (see below and Section 1A.37)
• Economic evaluation
• Appraisal: review of evidence in consultation with individuals, patient/carer groups, manufacturers, health
care professionals (e.g. through the medical Royal Colleges)
NICE Health Technology Appraisals make recommendations for when and whether the NHS should provide a
technology, and the expected impact on resources.
Reproduced from NICE (2004).

SERVICE DELIVERY AND ORGANISATION

Research in this field largely considers issues of health-care process, and how these will impact on outcomes: see
Box 1C.5.3.

Box 1C.5.3
Where Where is health care provided? Is the setting appropriate?
Who Who provides the health care? Could another professional provide care more cost-effectively? How do
different professionals work together?
How How is health care being delivered? Are there other more effective (or cost-effective) ways of providing
it? What factors facilitate or hinder changes to the way in which care is provided?

114

Public [Link] 114 3/6/08 [Link]

 1C Assessment and Evaluation

EXPERIMENTAL STUDY DESIGNS

While RCTs are considered the gold standard for many scientific studies, they have specific limitations when applied
to the local study of health service delivery and organisation: see Box 1C.5.4.

Box 1C.5.4
Resources Resources needed for a well-run RCT are usually unavailable for small-scale studies in health
services
Timescales Impact on health may not be apparent for decades, and the long-term follow-up needed
could be prohibitively expensive in an RCT
Changes in policy Changes in national policy may require changes to be made to local services
Multi-site studies Customisation of managerial interventions at a local level may make multi-site RCTs
impossible

Study designs other than RCTs may therefore be used, such as:
• Qualitative studies (interview, focus group, observation – see Section 1D)
• Surveys
• Documentary studies.
NICE ranks experimental studies according to the scale in Table 1C.5.1 (see also Section 1A.37):

Table 1C.5.1 Type and quality of evidence. The symbols are further indication of the quality of evidence, ++ is
stronger than + and – weaker than both ++ and +.
++
1 High-quality meta-analyses, systematic reviews of RCTs or RCTs (including cluster RCTs) with a very low risk
of bias
+
1 Well-conducted meta-analyses, systematic reviews of RCTs or RCTs (including cluster RCTs) with a low risk of
bias
–
1 Meta-analyses, systematic reviews of RCTs or RCTs (including cluster RCTs) with a high risk of bias
++
2 High-quality systematic reviews of non-RCT studies, or individual non-RCTs, case–control studies,
cohort studies, interrupted time series (ITS) and controlled before and after (CBA) studies with a very
low risk of confounding, bias or chance, and a high probability that the relationship is causal
+
2 Well-conducted non-RCTs, case–control studies, cohort studies, CBA studies, ITS and correlation studies
with a low risk of confounding, bias or chance, and a moderate probability that the relationship is causal
–
2 Non-RCTs, case–control studies, cohort studies, CBA studies, ITS and correlation studies with a high risk –
or chance – of confounding bias, and a significant risk that the relationship is not causal
3 Non-analytical studies, e.g. case reports, case series
4 Expert opinion, formal consensus
Reproduced from NICE (2006).

HEALTH SYSTEM COMPARISONS

Comparative studies consider variations in the structure, process and outcomes of health care and are particularly
useful for evaluating:
• Funding (see section 4). Examples include: contracting and commissioning systems

115

Public [Link] 115 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

• Performance. Examples include ‘benchmarking’ and assessments within health systems (e.g. Healthcare
Commission review, see Box 1C.5.5) and between systems (e.g. WHO, The world health report 2000 – Health
systems: improving performance 2000).

Eng Box 1C.5.5

Example: Healthcare Commission inspections of the NHS: annual health check
1. Online survey: NHS trusts complete a self-report of their progress in meeting core standards
2. Activities to cross-check organisations’ self-reports through:
3. Documents: supplied by the trust to support statements in the self-assessment
4. Surveys: patient survey, reports from partners, e.g. local authorities
5. Observations: visits made to a random selection of trusts
6. Performance against national targets: assessment uses data submitted to a range of organisations, e.g.
DH
7. Use of resources: local financial audits are used
Reproduced from [Link].

HEALTH NEEDS ASSESSMENTS

See Section 1C.1.

NATIONAL DISEASE AUDITS

This study design is helpful for evaluating the effectiveness of health care at a national level, as shown with the
example of the National Sentinel Stroke Audit (Box 1C.5.6).

UK Box 1C.5.6
Example: National Sentinel Stroke Audit
Conducted by the Royal College of Physicians of London, this is a biannual audit of stroke unit organisation. All
UK hospitals (except those in Scotland) that claim to operate a stroke unit are surveyed according to five core
criteria:
1. Consultant physician with responsibility for stroke
2. Formal links with patient and carer organisations
3. Multidisciplinary meetings at least weekly to plan patient care
4. Provision of information to patients about stroke
5. Continuing education programmes for staff
Reproduced from Hoffman et al (2004).

1C.6 STRUCTURE, PROCESS AND OUTCOMES

Measures of structure, process, service quality and outcome of health care
The performance of an activity may be assessed using the framework first described by Donabedian (1966): see Box
1C.6.1. This includes structure, process and output. An additional category of outcome is often added.

116

Public [Link] 116 3/6/08 [Link]

 1C Assessment and Evaluation

Box 1C.6.1
Donabedian’s framework
• Structure
• Process
• Output
• Outcome

STRUCTURE
This category consists of all of the inputs to the activity, such as:
• Staff
• Budgets
• Buildings
• Beds.
Advantages and disadvantages of considering this criterion are shown in Box 1C.6.2.

Box 1C.6.2
Advantage Disadvantage
Resource information is relatively Structural data may not be comparable
easy to measure between systems. For example, with regard to
staffing, all nurses may not be working at the
same level of responsibility or may not have
received the same level of training

PROCESS
These are the activities that constitute the intervention, and may be considered as follows:
• Strategies, plans and procedures
• Referral patterns
• Prescription practices
• Consultations (duration, number)
• Bed occupancy
• Waiting times.
Advantages and disadvantages are shown in Box 1C.6.3.

Box 1C6.3
Advantages Disadvantages
Relatively easy to obtain in centralised health- Processes do not necessarily predict health
care systems such as the NHS outcomes
Some processes are directly related to
outcomes, e.g. immunisation coverage

117

Public [Link] 117 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

OUTPUTS
These are the products of the activity. Examples include:
• Numbers of operations conducted
• Length of stay (an indicator of effectiveness)
• Waiting times (an indicator of access).

OUTCOMES
These are changes in health status that are attributable to the activity. Examples include:
• Death (mortality rates)
• Disability (and quality of life)
• Discharge (and complications, e.g. emergency re-admissions).
Advantages and disadvantages are shown in Box 1C.6.4. The challenges of using outcomes to evaluate health
services are illustrated with respect to smoking cessation in Box 1C.6.5.

Box 1C.6.4
Advantages Disadvantages
Ultimately, the aim of health service Not necessarily related to performance: affected by
interventions is to improve outcomes case mix, i.e. patient characteristics
Surrogate endpoints can indicate health Outcomes are relatively long term, so difficult to use
outcomes, e.g. CD4 counts in an HIV/AIDS in short-term trials
drug trial Often costly to collect or incomplete (apart from
mortality data – but these can be difficult to link to
interventions)

UK Box 1C.6.5
Example: stop-smoking services
Structure Numbers of staff employed as stop-smoking advisors
Budget for nicotine replacement therapy (NRT)/bupropion
Process Number of prescriptions of NRT or bupropion
Advisors’ training
Numbers of clients setting quit dates
Output Numbers of 4-week quitters
Outcome Ideally the morbidity and mortality from smoking-related
diseases. However, outcomes have to be attributable to the
intervention, and the effects of stop-smoking services on lung
cancer deaths may not be apparent for 20 years. Surrogate
endpoints could be used instead and include length of stay in
hospital for ex-smokers undergoing surgery

118

Public [Link] 118 3/6/08 [Link]

 1C Assessment and Evaluation

1C.7 MEASURING HEALTH

Measures of health status, quality of life and health care, population health outcome indicators
The broad definition of health used by public health practitioners means that specific definitions of individual
health status are required. At an aggregate level, a variety of health indicators is required to describe a population’s
health.

HEALTH STATUS
In 1980 the World Health Organization (WHO) published the International Classification of Functioning, Disability and
Health. This system (which distinguished impairment, disability and handicap) was updated in 2000 and is shown in
Table 1C.7.1.

Table 1C.7.1 World Health Organization classification of functioning, disability and health
WHO 1980 term Impairment Disability Handicap
WHO 2000 Term Body structure/function Activities Participation
Description Pathology and clinical Symptoms and health status Effect of disability on life
measures
Measurement Usually observed by Self-reported measure of Usually self-reported. This
clinicians or measured using what the person can do is the extent to which the
instruments condition affects the person’s
normal life
Examples of Blood pressure, temperature, Beck Depression Inventory Health-related quality of life
measures tumour size as assessed on
Activities of daily living
CT
scales
Example 1 Examination: enlarged Self-report: trips to the toilet Self-report: person cannot go
(prostate) prostate per day to cinema because they need
the toilet too often
Test: increased prostate-
specific antigen levels in
serum
Example 2 (hip) Clinical examination: reduced Self-report: pain or distance Self-report: isolation –
joint mobility that patient can walk cannot walk to the park
Investigation: joint
degeneration seen on
radiograph

HEALTH-RELATED QUALITY OF LIFE

Increasing importance is being placed on health-related quality of life (HRQoL) in clinical trials. This is because
for many potentially fatal conditions (e.g. lung cancer) or for many long-term conditions (e.g. epilepsy) treatments
may not affect life-expectancy but still have a profound impact on the quality of life. Studying HRQoL can also
reveal why a particular treatment is acceptable or not to patients. For example, a trial of cancer chemotherapy might
reveal that small treatment benefits (a few weeks’ extra life-expectancy) were outweighed by intractable nausea and
vomiting.

119

Public [Link] 119 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

Different theoretical models exist for measuring quality of life, and take into account the factors listed in Box
1C.7.1.

Box 1C.7.1
Expectations Differences between a person’s present experience and his
or her hopes and expectations. For example, a professional
pianist would be more affected by the loss of the use of a
finger than would a cleaner
Needs Ability or capacity of a person to satisfy basic human
functions, e.g. eating, sleeping, enjoying music
Normal living This is being able to do what the person wants to do (rather
than being free of disease or symptoms)

HRQoL scales contain several items that build up a composite picture of health status. Except in exceptional
circumstances (e.g. dementia or very young children), HRQoL scales should be completed by the patient rather than
by a carer or clinician.
Examples of health-related quality of life scales are listed in Box 1C.7.2.

Box 1C.7.2
Type of measure Examples
Generic Short Form 36 (SF36)
Nottingham Health Profile (NHP)
EQ5D (EuroQoL)
Disease related Functional Assessment of Cancer Therapy
Specific aspects of quality of Hospital Anxiety and Depression Scale;
life Multidimensional Fatigue Scale

MEASURES OF HEALTH CARE

Ways of assessing different measures of health care are described in Sections 1C.5 and 1C.6.
In addition to quality-of-life measures (above) and population outcome indicators (below), measures can also
include:
• Patient satisfaction (see Section 1C.15)
• Adherence to standards, compliance with guidelines and other established criteria (see Section 1C.11)
• Financial accountability (see Section 5).
See also Section 1C.9.

POPULATION HEALTH OUTCOME INDICATORS

These indicators measure the effect of health care on health status at a population level. They often draw on the
concept of avoidable mortality as an indicator of health-care performance. Assessment of avoidable mortality
involves studying the incidence of a collection of diseases with unnecessary or untimely outcomes.
Examples are listed in Box 1C.7.3.

120

Public [Link] 120 3/6/08 [Link]

 1C Assessment and Evaluation

Box 1C.7.3
Measure Disease and population Purpose
Standardised mortality ratios Diabetes: people aged under 45 Reduce premature deaths from
diabetes
Gastric, duodenal, and peptic ulcers: Reduce deaths from gastric, duodenal
people aged 25–74 and peptic ulcers
Procedures Lower limb amputations in diabetic Monitor and improve management of
patients diabetes
Termination of pregnancy Reduce the number of unwanted
pregnancies

Indicators reflect:
• Performance of health-care services
• Population characteristics.
Population health outcome indicators can be used to:
• Prompt the assessment of local health outcomes, particularly in relation to national targets
• Monitor variation in health care (particularly using small area analysis – see Section 1A.18)
• Monitor trends in health care, to answer questions such as: ‘Is effectiveness of health care improving?’ (see
Lakhani et al 2005)
• Monitor of quality of life as part of a population HNA.
Eng In England, data on population health outcome indicators are published in the Compendium of Clinical and
Health Indicators, produced by the National Centre for Health Outcomes Development.

1C.8 DEPRIvATION MEASURES

Deprivation measures describe a population according to social and economic disadvantage. They are used to
identify people or areas with the highest levels of disadvantage and deprivation, in order to target programmes and
resources, and may be measured at the individual or area level.

INDIVIDUAL INDICATORS
Indicators of deprivation at the individual or household level include:

UK NATIONAL STATISTICS SOCIOECONOMIC CLASSIFICATION (NSSEC)

The former system of social class (I–V) based on occupation has been replaced by an arrangement based on
occupation and economic group. The Office of National Statistics now uses this system in all official statistics and
surveys. A category is assigned according to the current or former occupation of the Household Reference Person (i.e.
the person responsible for owning or renting the household’s accommodation) using the scale in Table 1C.8.1.

121

Public [Link] 121 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

UK Table 1C.8.1
NSSEC categories
1 Higher managerial and professional occupations
1.1 Large employers and higher managerial occupations
1.2 Higher professional occupations
2 Lower managerial and professional occupations
3 Intermediate occupations
4 Small employers and own account workers
5 Lower supervisory and technical occupations
6 Semi-routine occupations
7 Routine occupations
8 Never worked and long-term unemployed
Reproduced from the Office for National Statistics (2007).

NZ NZIDEP – SOCIOECONOMIC DEPRIvATION FOR INDIvIDUALS

The New Zealand index of socioeconomic deprivation for individuals (NZiDep) does not depend on occupation or
education data. Instead, it assigns deprivation levels on a five-point scale to individuals based on responses to
eight simple questions, e.g. ‘In the last 12 months have you personally been forced to buy cheaper food so that you
could pay for other things you needed? [Yes/No].’

OTHER INDIvIDUAL INDICATORS

Alternative markers of socioeconomic status that are sometimes used in public health include:
• Income
• Occupation
• Years of education
• Housing – owner/renter, or occupancy per room
• Ownership of commodities (e.g. car, television).

AREA INDICATORS
Countries have adopted different indicators tailored to circumstances and availability of data.
Eng INDEx OF MULTIPLE DEPRIvATION (IMD)
The Department for Communities and Local Government (CLG) applies the IMD-2007 scale to the English Super-
Output Areas (SOA) see Box 3A.1.1 on page 327. Each SOA is scored and ranked for deprivation based on seven
weighted measures:
• Income
• Employment
• Health
• Education, skills and training
• Crime
• Housing and services
• Living environment.
Reproduced from Social Disadvantage Research Centre (2007).

122

Public [Link] 122 3/6/08 [Link]

 1C Assessment and Evaluation

Scot SCOTTISH INDEx OF MULTIPLE DEPRIvATION

The Scottish Index of Multiple Deprivation (SIMD) is used across local and national government for directing
resources, setting targets, and monitoring social and health inequalities. It is available at data zone level as an
overall deprivation index. It includes separate indices for different domains:
• Income
• Employment
• Housing
• Health
• Education
• Geographical access
• Telecommunications.
Reproduced from the Scottish Government’s Scottish Index of Multiple Deprivation (2004).

UK JARMAN
The Jarman score is based on factors that affect patients’ demand for primary care. It is a measure of GP practice
workload and is used to calculate additional deprivation payments to practices. Factors are determined from the
census and include:
• Elderly living alone
• One-parent families
• Children under 5 years old
• Unemployed (as percentage of economically active population)
• Overcrowded households
• Moved house within the last year
• Born in the New Commonwealth or Pakistan.
Reproduced from [Link]/indexmenu/jarman_desc.html.

Eng TOWNSEND
This is another census-based measure of comparative deprivation, and is defined by the proportion of households
that:
• Have more than one person per habitable room
• Have no car
• Are not owner occupied
• Include a person who is unemployed.

Wal WELSH TOWNSEND

Welsh Townsend is a direct measure of health needs rather than a measure of deprivation as such. It is included here
to avoid the potential for confusion with Townsend. The indicator that is used to allocate financial resources for
health care uses Welsh Health Survey data.

Scot CARSTAIRS
This is a measure of relative deprivation used in Scotland. It is based on four census variables:
• Overcrowding
• Male unemployment

123

Public [Link] 123 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

• Low social class

• No car.
Reproduced from Carstairs (1989) and Morgan (2006).

NI NORTHERN IRELAND MULTIPLE DEPRIvATION MEASURE (NORTHERN IRELAND STATISTICS

AND RESEARCH AGENCY, 2005)
The NIMDM contains seven domains of deprivation, namely.
• Income deprivation
• Employment
• Health deprivation and disability
• Education and skills, training
• Proximity to services
• Living environment
• Crime and disorder.

Ire IRELAND INDICES

Two indices of area level deprivation have been developed in Ireland, based on information from the census. The
Small Area Health Research Unit (SAHRU) index includes:
• Unemployment rate
• Percentage of households that are local authority owned
• Percentage of the population in social classes 5 and 6
• Percentage of households with no car
• Persons per room.
The Haase and Pratschke index includes persons per room and a more detailed classification of social class,
unemployment, age, education status, lone parents and change in population size since the previous census.
Underlying the index are the three dimensions of rural demographic decline, urban labour market deprivation and
social class disadvantage.

NZ NEW ZEALAND DEPRIvATION INDEx

The New Zealand small area index of relative socioeconomic deprivation, NZDep, is based on nine census variables,
and is widely used within the country for health resource allocation, needs assessment, research and advocacy. It
provides a deprivation score between 1 and 10, with 1 being the least deprived decile of areas, and 10 being the
most deprived decile of areas. NZDep is available in four versions according to the census on which it was based
(e.g. NZDep2006 was based on the 2006 Census). Scores are available for meshblocks (which correspond to a city
block) and census area units (which correspond to a neighbourhood), as well as larger geographical areas.
NZDep indices are based on the following nine measures generated from census variables:
• No access to a telephone
• Receiving a means-tested benefit, aged 18–59
• Unemployed, aged 18–59
• Household income below a threshold
• No access to a car
• Single-parent family, aged 18–59
• No educational qualifications, aged 18–59
• Living in rental housing
• Living in overcrowded housing.

124

Public [Link] 124 3/6/08 [Link]

 1C Assessment and Evaluation

SA PROvINCIAL INDICES OF MULTIPLE DEPRIvATION

The nine Provincial Indices of Multiple Deprivation (PIMD) for South Africa are the first stage of a larger project to
measure multiple deprivation at a small area level for the whole country. Each PIMD was created using the 2001
census: distinct, unidimensional domains of deprivation were combined, with appropriate weighting, into a single
measure of multiple deprivation.

COMPOSITE MEASURES
Measures such as the IMD, Townsend and Carstairs are composite indicators of deprivation.

Box 1C.8.1
Advantages of composite indicators Disadvantages of composite indicators
Readily available Rely on census data, therefore could be up
to 10 years out of date
Can shape health service and local authority Potential for autocorrelation (see IMD
planning and priorities (e.g. initiatives can above)
be targeted to areas of high deprivation or to
specific features of deprivation, such as crime)
Can be used for research into associations
between deprivation and health

1C.9 EvALUATION
Principles of evaluation, including quality assessment and quality assurance
In public health, an evaluation of a service is a thorough assessment of whether a health service meets its
objectives. This can be performed using routine or specially collected data, and may form part of an audit or
research exercise (Box 1C.9.1).

Box 1C9.1
Research Results that can be generalised outside the setting of the
evaluation
Audit Results are pertinent only to the setting being evaluated

The perspective of the evaluation must always be stated (e.g. the commissioner, the clinician, the patient or society
in general). A rigorous evaluation will address whether the service is cost-effective and also whether it is needed
and wanted by the local population.

FRAMEWORKS FOR EVALUATING SERVICES

Two standard frameworks used in public health are those of Donabedian and Maxwell.

DONABEDIAN
This framework of structure, process, output (and outcome) is discussed in Section 1C.6.

MAxWELL
The framework described by Maxwell (1984) consists of six dimensions of quality described in Table 1C.9.1.

125

Public [Link] 125 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

Table 1C.9.1 Maxwell’s dimensions of quality

Access Whether patients can see a clinician when they wish
Whether there are any tangible or intangible barriers to access. The latter reflect how
uninviting or unsuited the service is to particular groups
Tangible barriers
• Geography (services near to home or public transport)
• Finance (user charges, costs attached to attending services, e.g. lost income,
childminding fees)
• Opening hours, waiting times
Intangible barriers
• Languages used
• Female doctors available to see female patients
Relevance Whether services provided are appropriate to patients’ needs
Equity Whether services are provided fairly (not necessarily equally – see Section 1C.10)
Efficiency Whether the costs of providing interventions are justified by the benefits. Economic
evaluations (section 4) are used to address this question
Effectiveness Whether the outcomes of the interventions are reflected in improvements in health (see
Section 1C.6)
Acceptability Whether the patient (and in some cases the practitioner) is satisfied by the care provided
(for patient satisfaction measures, see Section 1C.15)

QUALITY ASSESSMENT AND QUALITY ASSURANCE

In public health, the terms quality assessment and quality assurance are often used interchangeably with the term
evaluation. However, in other fields the term quality assurance may relate to the process of ensuring that a service
meets a defined standard.

1C.10 EQUITY IN HEALTH CARE

Equity and equality are distinct concepts: see Box 1C.10.1.

Box 1C.10.1
Equity Equality
Definition Fairness Sameness
Distributive justice
Issue Differences in health due to Avoidable and unavoidable
avoidable inequalities (e.g. differences (e.g. genetic diseases
distance from a general practice and road traffic injuries)
surgery)

As with the concept of fairness, there are different viewpoints as to what constitutes equity.

126

Public [Link] 126 3/6/08 [Link]

 1C Assessment and Evaluation

VERTICAL EQUITY
Vertical equity requires unequal health care to be provided for unequal need. For example, in a system that
is vertically equitable, sicker people on a waiting list would be prioritised above others who were less sick
but had been waiting longer on the same list. Similarly, a progressive taxation system (where the rich face a
disproportionately large tax burden) is vertically equitable.

HORIZONTAL EQUITY
Horizontal equity requires equal health care for equal need. It can be considered in four ways outlined in Table
1C.10.1, although it is recognised that access and use overlap.

Table 1C.10.1 Horizontal equity

Type Definition Advantages Disadvantages
Equal Budgets are allocated according to Simple to measure because Health care for a particular
spending for health need. For example, resources spending can be clearly condition will cost varying
equal need would be targeted by means of identified amounts in different settings,
SMRs and area-level deprivation e.g. cost of a visit from a
indicators (see 4D3) district nurse may be greater in
rural than urban areas because
of a need to travel further
Equal access Access can be considered in Defining principle of Barriers to access may be
for equal need terms of removing the barriers services such as the NHS intangible (e.g. cultural) and
to service use. The barriers could therefore difficult to identify
Important for providers
be geographical (distance to the
to consider whether the
service), financial (cost to the user)
supply of services is
or related to time (whether services
equitable
are open only during office hours)
Equal use for This assumes that people should Overcomes the problem Assumes that people with the
equal need have the same health-care demands of intangible barriers to same symptoms have the same
for the same symptoms, so that access demand for health care
any differences reflect barriers to
access. One example often quoted
is that of coronary bypass rates in
areas of different SMR for coronary
heart disease. Note that many
studies that purport to measure
access in fact measure health-care
use
Equal health This is clearly the ultimate aim for Gold standard Subject to factors outside the
for equal need equity in health services control of health care, which
may be unavoidable (e.g.
genetics) or avoidable (e.g.
poverty)

In a seminal paper in 1971, Julian Tudor Hart described inequities, particularly as result of the introduction of
market forces, in the NHS. He coined the phrase, the inverse care law, to describe this phenomenon:
‘the availability of good medical care tends to vary inversely with the need of the population served’

127

Public [Link] 127 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

Tudor Hart’s evidence of the inverse care law came from the distribution of GPs in South Wales, but in subsequent
years empirical evidence of the existence of the inverse care law has been found in numerous areas of health.

1C.11 CLINICAL AUDIT

A clinical audit is a systematic review of care, as measured against explicit criteria. It is a central component of
clinical governance, the latter being a local system for assuring the quality of patient care and for making ongoing
improvements.
UK National audit programmes also exist, including:
• Stroke – the National Sentinel Stroke Audit Scope
• Coronary heart disease – Society of Cardiac Surgeons’
Cardiac Surgery Audit.

AUDIT CYCLE Repeat Measure

In order for lessons to be learnt from an audit exercise, the

process should always be seen as an ongoing cycle rather
than a one-off event (see Figure 1C.11.1).
The steps involved in conducting an audit are described in
Table 1C.11.1. Findings Implement

Figure 1C.11.1 The audit cycle

Table 1C.11.1 The audit process
Scope Begin by selecting criteria against which to audit. Useful sources of criteria include national
standards (e.g. National Service Frameworks) and national guidelines (e.g. NICE Health Technology
appraisals)
The criteria should be:
• Explicit statements that define what is being measured
• Elements of care that can be measured objectively
The scope of the audit can then be determined by deciding:
• Which patients should be included
• Over what time period they should be audited
• Which aspects of care should be considered
An appropriate sample should be chosen by using a strategy such as:
• Interval sampling (all patients visiting a clinic in a given period of time, e.g. January to
March)
• Two-stage sampling (a small sample is selected first; if unequivocal conclusions can be drawn,
then no more data are collected; if the results are ambiguous, then a larger sample is selected)
Measure Measure observed performance against the selected criteria. Data sources include:
• Registers – to identify patients
• Patient records – to examine aspects of care
• Electronic systems – for rapid retrieval of information
• Teams delivering the care – to find out if information is collected, and the extent to which
this information is suitable

128

Public [Link] 128 3/6/08 [Link]

 1C Assessment and Evaluation

Findings Use the findings to identify areas for change. Ensure that the findings are shared with those who
need to know, including:
• Teams working in the area of the audit
• Area service directors
• Boards
Produce short summaries of findings that are readily understandable
Implement Implement the findings of the audit by making improvements to the service
Consider adopting change management approaches (see section 5C)
Repeat Check that these improvements are in place and are making a difference by repeating the audit
cycle

SUCCESSFUL AUDIT
Baker and colleagues (1995) identified several factors influencing doctors’ participation in audit. These factors
and other situations identified by Johnston and colleagues (2000) that may influence the success of an audit are
described in Table 1C.11.3.

Table 1C.11.3 Drivers and barriers to success in audit

Drivers of success Barriers to success
Time Protected time Lack of time main reason for failure or
incomplete audit
Strategy Leadership and structured programmes Lack of an overall plan for audits
Skills Training for clinical staff Lack of expertise or advice in project
design and analysis
Availability of specialist audit staff
Resources Modern medical records and IT systems Lack of these factors
Dedicated audit staff
Funding to conduct the audit and
implement changes
Organisational culture Dialogue between purchasers and Problems between groups and group
providers members
Whole team participates in the audit and
implementation of change
Supportive learning environment with a
no-blame culture

1C.12 CONFIDENTIAL ENQUIRY PROCESSES

Confidential enquiries are national investigations into serious untoward incidents. They are an example of a national
investigation of serious concerns. Examples for England are described in Box 1C.12.1. Other examples of national
investigations with implications for health include:
• Clinical Standards Advisory Group (1991–2000)
• Bichard Enquiry into the Soham murders (made recommendations regarding child protection procedures
following the murder of two young girls).

129

Public [Link] 129 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

The purpose of all such investigations is to identify what went wrong and to draw general lessons that can be
shared.

Eng Box 1C.12.1

Confidential enquiries in England
• National Confidential Inquiry into Suicide and Homicide in Mental Health Services: [Link]-confidential-
[Link]
• Confidential Enquiry into Maternal and Child Health – [Link] – collects data on maternal deaths,
perinatal, neonatal and infant mortality and causes of stillbirth. Sudden infant deaths are routinely reported
to the Coroner’s Office
• National Confidential Enquiry into Patient Outcome and Death (NCEPOD): [Link]

CONDUCT OF ENQUIRIES
A case–control study design is often employed, and is used to detect risk factors that predispose to unexpected
deaths or other untoward incidents.
Both quantitative and qualitative data may be used, including:
• Routinely collected data (e.g. hospital episode statistics)
• Hospital records
• Confidential surveys of individuals (including hospital staff, HM Coroner, families, GPs, friends).

OUTCOME OF ENQUIRIES
Confidential enquiries produce reports that are circulated widely and make recommendations for health services on
how to minimise the risk of similar events happening again. The reports sometimes include self-audit tools that
health services may use to assess their own risk.

1C.13 DELPHI METHODS

This is an iterative technique for generating consensus about a particular issue without face-to-face meetings.
The process works as shown in Box 1C.13.1.

Box 1C.13.1
Step 1 A group of experts is contacted and surveyed
Step 2 Views of the group are shared anonymously among the same group, highlighting areas of
disagreement
Step 3 Respondents are asked if they wish to change their views in light of step 2
Step 4 Steps 2–3 are repeated several times until consensus is reached

The Delphi technique was originally used for making technology forecasts, but it has also proved useful in public
health, especially in fields where data are lacking, or as a way of engaging people who would not otherwise
be involved. Its advantages and limitations are shown in Box 1C.13.2. The application of the Delphi method is
described in an example from mental health services in Box 1C.13.3.

130

Public [Link] 130 3/6/08 [Link]

 1C Assessment and Evaluation

Box 1C.13.2
Advantages Limitations
Anonymity and written process avoids some of Written survey format may be more suitable for some
the pitfalls of face-to-face discussion, e.g. each respondents than others
contribution is valued equally, with reduced potential
for personality to influence responses
Time-efficient: does not rely on bringing together Open to manipulation from those administering it
disparate, busy groups
Encourages open critique and admission of errors by Does not always produce useful results. Sometimes
encouraging contributors to revise earlier judgements future developments are more accurately predicted by
unconventional thinking than by iterative consensus
from accepted experts

Box 1C.13.3
Example: Describing service models of community mental health practice
This study used the Delphi method to generate a ‘valid and reliable set of categories to describe the clinical
work practices of intensive case managers’. Eight case managers took part in the process, which involved a
combination of (a) the Delphi method to generate a list of categories to describe their working practices and
(b) face-to-face discussions to refine this list. The authors describe the Delphi method as: ‘… an effective,
straightforward, and time-efficient way of obtaining a workable consensus.’
Reproduced from Fiander and Burns (2000).

1C.14 ECONOMIC EvALUATION

See Section 4D.5.

1C.15 APPROPRIATE AND ACCEPTABLE SERvICES

Appropriateness and adequacy of services and their acceptability to consumers and providers
Both the appropriateness (suitability) and adequacy (sufficiency) of a health service can be seen to take several
dimensions, including those shown in Box 1C.15.1.

Box 1C.15.1
Humanity Degree to which patients are treated with respect
Access Whether consumers received care when they needed it:
• Acceptable waiting times
• Information about the service being provided in relevant languages
Environment Whether the service is offered in an appropriate setting:
• Building was clean and well maintained
• Adequate privacy was provided
• Setting was suitable for its patients (e.g. toys provided for children)
Information Whether patients are fully informed about their care:
• Translation services provided
• Lay terminology used to describe the treatment options and procedures

131

Public [Link] 131 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

See also Section 1C.5.

The exact requirements for appropriate services will depend upon the factors shown in Box 1C.15.2.

Box 1C.15.2
Individual Every person has different values, tolerances and ideas of what is acceptable
Patient group Old or young; long-term or day-case patients have different requirements for appropriate services
Service Inpatient versus community clinics have different dimensions and requirements for providing
humane and accessible care
Priorities There may be a trade-off between accessibility and humanity, e.g. the need to see a GP promptly
may take priority for some over seeing the same GP each visit or each consultation limited to
5–10 min

ASSESSING ACCEPTABILITY TO CONSUMERS AND PROVIDERS

Reasons for assessing acceptability include:
• Feedback should lead to service improvements.
• It guides consumers in their choice of care. For example, information about patient satisfaction can be
provided to patients who are choosing their hospital under the choose and book initiative.
• It should ultimately guide commissioners in their decisions of where to award contracts to health-care
providers.

Eng INFORMATION USED TO ASSESS ACCEPTABILITY

See Box 1C.15.3.

Eng Box 1C.15.3

Complaints and • Informal comments/complaints, e.g. through PALS (Patient
compliments Advice and Liaison Service), conversations with staff
• Formal complaints made directly to services
• Ombudsman’s reports, complaints reported to a third
person
Untoward incidents • Local analysis of near misses and mistakes in care
• National Confidential Enquiries
visits/enquiries • Regulatory bodies such as the Healthcare Commission,
GMC/NMC
• Local scrutiny
Surveys • National: patient and staff surveys (run by the Healthcare
Commission)
• General practice: undertaking annual approved patient
survey part of the quality requirements of the new GMS
contract
• Local: bespoke surveys

CONSIDERATIONS FOR ASSESSING ACCEPTABILITY

Considerations for assessing acceptability of services are described in Table 1C.15.1 and illustrated in the example of
cardiology services from Northern Ireland in Box 1C.15.5.

132

Public [Link] 132 3/6/08 [Link]

 1C Assessment and Evaluation

Table 1C.15.1 Assessing acceptability of health services

Who is surveyed Sample selection is important to determine and record.
Should people who have already complained be asked,
or all patients? Different degrees of satisfaction are
seen among different sections of the population: key
determinants include age, socioeconomic status and
ethnicity
When is the survey taken During care or after? There is a trade-off between
patients’ capacity to remember care accurately versus
asking someone in the vulnerable position of still
receiving care
Where survey is performed Satisfaction results are generally higher if answered
within health-care setting compared with outside, e.g.
the patient’s home
How consumers are surveyed Acceptability is highly dependent upon context and
individuals. Qualitative studies are good for describing
experiences, but are time-consuming to analyse.
Quantitative studies explore only what is asked, but are
useful when the questions that have to be addressed
are known (e.g. after some initial qualitative research
to identify key issues). For example, if cleanliness
arises as an issue in qualitative research, then
quantitative methods could assess how often wards are
dirty

NI Box 1C15.4
Example: Survey of cardiology patients
A group of patients was contacted immediately after angiography
to determine their views on what they believed should influence
priorities for cardiac revascularisation. The survey revealed that
patients felt that the following factors should be considered:
• Patients’ age
• Patients’ smoking status
• Whether patients have dependent relatives
Reproduced from Kee et al (1997).

1C.16 EPIDEMIOLOGICAL BASIS FOR PREvENTIvE STRATEGIES

The epidemiology of a disease can guide strategies for preventing it: factors that are found to be associated with
the disease, and which meet the criteria for causation, can provide useful targets for preventive strategies (see
Section 1A.13).

LEVELS OF PREVENTION
Preventive interventions are considered at three levels: primary, secondary and tertiary: see Table 1C.16.1.

133

Public [Link] 133 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

Table 1C.16.1 Levels of prevention

Level of prevention Definition Example
Primary prevention Aims to prevent or delay the onset of Low-dose aspirin given to those at
disease in those at risk but who do not sufficient risk of developing ischaemic
currently have the disease heart disease
Secondary prevention Aims to prevent or delay further An exercise programme offered to patients
progression or complications in people who have had a myocardial infarction
known to have the disease
Tertiary prevention Aims to reduce long-term disability that Removing allergens from the environment
would otherwise result from the disease of a patient with asthma

Note that primary prevention is not the same as screening: the latter determines who in a population is at sufficient
risk of the disease to warrant further investigation.

PREVENTIVE STRATEGIES
There are broadly two approaches to preventive medicine: the population approach and the high-risk approach: see
Table 1C.16.2.

Table 1C.16.2 Population and high-risk approaches to prevention

Strategy Description Example
Population The whole population receives health Legislation for wearing seat belts in cars
approach promotion and disease prevention
interventions to reduce risk throughout
because all are at risk
High-risk Health promotion and disease prevention Chlamydia screening targeted at people aged
approach is targeted at groups identified from 15–24 (who have the highest prevalence of
epidemiological studies to be most at risk the disease among the population)

Epidemiological strategies to identify high-risk groups include:

• Ecological studies to identify groups with highest incidence of disease (see Section 1A.17)
• Analytical studies to link risk factors and diseases (e.g. blood pressure, diabetes and coronary heart disease)
• Genetic studies to identify families and individuals with genetic risk factors (e.g. gene for
hypercholesterolaemia – see Section 2D.9)
• Population screening to identify individuals who have known risk factors (e.g. diabetes for heart disease, age
profile for chlamydia infection).
Geoffrey Rose (1992) outlined the rationale for a population-based preventive approach in his seminal work, The
Strategy of Preventive Medicine. This approach is useful for tackling common conditions that are normally distributed
and have a widespread cause – the classic example being deaths from coronary heart disease due to hypertension
(see Figure 1C.16.1 and Box 1C.16.1).
However, the prevention paradox (see Section 2H.4) significantly limits the effectiveness of the population
approach. Current thinking emphasises the need for a combination of both population and targeted approaches to
prevention (Anon 2006).

134

Public [Link] 134 3/6/08 [Link]

 1C Assessment and Evaluation

0.45 160
Shifting the population
distribution by just 1 SD
0.4 prevents 40% of the
disease 140

0.35
Low-risk population High-risk population 120

0.3

Relative risk of disease

100
Probability density

0.25
Risk of 80
disease
0.2

60
0.15

40
0.1
Treating people with
the top 5% of the risk
0.05 factor prevents just 20
12.5% of the disease

0 0
�5 �4 �3 �2 �1 0 1 2 3 4 5
Standard Deviation

Figure 1C.16.1 Effects of a population-based versus a high-risk-based approach to preventing heart disease

Box 1C.16.1
Example: controlling risk factors for heart disease
The blood pressure of the population is normally distributed, so a minority of people have very high blood
pressures. In Figure 1C.16.1, the top 5% of the population with the highest blood pressures are represented by
the white triangular area. These people are at high risk of coronary heart disease (CHD).
However, since the numbers of high-risk people are small, their burden of disease on the overall population is
modest. In this example, these 5% of people represent 12.5% of the burden of disease. Therefore, many more
cases of CHD occur in people with lower blood pressure, i.e. to individuals who are outside the white triangular
area. Targeting only the top 5% of the population with very high levels could at best reduce disease by 12.5%.
An approach that slightly reduced blood pressure across the whole population would, however, have a much
bigger impact, in this example reducing disease by 40%.
Adapted from Department of Health (2000), Emberson et al (2004), Manuel et al (2006) and Marshall and Rouse
(2002).

1C.17 HEALTH AND ENvIRONMENTAL IMPACT ASSESSMENTS

An environmental impact assessment (EIA) is a technique for assessing the potential effects of a proposed
development on the environment. All member states of the European Union are legally required to conduct an

135

Public [Link] 135 3/6/08 [Link]

 1C As s e s s me nt and Eval uation

EIA for any major civil engineering project. Some EIAs consider health as a criterion against which to assess the
proposal, but this is typically covered only superficially.
The WHO Ottawa Charter for Health Promotion (1986) proposed an analogous technique of health impact
assessments (HIAs) that would concentrate solely on the potential effects of proposed policies or plans on the
health of the population. Such policies include those shown in Box 1C.17.1.

Box 1C.17.1
Policy type Example
Transport strategies London Heathrow Airport Terminal 5
Civil engineering Wembley Stadium in North London
New landfill site
Urban regeneration Olympics regeneration in East London
Lobbying Manchester Airport second runway expansion
Developing countries Donor aid project appraisals

STEPS IN UNDERTAKING A HEALTH IMPACT ASSESSMENT

There are five key steps to an HIA: see Box 1C.17.2.

Box 1C.17.2
1. Screening Establish whether there is any health relevance to a policy, programme
or other development
2. Scoping For proposals with health relevance, the questions that the HIA should
address need to be decided upon, together with reporting arrangements
3. Appraisal Assess potential health impacts either through rapid appraisal (over a
few days or weeks) or through in-depth assessment (weeks or months)
4. Reporting Conclusions are made, followed by recommendations that minimise
the negative impacts and maximise the positive impacts of the
development
5. Monitoring Actual impacts are monitored to enhance the evidence base for future
HIAs

CHALLENGES
Challenges to conducting an HIA may include a dearth of evidence, time or resources.

EvIDENCE
A crucial point to consider is whether there is any evidence on which to base assessments. For new projects, the
health impacts may never have been considered, observed or measured before. Quantifiable evidence often proves to
be more influential than qualitative evidence, but may be not be feasible to collect.
The evidence for an HIA may originate from several sources, often coming from stakeholders with conflicting
perspectives. HIAs may thus require the contradictory information to be synthesised.

136

Public [Link] 136 3/6/08 [Link]

 1C Assessment and Evaluation

QUALITY vERSUS TIME AND RESOURCES

Some HIAs demand rapid results with few resources, in which case it may not be possible to conduct HIAs with the
same rigour as longer-term studies. This needs to be explicitly acknowledged at the outset of an HIA.

ACTION
A final, but decisive challenge is to ensure that the findings and recommendations of the HIA are actually
implemented. For this to happen, the decision-makers responsible for the programme need to be engaged in the HIA
at the reporting stage.

137

Public [Link] 137 3/6/08 [Link]

Public [Link] 138 3/6/08 [Link]
 1D Principles of Qualitative Methods

1D
Principles of Qualitative Methods

1D.1 Overview of qualitative methods 139 1D.4 Validity, reliability and generalisability −
1D.2 Appropriate use 141 data collection 141
1D.3 Ethical issues 141 1D.5 Common errors and their avoidance 143

Scientific studies often provide quantitative data about an issue. For example, they might measure the proportion
of people affected by a condition who have a particular risk factor. Sometimes, however, the information required
to answer a question pertains more to the character of an issue, rather than to any numerical value. Qualitative
methods aim to explore this second type of enquiry, i.e. those that relate to the kind or the quality of things.
These techniques are best used for answering how?, what? or why? questions. For example, qualitative research
questions in health include:
• What influences parents’ decisions to vaccinate their children?
• How do different patients view their illnesses?
• Why do some patients decline information about their condition?
In general, qualitative methods are more time-consuming than quantitative methods with regard to data collection
and data analysis. However, they can provide information that would be unobtainable by quantitative means.

1D.1 OvERvIEW OF QUALITATIvE METHODS

Overview of qualitative methods, including semi-structured and in-depth interviewing, focus groups, action research,
participant observation and their contribution to public health research and policy
The four principal methods for collecting qualitative data, summarised in Table 1D.1.1, are:
• Semi-structured and in-depth interviews
• Focus groups
• Action research
• Participant observation.

139

Public [Link] 139 3/6/08 [Link]

 1D P r i nc i p l e s of Q ual itative Methods

Table 1D.1.1 Four principal methods of qualitative data collection

Semi-structured and Focus groups Action research Participant observation
in-depth interviewing
Interviewer uses a topic Facilitator leads a The people under study Researchers actively take
guide, rather than a rigid small number of people (e.g. those using a part in a setting as they
set of questions. Issues (usually 5−10) in a service) are actively collect their data, either as
should arise naturally structured conversation. involved as researchers. a member of the community
What is done?

in the conversation. The focus is not just Action research enables (e.g. a nurse in a hospital),
In-depth interviews can on the responses to the researchers/actors or a supporter (e.g. hanging
be conducted over several facilitator but also on to make changes to around a criminal gang but
hours. Researchers may the reactions to other projects during the not taking part in criminal
also conduct several participants process, not just at the activities)
interviews at different end of the research
stages with the same
individuals
Uses: Uses: Uses: Uses:
• Generating new • Understand target • Involving people • Immersing the
concepts and ideas audiences who would researcher in the
around an issue • Exploring subjects otherwise not experiences of those
• Seminal study: that are difficult to normally take part under study
How is this used?

Jocelyn Cornwell – discuss individually, in research • Gaining insight into

Hard Earned Lives: e.g. sexual beliefs • Community subcultures not usually
explored people’s and behaviours development open to study or
accounts of health observation
care over several • Seminal study:
interviews. Revealed: Rosenbaum’s study of
hierarchies of health treatment of healthy
settings; what is people in psychiatric
‘counted’ as health; institutions when they
why people do not pretended to experience
seek care hallucinations
Qualitative data are analysed in different ways depending, in part, on the method of data collection that was used.
There are three principal methods: see Box 1D.1.1.

Box 1D.1.1
Thematic The simplest form of analysis, this usually involves reporting interviewees’ comments in a
content structured form. Useful for areas where not much is known and as a starting point for other
types of analysis
Grounded theory Used for developing theories. Data collection takes place in a cyclical process: themes
identified from the analysis are fed back into the topics to be discussed during interviews, and
new data are used to refine the analysis. The analysis involves organising content into codes
and continually refining these codes
Framework Used for developing practical strategies as a result of research
There are several ways of enhancing the validity and the credibility of qualitative analysis, including:
• Identification of disconfirming cases or deviant data, then accounting for these, i.e. not solely quoting
supportive data

140

Public [Link] 140 3/6/08 [Link]

 1D Principles of Qualitative Methods

• Developing a coding structure with other researchers

• Presenting back to interviewees to check that the material accurately represents what they thought
• Counting responses in themes to give an indication of how common a particular response was (e.g. ‘seven
participants said x’ or ‘most said y’)
• Being as transparent as possible in the presentation and including raw data if possible (e.g. quotes)
• Describing the process of analysis
• Comparing and contrasting findings with findings from other studies, attitudes at beginning of the study or
other data collected in this study
• Cross-checking findings using other study methods (e.g. questionnaires) to explore the effect of the
interviewer/facilitator on findings. For example, a health service employee may hear only positive comments
about the health service.

1D.2 APPROPRIATE USE

Each of the principal types of qualitative investigation has a different application (see Table 1D.2.1).

Table 1D.2.1 Applications of qualitative investigation

Semi-structured Focus groups Action research Participant observation
and in-depth
interviewing
Exploring people’s Exploring people’s Aims to change practice as well Exploring what people actually
accounts of accounts of activities and as study it do, not just what they say they
activities and beliefs do
Cyclical research design where
beliefs
Generating large amounts data gathering, planning, Capturing information on
of data in a relatively observing and reflecting all familiar/routine areas of
short time feed into the next planning life that enables views or
cycle. Participants are therefore understandings to become
Exploring how people
also the researchers explicit
come to a consensus or
viewpoint Particularly useful for:
• Evaluation
• Assessing health needs
Can use innovative research
methods (e.g. drama, charts,
creative arts)

1D.3 ETHICAL ISSUES

The different qualitative methods each raise potential ethical concerns (see Table 1D3.1).

1D.4 vALIDITY, RELIABILITY AND GENERALISABILITY – DATA COLLECTION

The quality of research is often considered according to three dimensions:
• Reliability − the degree to which the collection of data in a study was consistent and repeatable
• validity − how well a study’s findings represent the ‘true’ state of affairs
• Generalisability – the extent to which the findings from one setting can be applied to another.
Concepts of validity and reliability are underpinned by a positivist perspective on research, i.e. that there is one
true answer and that a well-designed study would be able to reproduce the same findings if conducted by different

141

Public [Link] 141 3/6/08 [Link]

 1D P r i nc i p l e s of Q ual itative Methods

Table 1D.3.1 Ethical issues in qualitative research

Semi-structured Focus groups Action research Participant observation
and in-depth
interviewing
Confidentiality is a particular problem in The nature of informed consent If the observation is covert (e.g.
studies with small samples. The problem in this context should be pseudo-patient or mystery shopper),
may be mitigated by ensuring that considered carefully, especially then the study participants will not
there are several participants with each since there may be a blurring have consented to the research.
characteristic of interest. Interviewees of roles between the researcher This may damage trust among those
should be allowed to choose the degree of and the participants involved, and reduce cooperation for
anonymity that they require future studies
The reporting of negative findings, or There should be strong grounds for
findings that are unfavourable to the conducting covert observations (i.e.
interviewees, may need to be managed the study is valuable and the same
data would not be obtainable if the
In focus groups in particular, it is
research were conducted openly)
important to ensure that the group
atmosphere is supportive

people at different times. Much qualitative research comes from a constructivist tradition, i.e. research will not
yield one ‘true’ view of an issue, but could yield several equally valid perspectives depending on who is conducting
the research and when the research was conducted. However, there are still ways to ensure that qualitative research
is as rigorous as possible through considering its reliability and validity.
Reflexivity is an important dimension considered in qualitative research. This recognises that the researcher affects
the environment and the people whom he or she is researching. In contrast to a positivist approach that would
seek to reduce the effect of the researcher, qualitative research seeks to record and acknowledge this effect of the
researcher on study findings.

RELIABILITY
Reliability can be improved by:
• Using the same data collection tool (interview schedule, observation or topic guide) for each interview, group
or observation
• Using one researcher (interviewer, facilitator or observer) for all interviews/groups/observations – or else
ensuring that all researchers adhere to the same guidance
• Providing the same training and written guidance to researchers to keep discussions on topic, but still
enabling participants to express their views
• Producing an accurate record of the discussion, interview or observations through:
# Electronic recording of interviews (audio or video) rather than note-taking, not only because it ensures an
accurate record but also because it enables the researcher to listen fully and take part in the discussion.
However, recording may not be appropriate or feasible in all situations, and it may dissuade some subjects
from participating. Explicit consent must always be obtained for recording data collection
# Noting/recording observations as soon as possible afterwards to ensure that they are recalled correctly
# Writing/recording clearly and consistently. In participant observation studies, the researcher may be making
notes ‘on the hoof’, and usually develops systems of short-hand or his or her own abbreviations. While this
maximises the chance that the notes will be comprehensive and accurate, clear and consistent notes reduce
the risk of ambiguity when notes are analysed later.

142

Public [Link] 142 3/6/08 [Link]

 1D Principles of Qualitative Methods

VALIDITY
Validity is reduced by biases – the nature of which depends on the methods of data collection and analysis (see
Table 1D.4.1).

Table 1D.4.1 Effect of bias on validity in qualitative research

Semi-structured and Focus groups Action research Participant observation
in-depth interviewing
Potential for bias if Setting (group Validity, reliability and One of the most subjective
the interviewees feel conversation) more generalisability all depend forms of qualitative data
constrained from saying closely reflects real on what methods are used collection, findings will be
what they actually think life than a one-to-one in the action research. influenced significantly by
and instead say what is interview. However, there These methods may also the researcher and his or
socially acceptable, or remains the potential include quantitative her relationship with the
what they think reflects for bias since some techniques person being researched
best on them group members may not
Interaction is not simply To improve validity,
say their ‘true’ feelings
Interviewee is less likely between the interviewer consider:
because of supposed or
to be constrained in a and interviewee, but • Relationships and
voiced opinions from other
comfortable setting between all group trust gained with
members of the group.
members. This, therefore, community members
Consider: The group can reach false
offers opportunities for • Breadth of
• Location of interview consensus through a need
participants to reflect and observations/contacts
• Personal to be agreeable with each
react to the views and in the community
characteristics of the other
experiences of other group • Length of time spent
interviewer (e.g. age,
members in the setting
sex, social class)
• Understanding of
• Conduct of the
the language and
interviewer
terminology of the
community

GENERALISABILITY
This is the extent to which the study sample and setting are informative about the population. Generalisability is
increased by:
• Using a credible sample frame. While this need not be representative (compare quantitative research), it does
need to be sufficiently broad. For example, if the entire sample comes from one professional group (e.g. nurses),
then the findings of the study may not reflect the opinions of other professionals (e.g. physiotherapists).
• Drawing out themes that are transferable to other settings. For example, a generalisable study might consider
opinions on asthma rather than on asthma treatment in Cardiff Bay.

1D.5 COMMON ERRORS AND THEIR AvOIDANCE

Typical pitfalls in the conduct of qualitative studies are shown in Box 1D.5.1.

PRACTICAL ISSUES OF DATA COLLECTION

Qualitative research should not be undertaken lightly. Practical matters that should be borne in mind include those
shown in Box 1D.5.2.

143

Public [Link] 143 3/6/08 [Link]

 1D P r i nc i p l e s of Q ual itative Methods

Box 1D.5.1
Semi-structured and in-depth Action research Participant observation
interviewing and focus groups
Assumes that what people say May focus on changing practice rather Research can be sidelined if
accurately reflects what they think than on adding generalisable knowledge to it is viewed as relevant only
and do the field to the setting in which it was
located.
Research can be sidelined because it Difficulties in securing sufficient
is not seen as generalisable community participation may be
experienced
Professionals may have difficulty in
appropriately devolving power in research
to the community

Box 1D.5.2
Semi-structured and Focus groups Action research Participant observation
in-depth interviewing
Relatively easy Can be difficult to Can require training and support of Time intensive to
to organise: the organise, and there is those involved in research for the collect data and to
interviewer only needs the danger that some first time to ensure that they have analyse
to fit with one other invited participants may the skills and confidence necessary
Expensive: researcher
person’s diary and can not appear to take part
may need to be in
travel to the location
Can become expensive: place full time for
preferred by the
may need to provide many months to collect
interviewee (e.g. his or
travel expenses, crèche sufficient data
her office)
facilities, refreshments,
book rooms, etc.

144

Public [Link] 144 3/6/08 [Link]

 Section 2
DISEASE CAuSATION AND PREVENTION;
HEALTH PROMOTION

In order to improve health, practitioners need a broad understanding of what affects health and the evidence base
behind approaches to improve it.
Section 2 covers a wide range of subjects, starting with an introduction to different ways of conceptualising the
study of disease in epidemiological paradigms. The fundamentals of how disease is caused and where threats to
health orginate are covered in sections on the epidemiology of specific diseases, environment, genetic factors,
communicable disease, and health and social behaviour.
Finally, theory and practice of protecting and improving health is covered in sections on diagnosis and screening,
health promotion and disease prevention.

145

Public [Link] 145 3/6/08 [Link]

Public [Link] 146 3/6/08 [Link]
 2A Epidem iological Paradigms

2A
Epidemiological Paradigms

2A.1 Life-course paradigms 147

2A.2 Academic paradigms 148

Scientific method is governed by different frameworks, which set the rules, assumptions and techniques that the
scientific community tacitly accepts as valid. These frameworks or models are referred to as paradigms. There is a
range of different epidemiological paradigms. In this chapter, epidemiology is considered in terms of life-course
and academic paradigms.
Different paradigms in epidemiology offer:
• Alternative ways of viewing the same condition or population group
• Novel approaches to addressing epidemiological problems
• Insight into the effectiveness of certain interventions.
Knowledge of different paradigms enables public health practitioners to conduct insightful critical appraisal
of research. By identifying the paradigm of the research method, it is possible to identify and challenge the
assumptions that underlie the scientific approach used. In this way, alternative methods − and conclusions − can
then be considered.

2A.1 LIFE-COURSE PARADIGMS

The three examples below are means of considering epidemiological influences during life.

PROGRAMMING
This epidemiological paradigm considers the long-term effects of environmental exposures during critical periods
of growth (such as in utero development) upon adult diseases. For example, the Barker hypothesis states that an
effect of in utero malnutrition is an increased risk of coronary heart disease in adulthood.
Programming is also known as the critical period model because it assumes that there is a specific period during
which certain exposures will have lasting effects on the structure or function of organs.

147

Public [Link] 147 3/6/08 [Link]

 2A E p ide m io l o gical Paradigms

ADULT RISK FACTOR

The adult risk factor approach, in use since the 1950s, considers the impact of behaviours, such as smoking, diet,
exercise and alcohol consumption, on the onset and progression of diseases. Their relative contributory effects can
be assessed using multiple regression.

LIFE-COURSE
The life-course model combines elements of the programming (see above) and the adult risk factor (see Section
2A.1) paradigms, by investigating how various biological and social factors affect health and disease in adult life
independently, cumulatively and interactively.
A limitation of this paradigm is that few researchers will have access to a sufficiently wide range of biological and
social data on cohorts of subjects that extend longitudinally between birth and the outcome of interest.

2A.2 ACADEMIC PARADIGMS

Paradigms may also be classified according to the academic subdisciplines, study tools and frameworks of the
science of epidemiology. The conventional subdivisions of epidemiology include:
• Infectious diseases
• Chronic diseases
• Clinical epidemiology
• Psychosocial epidemiology
• Health care
• Genetic epidemiology.
Note that the tools employed in psychosocial epidemiology are rather different from those commonly used in
physical and environmental epidemiology.

148

Public [Link] 148 3/6/08 [Link]

 2B Epidem iology of Specific Diseases

2B
Epidemiology of Specific Diseases

2B.1 ‘Important’ diseases 149

Epidemiology of specific diseases (and their risk factors) of public health significance
This section covers the occurrence and distribution of certain key diseases within the population. Clearly, it is
impossible to have an encyclopaedic knowledge of all diseases. However, an overview of the epidemiology of
diseases of particular public health relevance is essential for prioritising, organising and allocating resources.
These key diseases are those that are:
• Major causes of premature death (e.g. cardiovascular disease, cancers , trauma)
• Major causes of morbidity or high use of health-care resources (e.g. diabetes, asthma and chronic obstructive
pulmonary disease, depression, schizophrenia, Alzheimer’s disease)
• Sharply changing in their incidence or prevalence in an area of the world (e.g. obesity in developed countries)
• Preventable (e.g. lung cancer).

2B.1 ‘IMPORTANT’ DISEASES

Knowledge of the defining clinical features, distribution, causes, behavioural features and determinants of diseases
that currently make a significant impact on the health of local populations, with particular reference to those that are
potentially preventable, or require the planned provision of health services at individual, community and structural
levels, or are otherwise of particular public concern, e.g. mental health
The World Health Organization’s global burden of disease project provides an estimate of the relative importance
of all communicable and non-communicable diseases, together with intentional harms (e.g. suicide and war). The
global burden of disease does not account for the degree to which illnesses are preventable or can be treated, but
it does provide a useful guide to which illnesses have the greatest impact globally – and are thus of public health
importance.
UK As Figure 2B.1.1 shows, neuropsychiatric conditions, cardiovascular disease and cancers account for most of
the UK’s burden of disease. Note also that the burden of chronic renal disease is set to increase dramatically in the
UK over the coming years. This is due to several factors, including the projected increase in the number of elderly
people from susceptible ethnic groups.
149

Public [Link] 149 3/6/08 [Link]

 2B E p ide m io l o gy of Specific D iseases

Neuropsychiatric conditions
Cardiovascular diseases
Cancers
Respiratory diseases
Digestive diseases
Sense organ diseases
Musculoskeletal diseases
Figure 2B.1.1 Top 10: UK global burden
Diabetes
of non-communicable disease. Reproduced
Endocrine disorders
from [Link]/healthinfo/statistics/
Genitourinary diseases
[Link]

The relevance of disease depends very much on context. As Figure 2B.1.2 shows, communicable disease, and
maternal, perinatal and nutritional conditions are responsible for most of the mortality in Africa. In contrast, in
Europe, non-communicable disease accounts for over 80% of mortality.

100
90
80
70
Contribution (%)

60
50
40
30
20
10
0 Figure 2B.1.2 Graph of
AFRO EMRO AMRO EURO SEARO WPRO relative impact of injury and
Region disease on mortality in different
regions. Reproduced from www.
Communicable, maternal, perinatal and nutritional conditions
[Link]/healthinfo/statistics/
Non-communicable diseases Injuries
[Link]

AFRO, African region EMRO, eastern Mediterranean region; AMRO, the Americas; EURO, Europe; SEARO, south-east Asian
region; WPRO, western Pacific region.
Similarly, the burden of disease is different across different ages and for different diseases. For major non-
communicable conditions, there is some variation in the age at which disease is most common, but the burden is
consistently greatest in those aged over 70, as Figure 2B.1.3 demonstrates.

PUBLIC HEALTH DISEASE KNOWLEDGE

The aspects of disease description that are important to a public health practitioner are:
• Incidence and prevalence
• Morbidity and mortality
• Importance (e.g. population-attributable risk fraction [for a risk factor] or the global burden of disease)

150

Public [Link] 150 3/6/08 [Link]

 2B Epidem iology of Specific Diseases

100
Percentage by age band (%)

80

60

40

20

0
Figure 2B.1.3
Malignant neoplasms Cardiovascular diseases Respiratory diseases Deaths in males
from three major
Age in years noncommunicable
0–4 5–14 15–29 30–44 45–59 60–69 70–79 diseases

• Time (trends – whether the incidence is rising or falling)

• Place (whether there are areas where the disease is particularly common or rare)
• Person (the type of person who is most at risk, with regard to demographics, lifestyle, health status and
workplace)
• Prevention (primary, secondary [early detection through tests/screening] and tertiary).
The aspects of non-communicable diseases that have a notable impact on the burden of disease in developed
countries or are otherwise of significant public health importance are covered in Tables 2B.1.1–2B.1.11.

NEUROPSYCHIATRIC CONDITIONS

Table 2B.1.1 Depression

Clinical Syndrome (group of symptoms) reflecting sad mood exceeding normal sadness or grief
characteristics (in intensity and duration) and with functional disabilities. Symptoms include negative
thoughts, moods and behaviour, and changes in bodily functions (e.g. eating, sleeping and
sexual activity)
Known causes/ Debated: investigations carried out into neurochemical abnormalities, hormonal risk
aetiological factors, genetic characteristics and adverse life events/social conditions
mechanisms
Can develop as a side effect of some medications, e.g. b-blockers, or following physical
illness or injury, e.g. diabetes, cancer, back pain or during/following pregnancy (postnatal
or perinatal depression)
Public health • Most common psychological disorder
relevance • Leading cause of disability as measured by years lost due to disability (YLDs) and the
fourth leading contributor to the global burden of disease (DALYs) in 2000
• Can mostly be reliably diagnosed and treated in primary care
Prevalence in the Half of all women and a quarter of all men will have depressive episode in their lifetime
UK

Table contd overleaf

151

Public [Link] 151 3/6/08 [Link]

 2B E p ide m io l o gy of Specific D iseases

Table 2B.1.1 contd

Time Increasing incidence worldwide
Place People living in deprived industrial areas (E&W) are more likely to be treated for
depression than people living in other areas
Person Gender: women > men
Age: increasing incidence with increasing age
Ethnicity: Western construct ➔ under-detection in some cultures
Socioeconomic status: unemployed twice as likely as employed
Life events: depression can follow adverse life events, e.g. divorce, bereavement, job loss
Disease: can follow chronic or serious disease or occur during or after pregnancy
(perinatal depression)
Prevention Secondary: effective treatment to prevent long-term sickness, reduce risk of suicide

Table 2B.1.2 Alzheimer’s disease and dementia

Clinical Loss of memory, confusion and problems with speech and understanding
characteristics
Alzheimer’s disease – progressive, one of the most common forms of dementia. Other
common forms include vascular dementia and dementia with Lewy bodies
Known causes/ Loss of acetylcholine receptors and neurons in the brain occurs in Alzheimer’s disease
aetiological
mechanisms
Public health Growing in importance (and prevalence) as population ages
relevance
There are no cures for dementia. Treatments for Alzheimer’s disease (acetylcholinesterase
inhibitors) are currently recommended by NICE for mild-to-moderate disease
Dementia is costly to the NHS/Social Services in terms of drug treatment, personal care, and
to society
Prevalence in the One person in 20 aged over 65 years and one person in 5 over 80 years of age will develop
UK dementia
Time Prevalence steadily rising, as proportion of older people in the population increases
Place Prevalence greater in countries with an older population (i.e. western Europe greater than
South Asia or Africa)
Person Gender: risk slightly greater for women than men
Age: risk increases with age (see above)
Lifestyle: smoking, sedentary lifestyle, high-fat/salt diet increases risk of vascular dementia;
excessive alcohol linked to development of Korsakoff’s syndrome
Disease/disability: high blood pressure, high cholesterol and obesity; learning disability,
Down’s syndrome
Family: genetic risk of early and late-onset forms of Alzheimer’s disease with apolipoprotein
E4 (APOE4) allele

152

Public [Link] 152 3/6/08 [Link]

 2B Epidem iology of Specific Diseases

Person (contd) Other: memory and cognition exercises are protective (e.g. doing crosswords); head injury
can increase risk
Prevention Secondary: people with mild memory loss encouraged to use their memory/cognitive skills
to preserve them

Table 2B.1.3 Schizophrenia

Clinical characteristics A chronic, often lifelong psychotic condition or group of conditions characterised by
three types of symptoms:
• Positive: hallucinations, delusions
• Negative: flat affect, low mood, withdrawal from social life, lack of motivation
• Cognitive: memory, concentration problems
Symptoms develop gradually – before diagnosis, a patient’s behaviours and mood may
have been deteriorating for some time (years). Treatment can control many of the
symptoms but can often lead to severe side effects
Known causes/ Not known. Studies have investigated neuro-anatomical and -chemical abnormalities,
aetiological including larger ventricles, imbalance in serotonin and dopamine transmission
mechanisms
Twin studies, family studies indicate genetic element. Injury/infection during
pregnancy or early in life has also been investigated
Public health relevance High burden of disease – long-term chronic illness
Higher risk of mortality from physical illness and injury (notably suicide) in people
with schizophrenia
Diagnosis/definition and treatment controversial and sometimes applied under legal
section
Cause of stigma
Prevalence in the UK 1% lifetime prevalence
Time Greater in urban areas than rural
Place Worldwide incidence fairly similar
Person Gender: men > women
Age: onset usually late adolescence; in women sometimes around menopause
Ethnicity: high rates of diagnosis in African−Caribbean but could be artefact, possibly
due to institutional racism
Socioeconomic status: higher in low income groups but note social drift in people
with schizophrenia as a result of illness
Lifestyle: cannabis (but not yet causally related)
Disease: other forms of psychosis, e.g. post-pregnancy, drug-induced psychosis (note:
not schizophrenia)
Family: higher risk if family members affected
Other: symptoms may start to appear after stressful life events
Table contd overleaf

153

Public [Link] 153 3/6/08 [Link]

 2B E p ide m io l o gy of Specific D iseases

Table 2B.1.3 contd

Prevention Secondary: symptom control with antipsychotics; psychosocial therapies used, e.g.
cognitive–behavioural therapy (CBT)
Early intervention: to promote early detection of psychosis and treatment

Table 2B.1.4 Parkinson’s disease

Clinical characteristics Progressive neurodegenerative condition leading to death of the dopamine-
containing cells of the substantia nigra. The disease manifests through its effect
on movements, such as walking, swallowing and writing
Known causes/aetiological Not known. Genetic studies have found several genes linked to parkinsonism, e.g.
mechanisms parkin gene
Public health relevance Parkinson’s disease is a frequent cause of falls, fractures and hospital admission,
and is a costly disease, especially in the later stages
Diagnosis can be difficult – there are no laboratory tests. Symptoms develop
gradually and can be mistaken for normal ageing
Covered by National Service Framework for Long Term (Neurological) Conditions
(2005)
Prevalence in the UK Thought to affect 1 in 500 people (though misdiagnosis is relatively common and
there are undiagnosed cases)
Time No marked geographical variation. Prevalence remains stable but mortality for
patients under 75 years is decreasing
Person Gender: men slightly more likely to be diagnosed than women
Age: rising prevalence with age (up to 2% of the population aged 80 and over)
Around 1 in 7 cases is diagnosed below the age of 60 years; 10% onset at 40
years
Family: rare – in most cases sporadic
Other:
• Older antipsychotic drugs can induce symptoms of parkinsonism
• Head trauma (e.g. through boxing) can increase risk of Parkinson’s disease
• Environmental exposures, e.g. herbicides? Possibly linked

CARDIOvASCULAR DISEASES

Table 2B.1.5 Coronary heart disease

Clinical characteristics Angina (chest pain on exertion or under emotional stress)
Myocardial infarction (MI, ‘heart attack’ – chest/arm pain, sweating, shortness of
breath, nausea)
Heart failure (ineffective pumping by the heart resulting in breathlessness,
oedema, tiredness)

154

Public [Link] 154 3/6/08 [Link]

 2B Epidem iology of Specific Diseases

Known causes/aetiological Atherosclerosis (narrowing of the coronary arteries due to a build-up of fat) leads
mechanisms to angina and an MI occurs if the artery becomes occluded. Heart failure can
result from subsequent damage to the heart muscle
Public health relevance Biggest cause of premature death in England
Potentially preventable
Source of health inequalities: disproportionately affects people in deprived areas
National Service Framework for CHD in England (2000) set eight standards for
services around preventing and treating heart disease
Incidence in the UK 80−90 cases of MI per 10 000 population diagnosed each year
Time Mortality rates reducing in UK since 1970s (halved from 1994 to 2004)
Place Geographical inequalities: Scotland mortality greater than England; deprived
areas greater than affluent
Person Gender: men higher than women though risk for women increases after
menopause
Age: risk increases with age
Ethnicity: mortality greater for south Asian population in UK, low in Chinese,
African−Caribbean
Socioeconomic status: higher mortality (in under-75s) in people on lower
incomes
Lifestyle: smoking, sedentary lifestyle, high-fat/salt diet; heavy drinking (light
drinking protective)
Disease: diabetes, high blood pressure, high cholesterol, obesity
Family: some rare genetic forms, e.g. familial hypercholesterolaemia
Prevention Primary: lifestyle measures to reduce smoking, encourage physical activity,
balanced diet
Secondary: providing treatment for high-risk people, e.g. those who have already
had an MI, those with diabetes or high blood pressure

Table 2B.1.6 Stroke

Clinical characteristics Ischaemic stroke (accounting for most strokes) is caused by occluded blood
vessels preventing blood reaching the brain. Haemorrhagic strokes are caused by
burst blood vessels. Transient ischaemic attacks (TIAs) occur when blood supply
to the brain is interrupted for a short time
Early signs include: numbness, weakness or paralysis on one side of the body;
slurred speech or difficulty finding words; loss of sight/blurred vision; confusion
or unsteadiness; headache
Longer-term effects depend on the affected part of the brain, the severity of the
stroke and the health of the person affected
Table contd overleaf

155

Public [Link] 155 3/6/08 [Link]

 2B E p ide m io l o gy of Specific D iseases

Table 2B.1.6 contd

Public health relevance Third biggest killer in the UK and most common cause of disability
Potentially preventable with lifestyle changes and medication
Incidence in the UK Stroke: 174−216 people per 100 000 population per year
TIAs: 35 people per 100 000 population per year
Time Incidence declining in western Europe but numbers increasing due to ageing
population
Person Gender: men higher than women though risk for women increases after
menopause
Age: risk increases with age
Ethnicity: higher risk for south Asian, African, African−Caribbean
Socioeconomic status: higher mortality (in under-75s) in people on lower
incomes
Lifestyle: smoking, sedentary lifestyle, high-salt/fat diet, heavy drinking
(affecting blood pressure)
Disease: high blood pressure, heart disease, obesity, diabetes, high cholesterol,
previous stroke or TIA
Family: higher risk if family members have had strokes
Prevention Primary: lifestyle modification to increase exercise, eat a healthy diet, reduce
alcohol intake; blood pressure checks and control; cholesterol checks and control;
treatment to reduce blood clotting
Secondary (after a stroke): lifestyle modification, treatment to reduce blood
clotting, blood pressure checks and control, peer support

COMMON WESTERN FORMS OF CANCER

Table 2B.1.7 Common cancers

Cancer Breast Lung Colorectal/bowel Prostate Cervical
type
Public Major cause of High case fatality Known risk Relatively high Common cause of
health mortality in rate that has not factors. Better incidence in older cancer in women.
relevance women, rare decreased. Large prognosis with men. Controversy Early detection
in men. Better number of cases early detection over benefits of improves prognosis.
prognosis with preventable and treatment prostate-specific Risk factors linked
early detection antigen (PSA) to social exclusion
and treatment testing (deprivation,
sexually transmitted
infections)
Incidence 41 000 cases/ 37 000 cases/year 34 000 cases/year 30 000 cases/year 3000 cases/year
in UK year

156

Public [Link] 156 3/6/08 [Link]

 2B Epidem iology of Specific Diseases

5-year 80 Men: 6; women: 7 50 71 64

survival
rate (%)
Time/ Survival rate Survival rate not Survival rate Survival rate Second most common
place improving in the improving improving. Rare in improving; more cancer in women aged
UK Nigeria die with it than <35
from it
Person Gender: women Gender: men Gender: third Age: rare under Deprivation: higher
>> men >> women but most common 50 years incidence in deprived
increasing number cancer in men; than affluent women
Age: greater Ethnicity: most
of women affected second most
proportion of common in Health: HPV (human
common in women
cases in older Deprivation: men of African papillomavirus)
women death rates higher Age: 90% occur in descent, least infection strong
in more deprived over-50s common in Asian risk factor. Vaccine
Deprivation:
communities introduced in England
Higher incidence Health: polyp Genetics: rare but
for girls. Weakened
in more affluent Exposure: in the bowel or links to BRCA-2
immune system,
communities asbestos, radon previous bowel
Exposure: through smoking, HIV.
cancer; obesity
Genetics: ~5% Lifestyle: smoking radiation Poor diet increases
of cancers due (9/10 cases), Lifestyle: diet risk
to genetic passive smoking high in meat and
Lifestyle: many
factors, mainly fat, low in fruit
Lifestyle: diet sexual partners/sex at
BRCA-1 or -2 and vegetables.
high in fruit and an early age increases
Lack of exercise
vegetables may risk
increases risk
lower risk
Tests/ Screening None available Screening Screening via Screening programme
screening programme via (short lead time) programme: faecal PSA but not for women aged
(for more mammogram occult blood recommended in 20−64 years every 3−5
detail see for women aged test detects tiny England years via liquid-based
2C.1) 50−70 amounts of blood cytology
in faeces

RESPIRATORY DISEASES

Table 2B.1.8 Asthma

Clinical characteristics A chronic condition marked by spells of shortness of breath, coughing or wheezing
caused by inflamed and narrowed airways. Asthma can be triggered by a number
of different factors, from environmental pollution to stress or smoking. Triggers for
asthma vary between people
Public health relevance Most common chronic disease in childhood; rising prevalence
Emergency admissions for patients with a diagnosis of asthma are included in the
Compendium of Public Health Indicators as a measure of health-care performance
because, in general, admissions should be avoided through effective management in
primary care or at home
Table contd overleaf

157

Public [Link] 157 3/6/08 [Link]

 2B E p ide m io l o gy of Specific D iseases

Table 2B.1.8 contd

Prevalence in the UK Between 10 and 20% of children are diagnosed with asthma
Time Increasing prevalence (possibly in part due to changes in diagnosis from bronchitis
to asthma) but decreasing severity/disability
Place
No clear geographical pattern of asthma occurrence in the UK
Person Age: more common in children than adults
Gender: more common in boys than girls
Family/genetics: some genetic component. Higher risk if family members affected
Risk factors for asthma symptoms depend on the individual but commonly include:
Lifestyle: smoking can exacerbate asthma symptoms; some people have allergies
to certain foods, e.g. dairy products, preservatives. Activities such as sex and
exercise can also lead to symptoms
Health: colds, viral infections, e.g. influenza; stress
Exposures: air pollution due to ozone, traffic fumes, smoking; house-dust mite
faeces; damp environments; pollen
Prevention Avoiding triggers
People with asthma recommended to have the annual influenza vaccine
Effectively treated with inhaled steroids and bronchodilators

Table 2B.1.9 Chronic obstructive pulmonary disease

Clinical characteristics The term chronic obstructive pulmonary disease (COPD) is used to describe a
range of conditions where airflow is obstructed, including chronic bronchitis and
emphysema. The airflow obstruction is usually progressive, not fully reversible and
does not change markedly over several months
Public health relevance Potentially avoidable disease: most cases (80%) linked to smoking
Can remain undiagnosed for years until symptoms are severe
Up to 1 in 8 emergency hospital admissions may be due to COPD
High burden of mortality (sixth leading cause of death worldwide) and morbidity
Prevalence in the UK 900 000 people in the UK have been diagnosed as having COPD, and half as many
are thought to be living with COPD without a diagnosis
Time Prevalence stable in men, increasing in women
Place Mortality in the UK from respiratory disease (including COPD) almost double
European average

158

Public [Link] 158 3/6/08 [Link]

 2B Epidem iology of Specific Diseases

Person Gender: mortality increasing in women, falling in men

Age: prevalence increases with increasing age
Socioeconomic status: more prevalent in lower income groups
Lifestyle: smoking
Other: occupation, e.g. inhaling airborne particulates
Prevention Primary prevention: stop-smoking initiatives (reducing exposures at work)
Secondary prevention: medication to control symptoms, physiotherapy to
help clear excess sputum, exercise to reduce disability, influenza and anti-
pneumococcal vaccinations to reduce the risk from viral infections

OTHER DISORDERS

Table 2B.1.10 Sickle cell disease

Clinical characteristics Genetic condition leading to abnormally formed haemoglobin that in turn leads
to anaemia, severe pain, reduced immunity and sometimes chronic kidney or
bone damage. Carriers (people with sickle cell trait who have only one affected
haemoglobin gene) do not have symptoms but may have difficulty with activities
requiring exertion or high oxygen demand (e.g. scuba diving, mountain climbing)
Known causes/aetiological Genetic: recessively inherited
mechanisms
Public health relevance Health inequalities – in the UK mainly affects people of African and Caribbean
communities
Prenatal testing can be carried out by amniocentesis and chorionic villous
sampling (CVS)
Antenatal and neonatal screening recommended for people at risk (i.e. of ethnic
origins most likely to be affected)
Prevalence in the UK African and Caribbean descent: 1 in 10−40 has sickle cell trait, 1 in 60−200 has
sickle cell disease
150−200 babies born in the UK/year with this condition
Place Most common in people of African and Caribbean descent but also found in people
from the Middle East, eastern Mediterranean (e.g. Turkey), Asia.
Person Age: occurs from birth
Ethnicity: incidence in people originally from Africa, the Caribbean, the eastern
Mediterranean, Middle East and Asia
Family: recessively inherited
Prevention Identification of affected fetuses through screening programme

159

Public [Link] 159 3/6/08 [Link]

 2B E p ide m io l o gy of Specific D iseases

Table 2B.1.11 Diabetes

Clinical Too much glucose in the blood. Diagnosis according to WHO is: fasting plasma
characteristics glucose >7 mmol/l on two separate occasions. If not well controlled, leads to range of
morbidities (heart disease, kidney failure, blindness and amputation, reduced immunity)
and can be fatal. There are two types:
• Type 1 − always requires insulin for treatment (‘insulin-controlled diabetes’)
• Type 2 − most common type of diabetes (90% of all diabetes). Can be treated with
diet, tablets (‘tablet-controlled diabetes’) or insulin (‘insulin-requiring diabetes’)
Known causes/ High levels of glucose are related to the function or presence of insulin. In type 1
aetiological diabetes, there is a shortage or absence of insulin because islet cells in the pancreas
mechanisms (which produce insulin) have been destroyed. It is not known what triggers their
destruction
In type 2 diabetes, the body produces insulin but in insufficient quantities to act, or
there is resistance to insulin, preventing it from acting to reduce glucose levels
Public health Incidence increasing and younger onset as the incidence of obesity increases
relevance
National Service Framework for Diabetes set 12 standards for diabetes services.
Emergency admissions and deaths included in the Compendium of Public Health
Indicators as a measure of health-care performance (in general, admissions and mortality
should be avoided through effective management in primary care or at home)
Prevalence in the UK 4.3% in men, 3.4% in women (diabetes diagnosed by a doctor, Health Survey for
England)
Estimates of regional prevalence using PBS (PHO-Brent-ScHARR*) diabetes population
prevalence model
Diabetes registers in primary care provide estimate of diagnosed cases in GP-registered
population. People have type 2 diabetes for an average of 6−7 years before diagnosis
*Public Health Observatory – Brent – School of Health And Related Research
Time Worldwide incidence increasing but not in UK since 1999
Place Common in some communities, e.g. Pima Indians, Pacific Islanders, south Asian
communities in the UK
Person Age: type 1 onset in childhood, declining incidence thereafter; type 2 incidence in
adults usually aged 40 or more
Ethnicity: high in populations of south Asian and African−Caribbean origin (men −
Bangladeshi 4x, Pakistani, Indian 3x more prevalent than general population; women:
Pakistani 5x, Bangladeshi and African−Caribbean 3x, Indian 2.5x more prevalent than
general population)
Socioeconomic status: more common in those on lower incomes
Disease/health: obesity, particularly central obesity − strong association with type 2;
risk of gestational diabetes in pregnancy
Lifestyle: inactivity/calorific diet
Family: particularly for type 1. Also inherited forms of type 1 and 2

160

Public [Link] 160 3/6/08 [Link]

 2B Epidem iology of Specific Diseases

Prevention Primary: weight management through promoting a balanced diet, regular exercise can
reduce the risk of type 2
Secondary (morbidities arising from diabetes): exercise, self-management, medication

161

Public [Link] 161 3/6/08 [Link]

Public [Link] 162 3/6/08 [Link]
 2C Diagnosis and Screen ing

2C
Diagnosis and Screening

2C.1 Screening for diseases 163 2C.7 Ethical, economic, legal and social
2C.2 Statistical aspects of screening 168 aspects of screening 175
2C.3 Differences between screening and 2C.8 Informed choice 176
diagnostic tests 170 2C.9 Planning, operation and evaluation of
2C.4 Case finding 171 screening programmes 177
2C.5 Likelihood ratios 172 2C.10 Developing screening policies 179
2C.6 Pre- and post-test probability 174 2C.11 Ethical, social and legal implications of a
genetic screening test 181

Diagnosis and screening are linked processes that span from the individual patient through to entire populations.
Screening is an example of a practical public health intervention that has the potential to save thousands of lives.
As with any clinical intervention, however, before offering any screening programme the first priority is to decide
whether the proposal does more good than harm. This a question that is far more complex than might initially seem
to be the case.

2C.1 SCREENING FOR DISEASES

Principles, methods, applications and organisation of screening for early detection, prevention, treatment and control
of disease
The aim of screening is to reduce the harm caused by a disease or its complications. Screening is a service in
which members of a defined population at risk of a disease are asked a question or offered a test to identify those
individuals who are more likely to be helped than harmed by further tests or treatment. A screening test does
not diagnose the disease: this is usually done by a subsequent diagnostic test.
Screening programmes vary from country to country in terms of
• Diseases that are screened
• Target age groups
• Frequency of testing
• Tests used, etc.

163

Public [Link] 163 3/6/08 [Link]

 2C D ia g no s i s and Screening

In addition, GPs and other clinicians may conduct ‘opportunistic screening’, with or without formal reward, e.g. use
of urinalysis as a screening test for diabetes.

PRINCIPLES
In order to be successful, certain prerequisites regarding the condition, treatment and programme must be met
(Table 2C.1.1). See Section 2C.9 for full details

Table 2C.1.1 Principles of screening

Condition Treatment Screening programme
The condition is an important health The condition can be treated better There is robust evidence* that
problem (i.e. common and serious) at an early stage the screening programme reduces
mortality and morbidity
Its epidemiology is well understood There is a suitable and acceptable
diagnostic test There are facilities to perform the
It has a latent period
further diagnostic tests required
There is a favourable balance of
benefit against harm
The opportunity cost is justified by
the benefit
*Such as randomised controlled trials.
It must always be stressed to screened people that a negative test does not mean that they are necessarily disease
free.

METHODS AND APPLICATIONS

UK UK National Screening Programmes
See Tables 2C.1.2–2C.1.5.

164

Public [Link] 164 3/6/08 [Link]

 2C Diagnosis and Screen ing

Table 2C.1.2 Antenatal screening programme

Down’s syndrome The combined test is available to women who request screening in the first trimester
and understand the implications of doing so. It is based on combining an ultrasound
measurement of nuchal translucency (NT; thickness of fluid at the nape of the fetal neck
as assessed on ultrasound), human chorionic gonadotrophin (hCG), pregnancy-associated
plasma protein A (PAPP-A) and the woman’s age
The integrated test provides better performance than the combined test if the woman
is prepared to wait until the second trimester for the result. It integrates measurements
performed at different times during pregnancy into a single test result. Typically it refers
to the integration of NT and PAPP-A in the first trimester with the quadruple test in the
second
The quadruple test is offered to women who attend for screening in the second
trimester, and also forms part of the integrated test as described above. It is a second
trimester test consisting of a-fetoprotein (AFP), unconjugated oestriol (uE3), hCG,
inhibin-A and the woman’s age (the triple test does not include inhibin-A)
If the pregnancy is deemed to be of high risk, then the woman is offered amniocentesis
(or chorionic villous sampling [CVS] if <13 weeks’ gestation)
Fetal anomaly scan This is performed at 18−20 weeks’ gestation. It is not universally performed in Scotland
Other Screening for the following conditions is offered to all women:
• Anaemia
• Bacteriuria
• Blood group and rhesus D status
• HBV (hepatitis B virus) and HIV
• Syphilis
• Rubella immunity
• Spina bifida
• Pre-eclampsia
• Sickle cell disease and thalassaemia (not in Scotland)
Those women who are at high risk are offered screening for:
• Psychiatric illness
• Tay−Sachs disease

Table 2C.1.3 Newborn screening programme

Heel-prick test Used to test for:
• Phenylketonuria
• Congenital hypothyroidism
• Sickle cell disease and thalassaemia
• Cystic fibrosis
Hearing screening programme Oto-acoustic emissions test or attenuated brain-stem audiometry
Physical examination Offered during first 72 hours of life to detect, among others:
• Congenital heart disease
• Congenital cataract
• Congenital malformation
• Cryptorchidism
• Developmental dislocation of the hip

165

Public [Link] 165 3/6/08 [Link]

 2C D ia g no s i s and Screening

Table 2C.1.4 Children’s screening programmes

Growth Height and weight should be measured at the time of school entry
Hearing Screening for hearing loss is offered to school-aged children
vision Screening for visual impairment is offered at age 4−5 years

Table 2C.1.5 Adult screening programmes

Disease Screening test Subsequent test Age group Frequency
Abdominal aortic Abdominal ultrasound Repeat ultrasound Men aged 65 years One-off
aneurysm*
Breast cancer Mammography Fine-needle 50−70 (first 3-yearly
aspiration invitation always
before age 53)
Cervical cancer Smear test or liquid- Colposcopy In England:
based cytology • 25−49 years • 3-yearly
• 50–64 years • 5-yearly
In Scotland: 20−60 • 3-yearly
years
Bowel cancer Faecal occult blood Colonoscopy 60−69 2-yearly
Diabetes* Questionnaire and Urinalysis or oral 40+ years To be determined
body mass index glucose tolerance
(BMI) measurement test
Genital Chlamydia Nucleic acid Genitourinary clinic Young men and Opportunistic
trachomatis amplification test or outreach clinic women
appointment
Prostate cancer Prostate-specific Prostate biopsy Men aged 50+ Provided only on
antigen (PSA) request
vascular risk A Vascular Risk Management Programme is being developed in which the whole population
will be offered a risk assessment that could include blood pressure, cholesterol and glucose
measurements
*Being piloted in England.

PROGRAMMES IN INDIvIDUAL COUNTRIES

Wal Scot There is an all-Wales diabetic retinopathy screening programme, which builds on experience in a
programme in the Cardiff area. Scotland also has a national diabetic retinopathy screening programme.
Aus Details of Australian screening programmes are shown in Table 2C.1.6.

166

Public [Link] 166 3/6/08 [Link]

 2C Diagnosis and Screen ing

Aus Table 2C.1.6 Australian screening programmes

Australian screening Details
programme
Antenatal and newborn Very similar to UK
screening
Breast cancer Actively offered, by direct invitation, to all women aged 50−69 years at 2-yearly
intervals. Breast screening is also freely available to women from age 40 years
Cervical cancer Cervical screening is supported by a back-up record system that enables reminders,
recall and comparisons with previous histology. There is current debate about changing
to 3-yearly and the impact of the new HPV vaccine
Bowel cancer Use of an immunochemical faecal occult blood test has been successfully pilot tested.
Provided that adequate resources for diagnosis and treatment can be demonstrated, may
be rolled out over the coming years
Other Informal screening for melanoma (physical examination) and prostate cancer (PSA
testing and digital rectal examination) occurs in the primary care setting − but neither
meets the WHO criteria

NZ Screening activities in NZ include both screening programmes and opportunistic screening. The former are
distinguished by the fact that all steps along the screening pathway are planned, coordinated and resourced, and
are provided with quality assurance and programme monitoring.
Screening programmes in NZ
• Breast cancer screening: ‘BreastScreen Aotearoa’ (BSA)
• Cervical screening: National Cervical Screening Programme (NCSP)
• Newborn Baby Metabolic Screening (for phenylketonuria, maple syrup urine disease, galactosaemia, biotinidase
deficiency, congenital adrenal hyperplasia, congenital hypothyroidism, cystic fibrosis)
• Adult hepatitis B screening.
Opportunistic screening in NZ
This generally lacks the quality processes that are a feature of screening programmes and usually relies on the
population of people who present to the health service for other reasons. Examples of such screening:
• Screening for hearing impairment at school entry
• Antenatal screening: anaemia; rhesus incompatibility (to avoid newborn haemolytic disease); gestational
diabetes; serology for syphilis, rubella, hepatitis B; ultrasound screening for anatomical abnormalities, e.g.
neural tube defects; risk factors for HIV; chromosomal abnormalities, e.g. Down’s syndrome (NT + maternal
serum screening)
• Newborn physical examination, e.g. screen for congenital hip dislocation, undescended testes, cardiac
abnormalities
• Well Child screening for developmental delays
• Screening for complications of diabetes (retinal, foot and kidney)
• Screening for breast cancer with clinical breast examination
• Mammographic breast screening outside of BSA
• Diabetes screening
• Colorectal cancer screening
• Prostate cancer screening
• Cardiovascular disease risk factor screening (smoking, serum cholesterol, hypertension)
• Screening for alcohol and drug misuse among adolescents and adults

167

Public [Link] 167 3/6/08 [Link]

 2C D ia g no s i s and Screening

• Osteoporosis risk factor screening (which may include bone mineral density scanning)
• Screening for congenital hearing impairment.
Adapted from National Advisory Committee on Health and Disability (2003).

2C.2 STATISTICAL ASPECTS OF SCREENING

Statistical aspects of screening tests, including knowledge of and ability to calculate, sensitivity, specificity, positive
and negative predictive values, and the use of receiver operating characteristic (ROC) curves
The accuracy of a screening test is expressed in four dimensions (sensitivity, specificity, positive predictive value
and negative predictive value), calculated as described in Box 2C.2.1 and Table 2C.2.1. Note how the prevalence of
the condition affects some – but not all − of the test performance characteristics.

Box 2C.2.1
Does the person truly
have the condition?
YES NO
a (true b (false
POSITIvE
Test positive) positive)
result c (false d (true
NEGATIvE
negative) negative)

Table 2C.2.1 Screening test dimensions, calculations and effect of high prevalence
Dimension Definition Calculation* Effect of high
prevalence
Sensitivity Proportion of those people who have the disease a No effect
who are correctly detected by the test a+c
Specificity Proportion of those people who do not have the d No effect
disease who are correctly left undetected by the b+d
test
Positive Proportion of those testing positive who truly have a Increases PPV
predictive the disease a+b
value
(‘yield’)
(PPv)
Negative Proportion of those testing negative who are truly d Decreases NPV
predictive disease free c+d
value (NPv)
*Note that in an examination, candidates may need to redraw the 2 ¥ 2 matrix given so that it is in the same format
as shown above – otherwise these calculations will of course be incorrect.
The more sensitive the test, the less likely it is that a negative result will be a true positive − and hence the higher
the negative predictive value.
An example is shown in Box 2C.2.2.

168

Public [Link] 168 3/6/08 [Link]

 2C Diagnosis and Screen ing

Box 2C.2.2
Example: Cervical cancer screening test
88 084 women are screened for cervical cancer but 73 of the 86 569 women who tested negative turned out to
have the disease. What is the sensitivity, specificity, positive predictive power (PPP) and negative predictive
power (NPP) of the test?

Cervical cancer
Confirmed Refuted
Smear test Positive 267 1248 1515
Negative 73 86 496 86 569
340 87 744

Sensitivity = 267 ∏ 340 = 0.785 or 78.5%

Specificity = 86 496 ∏ 87 744 = 0.986 or 98.6%
PPP = 267 ∏ 1515 = 0.176 or 17.6%
NPP = 86 496 ∏ 86 569 = 0.999 or 99.9%

THRESHOLD SETTING
All screening and diagnostic tests require a threshold level to be defined (e.g. systolic blood pressure of 130 mmHg
for a diagnosis of hypertension). Most patients will be clearly normal or abnormal, but some will remain in the grey
area between the two and a line needs to be drawn by the investigator in this grey zone to determine how the test
will perform when it is used. The positioning of this line will always constitute a compromise between sensitivity
(i.e. picking up everyone who has the condition) and specificity (i.e. avoiding healthy people being labelled as
positive for the test). See Box 2C.2.3.

Box 2C.2.3
Reasons for setting the line towards high sensitivity Reasons for setting the line towards high specificity
• Serious disease with definitive treatment • Unpalatable treatment
• Risk of infectivity to others • Costly, risky subsequent diagnostic test
• Subsequent diagnostic test cheap and low risk

PARALLEL AND SERIAL TESTING

Two screening tests are often necessary. These may be offered at the same time (parallel testing) or sequentially
(serial testing). See Box 2C.2.4.

Box 2C.2.4
Parallel testing Serial testing
Procedure Two screening tests performed at the same time Second screening test is performed only if the
and the results are subsequently combined result of the first screening test is positive
Effect Higher sensitivity but lower specificity Improves specificity at the cost of lower
sensitivity

169

Public [Link] 169 3/6/08 [Link]

 2C D ia g no s i s and Screening

ROC CURVES
A receiver operating characteristic (ROC) curve is a plot of (sensitivity) versus (1 − specificity). The name derives
from its original use in radar technology (see Figure 2C.2.1). The dotted line shown in the figure represents a
useless test that has no discriminatory power.

100%
The size of the area between
Sensitivity

the ROC curve and the

dotted line reflects the ability
of the test to discriminate
between diseased and non-
diseased individuals across
the range of potential cut-offs

1 – specificity

Figure 2C.2.1 ROC curve

Another, perhaps more intuitive, way of displaying this same information is shown in Figure 2C.2.2.

Specificity
0.8
Sensitivity or specificity

0.6

0.4
Sensitivity

0.2

0
50 100 150
Value at which the test becomes positive

Figure 2C.2.2 Alternative depiction of ROC curve

2C.3 DIFFERENCES BETWEEN SCREENING AND DIAGNOSTIC TESTS

Screening tests are offered to asymptomatic people who may or may not have early disease or disease precursors
and are used to guide whether or not a diagnostic test should be offered. Diagnostic tests are offered to people
who have a specific indication of possible illness (a history, symptom, sign or screening test result) to determine
whether or not they have the disease in question.

170

Public [Link] 170 3/6/08 [Link]

 2C Diagnosis and Screen ing

COMPARISON BETWEEN SCREENING AND DIAGNOSTIC TESTS

See Table 2C.3.1.

Table 2C.3.1 Differences between screening and diagnostic tests

Screening test Diagnostic test
Result The cut-off is set towards high sensitivity. The cut-off is set towards high specificity,
As a result many of the positive results with more weight given to diagnostic
are falsely positive. This is acceptable, precision and accuracy than to the
particularly if the screening test is not acceptability of the test to patients
harmful or expensive
Cost Since large numbers of people will be Patients have symptoms that require accurate
screened to identify a very small number of diagnosis and therefore higher costs are
cases, the financial resources needed must be justified
justified carefully
Result of test The result of the test is an estimate The test provides a definitive diagnosis (e.g.
of the level of risk (e.g. risk of Down’s a definite diagnosis of Down’s syndrome
syndrome in antenatal screening is based through CVS)
on a combination of maternal age, AFP,
nuchal fold, etc.) and determines whether
a diagnostic test (e.g. amniocentesis) is
justified
Invasiveness Often non-invasive May be invasive

Population offered Those at some risk but without symptoms of Those with symptoms or who are under
the test disease investigation following a positive screening
test

2C.4 CASE FINDING

Case finding is a strategy for targeting resources at individuals or groups who are suspected to be at particular
risk of a disease. It involves actively searching systematically for high-risk people, rather than waiting for them to
present themselves to medical attention after symptoms or signs of active disease have occurred.
Note the similarities between case finding and screening: both seek to risk stratify the population using a simple
and cheap procedure, and assume that better outcomes can be achieved through identifying the early stages of
disease and offering prompt treatment.
Advantages and disadvantages of case finding are listed in Box 2C.4.1.

ExAMPLES OF CASE FINDING

Case finding may be used as part of the investigations into an outbreak of a communicable disease (e.g. syphilis) to
identify potential sources of the disease. It may also be employed during food-borne outbreaks to identify as many
at-risk individuals as possible.
Health data systems can be used to identify ‘missed’ risk groups (e.g. registered GP patients over 50 years of age
with a BMI >30 who may not be on the register of people at risk of coronary heart disease).

171

Public [Link] 171 3/6/08 [Link]

 2C D ia g no s i s and Screening

Box 2C.4.1
Advantages Disadvantages
Cheap Potential to widen health inequalities because
some high-risk groups are hard to reach
Low personnel demand
(homeless, refugees, etc.)
Case finding improves the positive predictive value of a
diagnostic test by targeting high-risk patients with higher
underlying prevalence
By targeting preventive care, case-finding tools can help
improve care of individuals and reduce costs for the state
Cost-effective method for identifying cases of familial
conditions such as familial hypercholesterolaemia

ENG The King’s Fund Patients at Risk of Re-Hospitalisation (PARR) and Combined Model case-finding tools use
patterns in routinely collected data to forecast which individuals in a population are at high risk of emergency
hospital admission in the forthcoming year.

2C.5 LIKELIHOOD RATIOS

The likelihood ratio combines the sensitivity and specificity of a test. It provides a unified estimate of the degree
to which a test result changes the odds of having a disease:
• The likelihood ratio for a positive result (LR+) indicates by how much the odds of the disease increase when a
test is positive
• The likelihood ratio for a negative result (LR-) indicates by how much the odds of the disease decrease when a
test is negative.
See Box 2C.5.1.

Box 2C.5.1
Likelihood ratio of a positive test result
True +ve rate Sensitivity
(LR+) = =
False +ve rate (1 − Specificity)

Likelihood ratio of a negative test result

False -ve rate (1 − Sensitivity)
(LR−) = =
True -ve rate Specificity

Bayes’ theorem (see Section 1B.19) states that the value of a diagnostic test is affected by the pre-test odds, i.e. by
the likelihood, prior to testing, that the patient has the disease. The likelihood ratio should therefore be considered
in conjunction with the following three pieces of information:
1. Prevalence of the disease
2. Characteristics of patient pool
3. Information about particular patient to determine the post-test odds of disease.

172

Public [Link] 172 3/6/08 [Link]

 2C Diagnosis and Screen ing

INTERPRETATION OF LIKELIHOOD RATIOS

The further away a likelihood ratio (LR) is from 1, the stronger the evidence for the presence or absence of disease.
LR >1 indicates that the test result is associated with the presence of the disease.
LR <0.1 indicates that the test result is associated with the absence of disease.
See Table 2C.5.1.

Table 2C.5.1 Interpretation of likelihood ratios

value of likelihood Interpretation
ratio
<0.1 Large pre- to post-test changes in probability, i.e.
provides strong evidence against the diagnosis in most
circumstances
0.1–0.2 Moderate pre- to post-test changes in probability
0.2–0.5 Small pre- to post-test changes in probability but may
be useful
0.5−2 No useful change in pre- to post-test probability
2–5 Small pre- to post-test changes in probability but may
be useful
5–10 Moderate pre- to post-test changes in probability
>10 Large pre- to post-test changes in probability, i.e.
provides strong evidence in favour of the diagnosis in
most circumstances

USE OF LIKELIHOOD RATIOS

Note that the calculation of likelihood ratios is not currently tested as a skill in the UK MFPH Part A examination.
The post-test odds (i.e. the likelihood that a particular patient has the disease) are estimated by multiplying the
pre-test odds by the likelihood ratio:
Post-test odds = pre-test odds ¥ likelihood ratio
Factors affecting pre-test and post-test odds are shown in Box 2C.5.2.

Box 2C.5.2
Factors affecting pre-test odds Factors affecting post-test odds
Prevalence of the disease Prevalence of the disease in the
catchment population
Patient-specific patient risk factors
(pre-test odds)
Diagnostic test itself (the likelihood
ratio)

The use of odds, rather than risks, makes the calculation of post-test odds slightly complex – but a nomogram can
be used to transform odds into probabilities: see Figure 2C.5.1.

173

Public [Link] 173 3/6/08 [Link]

 2C D ia g no s i s and Screening

Pre-test Post-test
probability probability
Mark the estimated probability of the
0.001 0.999 disease prior to testing (e.g. 0.1)
0.002 0.998
and mark the likelihood ratio for the
0.003 0.997 diagnostic test (e.g. 20), then draw
0.005 0.995 a line connecting the two. Extend
0.007 0.993
0.01 0.99 this line until it intersects with the
Likelihood post-test probability to find the new
0.02 ratio 0.98
0.03 1000 0.97 estimate (0.7).
500
0.05 0.95
0.07 200 0.93
0.1 100 0.9
50
0.2 20 0.8
10
0.3 5 0.7
0.4 2 0.6
0.5 1 0.5
0.6 0.5 0.4
0.7 0.2 0.3
0.1
0.8 0.05 0.2
0.02
0.9 0.01 0.1
0.93 0.005 0.07
0.95 0.002 0.05
0.97 0.001 0.03
0.98 0.02

0.99 0.01
0.993 0.007
0.995 0.005
0.997 0.003
0.998 0.002

0.999 0.001

Figure 2C.5.1 Fagan’s nomogram for calculating post-test probabilities. Reproduced from Fagan (1975) with
permission from the New England Journal of Medicine

2C.6 PRE- AND POST-TEST PROBABILITY

See also Sections 2C.5 and 1B.19.
By comparing the pre- and post-test probabilities, it is possible to determine whether surety of diagnosis has risen
(i.e. the post-test probability has increased) or fallen (i.e. post-test probability has decreased). In this way, it is
possible to provide comprehensive information about a screening test in order to enable informed choice.

PRE-TEST PROBABILITY (OR PREVALENCE)

This is the proportion of people in the population at risk who have the disease at a specific time or time interval,
i.e. the point prevalence or the period prevalence of the disease. In other words, it is the probability − before the
diagnostic test is performed − that a patient has the disease. Pre-test probabilities may be estimated from routine
data, practice data or clinical judgement.

POST-TEST PROBABILITY
This is the proportion of patients testing positive who truly have the disease. It is the same as the positive
predictive value.

174

Public [Link] 174 3/6/08 [Link]

 2C Diagnosis and Screen ing

Post-test odds
Post-test probability =
1 + Post-test odds

2C.7 ETHICAL, ECONOMIC, LEGAL AND SOCIAL ASPECTS OF SCREENING

Screening programmes are inherently attractive to the public and to the media because of their potential to ‘nip a
problem in the bud’. This can lead to undue pressure on policy makers to introduce a screening programme without
considering the opportunity costs and harms such as those from false results. For this reason, the ethical, economic
and social consequences should always be carefully considered prior to the introduction of a new screening
programme.

ETHICS
Beauchamp and Childress (2001) set out one of the most widely used frameworks in medical ethics first described in
1979. It consists of four principles, namely beneficence, non-malfeasance, justice and autonomy (Table 2C.7.1).
Screening programmes have the potential to violate each of these.

Table 2C.7.1 Beauchamp and Childress’s ethical principles applied to screening

Justice Screening programmes should be used only when all other primary preventive measures are in
place because primary prevention is likely to be more cost-effective than screening. In certain
circumstances (e.g. late-presenting cancers) a screening programme does offer a cost-effective
means of improving health. A challenge for those running screening programmes is to ensure
that uptake is as good among the deprived populations (who are typically at greatest risk) as it
is among more affluent populations
Screening programmes are examples of the prevention paradox (see Section 2H.4), which states
that the majority of the people receiving a preventive intervention (e.g. screening) will not
benefit from that intervention. It could be argued that this is unjust
Autonomy Communicating risk to patients is notoriously difficult, and it is questionable whether people
who partake in a screening programme truly understand the consequences of their participation
– especially as the parameters of a screening test are intentionally set at <100% accuracy
Beneficence All those screened are by definition symptom free but at some risk. Screening programmes have
large benefits for those who can be offered early treatment, but most will not benefit in this
way
Non- Potential harm from a screening programme includes:
malfeasance • Psychological harm from false positives in the interval before diagnostic testing
• Iatrogenic harm from the subsequent diagnostic test (which is often invasive)
• Unwarranted reassurance from false negatives (may cause people to belittle symptoms that
develop later)

ECONOMICS
A health economist may consider the opportunity costs (i.e. opportunities foregone) from the following
perspectives:
• Health sector
• Patients and their families
• Other sectors.
The evaluation of a screening programme would ideally take all costs from each of these perspectives into
consideration as part of a randomised controlled trial. Since this is not always possible (especially if the disease in

175

Public [Link] 175 3/6/08 [Link]

 2C D ia g no s i s and Screening

question manifests itself only much later), other forms of economic analysis are needed (e.g. decision-tree analysis).
All evaluations should be subjected to sensitivity analysis, and should take account of discounting (see Section
4D).
Economic analyses of adult screening programmes should specifically address the following points:
• At what age should individuals begin to be offered screening for this disease?
• How often should individuals be screened?

LEGAL ASPECTS
See Table 2C.7.2.

Table 2C.7.2 Legal aspects of screening

Access Most screening tests are accessed only via an authorised clinician
Accreditation In order to assure standards, diagnosticians (cytopathologists,
radiologists, etc.) must be registered and accredited by a statutory
body. To retain their accreditation, they must deal with a minimum
number of abnormal cases each year
vulnerable For certain diagnoses (e.g. genetic screening for Huntington’s
groups disease), the right of children not to be screened may be protected
in law. Likewise, the law offers protection to prisoners and to people
with learning difficulties against coercion into screening programmes
Confidentiality Data from screening programmes would potentially be valuable
to companies offering life insurance. For this reason the data are
carefully protected

SOCIAL ASPECTS
Social aspects of a screening include those factors that may affect participation, such as those shown in Box 2C.7.1.

Box 2C.7.1
Factors that increase participation in Factors that decrease participation in
screening screening
Perception of disease severity Disease phobia
Perception of susceptibility to the
disease

Health beliefs and attitudes are important influences on preventive behaviour, and they tend to vary between
subgroups of the population. These should be considered at the planning stage in order to ensure that any new
screening programme does not exacerbate health inequalities, e.g. between social classes or between different
ethnic groups.

2C.8 INFORMED CHOICE

An informed choice is a decision based on accurate information regarding rights, risks and benefits that has been
fully understood. In the context of screening, the decision of interest is whether or not to participate in the
programme.

176

Public [Link] 176 3/6/08 [Link]

 2C Diagnosis and Screen ing

REQUIREMENTS FOR INFORMED CHOICE

UK The General Medical Council advises that patients require the following information in order to make an
informed decision with regard to a screening programme:
• Purpose of the screening
• Likelihood of positive and negative findings
• Possibility of false results
• Risks attached to the screening process
• Any significant implications of screening for the particular condition: medical, social or financial
• Follow-up plans, including availability of counselling and support services.

2C.9 PLANNING, OPERATION AND EvALUATION OF SCREENING PROGRAMMES

The introduction of a new screening programme is a major undertaking. The infrastructure, logistics and workforce
issues all need to be planned carefully, and an ongoing audit of the programme should form an integral part of the
plans.

PLANNING AND OPERATION

The planning, operation and evaluation of a screening programme can all be considered with regard to the disease,
the screening test and the treatments offered (Figure 2C.9.1).

Test Treatment Disease

Figure 2C.9.1 Factors to consider in the planning, operation and evaluation of screening programmes

The WHO criteria for assessing the viability, effectiveness and appropriateness of a screening programme are based
on those of Wilson and Jungner (1968) and are summarised in Table 2C.9.1.

EVALUATION
Evaluation of a potential screening programme involves consideration of three main issues: feasibility, effectiveness
and cost.

FEASIBILITY
Feasibility will depend on:
• Relative ease with which the population can be organised to attend for screening
• Acceptability of the screening test
• Existence of facilities and resources to perform the necessary diagnostic tests and treatments post-screening.

EFFECTIvENESS
This is the extent to which implementing a screening programme affects the subsequent outcomes. This is difficult
to measure because of three biases that must be borne in mind when evaluating a screening programme: see Table
2C.9.2.

177

Public [Link] 177 3/6/08 [Link]

 2C D ia g no s i s and Screening

Table 2C.9.1 Wilson and Jungner’s (1968) criteria for assessing screening programmes
Disease Important health problem (i.e. common and serious)
Well-recognised pre-clinical stage
Natural history understood including development from latent to declared disease −
detectable risk factor, disease marker
Long latent period (i.e. time between first detectable signs and overt disease)
Diagnostic test valid (sensitive and specific)
Simple and cheap
Safe and acceptable
Reliable
Diagnosis and Agreed policy on the further diagnostic investigation of individuals with a positive test result
treatment and on the choices available to those individuals
Facilities are adequate
Evidence-based policies covering which individuals should be offered treatment and the
appropriate treatment to be offered
Effective, acceptable and safe treatment available
Cost-effective
Sustainable
Overall screening Evidence from high-quality randomised controlled trials that the screening programme is
programme effective in reducing mortality or morbidity
Evidence that the complete screening programme (test, diagnostic procedures, treatment/
intervention) is clinically, socially and ethically acceptable to health professionals and the
public
Opportunity cost of the screening programme (including testing, diagnosis and treatment)
should be economically balanced in relation to expenditure on medical care as a whole
Adequate staffing and facilities for testing, diagnosis, treatment and programme
management should be available prior to the commencement of the screening programme
Clear management, monitoring and quality assurance
Those offered screening must be able to make informed choices

COST
Resources for health care will always be scarce relative to competing demands. The cost-effectiveness of a screening
programme compared with other forms of health care (including public health interventions, such as stop-smoking
services) should therefore be considered. Costs relate not just to the implementation of the screening programme
but also to the further diagnostic tests and the subsequent costs of treatment. Moreover, the costs associated with
the absence of screening must be considered too: costs would be incurred in the treatment of patients with more
advanced stages of disease.

178

Public [Link] 178 3/6/08 [Link]

 2C Diagnosis and Screen ing

Table 2C.9.2 Types of biases in screening programmes

Type of bias Description
Selection bias Members of the public who participate in screening programmes differ systematically from
those who do not (e.g. in breast screening, those who present for screening are generally
from higher socioeconomic groups and have English as a first language). Performing an
intention-to-treat analysis in the evaluation eliminates this form of bias
Lead-time bias People in the screened group appear to survive longer after diagnosis than those in the
unscreened group. However, screened patients will be made aware of their disease earlier
(the lead time) and this must be subtracted from the total survival time when making
comparisons with the unscreened group. This is illustrated in the diagram below in which
the apparent survival time from diagnosis is longer in those detected by screening (5 years)
than for those detected clinically (3 years) – whereas in fact the date of death is the same
in both groups
Cancer
starts Clinically obvious Death
3 years

5 years
Screen
detected

Length-time bias Cases detected through screening will tend to have less aggressive forms of the condition.
This is because fast-growing tumours will have a shorter pre-clinical stage in which to be
detected by screening. Because such fast-growing tumours also have a worse prognosis,
it means that survival will appear to be better in cases detected by screening than those
detected clinically

2C.10 DEvELOPING SCREENING POLICIES

Evidence basis needed for developing screening policies and implementing screening programmes, including established
programmes (e.g. breast and cervix) and those currently in development, being piloted or subject to major research
activity (e.g. colon cancer, Chlamydia, antenatal/neonatal screening)
Evidence for screening programmes determines which of the policy options shown in Table 2C.10.1 should be taken
regarding established and proposed programmes.

Table 2C.10.1 Policy options for screening programmes

Current Current programme should continue unchanged
programmes Current programme should continue but be revised
Current programme should be stopped
Proposed Proposed programme should be introduced, provided that the
programmes resources, both financial and human, are available to ensure
adequate quality standards
Proposed programme should not be introduced (e.g. prostate
screening not recommended until further evidence shows there to
be a reliable test for screening purposes)

179

Public [Link] 179 3/6/08 [Link]

 2C D ia g no s i s and Screening

EVIDENCE BASE FOR SCREENING POLICIES

UK The National Screening Committee (NSC) advises the UK’s four chief medical officers on screening policy. It
considers evidence about the benefits, risks and costs of screening for conditions currently being investigated by
research, and evaluates this evidence according to the WHO screening criteria as set out by Wilson and Jungner (see
Section 2C.9). Final policy decisions on screening are based on rigorous assessment of the health technology – often
by means of randomised controlled trials.
Before being rolled out nationwide, pilots are first established to compare the theoretical benefits of the screening
shown in a research setting (‘efficacy’) with those in an ordinary service setting (‘effectiveness’). This is important
because there may be practical issues that were hitherto unapparent, e.g. staff in a service setting may not be as
committed and skilled as those in a research team.

UK SCREENING PROGRAMMES
UK UK screening programmes that have been established or are in development are described in Section 2C.1. The
evidence base can be accessed via the NSC website ([Link]).

NO CURRENT UK SCREENING
UK Some of the conditions in the UK for which routine screening is not currently recommended are listed below.

ANTENATAL
• Bacterial vaginosis
• Cystic fibrosis
• Diabetes
• Fragile X syndrome
• Genital herpes
• Hepatitis C
• HTLV-1
• Postnatal depression
• Preterm labour
• B group streptococci
• Thrombocytopenia
• Thrombophilia
• Toxoplasmosis.

NEWBORN
• Amino acid metabolism disorders
• Muscular dystrophy
• Neuroblastoma
• Organic metabolism disorders
• Thrombocytopenia.

ADULT
• Cancers: anal, bladder, lung, oral, ovarian, stomach, testicular
• Diabetes (being piloted in England)
• Hepatitis C
• Hyperlipidaemia
• Osteoporosis
• Renal disease

180

Public [Link] 180 3/6/08 [Link]

 2C Diagnosis and Screen ing

• Thrombophilia
• Thyroid disease
• Vision.

2C.11 ETHICAL, SOCIAL AND LEGAL IMPLICATIONS OF A GENETIC SCREENING TEST

While all screening programmes have important ethical, social and legal dimensions (see Section 2C.7), these issues
are brought to the fore in the context of genetic screening: see Table 2C.11.1.

Table 2C.11.1 Ethical, social and legal implications of a genetic screening test
Ethical Justice Equitable access to genetic tests needs to be ensured

Autonomy Autonomy may be compromised if parents test their

children for adult-onset diseases

Beneficence Use of genetic information can be valuable in

reproductive decision-making

Non- The potential for long-term psychological harm from the

malfeasance knowledge of a diagnosis needs to be considered

Social Knowing the genetic profile of an individual has the potential to cause
psychological impact including stigmatisation
Society must deal with the uncertainties associated with genetic
susceptibility tests for complex conditions (e.g. heart disease) that are
linked to multiple genes and gene−environment interactions
Legal Certain people are incapable of providing valid consent, e.g. people
with learning difficulties or prisoners. There are therefore legitimate
concerns about the validity of the consent provided by a third party
Privacy and confidentiality of genetic information
Legislation must ensure fairness in the use of genetic information by
insurers, employers, courts, schools, adoption agencies, the military,
etc., and outlaw the potential for discrimination

WORLD HEALTH ORGANIZATION RECOMMENDATIONS

See Table 2C.11.2.

181

Public [Link] 181 3/6/08 [Link]

 2C D ia g no s i s and Screening

Table 2C.11.2 WHO recommendations with regard to genetic screening (WHO 2007)
Disease Need for universal standard definitions
Test Regulatory framework and criteria for test development and use
Public information and education
Quality assurance
Professional development
Partnerships and collaborations
Informed consent
Consent procedures for children and vulnerable individuals in human genetic research
Outcome Protection from discrimination
Data protection: confidentiality, privacy and autonomy

182

Public [Link] 182 3/6/08 [Link]

 2D Genetics

2D
Genetics

2D.1 Human genetics 183 2D.6 Polygenic disorders 189

2D.2 Inherited diseases 186 2D.7 Gene−environment interactions 191
2D.3 Patterns of inheritance 186 2D.8 Genes in health and disease 192
2D.4 Penetrance 188 2D.9 Diseases in relatives 193
2D.5 Different genotypes and phenotypes 189 2D.10 Molecular biology 194

Public health genetics is the science of translating genome-based knowledge and technology into benefits at a
population level. Since almost all diseases have a genetic component, genetics has a profound influence on almost
all branches of public health – an influence that is set to grow rapidly in coming years as knowledge in this field
expands. Links between genetic mutations and disease have been identified and can now be used to identify
people at risk of certain conditions. They are also increasingly being used to develop revolutionary approaches to
treatment.
The Part A syllabus requires candidates to understand the basic principles relevant to public health genetics such
as patterns of inheritance and gene–environment interactions. This chapter covers these principles and includes
examples of practical relevance to public health.

2D.1 HUMAN GENETICS

Genetics is the study of how characteristics (including conditions and diseases) are inherited. The related field of
genomics considers the interactions of compete sets of genes and their products. Cells in the human body contain
genetic material in the form of DNA (deoxyribonucleic acid). DNA is most commonly found in chromosomes, located
in each cell’s nucleus (see Figure 2D.1.1). There are 23 pairs of chromosomes per cell, of varying sizes. Twenty-two
of these are autosomal (i.e. not sex linked) but there is one pair of sex chromosomes for each cell: xx in females
or xY in males. Each chromosome carries a number of genes (segments of DNA that encode a particular enzyme or
protein). It is estimated that there are between 20 000 and 25 000 genes in total in the human genome.

183

Public [Link] 183 3/6/08 [Link]

 2D G e ne t ic s

Chromatid Chromatid
Telomere

Nucleus
Centromere

Telomere

Cell – Most DNA is Chromosome – There are 46 DNA – Stored as a double

located in the nucleus, human chromosomes, with a helix and is stored in each
in chromosomes. telomere capping the ends, and chromosome.
proteins called histones
controlling the structure.

Sugar- Sugar-
phosphate Base pairs phosphate
backbone backbone

P A A
S
Hydrogen
S bonds P

P G G
S
S
P

P T T Each DNA sequence comprises DNA base

S
pairs – there are around 3 million, each
S
P comprising nucleotides, either purine (A and G)
A A or pyrimidine (C and G) bases with sugars
P (S) attached.
S
S
P

P C G
Base pair S
S
P

P G C
S
S
P
Nucleotide
Figure 2D.1.1 Storage of genetic material in a human cell
184

Public [Link] 184 3/6/08 [Link]

 2D Genetics

DNA consists of four chemical bases (also known as ‘nucleotides’) labelled A, G, C and T. A group of three base-
pairs (also known as a ‘triplet nucleotide sequence’ or a ‘codon’) represents instructions, which ultimately produce
proteins. DNA acts as a template for RNA (ribonucleic acid), produced by transcription. This in turn is translated
into amino acids. These are joined together to produce polypeptides, which themselves are joined to produce
proteins (see Figure 2D.1.2).

DNA

Replication

Transcription

RNA

Translation

Protein

Figure 2D.1.2 Information stored as DNA is transcribed into RNA, which is then translated into proteins

Interspersed between each sequence that codes for a protein (called an exon) is a long sequence of DNA (called an
intron) that does not code for proteins. Exons represent only a small proportion (2%) of the DNA in a chromosome.

VARIATIONS IN GENETIC MATERIAL

The vast bulk (99.9%) of human genetic material is identical, with only 0.1% of DNA differing between individuals
(Department of Health 2003). However, the differences in DNA that do exist are responsible for individual
characteristics, including propensity to different diseases.
Different forms of a gene are known as alleles. These are located at precise points on a chromosome and can lead to
inherited variations in characteristics. A genetic polymorphism describes a variation in the genetic sequence that
is present in at least 1−2% of the population. Most polymorphisms do not lead to disease, but some are clear risk
factors or protective factors (e.g. genes coding for haemoglobin affect the severity of malaria symptoms).
Gene mutations are changes in DNA sequence that can lead to disease. They can occur due to chance, age or
exposure to environmental hazards. Most genetic mutations are repaired by cells. If mutations occur on sex
chromosomes, they can be passed on to an individual’s offspring.
An individual carries two copies of each autosomal gene: one on each chromosome. If the person has two identical
copies of a particular gene, he or she is said to be homozygous for that gene. If the person has different copies of
the gene (i.e. different alleles), then he or she is said to be heterozygous for that gene.

185

Public [Link] 185 3/6/08 [Link]

 2D G e ne t ic s

2D.2 INHERITED DISEASES

Inherited causes of disease in populations
So-called genetic diseases may be considered in three categories summarised in Table 2D.2.1.

Table 2D.2.1 Inherited causes of disease

Single-gene disorders One altered gene is responsible for illness, e.g. Huntington’s disease,
cystic fibrosis
Common disorders with rarer For example, 95% of cases of breast cancer occur sporadically, with a
inherited forms lifetime risk for women in England of around 11%. However, for women
with BRCA-1 gene mutations, the lifetime risk of breast cancer is 80%
Known genetic abnormalities or In these diseases, certain genetic changes alter the risk of disease but
polymorphisms do not predict its occurrence absolutely. The concept is akin to biological
markers such as cholesterol, where occurrence of disease depends on a
range of other factors, including other genetic traits and environmental
exposures. Examples include hypertension and schizophrenia

2D.3 PATTERNS OF INHERITANCE

Inheritance is described as being either mendelian (i.e. following one of the simple patterns of inheritance that
were first described by Gregor Mendel) or non-mendelian.

MENDELIAN INHERITANCE
Mendelian inheritance may be autosomal or sex linked, and dominant or recessive. Table 2D.3.1 describes these
types of inheritance. The application of mendelian inheritance principles for maximising the effectiveness of
treatment is illustrated with respect to determining fetal sex in Box 2D.3.1.

Box 2D.3.1
Example: Non-invasive method of determining fetal sex in early gestation
Fetal sex is of clinical importance in cases where the mother is a carrier of an X-linked recessive disease gene
(such as haemophilia or some muscular dystrophies), since a female child will be unaffected. The sex of an
unborn child can be diagnosed reliably with ultrasound only from around 12 weeks. Earlier diagnosis has to be
made with CVS, which is invasive and carries a small chance of miscarriage. In the UK, polymerase chain reaction
(PCR) testing is now becoming routine clinical practice for determining fetal gender safely and accurately
Technology advances: in 1997, Lo and colleagues reported detection of fetal DNA (fDNA) in maternal plasma
using PCR. PCR has since been used to detect two genes found on the Y chromosome, the SRY gene and the
marker sequence DYS14 on the TSPY gene. It is highly accurate (95−100% sensitivity) with an extremely low
false-positive rate (~100% specificity) from as early as 6 weeks’ gestation.
Other applications and ethical considerations: clinical applications of diagnosing fetal sex can extend
beyond the risk of X-linked conditions. For example, it can target treatment with steroids of congenital adrenal
hyperplasia (CAH), which otherwise leads to the development of ambiguous genitalia in female babies. Early
identification of fetal sex enables clinicians to avoid unnecessary steroid use where the fetus is male.
Reproduced from Avent and Chitty (2006).

186

Public [Link] 186 3/6/08 [Link]

 2D Genetics

Table 2D.3.1 Mendelian inheritance

Inheritance Description Example
pattern
Autosomal Dominant genetic diseases require only one copy of the gene to be abnormal Huntington’s
dominant in order to cause illness. If one parent has the disease, then there is a 50% disease
chance of each child inheriting the disease. Under mendelian inheritance
laws, there is no ‘carrier’ state for dominant conditions, because the disease
affects all individuals who have an abnormal copy of the disease gene
Autosomal Both copies of the gene are required to be abnormal in order for the Cystic fibrosis
recessive individual to be affected by the disorder. A normal gene is able to
‘compensate’ for the abnormal gene; therefore, individuals with one copy of
the abnormal gene are not affected by the gene mutation, but are described
as asymptomatic carriers. Where both parents carry one abnormal copy of the
gene, there is a 25% chance of each child inheriting the disease. In addition
there is a 50% chance of the child’s being a genetic carrier for the disease
x-linked Here the mutation is on the X chromosome. Males have only one X Haemophilia
recessive chromosome and therefore a mutation in a gene on the X chromosome is
X-linked colour
more likely to cause disease in men. In contrast, a single recessive mutation
blindness
on a gene on the X chromosome in women is compensated for by the normal
allele on the other X chromosome so that the disease does not occur.
Therefore, males are far more likely to be affected with X-linked recessive
disorders, and women are more likely to be carriers. In the next generation,
all daughters of affected men will be carriers of the condition, while none of
their sons will be affected
x-linked X-linked dominant conditions are all rare but occur when a single copy of Coffin−Lowry
dominant the gene on the X chromosome is mutated syndrome
Y-linked Y-linked disorders are extremely rare Male infertility

NON-MENDELIAN INHERITANCE
Most human diseases do not strictly follow ‘mendelian’ rules of inheritance. Examples of the ways in which they
differ are described in Table 2D.3.2.

187

Public [Link] 187 3/6/08 [Link]

 2D G e ne t ic s

Table 2D.3.2 Non-mendelian inheritance

Inheritance Description Example
pattern
variable The severity of the disease differs according to the number of abnormal Hypercholesterolaemia
severity genes inherited. For example, hypercholesterolaemia is more severe in
homozygotes than in heterozygotes
variable Some dominantly inherited gene mutations are not always expressed as BRCA-1
penetrance disease. For example, only 80% of those with BRCA-1 gene mutations
will develop breast cancer even though this is a dominant gene. See
2D.4
Mitochondrial As well as being present in chromosomes, DNA is also found in Leigh’s disease
inheritance mitochondria. Mitochondrial DNA encodes enzymes that are responsible
for energy production in a cell. Mitochondrial DNA mutations can lead
to a wide range of symptoms associated with energy deficiency in cells,
such as cardiac disorders and exercise intolerance
Mitochondrial mutations can only be maternally inherited because
mitochondria are rarely passed on in sperm. There is a threshold effect
with mitochondrial inheritance whereby a child has to inherit a certain
proportion of mutated DNA in order to be symptomatic. As a result,
there is no predictable inheritance pattern: mothers with mutated
mitochondrial DNA can give birth to both affected and non-affected
children
Polygenic See Section 2D.6 Asthma, diabetes
disorders
Gene– See Section 2D.7 Atherosclerosis
environment
interactions

2D.4 PENETRANCE
Penetrance describes the proportion of those who have a gene for a particular disease who develop the disease.
Penetrance varies across different genes, as shown in Box 2D.4.1

Box 2D.4.1
Examples of variable gene penetrance
Huntington’s disease – 100% penetrance
Certain alleles associated with Huntington’s disease have full penetrance. If people carry one HD allele with 40
or more CAG triplet repeats, they will definitely develop the disease.
BRCA-1 or BRCA-2: female breast cancer – 80% penetrance
Hereditary breast cancer is linked to two genes, BRCA-1 and BRCA-2. These have 80% penetrance, and so four of
five women carrying either of these genes will develop breast cancer.

188

Public [Link] 188 3/6/08 [Link]

 2D Genetics

2D.5 DIFFERENT GENOTYPES AND PHENOTYPES

The presence of a particular genetic composition (genotype) does not necessarily mean that the individual will
manifest the particular condition (phenotype). Variable phenotypes arise for many reasons, including:
• Incomplete penetrance (see Section 2D.4)
• There is more than one gene relating to one protein or one characteristic (for example, cystic fibrosis; see Box
2D.5.1)
• There is more than one mutation on one gene associated with disease
• Gene−environment interactions mean that the gene is expressed only if certain environmental conditions are
satisfied.

Box 2D.5.1
Example: Cystic fibrosis
Cystic fibrosis is a recessively inherited disorder that involves a mutation on the gene that codes for a cell
membrane channel, called the cystic fibrosis transmembrane conductance regulator (CFTR) protein. Abnormalities
in this protein cause it to malfunction, leading to symptoms including:
• Respiratory symptoms due to thickened respiratory secretions
• Pancreatic insufficiency (caused by an accumulation of thick mucus) causing malabsorption of food
• Elevated levels of chloride in the sweat
• Sterility in men
Hundreds of different mutations have been described on the CFTR gene, the most common being DF508 (i.e. a
mutation at point 508 on the cystic fibrosis gene). The severity of disease (the phenotype) varies considerably
between individuals affected by different mutations − particularly with regard to lung function. Some
mutations can be non-sense mutations, i.e. the mutation is so severe that none of the protein is produced.
Other mutations are mis-sense, i.e. some protein is produced but in small amounts or with altered function.
Gene−environment interactions can play a part in cystic fibrosis: factors such as smoking or malnutrition are
thought to influence the severity of lung disease. In the future, gene therapy may offer an effective treatment
for cystic fibrosis by correcting the CFTR gene.
Reproduced from National Human Genome Research Institute (2007).

2D.6 POLYGENIC DISORDERS

Family studies have implicated a strong genetic component to many disorders not typically thought of as ‘genetic’
diseases. As well as rare single-gene causes of common diseases (such as diabetes and Alzheimer’s disease),
polygenic forms of these diseases are far more common.
Identifying the exact genes that underlie the polygenic forms of these disorders is useful for:
• Improving the understanding of disease aetiology
• Predicting those at risk of illness
• Targeting disease prevention or health promotion
• Targeting treatments to individuals with genetic variants most likely to benefit
• Identifying new sites or biological processes to target therapies.
The patterns of heredity for polygenic disorders − where several gene variants are needed for the disorders to
develop – are complex. The mechanisms by which the genes interact are, in general, poorly understood. Simple
models assume either additive properties or that different subsets of genes cause different types of disease. In
reality, the ways that genes interact to cause disease are far more complicated and involve highly convoluted
gene−environment interactions.

189

Public [Link] 189 3/6/08 [Link]

 2D G e ne t ic s

Genetic studies (see Section 1A.41) have made some progress in identifying genes implicated in polygenic disorders.
However, few gene variants have been conclusively linked with particular disorders. Scientific progress has been
hampered by factors such as those outlined in Table 2D.6.1. The example of type 2 diabetes (Box 2D.6.1) shows how
a greater biological understanding of diabetes can help to explain the findings of genetic studies.

Table 2D.6.1 Challenges to identifying genes in polygenic disorders

Many susceptibility In complex polygenic disorders there are many
genes susceptibility genes, each contributing a small effect
to overall disease susceptibility. In order to detect
statistically significant gene effects, very large
sample sizes are needed
Population Studies conducted on different populations with
heterogeneity the same disorder may not identify the same gene
variants. This could be because of variations between
different population gene pools but could also be
because these populations are exposed to different
environmental circumstances, e.g. diet, pollution
Incomplete Candidate gene studies (where particular genes are
understanding of selected for study because they code for biological
disease biology functions connected with the disease) have been
a fruitful approach to finding genes implicated in
disorders. However, the selection of candidate genes
requires an understanding of disease pathophysiology
that is currently underdeveloped. Genome-wide
scans are now possible, however, and do not require
the same understanding of disease biology

Box 2D.6.1
Example: Type 2 diabetes: identifying susceptibility genes can inform disease aetiology
The incidence of type 2 diabetes has increased in recent years, thereby implicating environmental factors in
the aetiology. However family studies indicate that diabetes also has a strong genetic component. Studies have
identified monogenic forms of the disorder such as maturity-onset diabetes of the young (MODY types 1 and 2),
but these account for <5% of all cases of diabetes. Most cases of diabetes are therefore due to a combination of
factors, including intrauterine conditions, adult lifestyle habits and a range of gene−environment interactions.
Candidate gene studies have identified variations in some genes for functions known to be connected with
diabetes. For example, the PPARG gene encodes a cell receptor involved in adipocyte development, which is the
target for thiazolidinedione diabetes drugs. However, incomplete understanding of the biology of diabetes has
limited the identification of candidate genes that could be involved in the disorder.
Genome-wide scans have identified other potential regions on several chromosomes, but some of these have
not been replicated in other studies. One important gene to emerge from genome studies is CAPN10, coding for
the calpain 10 protease. When calpain was first identified, there was no known biological link between calpain
function and diabetes; since then, biochemical and pharmacological studies have implicated calpain in insulin
secretion.
Adapted from McCarthy (2004) and Elbein (2002).

190

Public [Link] 190 3/6/08 [Link]

 2D Genetics

2D.7 GENE–ENvIRONMENT INTERACTIONS

Purely genetic diseases, such as Huntington’s disease, are very rare. In contrast, the occurrence of almost all
common diseases results from a combination of a range of genetic susceptibility factors that interact with
environmental risks. Such diseases are described as multifactorial. Environmental factors include:
• Infections
• Chemicals
• Physical hazards
• Nutritional exposures
• Behaviours.
This implies that few people inherit a disease state genetically; rather, they inherit susceptibility to a disease,
which is modified by their exposure to environmental factors. Understanding the interaction between genes and the
environment will enable the identification of high-risk individuals and thus provide opportunities to target health
promotion and disease prevention more effectively.

GENETIC DISEASE PREVENTION

Some genotypes are expressed only under particular environmental conditions, e.g. phenylketonuria (see Box
2D.7.1). If these environmental conditions are avoided, then individuals with this genotype will avoid the disease.

Box 2D.7.1
Example: Classic phenylketonuria (PKU)
PKU is a complete (or near-complete) deficiency of phenylalanine hydroxylase activity.
It results in dietary intolerance to the amino acid phenylalanine. The disease is
autosomal recessive, i.e. both parents need at least one copy of the faulty gene for a
child to have a 25% chance of having the disease and a 50% chance of being a carrier.
Babies with PKU are normally detected by the universal screening of newborns, but can
also be detected antenatally through genetic testing.
If expressed, PKU can cause severe and irreversible learning disability. However, those
with the PKU gene who limit their dietary intake of phenylalanine between birth and
adolescence do not develop symptoms.

HEALTH PROMOTION AND GENETIC DISEASE

Identification of the genetic factors associated with common diseases, and the interplay between environmental
risks, advances our understanding of the aetiology and epidemiology of common diseases. Study of which genes
increase the susceptibility to disease and the gene−environment interactions enables practitioners to:
• Identify individuals at high genetic risk of disease
• Understand which environmental factors place individuals or groups (e.g. particular ethnic groups) at greater
risk
• Target health promotion messages and disease prevention interventions to those most likely to benefit.
The potential for genetic information to influence heart disease prevention is demonstrated in Box 2D.7.2.

191

Public [Link] 191 3/6/08 [Link]

 2D G e ne t ic s

Box 2D.7.2
Example: Preventing heart disease
Health promotion to prevent heart disease usually focuses on universal messages, e.g. advising people to
give up smoking, take regular exercise and eat a healthy diet. However, many people can cite individuals who
followed this advice and still suffered from ischaemic heart disease in middle age, and other individuals who
smoked, rarely exercised but lived well into their 80s. Such examples can undermine public acceptance of health
promotion messages.
People who develop ischaemic heart disease often have a family history of the disease, and research implicates
several genes in its aetiology. There are rare single-gene mutations (e.g. mutated genes that cause familial
hypercholesterolaemia) that increase an individual’s risk of dying prematurely from heart disease. Clearly, it is
important to lower cholesterol levels in these individuals through diet and drug treatment, in order to reduce
their risk of heart disease.
However, in addition to known ‘high-risk’ genotypes, a range of genetic variations may predispose individuals to
heart disease and in turn identify candidates for targeted health promotion approaches. While smoking cessation
messages are appropriate for the whole population, one study has suggested that men with a positive family
history of coronary heart disease who stop smoking have the potential to decrease coronary heart diease to a
greater extent than men without a positive family history.
Adapted from Hunt et al (2003) and McConnachie et al (2001).

2D.8 GENES IN HEALTH AND DISEASE

In the future, gene characteristics may be used to promote health and tackle disease. The field of
pharmacogenomics uses individuals’ genetic characteristics as a basis to understanding the relative effectiveness
of pharmaceutical treatments. In particular, people metabolise drugs at different rates, and this may explain why a
certain dose may be toxic for some people yet subtherapeutic for others. This may offer the opportunity to target
treatment regimens based on individual genotypes. Box 2D.8.1 describes how this technology may be applied to
clinical trials to understand non-response or side effects associated with particular drug doses.

Box 2D.8.1
Example: Testing for genetic variants of the cytochrome P450 CYP gene
CYP enzymes in the liver are primarily responsible for metabolising compounds such as drugs. Drug interactions
and adverse events associated with variation in CYP enzymes are relatively common. Genetic differences in CYP
may explain why one person experiences adverse events, while another does not.
In the USA, a CYP genotyping test is available which provides information on individual variants of the CYPC19
and CYP2D6 genes. This may enable clinicians to identify individuals who will rapidly or slowly metabolise a
range of treatments and could also be added to pharmaceutical clinical trials to account for the non-response or
occurrence of side effects in trial participants.
Reproduced from Davis and Khoury (2006).

GENE THERAPY
Gene therapy is an emerging field where vehicles such as viruses or plasmids are used to insert genetic material into
the cells of people with a particular disease. One of the few diseases in which this technology has been applied is
severe combined immunodeficiency (see Box 2D.8.2).

192

Public [Link] 192 3/6/08 [Link]

 2D Genetics

Box 2D.8.2
Example: Severe combined immunodeficiency (SCID)
SCID is caused by a single inherited gene abnormality in the adenosine deaminase (ADA) gene, leading to the
absence of the ADA enzyme. This leaves children with severely impaired immunity to infections. Gene therapy is
being developed to treat this disorder by inserting modified stem cells into the body that generate the missing
enzyme.

2D.9 DISEASE IN RELATIvES

Aetiology, distribution and control of disease in relatives
Relatives of people suffering from a disease that has a genetic component may wish to know whether or not they
are carrying an abnormal gene. Two types of test are particularly relevant in these circumstances: see Box 2D.9.1.

Box 2D.9.1
Method Description Example
Predictive genetic This is of value where an individual has Huntington’s
testing a family history of a disease with high disease
penetrance that develops in adulthood.
Inherited forms of
It can enable them to make informed
cancer
decisions about having children and,
where possible, to institute measures to
reduce their risk from disease
Individual carrier This can be offered to asymptomatic Cystic fibrosis
testing individuals who have a relative who is
affected by an autosomal or X-linked
genetic disorder

For commoner diseases, population carrier screening programmes are administered.

UK In the UK antenatal screening programme, pregnant women in high prevalence areas are offered a blood test
to identify whether they are carriers for sickle cell disease or thalassaemia (see Section 2C.11). If the mother is
identified as a carrier, then the father is also offered testing.

CONTROL OF DISEASE
In healthy relatives, control measures may include:
• Pre-implantation genetic diagnosis in assisted conception: one to two cells from embryos created by in
vitro fertilisation (IVF) are tested for genetic diseases. Only genetically disease-free embryos are used for
implantation.
• Antenatal testing and screening: can enable individuals to make informed decisions about whether to
continue with a pregnancy if the fetus has a severe genetic disorder, e.g. Down’s syndrome.
• Earlier or more frequent cancer screening and prophylactic surgery, e.g. for relatives of people with breast
cancer who have BCRA-1 or -2 mutations or for relatives of people with colorectal cancer (see Box 2D.9.2).
• Treatment to reduce disease risk, e.g. cholesterol lowering drugs for those with familial
hypercholesterolaemia.
• Genetic counselling to help people understand the disease and its potential impact, and to support people in
making decisions about the future.

193

Public [Link] 193 3/6/08 [Link]

 2D G e ne t ic s

Box 2D.9.2
Example: Colorectal cancer
Around 30% of people with colorectal cancer have a family history of the disease. Two syndromes have been
identified, accounting for 1−2.5% of inherited colorectal cancer: familial adenomatous polyposis (FAP, also
known as familial polyposis coli) and hereditary non-polyposis colorectal cancer (HNPCC)

FAP HNPCC (Lynch syndrome)

Incidence 1 in 7000 1 in 500

Genetics Dominant mutation in the Four genes involved in mismatch

adenomatous polyposis coli (APC) repair (MLH1, MSH2, MSH6 and
gene, a tumour-suppressor gene PMS2), most commonly MLH1,
MSH2
Characteristics Large numbers of colorectal polyps Very few colorectal polyps, which
develop usually by age 30. Most are are often malignant
benign but a high number of polyps
Average age of diagnosis of
means that there is a very high
colorectal cancer is 60 years
chance that one polyp will develop
a mutation in APC leading to a
malignancy
Chance of Without surgery, almost all will 70−80% lifetime risk
developing cancer develop cancer by age 40
Women also have 30−40%
lifetime risk but a 42% chance of
endometrial cancer
Control of disease Regular colonoscopies to monitor Regular colonoscopies to monitor
in relatives polyps and detect early cancers polyps and detect early cancers
Genetic testing Genetic testing

Reproduced from Kohlmann and Gruber (2004, updated 2006).

2D.10 MOLECULAR BIOLOGY

Elementary molecular biology as related to genetic epidemiology and microbiology
Molecular biology is the study and manipulation of biological processes and structures at a molecular level. This
includes the interrelationship of DNA, RNA and proteins − and it therefore overlaps with the fields of genetics and
biochemistry (see Figure 2D.10.1).

MOLECULAR BIOLOGICAL TECHNIQUES

Molecular biological techniques enable scientists to study DNA sequences, RNA, protein synthesis and cell function.
Three major processes for manipulating, identifying and visualising DNA are described below.

194

Public [Link] 194 3/6/08 [Link]

 2D Genetics

Function

Gen
istr
hem

etic
s
c
Bio

Proteins Genes Figure 2D.10.1 Interrelationship of genetics, biochemistry

and molecular biology. Reproduced with permission from
Molecular biology [Link]/wiki/Image:Schematic_relationship_between_
biochemistry%2C_genetics_and_molecular_biology.svg

PRODUCING FRAGMENTS OF DNA

DNA can be cut into smaller pieces using restriction enzymes, which recognise and cleave DNA at particular
sequences of nucleotides (see Figure 2D.10.2). The process of cutting DNA into sequences can have several uses:
• Cleaving midway through a gene will interfere with its function, and so, by observing the difference in activity
after the restriction enzyme has been added, the function of that particular DNA sequence can be identified
• DNA fragments produced by restriction enzymes can be added to other organisms using vectors such as
plasmids, then used to change the host’s DNA and its functions.

Enzyme recognises
particular DNA sequence

CATG

CATG

Enzyme cuts at particular

DNA sequence

CATG

CATG
A ‘ligase’ enzyme can join
different DNA sequences
cut at the same point

CATG

Figure 2D.10.2 Producing DNA fragments

195

Public [Link] 195 3/6/08 [Link]

 2D G e ne t ic s

COPYING DNA
The most versatile technique for cloning or making copies of DNA is PCR (the polymerase chain reaction), as
shown in Figure 2D.10.3.
PCR reactions start with a mixture in a buffer solution comprising:
• DNA that is to be copied
• DNA polymerase (an enzyme that replicates DNA)
• Sequence of ‘primer’ DNA (short sequence of 15−20 nucleotides of DNA that sticks to parts of the DNA strands to
be replicated but not to itself or copies of itself)
• Quantity of individual DNA nucleotides.
The process involves three steps, outlined in Table 2D.10.1.

Table 2D.10.1 Steps involved in copying DNA

Step Description Temperature (°C)
1. Denaturation DNA strands are separated 93
2. Annealing Primers attach to the separate DNA 40−55
fragments
3. Extension Nucleotides make new strands: the 72
polymerase reaction

This process is repeated for 30−35 cycles (after 35 cycles more unwanted byproducts are produced than useful DNA).

(1)
(1) Denaturing: DNA strands are separated by
heat.

(2)
(2) Annealing: Primers with specific DNA
sequences are added with DNA polymerase to
attach to single DNA strands.

(3)
(3) Extension: Individual nucleotides bind to
strand fragments to make new double strands.

Figure 2D.10.3 Copying DNA: (1) denaturation; (2) annealing; (3) extension

196

Public [Link] 196 3/6/08 [Link]

 2D Genetics

SEPARATING AND vISUALISING DIFFERENT DNA SEQUENCES

Once DNA has been replicated, e.g. by PCR, it can be sequenced to identify the order of nucleotides (A, C, G or T):
see Figure 2D.10.4.
Strand to be sequenced

A
T T A G G A A
C A C T
G C

Figure 2D.10.4 Strand to be sequenced

The most common method of sequencing uses chain termination. A primer, containing DNA with tagged or
terminator nucleotides, is added to denatured DNA strands (see Figure 2D.10.5).

Primed DNA
�

A
G C

T Prim
er T
C
A
A G T C

T A
A G C
Figure 2D.10.5 DNA primer Primer
T C G T

The primer bases bind to the relevant bases on the DNA C

to form base-pairs. However, every time a terminator C
T A A G T
sequence binds, the strand will end, thereby creating T C G
many different lengths of sequences, all ending in the
er
same base-pair (Figure 2D.10.6). Prim
Gel electrophoresis can be used to separate and Figure 2D.10.6 DNA sequences of varying lengths
identify different DNA fragments. DNA fragments are
‘run’ on an agarose gel under an electric field (DNA is negatively charged so it will run towards a positive electric
charge). Fragments of different sizes run along the gel at different speeds. Types of DNA can then be identified
according to their final position on the gel (see Figure 2D.10.7). The same process can also be used with RNA or
protein fragments.
DNA molecules

Longest DNA fragments

Shortest DNA fragments

Figure 2D.10.7 Appearance of DNA fragments on gel electrophoresis

197

Public [Link] 197 3/6/08 [Link]

 2D G e ne t ic s

visualising the fragments: fluorescent or radioactively labelled DNA probes are used to bind to the fragments on
a gel and can then be visualised using autoradiography (for radioactively labelled probes) or phosphoimaging (for
fluorescent probes).

MOLECULAR BIOLOGY APPLIED TO MICROBIOLOGY

Molecular biological techniques are invaluable in the investigation of microbial agents, their epidemiology and
communicable disease control. Particular uses have included:
• Identification of different genetic strains of particular pathogens that affect their clinical consequences, e.g.
disease severity, susceptibility to medication. For example, the Health Protection Agency’s report, Hepatitis C
in England (2005), noted that six genotypes of hepatitis C have been identified. The most common in the UK −
genotype 1 − is associated with a lower response to treatment than the other genotypes.
• Improved understanding of disease aetiology through the discovery of new pathogens.
• Detection of trends in the occurrence of new strains of disease through the development of molecular
databases and surveillance systems (see Box 2D.10.1).
• Linkage of infectious disease cases to assist the investigation of suspected or actual outbreaks.

Box 2D.10.1
Example: Applying molecular biology to identify tuberculosis strains
UK The Health Protection Agency (HPA) has introduced molecular typing for every culture of tuberculosis (TB)
bacteria that tests positive in the laboratory. Positive cultures of TB are sent by local hospitals to the National
Mycobacterium Reference Unit, where PCR is used to clone DNA from 15 loci on the TB genome. The analysis of
the PCR products is used to:
• Confirm the bacterium as TB (or another related bacterium)
• Check for antibiotic resistance
The technology, which provides results within a few weeks, enables local HPA teams to link TB patients who
have the same strain of TB. This can be used to trigger an outbreak investigation if linked patients are found
to have strains with the same molecular profile. It can also be useful in detecting false-positive TB results,
which can occur as a result of laboratory cross-contamination, and so prevent unnecessary treatment.
The HPA is developing a national database (the National Microbial Typing Database for Mycobacterium
tuberculosis) to create a national store of all the results. The database will be useful for linking strains of TB
geographically, for identifying clusters of related strains and to start linking molecular characteristics with drug
susceptibility.
Reproduced from Health Protection Agency (2006).

198

Public [Link] 198 3/6/08 [Link]

 2E Health and Social Beh aviou r

2E
Health and Social Behaviour

2E.1 Nutrition 199 2E.4 Dietary recommendations 204

2E.2 Nutritional interventions 201 2E.5 Lifestyle 208
2E.3 Choice of diet 203 2E.6 Complex interventions 208

The impact of different lifestyles on health is well recognised. One of the most important lifestyle choices that a
person makes in this regard relates to diet: the health implications of an unhealthy diet are severe and of growing
public health importance worldwide.
Influencing behaviour is a complex task that extends well beyond the boundaries of health services into home,
occupation and leisure activities. Public health has a key role to play with other agencies – national and local – in
interventions that improve health through social behaviour. However, it is not enough for public health practitioners
just to know what constitutes a healthy lifestyle. In order to change behaviours, practitioners also need to
understand:
• The barriers and motivations to adopting health lifestyles
• The evidence for recommendations related to diet and lifestyle
• How health and lifestyle are measured, particularly the potential strengths and pitfalls of these methods.

2E.1 NUTRITION
Principles of nutrition, its assessment in populations and its short- and long-term effects, and the influence of
malnutrition in disease aetiology and in growth and development
Appropriate nutrition is essential for health. Most of human history was spent as hunter−gatherers and therefore
evolution is thought to have rendered the body best suited to this type of diet.

ASSESSMENT OF NUTRITION IN POPULATIONS

Several techniques are available to assess a population’s food intake, as outlined in Box 2E.1.1.

199

Public [Link] 199 3/6/08 [Link]

 2E He a l t h a nd Social B ehaviour

Box 2E.1.1 Study designs to assess population nutrition

Ecological studies
Dietary comparisons are made between countries (or in the same county over time) with respect to disease
outcomes. The potential for confounding makes conclusions difficult to draw, but these studies are a fruitful
source of hypotheses.
Retrospective case–control studies
Patients with a given disease are compared with non-diseased controls regarding dietary influences. Like any
retrospective case−control study, these have the potential for recall bias and for disease onset to influence diet.
Cohort studies
Data on food intake and other factors are collected from a cohort of initially healthy people. Comparisons are
made between individuals who develop disease with those who do not, with adjustment made for confounding
factors.
Randomised trials
A nutrient or other food constituent is given to one group only. Disease outcomes in the two groups are
compared. As well as being very expensive, the effect sizes seen are usually small – therefore meta-analyses are
often necessary.
Dietary surveys
In the UK, the Food Standards Agency conducts food surveys such as the Low Income Diet and Nutrition
Survey, and the National Diet and Nutrition Survey. These involve 24-h food diaries, physical measurements
(e.g. BMI and blood pressure), blood analyses, questionnaires and interviews. Food diaries are potentially
unreliable (participants underestimate consumption) and biased (obese people underestimate to a greater
degree).
Biomarkers
When assessed as part of a dietary survey, biological parameters provide are more objective measure. Examples of
biomarkers include serum cholesterol, serum folate, serum vitamin B12, urinary iodine and urinary sodium.

SHORT-TERM NUTRITIONAL EFFECTS

Acute changes in diet may have the following effects:
• Sugary food  dental caries (the effect is particularly marked in young children)
• Reduction of salt intake  fall in blood pressure within a few weeks
• Lack of carbohydrate  ketosis.

LONG-TERM NUTRITIONAL EFFECTS

In the longer term, nutritional influences may not manifest themselves for several decades:
• Central obesity  type 2 diabetes
• Lack of breastfeeding  type 2 diabetes (immune-modulated response)
• Lack of vegetables and fruit  lung and colon cancer
• Alcohol + carcinogens ingested from smoking  oropharyngeal cancers
• Lack of dietary calcium  osteoporosis.
With regard to cancer causation, dietary factors may have an influence (either beneficial or harmful): see Figure
2E.1.1.

200

Public [Link] 200 3/6/08 [Link]

 2E Health and Social Beh aviou r

Normal INITIATION Pre-malignant PROMOTION Malignant

cell Random cell Several years’ exposure to cell
mutation or environmental factors
carcinogen

Figure 2E1.1 Influence of environmental factors on development of cancer

GROWTH AND DEVELOPMENT

Pregnancy and early childhood are both critical periods in life for nutrition.

PREGNANCY
The UK Food Standards Agency recommends that pregnant women take a daily intake of 400 mg of folate throughout
the first trimester. This is known to reduce the incidence of neural tube defects, including spina bifida.

EARLY CHILDHOOD
Growth references for children are widely used in public health and paediatrics. The WHO has developed child growth
standards to assess how well children worldwide are growing. The standards require that children be measured
for their height and weight regularly and set healthy ranges for children’s weight, height/length, BMI and motor
development from birth up to age 5 years. They are based on the WHO’s recommendations for:
• Child nutrition (e.g. breastfeeding should be promoted for babies wherever possible)
• Antenatal health (e.g. pregnant women should avoid tobacco, one of the risk factors for low-birthweight babies)
• Infant health care (e.g. children should receive vaccinations and health care sufficient to protect and maintain
their health).

2E.2 NUTRITIONAL INTERvENTIONS

Nutrition and food, and the basis for nutritional interventions and assessment of their impact
Protein–energy malnutrition (PEM) accounts for upwards of 5 million deaths per year worldwide. At the other end
of the malnutrition scale, obesity represents a key public health challenge to developed and transition countries,
where cheap food is energy rich.
Deficiencies in iodine, vitamin A and iron are responsible for a high global burden of disease. While classic
deficiency syndromes (such as scurvy or pellagra) are uncommon in developed countries, subclinical forms do exist
among high-risk groups: see Table 2E.2.1.

Table 2E.2.1
vulnerable group Nutritional deficiency
Strict vegetarians/vegans Vitamin B12 deficiency and iron deficiency (because non-
haem sources of iron have low bioavailability)
Immigrants Vitamin D deficiency
People with an alcohol dependency Vitamin B12 and folate deficiency
Very young and very elderly people; pregnant and General deficiencies
lactating women

201

Public [Link] 201 3/6/08 [Link]

 2E He a l t h a nd Social B ehaviour

The purpose of these interventions is to reduce morbidity and mortality through dietary change. These tend to be
relatively cheap interventions (low cost per QALY). Examples include:
• Fluoridation of water
• Interventions to reduce cardiovascular disease
• Workplace campaigns (e.g. for more vegetables in canteen meals)
• Television series about healthy lifestyles (e.g. Jamie Oliver’s school meals campaign)
• Campaign to promote home-grown fruit and vegetables
• Collaboration with supermarkets and with the food industry (low-fat meat products, healthier vegetable-oil
spreads, salt reduction in processed foods, improved labelling of food products)
• Free fruit in schools.
A range of these interventions was employed in the North Karelia Project (see Box 2E.2.1) which reduced rates of
heart disease in eastern Finland. Specific interventions may be indicated for geographical deficiencies (e.g. lack of
iodine in the Alps; lack of selenium in New Zealand’s South Island).

Box 2E.2.1
Example: Reducing deaths from heart disease – the North Karelia Project
The North Karelia Project was established in 1972 in response to excessively high rates of heart disease in
eastern Finland. The area was a traditional community, with a major dairy farming industry and a range of
socioeconomic problems.
The project included a wide range of interventions focusing on diet, reducing smoking and taking more exercise.
It worked at the individual patient level through health practitioners in providing disease prevention and health
promotion advice, and at the population level through TV media campaigns and policy changes. Examples of
interventions to affect nutrition included:
• Cholesterol-lowering competitions between villages and youth and school projects
• Collaborations with the food industry to promote low-fat dairy products and sausages and reduced salt
foods
• Encouraging farmers to grow fruit and vegetables
Surveys conducted every 5 years since 1972 show significant changes in local diet:

1972 dietary habits 1992 dietary habits

90% used butter on bread 10% used butter on bread
Vegetable oil rarely used for cooking 30% used vegetable oil regularly
Most people used full-fat milk Most used low fat or skimmed milk
Annual consumption of vegetables was about 20 kg per Annual consumption of vegetables increased to
person about 50 kg
Health outcomes included:
• Reduction in risk factors (cholesterol, blood pressure levels and percentage smoking)
• Reduction in mortality from heart disease among adults aged 35−64 years
Adapted from Puska et al (1983), Henkel (undated) and World Health Organization (2003a).

202

Public [Link] 202 3/6/08 [Link]

 2E Health and Social Beh aviou r

ASSESSMENT OF IMPACT
The impact of nutritional interventions can be assessed using several modalities:
• Food sales (e.g. full-fat milk compared with skimmed or semi-skimmed milk)
• Clinical markers (BMI, blood pressure, dental caries)
• Biological markers (serum cholesterol, urinary sodium)
• Cardiovascular endpoints (myocardial infarctions, mortality).

2E.3 CHOICE OF DIET

Social, behavioural and other determinants of choice of diet

POVERTY
Low-income families often have little disposable income to spend on food. As a result they tend to have poorer
nutrient intake and less food variety, and make less healthy food choices. Particular problems include:
• Discount stores often have a smaller range of foods available with less turnover of stock, resulting in older food.
• Tinned and frozen food are cheaper than fresh food.
• Lean cuts of meat are pricier than fatty ones.
• Poorer families often eat food from packets or ‘something on toast’.
• People with little money for food cannot afford to experiment with new recipes: a culinary disaster would result
in hunger.
Traditionally, poverty has been more closely associated with poor diet in adults than in their children: parents
protected their children, and school meals provided wholesome food. During the last 20 years, however, cooking
skills and confidence have waned, as has the nutritional content of school meals.

ETHNICITY, CULTURE AND RELIGION

While many black and minority ethnic people in the UK follow Western dietary habits, eating patterns often remain
influenced by:
• Traditional cuisine
• Festivals and fasts
• Proscription of certain foods
• Demographic and socioeconomic factors affecting the population as a whole.
As illustrated in Box 2E.3.1, in England recorded ethnic variations in eating habits were fond in the Health Survey
for England.

Box 2E.3.1
UK Example: Differences in reported eating habits by ethnic group: the 2004 Health Survey for England
The Health Survey for England asks a sample of the population about their eating habits. The 2004 survey found
that:
• A higher proportion of all black and minority ethnic groups, except the Irish, consumed the guideline five
portions of fruit and vegetables than the general population
• All black and minority ethnic groups had a lower fat intake than the general population
• All black and minority ethnic groups, except the Irish, were more likely to use salt in cooking
Reproduced from Sproston and Mindel (2006).

203

Public [Link] 203 3/6/08 [Link]

 2E He a l t h a nd Social B ehaviour

2E.4 DIETARY RECOMMENDATIONS

Current dietary goals, recommendations and the evidence for them
The media often give conflicting, sensationalist messages about healthy diets; therefore, authoritative
recommendations are crucial.

GOALS
UK The Food Standards Agency issues guidance on behalf of the government for England, Northern Ireland,
Scotland and Wales.

DIETARY REFERENCE vALUES

Dietary reference values issued by the Food Standards Agency are designed for policy-makers and professionals,
rather than the general public:
• Population averages for macro-nutrients and total energy
• Reference nutrient intakes (previously called ‘recommended daily amount’, RDA) for vitamins and minerals,
i.e. 2 standard deviations above the average requirement for each nutrient
• Upper safe level for each nutrient.

RECOMMENDED DIETARY ALLOWANCES

Ire The Food Safety Authority of Ireland (FSAI) published recommended dietary allowances (RDAs) in 1999,
providing indications of the level of nutrients required to meet the needs of healthy people. The FSAI adopted the
EU population reference intake (PRI) values for most nutrients. However, in the case of folate, iron, calcium and
vitamin C, the values for Ireland differ from the EU PRI on the basis of more recent research and consideration of
prevailing Irish conditions (e.g. high incidence of childhood anaemia).

ADEQUACY OF NUTRIENT INTAKE

Surveys of nutrient intake are conducted by weighing and recording all food consumption over a 3-day period, and
then assessing nutrient intake by using food composition tables. Nutrient intake can then be compared against
dietary reference values.

GOOD DIETARY vARIETY

Food usage questionnaires can be used to determine a person’s variety frequency score. This score is based upon
both overall food variety and variety within each food group.

HEALTHY DIETARY PATTERNS

Indicators of healthy dietary patterns include:
• Proportions of energy in the diet that are derived from fat and saturated fat
• Healthy diet score (derived from the Frequency of Food Usage Questionnaire)
• Frequency of eating five or more fruits and vegetables per day.

IMPLEMENTATION OF THE DIETARY GOALS

Scot The Scottish Diet Action Plan was launched in July 1996. It set out 71 recommendations to improve the diets
of Scots and established a series of dietary targets for achievement by 2005. The 71 recommendations were the basis
on which food and health action in Scotland were shaped in the 10 years following publication of the plan.

204

Public [Link] 204 3/6/08 [Link]

 2E Health and Social Beh aviou r

In September 2006, Health Scotland reported on the progress, impacts, successes and challenges since the 1996
Scottish Diet Action Plan. The review found:
• A major shift in food production and attitudes to diet and health
• Considerable impact on shopping, buying and eating patterns
• To achieve a better diet, changes are still required across the supply chain and in many policy areas that shape
food production and consumption.
Wal In Wales, the use of the dietary goals to reduce health inequality is promoted through the implementation of a
nutrition strategy for Wales. The Food and Fitness Implementation Plan (FFIP) seeks to address issues in relation to
children, and makes seven key recommendations and actions to achieve them, namely:
1. Extend the Welsh Network of Healthy School Schemes (WNHSS)
2. Improve the food and drink consumed throughout the school day (see Box 2E.4.1)
3. Provide high-quality physical education, health-related exercise and practical cookery skills
4. Provide an environment that will encourage children and young people to take up opportunities for physical
activity and healthier foods
5. Develop skills to enable children and young people to take part in physical activity and to prepare healthier
foods
6. Develop and deliver training on food and fitness for those working with children and young people
7. Ensure that actions are evidence based, or innovative with evaluation, and that findings are shared.

Wal Box 2E.4.1

Appetite for Life
As one example of action from the FFIP, the Appetite for Life report sets more stringent standards for school
lunches and minimum standards for all food and drink consumed throughout the school day. It is particularly
important in implementing nutritional standards that social aspects of consuming food and drink are considered:
mealtimes are not just about nutrition.

Ire Implementation of Irish national recommendations regarding salt intakeis summarised in Box 2E.4.2.

Ire Box 2E.4.2

Implementing nutritional recommendations: salt intake in Ireland
The recommended dietary allowance in Ireland is approximately 4 g salt/day for adults. However, the average daily
salt intake is 10 g in adults and exceeds 5 g in children aged 4−6 years, and 6 g in those aged 7−10 years, i.e. all
well in excess of physiological requirements. Typical consumption is accounted for as follows:
• 15−20% of total dietary sodium intake from salt added in cooking or at the table
• 15% from naturally occurring sodium in unprocessed foods
• 65−70% from manufactured foods
Meat, fish and bread together account for over 50% of salt intake from foods. The FSAI has worked with bakers to
reduce the salt content in bread and the Food Safety Promotions Board has run advertising campaigns to reduce
salt consumption.
Reproduced from the Review of the Scientific Evidence and Recommendations for Public Policy in Ireland (2005),
available online at: [Link].

205

Public [Link] 205 3/6/08 [Link]

 2E He a l t h a nd Social B ehaviour

RECOMMENDATIONS
UK Advice to the general public from the UK Food Standards Agency is shown in Box 2E.4.3.

UK Box 2E.4.3
UK Food Standards Agency: practical advice for the general public
• Fruit and vegetables should constitute about a third of the food eaten each day. The five-a-day target
encourages both increased variety and increased amount of fruit and vegetables eaten
• Starchy foods should constitute another third of the food eaten. ‘Low-carbohydrate’ diets are discouraged
• Fibre is promoted − both insoluble (wholegrain products, fruit and vegetables) and soluble (oats and
legumes). The former bulk the stool and the latter reduce blood cholesterol
• Fish is promoted since it contains protein, selenium and iodine. White fish is low in fat; oily fish is rich in
omega-3 fatty acids, vitamins A and D. Shellfish are good sources of selenium, zinc, iodine and copper. Game
fish should be restricted to one portion a week (they contain mercury) and fish liver oil/fatty fish should not
be eaten to excess, as over-consumption over many years is harmful. There is special advice for pregnant and
breastfeeding women and for children
• Meat should be eaten lean and cooked with little fat. Liver or liver products should be consumed no more
than once a week (again due to the risk of vitamin A excess)
• Low-fat diet is encouraged, and saturated fat should replace unsaturated fat. Trans-fats (contained in some
hydrogenated vegetable oils) may be even worse than saturated fats
• Sugar intake should be restricted since it contains calories but few other nutrients and can cause tooth decay
• Salt should be restricted to 6 g/day (salt = 2.5 ¥ sodium). The Food Standards Agency is working to improve
labelling and wishes ‘salt’ to be listed rather than ‘sodium’
• Water is encouraged: approximately 1.2 l of water should be drunk each day (6−8 glasses) − more in hot
weather
• Alcohol intake should be restricted to 2−3 (women) or 3–4 (men) units a day, spread over the week.
Postmenopausal women and men aged >40 should consider drinking 1−2 units of alcohol/day to reduce
cardiovascular risk
• vitamins, minerals and trace elements are usually found in sufficient amounts without recourse to
supplements. The danger of over-consumption from supplements is highlighted

Ire ‘Recommendations for a Food and Nutrition Policy for Ireland’ (the Nutritional Advisory Group of the FSAI
1995) provided guidance on promoting healthy eating similar to that of the UK advisory agencies. The Food
Pyramid has been used for health education, encouraging consumption of bread, cereal and potatoes (6+ portions/
day), fruit and vegetables (5 portions/day), milk, cheese and yogurt (3 portions/day), meat, fish, eggs and
alternative (2 portions/day) and foods high in fat and/or sugar (choose small amounts).

Aus In Australia, as elsewhere, there is an abundance of dietary advice in magazines − mostly not evidence based
− against which even the most authoritative recommendations find it hard to compete. The Commonwealth Scientific
and Industrial Research Organisation has recently published The CSIRO Total Wellbeing Diet which has become a best-
selling book. This offers scientifically proven, practical advice, and contains recipes and 12 weeks of menu plans.

EVIDENCE
The evidence for nutritional recommendations comes from a range of sources. However, there are challenges to:
• Obtaining undisputed evidence for nutritional goals for health
• Ensuring that evidence of health risks and benefits is widely accepted
• Implementing accepted evidence to change consumption in the population.

206

Public [Link] 206 3/6/08 [Link]

 2E Health and Social Beh aviou r

EvIDENCE FOR HEALTH-RELATED NUTRITIONAL GOALS

Major studies that have generated evidence for the health effects of food substances are listed in Appendix C.
Selected studies are also described in more depth as examples in this chapter.
There are significant ethical and practical obstacles to conducting randomised controlled trials involving alteration
in the diet of widespread food substances such as fats. Many of the outcomes of interest (such as cardiovascular
events or deaths) are long term, making trial designs expensive and impractical.
International ecological studies, such as Intersalt (see Box 2E.4.4) and Seven Countries (see Box 2E.5.1), have
capitalised on the differences in dietary habits that exist between different cultures. They have been useful for
indicating which factors affect health (see example boxes). However, as with all ecological studies, they are
vulnerable to ecological fallacies (see Section 1A.7) and, as with all observational studies, there are several
potentially confounding factors (e.g. different genetic susceptibility to disease, and other lifestyle variations that
could explain the observed observations).

Box 2E.4.4
Example: Evidence that salt intake is linked to blood pressure
Intersalt is a prime example of a large-scale, cross-sectional study. It investigated the relationship between
salt intake (measured by sodium excretion) and blood pressure. The study involved 10 079 men and women in
52 different centres around the world. For each individual, sodium excretion was measured over 24 h and was
related to the individual’s blood pressure.
Regression analyses performed on Intersalt data across the 52 centres and published in 1988 indicated that:
• Urinary sodium excretion was related to blood pressure (at an ecological level)
• BMI and alcohol intake were both independently and strongly associated with raised blood pressure
Intersalt’s findings were controversial: commentaries (including those published by the Salt Institute) questioned
the study’s validity. However, subsequent analysis by the Intersalt group (and by other researchers) found an
even stronger association between blood pressure and sodium levels, and further explored the relationship
between alcohol and blood pressure.
A range of evidence now clearly supports a reduction in dietary salt intake as a means of reducing blood
pressure; a recent Cochrane systematic review of randomised trials to determine the effect of salt reduction on
blood pressure found that, ‘reduction in salt intake lowers blood pressure both in individuals with elevated blood
pressure and in those with normal blood pressure’.
Adapted from Intersalt Cooperative Research Group (1988), Hanneman (1996) and He and MacGregor (2004).

ENSURING THAT EvIDENCE OF HEALTH RISKS AND BENEFITS IS WIDELY ACCEPTED

Even where evidence is strong, peer-reviewed research may be against competing commercial interests.
Information about nutrition comes to the public from a range of sources, but rarely from peer-reviewed research
sources. Common sources include publicity by manufacturers and retailers of food substances, diet books and
magazines. Each of these has a vested interest in maintaining or promoting sales in a product. Advertising for
foodstuffs (e.g. showing sports people consuming the product) is often far more powerful an influence on eating
habits than published research is.

IMPLEMENTING ACCEPTED EvIDENCE TO CHANGE CONSUMPTION IN THE POPULATION

People’s decisions on what they eat and drink are not always based on long-term health goals. Other priorities
such as convenience, taste, price, satiating hunger rapidly or achieving weight loss may all come above health in
influencing what they consume.

207

Public [Link] 207 3/6/08 [Link]

 2E He a l t h a nd Social B ehaviour

2E.5 LIFESTYLE
Effects on health of different diets (e.g. ‘western diet’), physical activity, alcohol, drugs, smoking, sexual behaviour and
sun exposure
In general, it is difficult to attribute health effects to a single aspect of lifestyle but Table 2E.5.1 outlines some
of the relationships observed between certain factors and health. An example – cholesterol and heart disease – is
given in Box 2E.5.1.

Box 2E.5.1
Example: Cholesterol and heart disease – Ancel Keys’ Seven Countries Study
The origins of the Seven Countries Study lay in observations, mainly by its lead, Ancel Keys, that heart disease
was a greater problem in the USA than in countries such as Italy. Keys posed the hypothesis that ‘differences
among populations in the frequency of heart attacks and stroke would occur in some orderly relation to physical
characteristics and lifestyle, particularly composition of the diet, and especially fats in the diet’.
Settings and methods: the Seven Countries Study involved populations of men aged between 40 and 59. They
were surveyed from 1958 to 1970 on their diet and a range of risk factors for heart disease. Men were selected
from 18 areas of 7 countries with contrasting dietary habits and rates of heart disease.
Results: the death rates and health of a cohort of men surveyed since 1958 were followed up. In countries
where animal fat was a major component of the diet, notably Finland, deaths from heart disease were higher
than in countries where olive oil was the main source, e.g. Crete. In addition to academic publications, Ancel
Keys also publicised his findings in books such as How to Eat Well and Stay Well: The Mediterranean Way (1975).
Legacy: debate about the links between heart disease and dietary fat has continued since the Seven Countries
Study and there are now links reported between other types of fat (trans-fats) and heart disease. However, the
Seven Countries Study was notable for demonstrating links between heart disease and dietary fat at an ecological
level, and was extremely influential in changing eating patterns across much of the USA and western Europe.
Adapted from Blackburn (1999) and Keys (1980).

2E.6 COMPLEx INTERvENTIONS

Combating complex problems using a wide range of approaches, including health service interventions and broader
cultural interventions
Several methods have been adopted to tackle complex issues such as poor diet. They include the so-called medical,
behavioural and socioenvironmental approaches.

MEDICAL APPROACH
This approach focuses on disease, with a narrow conception of the causes of disease and the determinants of health.
Illness is considered in microbiological or physiological terms. Prevention focuses on known risk factors, e.g. high
cholesterol and hypertension as risk factors for cardiovascular disease. Risk reduction focuses on pharmacological
interventions (e.g. statins or antihypertensives), and health is equated with the absence of disease and the
provision of health services.

208

Public [Link] 208 3/6/08 [Link]

 2E Health and Social Beh aviou r

Table 2E.5.1 Relationship between lifestyle and health outcomes

Factor Effect on health
Diet Western diet
• High total energy Energy  obesity
• High saturated fat (butter, red meat) Fat  breast cancer, obesity
• Low fibre Low fibre  colorectal cancer
• High salt Salt  stroke
Mediterranean diet
• High unsaturated fat (olive oil)
Lower cholesterol/heart disease
• High fruit and vegetables
Obesity
• Low red meat
Lower cancer risk
• Moderate red wine
South Asian diet (in the UK)
• High saturated fat (ghee) Obesity
• High fruit and vegetables Coronary heart disease
Stroke
Type 2 diabetes
Alcohol Moderate alcohol intake Probable cardiovascular protection
Acute excess alcohol Social problems
Accidents
Violence
Road traffic accidents
Peptic ulcer
Liver disease
Impotence
Sexually transmitted infections
Chronic excess alcohol Cirrhosis
Hepatitis
Pancreatitis
Suicide
Drugs Intravenous drug misuse  blood-borne
viruses (e.g. HIV/HCV)
Psychosocial problems
Smoking Cardiovascular disease and stroke
Lung cancer
Chronic respiratory disease
At least 50 other conditions linked
Sexual Sexually transmitted infections
behaviour Unwanted pregnancy
Sun Basal cell carcinoma
exposure Squamous cell carcinoma
Malignant melanoma

209

Public [Link] 209 3/6/08 [Link]

 2E He a l t h a nd Social B ehaviour

BEHAVIOURAL APPROACH
This approach values education and free choice, rather than legal or fiscal coercion. Disease prevention can be
achieved through the provision of information to populations about lifestyle risk factors (e.g. smoking, drinking
or diet). In particular, many campaigns now make use of social marketing to encourage people to adopt healthier
eating habits (see Section 2I.3). Social marketing aims to take into account the priorities and perspectives of
particular sectors of a target group for health messages and to tailor messages and health promotion campaigns
accordingly.
However, the behavioural approach tends to disregard sociocultural influences on behaviour − such as the way
in which dietary choices are influenced by advertising and income. Dietary education may therefore have little
impact on poorer families who have no access to affordable fresh fruit and vegetables. This approach has also been
criticised for blaming problems on ignorance or personal choices through the attribution of guilt.

SOCIOENVIRONMENTAL APPROACH
This approach seeks to improve health by means of strategies that modify the social, political and economic
environment via government and community action. Health education involves recognition of the aspects of home,
workplace and community life that are detrimental to health.
UK In recent years, the UK government has accepted this approach. It has published several documents that call
for greater efforts with respect to disease prevention and health promotion through a balance of individual and
government actions. An example is the 2004 Public Health White Paper entitled Choosing Health: Making Healthy
Choices Easier.

210

Public [Link] 210 3/6/08 [Link]

 2F En viron ment

2F
Environment

2F.1 Environmental determinants of disease 211 2F.7 Use of legislation in environmental control 223
2F.2 Hazard and risk 218 2F.8 Health and safety at work 225
2F.3 Climate change 220 2F.9 Occupation and health 227
2F.4 Sustainability 220 2F.10 Transport and the environment 228
2F.5 Housing 221 2F.11 Chemical incident management 230
2F.6 Monitoring and control of environmental
hazards 223

There is a growing focus on the impact of human activity on the environment. Climate change, for example, is now
widely considered as the most important threat to the planet. However, the impact of the environment − where
we live, where we work, our air, our water, our food − is itself of profound importance to health. Public health
practitioners therefore need to appreciate the environment from both perspectives:
• The human impact on the environment, i.e. the effects of pollution, climate change, policy levers to reduce
these impacts
• Environmental effects on humans, i.e. the determinants of health, effects of environmental change and the
range of policy levers that can be deployed to improve the environment – ranging from housing legislation to
water monitoring.

2F.1 ENvIRONMENTAL DETERMINANTS OF DISEASE

As was seen in Section 2E, most diseases can be best thought of as resulting from an interaction between hereditary
and environmental causes. The latter may be subdivided into the categories listed in Table 2F.1.1.

POLLUTION
Pollution is contamination of air, water or soil with harmful substances. It may be accidental or deliberate, direct or
indirect.

211

Public [Link] 211 3/6/08 [Link]

 2F E n v i ro n me nt

Table 2F.1.1 Environmental determinants of disease

Pollution Processes or activities generating pollution include energy generation and use, industry,
agriculture, transport (see Section 2F.10) and waste disposal
Resources Quality and availability of resources such as water, land, food and air
Atmosphere Characteristics of the atmosphere including climate change (see Section 2F.3), air pollution,
extreme weather
Behaviour Human behaviours such as diet, cigarette smoking, coughing, sneezing, etc

ENERGY USE
Rising energy use is a major cause of air pollution and climate change (see Section 2F.3).

AGRICULTURE
Agricultural practices can lead to deleterious environmental effects through:
• Release of farm slurry effluent into waterways. This represents far more of a pollution hazard that does domestic
sewage
• Pesticides (chemical agents used to destroy agents that affect the growth of crops) can enter the human body
by inhalation, ingestion or through the skin
• Soil pollution through additives to the soil.

WASTE DISPOSAL
The waste hierarchy originated in the 1970s in the European Union’s waste
management framework. It orders options for disposing of waste from the most to Reduction
the least desirable. As Figure 2F.1.1 indicates, the most desirable option is to reduce
the amount of waste created in the first place. Re-use
There are various ways and various levels whereby the aims of the hierarchy may
be achieved. For example, manufacturing industry can adopt low-waste processes, Recycling
and households can use composting to dispose of wasted organic material. Where
reducing use or production is not possible, re-using products should be considered Disposal
(e.g. domestic re-use of carrier bags). In circumstances where products cannot be
re-used, they should be recycled. Only where the other options in the hierarchy
are unavailable should incineration or landfill be considered. In the UK, most solid
waste is buried or incinerated, with only about 17% of domestic waste recycled. This
contrasts starkly with other European countries, such as Germany, where up to half
of all waste is recycled. Figure 2F.1.1 The waste
hierarchy
INCINERATION
Approximately 14% of UK waste is incinerated. Incineration may be considered more efficient than landfill because
the combustion of waste at high temperatures produces fewer harmful chemicals. However, it is questionable
whether even incinerators that generate electricity are environmentally sustainable: recycling saves considerably
more energy because it entails less use of raw materials. Moreover, the ash from incinerators often contains harmful
toxins, some of which (e.g. dioxins) may be carcinogenic.

LANDFILL
Eighty per cent of waste in Britain is disposed of in landfill sites. Landfill presents a number of potentially harmful
environmental effects:

212

Public [Link] 212 3/6/08 [Link]

 2F En viron ment

• Rotting rubbish emits explosive gases such as methane

• Noxious liquids from dumps risk polluting local waterways
• Additional nuisances include lorry traffic and noise, odours, smoke, dust, airborne litter and pests (e.g. rats and
seagulls).

HAZARDOUS WASTE
A number of waste items are at risk of causing harm when they are discarded. These include batteries, paint tins,
fertiliser, health-care waste (e.g. body parts, medicines, sharps). Inappropriate disposal of radioactive and chemical
waste leads to the generation of contaminated land.

UK REGULATION
In the UK, the Environment Agency (EA) and the Scottish Environment Protection Agency (SEPA) enforce a waste
management licensing system to ensure that landfill and incinerator systems for waste disposal do not have an
adverse effect on health, the environment or local amenities.

RESOURCES
WATER
Water is a vital resource for agriculture, industry, drinking, food preparation and sanitation. Its quantity and quality
strongly influence the risk of disease. Availability of water is an issue not just in arid countries but also in the UK
(see Section 2F.5).
Water pollution may damage aquatic animal and plant life, and makes drinking water more difficult to treat.
Pollutants in water are summarised in Table 2F.1.2.

Table 2F.1.2 Water pollutants

Fertilisers Nitrates and phosphates, mostly from agricultural processes, lead to the eutrophication of
waterways, i.e. the overgrowth of algae causing oxygen depletion
Metals Aluminium may be naturally occurring, and aluminium sulphate is used in water purification to
improve the taste of water
Lead is usually derived from domestic plumbing systems
Other heavy metals may be naturally occurring or may have leached from contaminated soils
Slurry Organic waste includes slurries, silage liquor, surplus crops, sewage sludge and industrial wastes.
These may enter the water course if they are poorly stored or are disposed or spread onto land
Sewage In the UK, acceptable sewage regulations are enforced by the EA or SEPA

WATER MONITORING AND CONTROL

UK In the UK, the Drinking Water Inspectorate (DWI) regulates the quality of water supplied to customers. To
ensure that it is safe and acceptable to drink, water undergoes tests for a number of quality characteristics listed in
Table 2F.1.3.
Eng Wal The EA promotes water availability through:
• Monitoring water levels in ground water, reservoirs, lakes, rivers and coastal waters
• Requiring water companies to produce a long-term water resource plan
• Issuing of abstraction licences to regulate those who can take water from water bodies.

213

Public [Link] 213 3/6/08 [Link]

 2F E n v i ro n me nt

Table 2F.1.3 Water tests

Physical characteristics Taste, colour, smell
Chemical composition The concentration of nitrogen is indicative of decomposition of organic matter
Chlorine levels and presence of faecal coliforms suggest contamination with
sewage
Calcium and magnesium are markers of hard water
The absence of oxygen signals stagnant water which can indicate heavy pollution
Bacteriology Bacteriological examination is performed to detect faecal organisms such as E. coli,
cryptosporidia and other coliforms

Individual/domestic measures to protect water as a resource include:

• Hosepipe bans
• Short showers not baths
• Short flush toilets or ‘hippo’ units in cisterns to reduce the amount of water used during flushes
• Fixing dripping taps and broken pipes
• Re-using water (e.g. from washing vegetables to water plants or flush toilets)
• Collecting rainwater using water butts
• Fitting water meters.

FOOD
The quality and availability of food can lead to or prevent disease. Issues concerned with nutritional diseases are
covered in Section 2E. Food can also cause disease due to:
• Residues on food from its production or packaging, e.g. pesticides
• Additives deliberately included in food
• Food production methods
• Microbes (see Section 2G).
UK The Food Standards Agency monitors or commissions monitoring on levels of contamination in food.

LAND
There are four major environmental concerns concerning land use and characteristics:
• Housing: see Section 2F.5.
• Contamination: harmful substances may have been introduced onto the land as part of its current or former
uses (e.g. former industrial sites). Contaminated land is a particular problem in urban areas due to increased
pressure to build on brown-field sites rather than green-field sites.
• Radiation: radon gas leaches naturally from the bedrock in certain regions (e.g. granite in Cornwall).
• Deforestation: the clearing of forests for house building or agricultural purposes has profound effects on the
climate, soil erosion and water contamination.

ATMOSPHERE
The largest cause of carbon dioxide (CO2) production is the burning of fossil fuels for electricity, heating and
transport. Options to reduce CO2 production are described in Table 2F.1.4.

214

Public [Link] 214 3/6/08 [Link]

 2F En viron ment

Table 2F.1.4 Options to reduce carbon dioxide production

Nuclear Nuclear energy produces minimal greenhouse gases. However, this option has its own
environmental problems, namely the disposal of toxic waste produced by generators, the risk of
harm from explosions or malfunctioning power stations (e.g. Chernobyl) and the risk that power
stations could be used as terrorist targets
Efficiency More efficient use of energy sources in homes and industry through increasing insulation (though
see indoor air pollution), cleaner fuels, waste minimisation, use of more efficient appliances and
engines, reductions in travel (especially by air and in sole-occupancy cars)
Renewables Develop renewable energy sources such as solar, wind, hydroelectric, geothermic and biomass
(where natural materials, such as wood or manure, are burnt or converted into gas to provide
energy). These sources generate less greenhouse gas, they are regarded as being safer than
nuclear energy and they will not run out − unlike fossil fuels

INDOOR AIR POLLUTION

Pollution from open solid-fuel stoves is a major public health problem in the developing world. Current efforts to
improve insulation are tending to decrease ventilation, thereby exacerbating the problem. Passive exposure to
cigarette smoke is an enduring problem, although many countries have now banned exposure in the workplace and
other enclosed public spaces.

ATMOSPHERIC AIR POLLUTION

Table 2F.1.5 describes a number of factors responsible for atmospheric air pollution.
EU Attempts have been made to quantify the mortality and morbidity associated with vehicle pollution in
European cities (see Künzli et al 2000).

MONITORING AND CONTROL

UK Air quality is monitored routinely across the UK. The UK National Air Quality Information Service provides
information on air pollution nationally and regionally, and issues forecasts with health advice, indicating days when
levels are above standard levels. The government’s Air Quality Strategy for England and Wales sets acceptable levels
of the major pollutants. Local authorities are responsible for assessing the air quality in their area, and are required
to produce an action plan to address any areas where pollution is higher than health-based standards.
The UN Economic Commission for Europe (UNECE) supports countries’ efforts to reduce pollution levels and
manage their natural resources. It has also negotiated treaties that require signatory countries to reduce pollution
(e.g. the Kyoto Protocol for greenhouse gases and the Montreal Protocol for volatile organic compounds).

RADIATION
Radiation – both naturally occurring and fabricated − is a cause of air pollution. There are two types of radiation.
Ionising radiation has the ability to ionise atoms (i.e. strip them of electrons) and is characterised by high energy.
Ionising radiation is emitted as alpha particles, beta particles or gamma rays. Non-ionising radiation does not
produce sufficient energy to ionise particles. Sources of the latter include sunlight, power-lines and electrical
equipment.

IONISING RADIATION
The most common source of naturally occurring ionising radiation is radon, a gas arising from uranium in rocks and
soils that accounts for about half of UK residents’ radiation dose. Naturally occurring levels are generally low and
are not associated with harm. Higher radon levels are found in the south west of England due to the granite geology

215

Public [Link] 215 3/6/08 [Link]

 2F E n v i ro n me nt

Table 2F.1.5 Causes of air pollution

Carbon monoxide Carbon monoxide (CO) is produced when fossil fuels are burnt in insufficient oxygen. Traffic
exhaust fumes and cigarette smoke both contain CO. An odourless and invisible gas, CO is
highly poisonous and causes many deaths each year from faulty gas appliances or blocked
ventilation. Low-cost CO meters are available, but CO poisoning is best avoided by correct
installation, maintenance and ventilation of gas appliances
Ozone Ozone is found in two regions of the atmosphere:
• Stratospheric ozone (the ‘ozone layer’) is naturally occurring, and by blocking ultraviolet
radiation prevents skin damage and skin cancers
• Ground-level ozone from anthropogenic sources can cause respiratory symptoms and
harms plant life. Ground-level ozone is produced in reactions between nitrogen oxides or
volatile organic chemicals and sunlight
Nitrogen dioxide This is a component of vehicle exhaust gases and power station emissions. Oxides of
nitrogen may cause respiratory symptoms, increase ground-level ozone (see above) and
contribute to acid rain (which damages plant life). Environmental levels remain relatively
high in urban areas due to traffic congestion
Sulphur dioxide Sulphur dioxide is mostly produced by coal- and oil-fired power stations. Sulphur dioxide
causes acid rain, which damages vegetation and can exacerbate respiratory illness.
Concentrations have reduced in the UK since the 1960s
Lead Lead is produced both through lead extraction and as a by-product of other industrial
processes. Lead was formerly added to petrol, but now only unleaded and low-lead fuels are
available. Lead toxicity can cause cognitive impairment, renal disease and abdominal pains.
Levels in the air have decreased markedly as a result of unleaded petrol
Fine particles Particles 1000th of a millimetre in diameter are generated naturally, by industrial processes,
and by traffic emissions – particularly diesel. They can exacerbate respiratory and
cardiovascular disease. In the UK, levels frequently exceed health-based standards
volatile organic These include benzene and 1,3-butadiene, which are components of traffic emissions. These
compounds chemicals are carcinogenic and they contribute to ground-level ozone and smog
Radon See page 215

in the area and have been linked to an increased risk of lung cancer (especially in miners). Householders in areas
with high radon levels should monitor and reduce the levels of radon in the home. The Health Protection Agency
recommends that radon levels should not exceed 200 Bq/m3. Actions to reduce indoor radon concentrations usually
involve improving ventilation to expel radon into the atmosphere.
Naturally occurring ionising radiation accounts for 85% of the radiation that people are exposed to. The remainder is
generated from three sources: see Box 2F.1.1.

Box 2F.1.1
Health care Health care (radiotherapy and diagnostic radiology) 14%
Industry Mainly from measurement purposes and for producing electricity <1%
Nuclear weapons Fallout from previous nuclear weapon explosions and other accidents and <1%
incidents worldwide

216

Public [Link] 216 3/6/08 [Link]

 2F En viron ment

NON-IONISING RADIATION
Sunlight is the main source of ultraviolet radiation and consists of ultraviolet A and B radiation: Box 2F.1.2.

Box 2F.1.2
UvA Skin ageing
UvB Sunburn

Exposure to both types increases the risk of developing cataracts and skin cancer (melanoma, squamous cell skin
cancer and basal cell cancer).
There are suggestions that some electromagnetic radiation may be associated with the development of cancer. Most
research evidence to date suggests that any risks, if they exist at all, are probably very small. Two major areas of
research (and public concern) have been mobile telephones and power lines: Box 2F.1.3.

Box 2F.1.3
Mobile The increasing use of mobile phones has been accompanied by concerns
telephones about possible harmful effects on health arising not only from exposure to
the radio waves that are produced by the phones held close to the head
but also from the base stations that serve the telephones
Power-lines High voltage power-lines emit extremely low-frequency (ELF) (50–60 Hz)
magnetic fields. ELF radiation may possibly increase the risk of leukaemia
in children in homes with high levels of exposure

MEASURING RADIATION
As described by the US Centers for Disease Control (2003), radiation is quantified using different methods and
different units, according to the amount of radiation released, the absorbed dose or the risk to health. The main
methods are shown in Table 2F.1.6

Table 2F.1.6 Measuring radiation

Unit Symbol Measure Illustrative examples
Becquerel Bq Amount of radioactivity in a material (unit relates 10 Bq/cm2: threshold level for
to disintegrations per second) contamination of surfaces requiring
remediation on health grounds
Gray Gy Absorbed dose: energy deposited in each gram of 10 Gy: destroys bone marrow
tissue, indicating acute radiation damage to organs
Sievert Sv Risk of exposure or effective dose: risk of cancer 0.1–2 mSv: 1 chest X-ray
from chronic or low doses of radiation. Adjusts for
2 mSv: average exposure to natural
the fact that the same absorbed dose from different
background radiation in 1 year
types of radiation has different capacities to cause
cancer 20 mSv: annual dose limit for a
radiation worker
1 Sv: increases lifetime risk of
cancer by approximately 5%

217

Public [Link] 217 3/6/08 [Link]

 2F E n v i ro n me nt

MONITORING AND CONTROL

UK The Radiation Protection Division of the Centre for Radiation, Chemical and Environmental Hazards at the
Health Protection Agency monitors and researches the effects of radiation in England; the Environment Agency
is a major source of guidance and regulation for a wide range of environmental issues in England and Wales; and
DEFRA (the Department for the Environment, Food and Rural Affairs) is responsible for implementing legislation and
regulation of chemicals and environmental threats to health in the UK. However, current UK policy is largely set by
EU directives.

HUMAN BEHAVIOUR
Human behaviour with respect to individual health is described in Section 2E. With respect to the environment, it
can be tempting to view individual human behaviour as insignificant compared with industrial impacts and global
threats. However, human behaviour is of significance in three ways: see Table 2F.1.7.

Table 2F.1.7 The relationship between human behaviour and the environment
Individuals as polluters Domestic consumption of energy has an impact on climate
change. Individuals can reduce their own consumption of fossil
fuels
Individual vulnerability to Individuals are exposed to environmental hazards such as air
environmental hazards pollution. In many instances, it is possible to control exposure
to hazards, e.g. using sun block to limit UV exposure
Individuals as lobbyists and Individuals can lobby governments and policy-makers on
influencers policy change to reduce industrial and global pollution and
consumption of natural resources. Industries also rely on
individuals as employees and consumers. Individuals can also
make decisions about the goods that they consume based on
their environmental impact

2F.2 HAZARD AND RISK

Risk communication is an important aspect of public health, and is arguably the most important area of
communication for public health practitioners. Successful risk communication relies on having an understanding
of the different concepts of risk, and the skills and integrity to communicate risk information appropriately and
honestly. Poor risk communication not only affects the issue under discussion but may also irreparably damage the
trust of the public.
If all other factors are held constant then, technically: Risk = Hazard x Exposure.
However, Sandman (1987) proposes an alternative concept of risk that takes into account the public’s response to a
hazard. He defines risk as being, ‘What people feel to be the likelihood of an event’ rather than the epidemiological
‘likelihood of an event’. He divides the ‘risk’ that people are worried about into two components:
• The technical (epidemiological) aspect of the risk. This relates to the magnitude and probability of the
undesirable outcomes (e.g. an increase in the cancer rate, a catastrophic accident, number of dead fish in the
river or even a decline in property values). He calls this the ‘hazard’. Thus a hazard can influence how society
perceives risks.
• The non-technical aspects of the risk. These are the perceived negative features of the situation itself (as
opposed to those of the outcomes). Such features are listed below and together they constitute the ‘outrage’.
Sandman’s concept is the sum of these two aspects: Risk = Hazard + Outrage.

218

Public [Link] 218 3/6/08 [Link]

 2F En viron ment

In his opinion, risk communication depends on securing an appropriate degree of ‘outrage’ in the public, so that
they are neither frightened unnecessarily nor apathetic about real problems. Sandman (1987) lists nine factors that
can increase or decrease the level of outrage: see Table 2F.2.1.

Table 2F.2.1 Sandman’s factors affecting public outrage and hazard perception
voluntariness A voluntary risk is much more acceptable to people than a coerced risk, because it
generates no outrage. Consider the difference between being pushed into a swimming
pool and diving into a swimming pool
Control Most people feel safer driving a car than sitting in the passenger seat. When
prevention and mitigation are in the individual‘s hands, the risk (though not the
hazard) is seen as being lower than when they are in the hands of a government
agency
Fairness When people feel that they are subjected to greater risks than their neighbours,
without access to greater benefits, they are naturally outraged – especially if
the rationale for the extra burden is perceived to be due to political, rather than
scientific, reasons
Process Outrage is affected by public perception of the authority or government body, i.e.
whether it is perceived as being trustworthy or dishonest, concerned or arrogant.
Perceptions can be improved by proactively informing the public and by responding to
the concerns of the community
Morality Where a risk has acquired a moral dimension (e.g. childhood cancer), discussion of
cost−risk trade-offs is unacceptably callous. Imagine a public health specialist arguing
that an occasional child death is an ‘acceptable risk’
Familiarity Novel or exotic risks provoke more outrage than do familiar risks
Memorability Memorable incidents (e.g. Chernobyl, Bhopal, Three Mile Island) make a risk easier to
imagine, and thus more risky (as defined above)
Dread Some illnesses are more dreaded than others. Compare cancer with, say, heart failure.
The long latency of most cancers and the surreptitiousness of most carcinogens add to
the dread
Diffusion in time and If hazard A kills 50 anonymous people a year across the country, and hazard B has
space one chance in 10 of wiping out an entire town of 50 000 people sometime in the next
century, then a risk assessment tells us that the two have the same expected annual
mortality of 50 deaths per year. However, an ‘outrage assessment’ would show that
hazard A is probably acceptable but hazard B is certainly unacceptable

Reproduced from Sandman (1987).

Practitioners can deal with public communication of risk more effectively by listening to the concerns expressed by
the public, including those of pressure groups, in order to:
• Take into account the factors that influence risk perception
• Understand the strength of feeling and the points of view
• Use appropriate media and language to communicate relevant information in a meaningful form.

219

Public [Link] 219 3/6/08 [Link]

 2F E n v i ro n me nt

2F.3 CLIMATE CHANGE

Effects of climate change and global warming
The main causes of human-induced climate change are increased levels of greenhouse gases, especially CO2 and
methane. Most of the heat coming from the sun is reflected off the earth back into space. But so-called ‘greenhouse
gases’ trap the warmth. As the concentration of greenhouse gases increases, so the planet’s temperature is predicted
to rise. Increased atmospheric levels of CO2 result from deforestation (trees would normally convert CO2 to oxygen
through photosynthesis) and from increased combustion of fossil fuels.

GLOBAL WARMING
Average global temperatures have risen over the past century due to both natural and human factors. Climatologists
predict that extreme weather (e.g. floods, hurricanes and droughts) will become more frequent because of climate
change. In the UK, the 1990s were the warmest decade in the last 100 years, and by 2080, the average annual
temperature is predicted to rise by between 2 and 3.5oC.

EFFECTS OF CLIMATE CHANGE ON HEALTH

Effects may be direct or indirect: see Box 2F.3.1.

Box 2F.3.1
Direct effects Indirect effects
• Extreme heat waves have led to excess mortality, • More floods, promoting the spread of water-borne
particularly among old people or those in diseases
vulnerable sectors of society
• More asthma and respiratory illnesses as a result of
• Changing epidemiology of infectious diseases, e.g. increased atmospheric pollution
spread of malaria as animals/insects migrate to
• Increasing numbers of refugees as some areas are
different geographical areas
submerged under rising sea levels
• Fewer deaths related to hypothermia (although
• Droughts and changing food patterns lead to
paradoxically the temperature in the British Isles
starvation, famine and increasing numbers of
may fall due to the effects on the Gulf Stream)
refugees

MONITORING AND CONTROL

Efforts to restrict the burning of fossil fuels and stop deforestation are among many measures to limit climate
change contained in the 1997 Kyoto Protocol. The protocol is a legally binding agreement for countries to cut
emissions of the six main greenhouse gases. It came into force in 2005, but while 141 countries have ratified the
protocol, two of the biggest producers of greenhouse gases − the USA and Australia − rejected the treaty. Many
countries that signed up to the treaty are experiencing difficulties in meeting its targets.

2F.4 SUSTAINABILITY
Principles of sustainability
Environmental sustainability encompasses the concern to balance the needs of the present generation with those
of the future. It involves long-term considerations about resource use, set in the context of the current high rate of
resource use and production of pollutants.

220

Public [Link] 220 3/6/08 [Link]

 2F En viron ment

UK The Sustainable Development Commission adopts a broad view of sustainability, considering not just
environmental, but also social and economic developments. It outlines five principles of sustainable development:
see Table 2F.4.1

Table 2F.4.1 Principles of sustainable development

Environmental limits Respecting the limits of the planet’s environment, resources and
biodiversity to improve the environment and ensure that all the
natural resources needed for life remain for future generations
Healthy and just society Meeting the diverse needs of all people in existing and future
communities; promoting wellbeing, social cohesion and inclusion;
and creating equal opportunity for all
Good governance Actively promoting effective, participative systems of governance
in all levels of society – engaging people’s creativity, energy and
diversity
Responsible use of science Ensuring that policy is developed and implemented on the basis
of strong scientific evidence while taking into account scientific
uncertainty through the precautionary principle (see Section 2F.7) as
well as public attitudes and values
Sustainable economy Building a strong, stable and sustainable economy that provides
prosperity and opportunities for all, an economy in which
environmental and social costs fall on those who impose them
(polluter pays, described in Section 2F.7) and in which incentives
are in place to promote efficient resource use

2F.5 HOUSING
Health problems associated with poor housing and home conditions, inadequate water supplies and sanitation
Poor housing and home conditions can be associated with health problems. Specific features of poor housing and
home conditions are summarised in Table 2F.5.1.

Table 2F.5.1 Housing conditions affecting health

Pollutants Examples include asbestos, carbon monoxide, radon, lead
and volatile organic chemicals
Damp Causes moulds and spores
Temperature Cold and heat extremes
Design Poor housing design or layout
Noise Intrusive or excessively loud noises or noise made at
antisocial hours
Social and behavioural environment Overcrowding, sleep deprivation
Housing allocation Homelessness, temporary accommodation, housing tenure,
housing investment

221

Public [Link] 221 3/6/08 [Link]

 2F E n v i ro n me nt

Eng The amenities available to a neighbourhood (such as access to transport, shops and open spaces) have
a profound impact on health. These are mediated though their effects on safety, social cohesion and crime. In
England, government policy explicitly recognises the benefits of open spaces for enabling local people to be active
and maintain social links. Councils have a duty to consult locally on plans for maintaining open spaces, sports
facilities and buildings.

HEALTH EFFECTS OF POOR HOUSING

The relationship between housing and health is complex: individual exposures may not relate directly to particular
morbidities. People living in poor housing conditions typically also experience other forms of deprivation (poor
education, unemployment, ill health, social isolation, etc.), and this makes it difficult to assess, isolate or modify
the overall health effects. In general, housing has both physical and mental effects on residents’ health.

PHYSICAL HEALTH PROBLEMS

Poor housing is associated with asthma, skin allergies and respiratory diseases. It is also linked to physical accidents
and injuries. Exposure to specific toxins has particular effects, e.g. radon is associated with cancer, CO causes
asphyxiation and asbestos causes asbestosis and malignancy.

MENTAL HEALTH PROBLEMS

Poor housing can lead to depression, isolation, anxiety or aggression. Noise-related stress, from poor sound
insulation, is associated with lack of sleep, mental distress and depression: for an example see Box 2F.5.1.

Box 2F.5.1
Example: WHO LARES Survey
The WHO’s LARES (Large Analysis and Review of European housing and health Status) was a survey across eight
countries involving 8500 people, which studied links between housing conditions and physical and mental ill
health. It found, for example, that people were more depressed and anxious when their housing did not protect
against noise, or they were subjected to overcrowding or to social isolation. Dampness and visible mould growth
were related to asthma, nasal allergies and eczema.
Reproduced from Bonnefoy et al (2004).

HEALTH PROBLEMS ASSOCIATED WITH INADEQUATE WATER SUPPLIES AND SANITATION

Inadequate water supplies remain a major problem for around 15−20% of the world’s population, largely in Asia
and sub-Saharan Africa. Inadequate water supplies and sanitation problems can result from natural disasters (flood,
earthquake or drought) or from pollution. Human interventions to reduce water shortages can themselves exacerbate
health problems (e.g. malaria associated with irrigation channels).
Inadequate water supplies can lead to a range of infectious diseases. These may be due either to microbial agents
in the water or to water-related vectors. WHO (2007b) outlines the most significant diseases associated with
inadequate water supply and sanitation (see Table 2F.5.2).

222

Public [Link] 222 3/6/08 [Link]

 2F En viron ment

Table 2F.5.2 Infectious diseases associated with water

Diarrhoeal illness 88% of diarrhoeal illness is due to unsafe water supply, inadequate sanitation and poor
hygiene
• Improved water supply reduces morbidity by 6−25%
• Improved sanitation reduces morbidity by 32%
Malaria There are 396 million malaria episodes per year worldwide, with most of the disease
burden in sub-Saharan Africa. Intensified irrigation, dams and other water-related projects
contribute to disease burden
Better management of water resources reduces transmission of malaria and other vector-
borne diseases
Chemical Pollutants can also cause disease, e.g.:
pollutants • Arsenic: contamination of ground water is found in many countries, including Argentina,
Bangladesh and the USA. Arsenic leads, among other things, to skin lesions
• Fluorosis: naturally elevated fluoride levels in drinking water usually occur due to the
geology of the area. It can lead to skeletal and dental fluorosis
• Aluminium: aluminium sulphate contamination of the Camelford reservoir in Cornwall in
the 1980s may have caused acute symptoms, and possibly long-term neural damage

The UN’s Millennium Development Goals include aims to reduce water shortages and improve sanitation. Reducing
the proportion of people without access to safe drinking water and basic sanitation improves wellbeing directly. It
will also have an indirect effect on goals associated with reducing morbidity and mortality, by affecting the risk of
illnesses associated with water supply, e.g. diarrhoea.

2F.6 MONITORING AND CONTROL OF ENvIRONMENTAL HAZARDS

Methods for monitoring and control of environmental hazards (including food and water safety, atmospheric pollution,
and other toxic hazards, noise, and ionising and electromagnetic radiation)
See Sections 2F.1 and 2F.3.

2F.7 USE OF LEGISLATION IN ENvIRONMENTAL CONTROL

The control of environmental pollution and waste production by businesses often requires legislation to be effective.
This is because there is commonly a perceived or real conflict between economic priorities and the environment.
The issue is particularly pertinent for small businesses, and also in developing countries where the information and
expertise to implement environmentally friendly measures may be lacking.
Many environmental hazards are not limited to geographical boundaries. For example, acid rain caused by UK power
plants in the 1970s and 1980s mainly fell in continental Europe. Therefore, legislation and strategies to tackle
pollution and waste production should be set at an international level. European legislation may be primary (e.g.
treaties and directives) or secondary (e.g. regulations and statutory instruments).

EU EUROPEAN UNION LEGISLATION

EU European environmental legislation is based on two guiding principles: ‘polluter pays’ and the ‘precautionary
principle’: Box 2F.7.1.

223

Public [Link] 223 3/6/08 [Link]

 2F E n v i ro n me nt

Box 2F.7.1
Polluter pays This is the principle that the party responsible for creating pollution should provide the
investment needed to meet higher environmental standards or to create a system to
recover, recycle or dispose of products after use. The payment may also be a tax on business
or consumers for using an environmentally unfriendly product, such as some types of
packaging
Precautionary Credible but unproved health hazards should be treated as real threats until shown
principle otherwise. Decisions affecting the environment should not be delayed while scientific data
are collected. Instead, policy-makers should err on the side of protecting the environment

European legislation aims to restrict levels of pollution, and to ensure the safe disposal of waste through measures
outlined in Table 2F.7.1.

Table 2F.7.1 Measures in European legislation to restrict pollution

Permits or Issued to those using natural resources, disposing of waste or producing pollution (e.g.
licences sewage companies, water companies and manufacturing industries). Conditions attached to
the permit typically set limits on the amount of pollution that may be produced, or stipulate
the use of specific procedures and best available techniques (BATs)
Taxes or levies By reflecting the environmental costs of production, these aim to discourage production of
emissions. An example is the climate change levy
Trading schemes Companies participate in a scheme that is committed to reducing emissions. Firms may ‘sell’
or ‘buy’ excess emission allowances and thus compete on the basis of an overall reduction in
emissions
Product labelling ‘Eco-friendly’ or energy efficiency ratings are displayed on household and office appliances
schemes

Compliance with the terms of permits and regulations can be enforced through:
• Inspections, warning letters or formal cautions
• Enforcement or prohibition notices
• Suspension, revocation or modification of permits
• Fines and imprisonment.

REGULATORY AGENCIES
UK In the UK, the Environment Agency regulates large and complex industrial processes, while local authorities
regulate smaller-scale industries. Other relevant regulatory agencies include the Health and Safety Executive,
Trading Standards and the Food Standards Agency.
Environmental health officers from the local authority enforce the regulations on the Clean Air Act, and prevent the
sale of contaminated foods.

EFFECTIVENESS OF LEGISLATION
Compliance with environmental legislation is variable. There are many reasons for this, including:
• Business are not always aware of legislative requirements
• Monitoring in some countries is more rigorous than others

224

Public [Link] 224 3/6/08 [Link]

 2F En viron ment

• Penalties for non-compliance are comparatively weak, e.g. compared with health and safety offences
• Compliance is not always perceived as central to a firm’s survival.

2F.8 HEALTH AND SAFETY AT WORK

Appreciation of factors affecting health and safety at work (including the control of substances hazardous to health)
Factors affecting health at work may be associated with the work environment, occupational equipment or other
employees. They can be either beneficial or harmful, and be mediated through either direct or indirect effects (see
Section 2H.6).
Workplace exposures may be considered as physical, substance related and psychological.

PHYSICAL EFFECTS
Back disorders are the commonest cause of workplace ill health, and musculoskeletal problems in general are a major
cause of morbidity. They are particularly associated with:
• Sedentary occupations
• Lifting and manual handling of heavy objects
• Driving for long periods
• Repetitive tasks.

SUBSTANCES
UK The Control of Substances Hazardous to Health (COSHH) (2002) Regulations cover the following types of
hazardous substance:
• Substances used directly in work activities (e.g. adhesives, paints, cleaning agents)
• Substances generated during work activities (e.g. fumes from soldering and welding)
• Naturally occurring substances (e.g. grain dust)
• Biological agents such as bacteria and other microorganisms (e.g. leptospirosis in sewage treatment workers).
The effects of these substances may be atopic, infectious or carcinogenic, and they may manifest in a range of
different timeframes: see Box 2F.8.2.

Box 2F.8.2
Acute effects Losing consciousness as a result of being overcome by toxic fumes
Early effects Skin irritation or dermatitis as a result of skin contact (particularly common among beauticians
and hairdressers)
Late illnesses Cancer may develop many years after the exposure to the carcinogen (e.g. asbestos)

PSYCHOLOGICAL FACTORS AFFECTING HEALTH

Work is the single biggest cause of stress in Britain. Employment-related factors that cause stress and mental ill
health include:
• Bullying
• Harassment
• Discrimination
• Lack of control
• Lack of reward
• Long working hours.

225

Public [Link] 225 3/6/08 [Link]

 2F E n v i ro n me nt

People at particularly high risk are:

• Temporary workers
• Women (a higher proportion are in lower status jobs)
• Those working illegally (e.g. asylum seekers)
• Lone workers
• Staff in small businesses (small businesses can have the most difficulty complying with workplace regulations)
• Pregnant women.
Depression and anxiety are the most common causes of long-term incapacity. Stress is also linked to premature
death from coronary heart disease. Workplace exposures can include:
• Stress
• Inadequate team support
• Inadequate reward
• Lack of control over roles and responsibilities in the workplace.
In certain circumstances, employees can be a risk to the health of other people, e.g.:
• Health-care workers with infections. Safeguards include restriction of exposure-prone procedures for those with
blood-borne infections, e.g. hepatitis B
• Food handlers, carers of at-risk groups with gastrointestinal illnesses. Safeguards include removing from work
during illness and until the infection has cleared
• Drivers, e.g. public transport, goods: safeguards include strict regulations and regular monitoring over
substance (drugs, alcohol) intake.

HEALTH AND SAFETY AT WORK

Effective management of health and safety issues at work not only maximises the benefits to health but also makes
economic sense for companies and society. This is because it reduces absenteeism, increases productivity and avoids
compensation claims.
UK Key agencies involved in occupational health are listed in Table 2F.8.1.

Table 2F.8.1 UK agencies involved with occupational health

Agency Purpose
Health and Safety • Sets regulations
Executive • Monitors workplaces
• Investigates incidents
• Enforces workplace regulations
Trade unions • Campaign for workers’ rights to healthy working conditions
• Survey employees’ working conditions
• Provide advice on legal issues concerning health and safety
• Represent employees with employers on health and safety issues
Local authorities • Local inspection, enforcement of regulations
Employers • Implement regulations
• Provide safe working conditions
Occupational health • Undertake pre-employment health checks
departments • Advise employers on workplace hazards and their control
• Advise employees on health and safety issues
• Monitor sickness absence
• Offer immunisations if appropriate, e.g. in health and social care jobs

226

Public [Link] 226 3/6/08 [Link]

 2F En viron ment

OCCUPATIONAL HEALTH REGULATIONS

UK Pertinent regulations include:
• Control of Substances Hazardous to Health Regulations (COSHH) (2002)
• Reporting of Injuries, Diseases and Dangerous Occurrences Regulations 1 (RIDDOR) (1995): required for all
serious work-related incidents
• WHO occupational limits for pollutants
• European Working Time Directive (among other provisions, this restricts the working week to 48 h).

SAFETY AT WORK AUDITS

Procedures to monitor and control health and safety factors at work can be considered in several sequential steps:
see Box 2F.8.3.

Box 2F.8.3
Assessment Assessment of hazards in the environment
Identification Identification of those at most risk (e.g. through pre-employment
health checks)
Surveillance Surveillance of hazards, incidents and health-related morbidity
(through systems of incident reporting, feedback and analysis)
Mitigation Action to reduce exposure to hazards (through training, provision of
protective equipment to exposure and emergency planning)
Monitoring Monitoring of the effect of interventions to reduce effects of hazards
(e.g. through incidence, sickness and absence records)

2F.9 OCCUPATION AND HEALTH

The effects of occupation on health vary according to the:
• Type of occupation
• Personal risk factors
• Levels of social support.
In general, there are substantial health benefits to being in work, but certain occupations can expose employees to
particular risks (see Section 2F.8).

UNEMPLOYMENT
Being out of work is associated with physical, mental and social effects. These are related partly to the length of
time spent unemployed. Although being out of work may cause morbidity, the causal relationship also works the
other way around: those who are physically or mentally unwell may be more likely to leave jobs or have difficulty
working. Studies of factory closures indicate that ill health and health-care use are also associated with job
insecurity, and particularly with the anticipation of job losses.
There are consequences from unemployment on the individual, on families and on society at large: see Box 2F.9.1.

227

Public [Link] 227 3/6/08 [Link]

 2F E n v i ro n me nt

Box 2F.9.1
Individual Suicide rates among the unemployed increase within a year of job loss
Cardiovascular mortality rises over 2 or 3 years, and continues for the next 10−15 years
Family Spouses of people out of work can also experience poor health. For those whose partner has
been forced to give up work due to illness, they may be burdened with caring responsibilities. In
families where individuals are seeking work, the effects of poverty appear most severe in the short
term. In the longer term, families may adapt to unemployment
Society Increased unemployment places a greater burden on society through fewer people making tax
contributions, and due to the increased social security and health service needs of those without
work

2F.10 TRANSPORT AND THE ENvIRONMENT

Transport policies and health impact assessment for environmental pollution
Transport policies need to acknowledge the potential benefits from increased transport and accessibility. These
improve the prospects of employment, availability of goods and greater choice of social activities. It is also
important to recognise the health benefits associated with walking or cycling, as well as the risks associated
with other forms of transport − particularly via road and air, for example; the effects of transport on the health of
Londoners has been the subject of a health impact assessment (see Box 2F.10.1). The main problems associated
with transport are:
• Air pollution and global warming
• Noise pollution
• Accidents
• Effects on social cohesion (e.g. from a busy dual carriage-way bisecting a community).
Road transport is one of the greatest contributors to air pollution, particularly in towns and cities (see Section
2F.1). However, while emissions from most sources are decreasing, the impact of air travel is increasing sharply.
According to the European Commission, EU emissions from international aviation increased by almost 70% between
1990 and 2002.

HEALTH IMPACT ASSESSMENTS

See also Section 1C.17.
A health impact assessment is an important tool for informing policy-makers about the health effects of potential
transport policy decisions. However, there are many challenges in conducting this type of research. In general,
the evidence regarding the quantitative effects on health is variable. For example, while the mortality data from
traffic accidents are uncontested, there are methodological disputes about how to measure the effects of noise
disturbance, the health benefits of being physically active, transport effects on respiratory illness, etc.

SUSTAINABLE TRANSPORT
Sustainable transport policies focus on the following elements.

ENvIRONMENTALLY FRIENDLY TRANSPORTATION

Alternatives to cars can be encouraged by sustainable transport policies: see Box 2F.10.2.

228

Public [Link] 228 3/6/08 [Link]

 2F En viron ment

Eng Box 2F.10.1

Example: On the move: Health Impact Assessment of Transport in London (2000)
In London, traffic is responsible for the production of 99% of carbon monoxide, 76% of nitrogen oxides and
90% of hydrocarbons. As part of London’s health strategy, the NHS commissioned a health impact assessment of
transport in London.
Scope: the assessment focused on:
• Traffic accidents
• Air pollution
• Noise
• Health benefits from physical activity
• Community severance, mental health and inequality effects
The positive effects of transport from increased access to goods and services and ability to move around London
were outside the scope of the assessment.
Methods: researchers assessed and quantified health effects by:
• Collating and interpreting routine data
• Reviewing evidence from published and grey literature
• Consulting stakeholders
Figures of accidents indicated that road traffic accidents led to 226 fatalities in the previous year. In comparison,
the authors estimated that pollution would lead to 380 premature deaths and 350 hospital admissions due to
respiratory disease.
Findings: this health impact assessment indicated that pollution is as important as (and perhaps more important
than) accidents with respect to the health of Londoners.
Reproduced from Watkiss et al (2000).

Box 2F.10.2
Transport option Promotion
Walking Pedestrian zones
Bicycles Bicycle lanes
Public transport Subsidised public transport systems

TAxATION
Taxation can be used to encourage industry and consumers to adopt more fuel-efficient vehicles and thereby reduce
vehicle emissions.

REDUCING CAR USAGE

This can be achieved through:
• Levies or taxes for road use (e.g. London Congestion Charge)
• Parking restrictions and charges
• Subsidising the cost of public transport
• Encouraging car sharing through introducing share-only lanes and high-volume-occupancy lanes
• Reducing the transport of freight on the roads and increasing rail transport.

229

Public [Link] 229 3/6/08 [Link]

 2F E n v i ro n me nt

REDUCING THE ENvIRONMENTAL IMPACT OF TRANSPORT

Carbon-offsetting initiatives seek to compensate the environment for emissions of carbon dioxide. Processes that
emit carbon dioxide are matched with a proportionate investment in projects that either reduce the emission of
carbon dioxide or remove an equivalent amount of the gas from the air. For example, air passengers can make a
donation to a non-governmental organisation that will fund projects designed to reduce emissions.
Wilkinson (2006) describes how researchers are exploring the option of carbon capture as an alternative approach
to reducing the impact of carbon dioxide production. The damage to the environment results from carbon dioxide
in the atmosphere. Carbon capture relies on finding locations to store carbon dioxide so that it does not reach the
atmosphere. Depleted oil and gas fields could be ideal carbon capture locations: they have increasing storage space,
and the insertion of carbon dioxide could make the continued extraction of gas and oil more straightforward. Note
that carbon capture would not result in a reduction in the use of hydrocarbons, or in the production of carbon
dioxide. It may be best regarded as a relatively inexpensive short-term solution the value of which lies in buying
time to develop alternative fuel sources. An example is air transport (see Box 2F.10.3).

EU Box 2F.10.3
Example: Air transport: an international perspective to transport policy
In Europe in the last 20 years, air transport policies have focused on expanding the air travel market, rather than
on reducing or containing its environmental impact. One of these policies is the exemption of aircraft fuel for
international flights from all taxes. While individual countries in the EU have the option to charge fuel tax for
domestic flights, only the Netherlands currently does so.
As a result, air transport has increased rapidly. Although planes are becoming more fuel efficient, this is more
than offset by air traffic. The European Commission recommends that the aviation industry should become part
of the EU Emissions Trading Scheme. Under this scheme, the European Commission places an overall ‘cap’ on
emissions per year. Within these limits, companies are given an emissions allowance. If they are in danger of
exceeding their allowance, they have two options: either to reduce their emissions (e.g. by investing in new
technology, changing production) or to ‘buy’ the allowances of their competitors on the open market.
Adapted from [Link]/comm/environment/climat/pdf/ia_aviation.pdf, [Link]/comm/
environment/climat/aviation_en.htm.

2F.11 CHEMICAL INCIDENT MANAGEMENT

Damage from chemical incidents can arise from acute events (such as a chemical spill or an explosion) or due to
chronic processes (such as leeching into the soil or water supply). The detection and management are different
between the two, but some of the principles and possible consequences are similar. Examples of chemical incident
management are described in Boxes 2F.11.1 and 2F.11.2 on pages 233 and 234.
The most common environmental effects of a chemical incident are:
• Contaminated land (particularly from leaks, slow accumulation of chemicals)
• Contaminated water supplies
• Air pollution, e.g. from industrial fires.
UK The Health Protection Agency provides several checklists that outline the management of different types of
chemical incident. While the details differ depending on the nature of the incident, the general steps involved are
outlined below (the order of the steps will depend on the nature of the incident).

230

Public [Link] 230 3/6/08 [Link]

 2F En viron ment

INCIDENT ASSESSMENT
This involves confirming which chemical or chemicals are involved, the extent of contamination and the potential
for further contamination. The pathway through which chemicals are leaking or leeching must be determined.
Samples of soil, water and air from the local environment are taken.

HEALTH ASSESSMENT
Appropriate epidemiological study designs to assess health effects are generally either a case–control study
(requires a case definition and for cases to have been identified) or a cohort study (if there is a defined population
that was potentially affected).
Both study designs could involve biological sampling of those exposed and with symptoms (e.g. blood or urine
analysis), and a survey of all exposed people to quantify physical and psychological symptoms. After an incident it
may also be appropriate to consider whether to instigate long-term epidemiological surveillance, or to follow up a
selected cohort (measuring the nature and extent of any long-term health effects).

RISK MINIMISATION
After an incident, these steps usually take the form of making arrangements for moving people or materials and are
designed to:
• Provide advice to exposed and potentially exposed populations
• Ensure appropriate shelter: this may involve simply advising people to stay inside but could also require
arranging alternative accommodation if homes or businesses are damaged or unsafe
• Manage evacuation: as well as identifying a suitable place for people to be evacuated to, it is also important to
consider how to inform, prepare and move the public
• Clear up contamination where feasible.

COMMUNICATION AND COORDINATION

In a chemical incident, a range of agencies will be involved in minimising the damage and clearing up the incident.
It is therefore important for all organisations involved to understand their respective roles and responsibilities.
Clear lines of communication are required in order to ensure that the most current information is available to all –
particularly with regard to the potential hazards faced. Different agencies will be involved depending on the scale
and nature of the incident, but Tables 2F.11.1 and 2F.11.2 indicate agencies likely to be involved in the UK.
Aus There is a National Disaster Organisation, but state police control most incidents. The police work in
association with fire, health and environmental protection authorities. Other organisations (e.g. water utilities) are
involved as appropriate. Officers from all of these agencies have well-defined, statutory duties in such situations.

PUBLIC INFORMATION
Those affected by the incident need to be kept well informed. The general public may have been directly or
indirectly exposed to the effects of the incident, and many will be concerned that they could be affected. Useful
information includes:
• Actions to take to minimise personal risk
• Minimising risk to others
• Where to get help and advice.

231

Public [Link] 231 3/6/08 [Link]

 2F E n v i ro n me nt

Table 2F.11.1 UK national agencies involved in reducing the impact of environmental incidents
Eng National agencies Wal National agencies Scot National agencies UK National agencies
Environment Agency Environment Agency Scottish Environment Food Standards Agency (if
Protection Agency (SEPA) potential for contaminated
DEFRA DEFRA
food stuffs)
Scottish Executive
Environment and Rural Health and Safety
Affairs Department Executive
(SEERAD)
National newspapers, TV
and radio
Internet
Specialist press
Health Protection Agency National Public Health Health Protection Scotland
Service for Wales

Table 2F.11.2 UK local agencies involved in reducing the impact of environmental incidents
UK Local agencies
• Emergency services (police, fire, coastguard)
• Local authority
• Environmental health department
• Health and safety departments
• Pollution control
• Health protection units, departments of public health
• Water companies (if potential for contaminated water supplies)
and local Environment Agency office
• Health and Safety Executive (local branch)
• Eng NHS: A&E, acute hospital trusts; PCTs; NHS Direct
• Wal NHS: A&E, local health boards; NHS Direct
• Scot NHS: A&E, local health boards; NHS 24
• Media: local newspapers, regional TV and radio

Useful media for inform people include:

• Local radio
• Helplines
• Special mailings
• Websites
• Press releases to local, specialist and national media.

LEARNING
Once the process has been completed, all the agencies involved should identify lessons learnt through the incident.
Based on those experiences, plans should be modified to prepare for managing future incidents and to prevent
similar incidents from occurring.
Boxes 2F.11.1 and 2F.11.2 give examples from the UK and New Zealand.

232

Public [Link] 232 3/6/08 [Link]

 2F En viron ment

UK Box 2F.11.1
Example: Management of the Buncefield Oil Depot Fire
The incident: On 11 December 2005, Buncefield Oil Depot near Hemel Hempstead (south-east England) was
the site of ‘possibly the largest explosion in peacetime Europe’. A fog of petrol fumes caught alight, leading to
explosions for the next 24−48 h. Over the next few days, London and significant parts of south-east England were
covered by the smoke plume, which spread to parts of northern France and Portugal. The site housed hundreds
of businesses, employing several thousand staff. Nearby, there was a traveller community and a substantial
residential community. Over 2000 people were evacuated from their homes and sections of the nearby motorway
were closed to prevent exposure to the fumes produced by the explosions.
Agencies: the police led the response to this incident, with input from a range of agencies, including Dacorum &
Watford and Three Rivers Primary Care Trusts, the Health Protection Agency, the Environment Agency, the Health
and Safety Executive and the fire service.
Environmental effects: the fire actually generated few pollutants – its high temperature meant that organic
chemicals in the fuel were destroyed. The Met Office provided information on the plume direction and spread
using observations, satellite images and computer modelling. This indicated that smoke from the fire rose high in
the atmosphere.
Health and social effects: data collected from NHS Direct and A&E in the hours and days after the incident
indicated that Buncefield was not a major incident for health; there were no deaths and few injuries.
A survey was also used to find out residents’ experience of the fire, acute exposure and longer-term health
effects.
Longer term: risks under investigation include:
• The risk of contamination of water supplies from foam
• Longer-term health effects – both physical and psychological
• Employment and housing in the Buncefield area
Adapted from HPA conference, 7 March [Link], 12 December 2005: Massive blaze rages at fuel depot, The
public health impact of the Buncefield oil fire, 2006: [Link]/publications/2006/buncefield/[Link],
DEFRA, Initial review of air quality aspects of the Buncefield oil depot explosion: [Link]/environment/
airquality/buncefield/[Link].

233

Public [Link] 233 3/6/08 [Link]

 2F E n v i ro n me nt

NZ Box 2F.11.2
Example: New Zealand management of hazardous substances (Hazsub) events
Hazsub are mainly accidental releases from transportation-related events or from fixed facilities such as factories
and storage plants. Rarely, they may be intentional releases from criminal or terrorist activity or from natural
hazards such as a volcanic eruption. Responding to such incidents involves the coordinated actions of multiple
agencies.

District health boards (DHBs) (covering hospitals, Plan the response to such incidents, with Hazsub events
public health services and contracted primary care included in the DHB Major Incident and Emergency Plan.
providers) Public health services have a key role in preparedness
planning
Fire service Scene management, containment of released hazardous
substance and decontamination of individuals at the
scene
Ambulance service Assessment, triage, initial treatment and transport of
people injured during a Hazsub release
Police Securing and managing the scene and investigating the
incident if the release results from criminal or terrorist
act, potentially with support from the New Zealand
Defence Force
Hazardous substance technical liaison committees Advise and support the fire service when dealing with
hazardous chemical incidents
National Poisons Centre Maintains a database of chemicals, health effects and
recommended treatment
National Radiation Laboratory Expert advice and technical support to the emergency
and health services if ionising radiation involved
Civil defence and emergency management groups Consortia of local authorities, emergency services,
major utilities and others to ensure that emergency
management principles are applied at the local level
Ministry of Health Produces the National Health Emergency Plan (NHEP)
Provides advice to DHBs, public health services and other
agencies to prepare their own action plans
Ministry of Civil Defence and Emergency Management Responsible for the administration of the Civil Defence
(MCDEM) Emergency Management Act 2002. Uses national
standard: the Coordinated Incident Management System
(CIMS)

234

Public [Link] 234 3/6/08 [Link]

 2G Com mun icable Diseases

2G
Communicable Diseases

2G.1 States in the development of infectious 2G.6 Outbreak investigations 244

diseases 235 2G.7 Important infectious diseases 245
2G.2 Surveillance – national and international: 2G.8 Organisation of infection control 267
its evaluation and use 237
2G.9 Microbiological techniques 269
2G.3 Methods of control 241
2G.10 International aspects of communicable
2G.4 Design, evaluation and management of disease control 271
immunisation programmes 241
2G.5 Choices in developing an immunisation
policy 244

Infectious diseases represent the largest cause of childhood and adolescent deaths worldwide. They account for over
13 million deaths a year and are the cause of over half of all deaths in developing countries. In developed countries
infectious diseases are of importance for the following reasons:
• Re-emergence of old scourges (e.g. TB)
• Novel infections (e.g. SARS)
• Threat of pandemic influenza
• Hospital-acquired infections (e.g. MRSA)
• Burden of long-term conditions (e.g. HIV/AIDS)
• Recognition that certain cancers are caused by viruses (e.g. cervical cancer).

2G.1 STATES IN THE DEvELOPMENT OF INFECTIOUS DISEASES

Definitions (incubation, communicability, latent period, susceptibility, immunity and herd immunity)
See Table 2G.1.1.

235

Public [Link] 235 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

Time

Discharge of Discharge of
pathogen begins pathogen ends

Latent period Period of

communicability

Infectious

Susceptible Immune
Infected

Incubation period

Infection acquired Start of symptoms End of infection

(immune/death)

Figure 2G.1.1 Temporal measures in infectious disease

Table 2G.1.1 States in the development of infectious diseases

Susceptibility A susceptible person has insufficient resistance against the pathogen to avoid
infection
Incubation period Also known as the ‘subclinical period’, this is the time between infection and the onset
of clinical symptoms. Its duration may be affected by the infecting dose
Establishing the incubation period enables the following to be determined:
• When infection occurred
• Who might be a contact and how long they should be quarantined for
Latent period Time between initial infection and the start of infectiousness
Infectious period Time during which the infected person is capable of transmitting the infective agent
(often begins before the onset of symptoms)
Degree of infectivity This will usually vary during the infectious period. It determines:
• The extent to which the disease will be transmitted in the population
• For how long patients should be isolated or quarantined
Index case The first case of the disease to come to the attention of the investigator
Primary case Case that occurred due to infection from the same source as the index case
Secondary case Case that occurred due to infection from a primary case
Carrier Person who becomes infected and infectious but not ill
Serial interval Time between the onset of symptoms in a primary case to the onset of symptoms in a
secondary case

236

Public [Link] 236 3/6/08 [Link]

 2G Com mun icable Diseases

Secondary attack rate Probability of disease among known (or presumed) susceptible people following
contact with a known primary case
Basic reproductive Average number of new infectious cases in a completely susceptible population
number (R0) produced by a single case during its entire period of infectiousness

IMMUNITY
Immunity is the state of having sufficient biological defences to avoid infection. It can be considered as passive or
active: see Table 2G.1.2.

Table 2G.1.2 Active and passive immunity

Definition Natural acquisition Artificial acquisition Duration
Passive Passive immunity is Transplacental transfer Inoculation of specific protective Short: days
immunity acquired by transfer from the mother antibodies (from immunised to months
of antibodies (or T animals, or from human
lymphocytes) from convalescent serum)
another individual
Active Active immunity is Infection (with Inoculation (of either the agent Long: usually
immunity acquired through an or without clinical itself in killed, modified or several years
encounter with an manifestations) variant form, or else of fractions
antigen or products of the agent)

HERD IMMUNITY
If most people in a community are immunised against an infection, the spread of that infection is significantly
reduced and even unvaccinated people are at much lower risk of catching the illness. For childhood infections such
as measles, about 95% of children must be vaccinated in order to achieve this level of protection.

2G.2 SURvEILLANCE – NATIONAL AND INTERNATIONAL: ITS EvALUATION AND USE

Surveillance is the ongoing systematic collection, collation, analysis and interpretation of data, and the
dissemination of information (to those who need to know) in order that action may be taken: in summary it provides
information for action. See Table 2G.2.1.

Table 2G.2.1 Principles of surveillance

Case definition Using clinical or microbiological criteria
Cases identified through variety of sources Clinicians or laboratories
Primary, secondary or tertiary care
Data sources Electronic database, e.g. CoSurv
Special forms, e.g. legionella questionnaire
Systematic collection of data for cases that satisfy the
case definition
Analysis of data and summary statistics
Feedback to data providers and distribution to those Disseminate information – newsletter/bulletin/articles/
who require it for action reports (e.g. CDR weekly/Health Protection Report)
Continuing surveillance for evaluation of interventions

237

Public [Link] 237 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

NATIONAL SURVEILLANCE
UK The Health Protection Agency’s (HPA) Centre for Infections (CfI) is responsible for coordinating communicable
disease surveillance across England and Wales. In Scotland, this role is performed by Health Protection Scotland
(HPS). Certain key infectious diseases are kept under constant surveillance in order to:
• Detect trends
• Evaluate prevention and control measures
• Alert appropriate professionals and organisations to infectious disease threats.
See Table 2G.2.2.

Table 2G.2.2 The main sources of information for national surveillance

Notifiable diseases By UK law, certain diseases are deemed ‘notifiable’, meaning that doctors have a
statutory duty to report them. These reports should be made on the basis of clinical
suspicion, rather than waiting for laboratory confirmation
The HPA produces a NOIDS Weekly Report (Notifications of Infectious Diseases) and
WARNER Reports (Weekly Analysis Report of Notifications above Expected Rates)
See Section 2G.8 for a list of the UK notifiable diseases
Reports/special Laboratory
forms/returns
Case report – rare disease
Incident report
Royal College of General Practitioners weekly returns (e.g. for influenza, whooping
cough, asthma)
Genitourinary/sexual health clinic disease statistics (called KC60 returns in England and
Wales)
Vaccine coverage: COVER (Cover Of Vaccination Evaluated Rapidly); ISD (Information
Services Data, Scotland) statistics
Death registration
Hospital episode statistics
NHS Direct/NHS 24
Surveillance systems:
• Sentinel surveillance, e.g. HPA/Nottingham University Primary Care Surveillance
Scheme (Q Flu, Q Research)
• Enhanced TB Surveillance (TB register in London)/Enhanced Surveillance of
Mycobacterium Infection (ESMI)
• Enhanced Meningococcal Surveillance
Surveys Survey of Prevalent HIV Infections Diagnosed (SOPHID)
Seroprevalence surveys
Unlinked anonymous surveys, e.g. Blood-Borne Viruses (BBV) in injecting drug users,
BBV in antenatal women

See Section 1C.3 for details of disease surveillance specific to the four UK countries.

238

Public [Link] 238 3/6/08 [Link]

 2G Com mun icable Diseases

Enhanced surveillance schemes use multiple data sources to gather additional detailed information on worrisome
diseases (e.g. meningococcal disease, TB). The additional information is used to:
• Advise policy-makers whether action is required to address an emergent threat such as pandemic influenza
• Detect new or imported infections, such as SARS.
Ire In Ireland, the Health Protection Surveillance Centre (HPSC) is a unit within the Population Health Directorate
of the Health Service Executive. The HPSC operates the Computerised Infectious Disease Reporting (CIDR) system to
manage the surveillance and control of infectious diseases in Ireland and to monitor antimicrobial resistance. Health
personnel electronically return notification, outbreak and enhanced surveillance forms. The data provide the basis
for the weekly infectious diseases and outbreak reports published by the HPSC.
An example of enhanced surveillance (for pandemic influenza) is shown in Box 2G.2.1.

Box 2G.2.1
Example: Enhanced surveillance for pandemic influenza
The WHO runs a network of over 100 National Influenza Centres that monitor influenza incidence and analyse
virus isolates. Any unusual activity and/or isolates are reported immediately through the WHO Global Influenza
Programme. The current early warning system has potential weaknesses in that some countries lack the ability to
test for H5N1, cases tend to occur in rural areas and other illnesses with similar symptoms are common.
The WHO plans to provide infrastructure to improve case detection, actively search for human cases where an
outbreak in animals is identified, support epidemiological investigation and promote collaboration between Asian
countries.

Interpandemic Phase 1 No new influenza virus subtypes have been detected in humans.
period An influenza virus subtype that has caused human infection
may be present in animals. If present in animals, the risk of
human infection or disease is considered to be low
Phase 2 No new influenza virus subtypes have been detected in humans.
However, a circulating animal influenza virus subtype poses a
substantial risk of human disease
Pandemic alert Phase 3 Human infection(s) with a new subtype but no human-to-
period human spread, or at most rare instances of spread to a close
contact
Phase 4 Small cluster(s) with limited human-to-human transmission but
spread is highly localised, suggesting that the virus is not well
adapted to humans
Phase 5 Larger cluster(s) but human-to-human spread still localised,
suggesting that the virus is becoming increasingly better
adapted to humans but may not yet be fully transmissible
(substantial pandemic risk)
Pandemic Phase 6 Pandemic: increased and sustained transmission in general
period population
Adapted from [Link]/csr/disease/influenza/pandemic/en/, [Link]/csr/resources/publications/influenza/
WHO_CDS_CSR_GIP_05_8-[Link], [Link]/flu/pandemic/[Link].

239

Public [Link] 239 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

INTERNATIONAL SURVEILLANCE
The value of global surveillance and of information sharing (such as emergence of resistance in key pathogens) lies
in:
• Guiding infection control policies
• Engaging and prompting dialogue with policy-makers
• Developing advocacy and educational programmes
• Stimulating research.
Global surveillance serves as an early warning system for epidemics and provides the rationale for public health
intervention. Early detection of communicable diseases and immediate public health intervention can curtail the
numbers of communicable illnesses and deaths, and reduce the negative effects on international travel and trade.
Global surveillance depends on strong national surveillance systems. The UK’s HPA houses the WHO reference
laboratories for various infections (e.g. influenza, TB, SARS) and it is the European Co-ordinating Centre for the
Global Programme on Drug Resistant Tuberculosis.
Box 2G.2.2 shows an example of a global programme.

Box 2G.2.2
Example: The WHO Information for Action – global programme for vaccines and immunisation
The programme monitors and assesses the impact of strategies and activities for reducing morbidity and mortality
of vaccine-preventable diseases. Its global goals by 2005 included:
• Polio eradication
• Measles mortality reduction
• Maternal and neonatal tetanus elimination (MNTE)
Reproduced from WHO, Immunization surveillance, assessment and monitoring, available online at: [Link]/
immunization_monitoring/en.

EVALUATING A SURVEILLANCE SYSTEM

An evaluation of a surveillance system should describe:
• The public health importance of the disease under surveillance
• Details of the surveillance system: its objectives, components and case definition (which should be clear and
consistent)
• Details of the analysis and reporting system
• Conclusions and recommendations to improve the system.
It should consider each of the attributes shown in Box 2G.2.3.

USEFULNESS OF SURvEILLANCE
One way to assess the impact of a surveillance system is to list the actions that have been taken as a result of its
findings. These might include:
• Early detection of outbreaks
• Monitoring of trends
• Early warning of changes of incidence
• Guidance to public health programmes (e.g. identification of high-risk groups to receive neonatal BCG to
prevent childhood TB)
• Collection of cases for further studies.

240

Public [Link] 240 3/6/08 [Link]

 2G Com mun icable Diseases

Box 2G.2.3
Qualitative Simplicity
Flexibility
Acceptability
Representativeness
Completeness
Quantitative Sensitivity
Positive predictive value
Timeliness
Resource use (direct costs)

2G.3 METHODS OF CONTROL

Communicable disease control consists of activities aimed at preventing the spread of infection. In a health-care
setting it is termed ‘infection control’ and aims to avoid cross-contamination between patients, from health-care
workers to patients, and from patients to health-care workers. See Table 2G.3.1.

Table 2G.3.1 Methods of communicable disease control

Universal precautions Hand washing (e.g. hand hygiene campaigns)
Infection control standard, contact, droplet and airborne precautions, e.g. facemasks,
personal protective equipment
Isolation Single room isolation, e.g. MRSA
Negative pressure room for source isolation, e.g. TB
Positive pressure room for protective isolation, e.g. for immunocompromised patients
such as those receiving a bone marrow transplant
Decontamination Decontamination of persons
Disinfection of equipment and the environment
Quarantine Quarantine of contacts, e.g. plague
Immunisation Vaccine prophylaxis of exposed individuals
Chemoprophylaxis Antibiotics offered to people who have come into contact with an infectious disease
(e.g. meningococcal meningitis or anthrax)
Source removal Biociding the water in a cooling tower to control an outbreak of legionnaires’ disease
Product recall
Closure of a restaurant

2G.4 DESIGN, EvALUATION AND MANAGEMENT OF IMMUNISATION PROGRAMMES

Note that the terms vaccination and immunisation, while used interchangeably in common parlance, are in fact
slightly different: see Box 2G.4.1.

241

Public [Link] 241 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

Box 2G.4.1
vaccination Administration of a vaccine
Immunisation Administration of a vaccine plus
the development of an immune
response by the body

DEVELOPMENT OF A VACCINATION PROGRAMME

In developing a new vaccination programme, the following areas of work must be considered:
• Scientific evidence
• Programme strategy
• Administration
• Finance
• Vaccine purchase and distribution
• Communication
• Informatics.
UK In the UK, the Department of Health (DH) has overall responsibility for immunisation policy. It is supported
by the HPA, which undertakes vaccine research, epidemiology and surveillance. The Joint Committee on Vaccination
and Immunisation (JCVI) is an independent expert advisory committee that advises the government on matters
relating to communicable diseases. This committee receives papers and hears presentations and must make its
recommendations in light of a cost−benefit analysis.
Ire In Ireland, the Health Service Executive (HSE) National Immunisation Office website has up-to-date
information for parents and professionals on the childhood immunisation schedule and on other topics such as
immunisation for travel ([Link]/en).
The DH and the HSE take guidance from the Immunisation Advisory Committee of the Royal College of Physicians of
Ireland. That committee has representatives from the Department, the HSE, the RCPI, the DH and Social Services,
Northern Ireland and other relevant organisations. The Committee’s guidance on immunisation is available on the
Health Protection Surveillance Centre website ([Link]/hpsc).

VACCINATION STRATEGIES
WHO recommendations are shown in Table 2G.4.1.

Table 2G.4.1 WHO recommendations for immunisations as part of national strategies:

All countries Diphtheria
Hepatitis B
Measles
Pertussis
Poliomyelitis
Tetanus
High-risk countries Hib vaccine
BCG vaccine
Yellow fever

242

Public [Link] 242 3/6/08 [Link]

 2G Com mun icable Diseases

Immunisation schedules vary by country, depending on local epidemiology, funding, a consideration of the risk and
efficacy of vaccines at different ages. The annual WHO/UNICEF joint reporting form lists the vaccination schedules of
all countries ([Link]/immunization_monitoring/data/schedule_data.xls).
In addition, targeted vaccination strategies may be implemented in certain circumstances: see Box 2G.4.2.

Box 2G.4.2
Circumstances Example
Travel Hepatitis A
Occupational Hepatitis B
Outbreak Meningitis C
Mass infection Eradicate Smallpox
Eliminate Polio
Contain Deliberate release of smallpox

IMPLEMENTATION OF A VACCINATION POLICY

Funds are normally secured centrally for implementation of the vaccine programme locally. Communication of the
new policy is achieved through:
• Letters to all registered doctors, nurses and pharmacists from the Chief Medical Officer/Chief Nursing Officer/
Chief Pharmacist
• Website publicity
• ‘Green Book’ updates
• DH manages a network of ‘immunisation coordinators’
• Public promotion (see below).
After resources have been secured, vaccine manufacturers are invited to submit bids through a competitive
procedure. Criteria for successful bidding are safety, efficacy, availability, price and record of the company against
previous contracts. Wherever possible, more than one supplier is chosen to minimise the likelihood of vaccine
shortages.

PUBLIC RELATIONS
A new promotion programme is developed for the public, which may include the following: leaflets; fact sheets;
press, television and radio advertisements; videos; and internet materials (including ‘Question & Answer’ formats,
frequently asked questions (FAQs) and mail-in facilities for internet questions).
All new materials are pre-tested with the appropriate target audience and amended in light of consumers’ comments,
and their impact is monitored.

OTHER IMPLEMENTATION ACTIvITIES

• New surveillance arrangements to measure the impact of a new policy through laboratory-based data or disease
notification data
• Sero-epidemiological surveillance for population impacts
• Market testing is undertaken to evaluate the impacts of any advertising accompanying the introduction of new
vaccines
• New data collection arrangements are set up in advance so that the vaccine coverage of a new initiative can be
monitored

243

Public [Link] 243 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

• In the case of recent new vaccine policies, expert groups have been set up in advance to monitor and
investigate reports of serious adverse events
• Special studies are also prepared that link hospital or primary care records of clinical events with immunisation
data so that risks of adverse outcomes can be assessed.

LOCAL IMPLEMENTATION
• Will need local implementation group, usually lead by immunisation coordinators but may include consultant in
communicable disease control (CCDC), pharmacists, health visitors, child health, community paediatricians, PCT
primary care leads
• A local training programme for health professionals.

2G.5 CHOICES IN DEvELOPING AN IMMUNISATION POLICY (SALISBURY 2005)

Issues for policy-makers to consider when developing immunisation policy include:
• Outbreak response (including whether to create a vaccine stockpile)
• Surveillance
• Containment
• Investment in future research.
• Choice of selective versus universal or mass vaccination.
Examples of policy decisions are given in Table 2G.5.1.

Table 2G.5.1 Examples of recent UK policy development decisions

Pertussis The change from whole-cell pertussis vaccine to acellular pertussis vaccine was not made
until 2004 − later than in other industrialised countries. This was a deliberate decision since
protection from the whole-cell vaccine was excellent, and reactions when the vaccine was
given at 2, 3 and 4 months of age are fewer than when the vaccine is given later (e.g. as in
other countries). In 2004, a ‘five-component’ acellular pertussis vaccine became available with
efficacy shown to at least equal to that of the previously used whole-cell vaccine
varicella Varicella vaccine is not currently recommended for the routine UK childhood programme,
although it is a routine vaccination in the USA. This policy will not be changed until there is
confirmation that its introduction in childhood will not increase the burden of varicella-zoster
in older age groups
Pneumococcus Pneumococcal conjugate vaccine (PCV), which is in routine use in certain countries such as the
USA, has been introduced into the UK childhood programme

2G.6 OUTBREAK INvESTIGATIONS

Outline the steps in outbreak investigation, including the use of relevant epidemiological methods
An outbreak is a localised epidemic (the latter being the occurrence of more cases of disease than expected in a
given area, or among a specific group of people, over a particular period of time).
The objectives in controlling an outbreak will be to:
• Minimise the number of primary cases of illness through prompt recognition of the outbreak, and through the
identification and control of the source of the infection or contamination
• Minimise the number of secondary cases of infection by identifying cases and taking appropriate action to
prevent any spread
• Prevent further episodes of illness by identifying continuing hazards and eliminating them or minimising the
risk that they pose

244

Public [Link] 244 3/6/08 [Link]

 2G Com mun icable Diseases

• Introduce measures to prevent future outbreaks.

OUTBREAK CONTROL PLANS

Outbreak control plans should include:
• Description of the roles and responsibilities of each of the organisations and individuals
• Arrangements for informing and consulting the key personnel (e.g. directors of public health, the regional
epidemiologist, relevant reference laboratories, senior managers from the health service and health protection
agency, and the DH)
• Arrangement for liaison with local government, hospitals and health authorities
• Facilities required to manage an incident (e.g. an incident room equipped with telephones, fax machines and
other efficient electronic communication systems) – including arrangements for outside normal working hours.

OUTBREAK CONTROL GROUP

An outbreak control group should generally be convened when an outbreak occurs and any of the following apply:
• The disease poses an immediate health hazard to the local population
• There are a large number of cases
• Unexpected cases appear in several districts
• The disease is unusual and severe.

INVESTIGATION OF AN OUTBREAK
The management of an outbreak consists of five tasks that should be conducted concurrently – one of which is an
epidemiological sequence that should be conducted serially (Figure 2G.6.1).

2G.7 IMPORTANT INFECTIOUS DISEASES

Knowledge of natural history, clinical presentation, methods of diagnosis and control of infections of local and
international public health importance (including emerging diseases and those with consequences for effective control)
Tables 2G.7.1–2G.7.5 group diseases as follows:
• Food-borne diseases
• Viral hepatitides
• Vaccine-preventable diseases
• Nosocomial infections
• Sexually transmitted infections.
UK Note that the Health Protection Agency identifies four groups of people as posing an increased risk of
spreading gastrointestinal infection: see Box 2G.7.1.

Box 2G.7.1
Group A Any person of doubtful personal hygiene or with unsatisfactory toilet, hand-washing or hand-drying
facilities at home, work or school
Group B Children who attend pre-school groups or nurseries
Group C People whose work involves preparing or serving unwrapped foods not subjected to further heating
Group D Clinical and social care staff who have direct contact with highly susceptible patients or persons in
whom a gastrointestinal infection would have particularly serious consequences
Reproduced from PHLS Advisory Committee on Gastrointestinal Infections (2004).

245

Public [Link] 245 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

Control measures Five concurrent tasks Epidemiological sequence

• Control the 1. Epidemiological sequence 1. Establish case definition
source (animal, 2. Confirm that cases are real
human or
3. Determine background rate of disease
environmental) 2. Control measures
4. Assess whether this is an outbreak
• Control the
mode of spread 5. Find cases
3. Collect environmental samples
6. Describe epidemiological
• Protect
characteristics of cases – time,
persons at risk
4. Convene outbreak control group place, person, clinical characteristics,
• Continue laboratory information
surveillance of 7. Plot the epidemic curve
control measures 5. Communication
8. Generate a hypothesis
• Declare the
9. Test the hypothesis by means
outbreak over
of an analytical study
once the
number of new 10. Consider further studies
cases has Communication 11. Generate conclusions
returned to
background levels • Consider the best media of
communication with
• Introduce
colleagues, patients and the
measures to
public
prevent future Outbreak control group
outbreaks • Ensure accuracy and
(e.g. prosecute, timeliness The membership will vary but may include:
legislate) • Include all those who need 1. Consultant in Communicable
to know Disease Control
• Use the media constructively 2. Environmental Health Officer
• Prepare written report for 3. Consultant Microbiologist or Virologist
local, regional and national 4. Administrative and secretarial support
authorities
5. Food chemist
6. Member of State Veterinary Service
7. Toxicologist
8. Food microbiologist
9. Regional epidemiologist
10. Director of Public Health
11. Press officer
12. Representative from the HPA CFI
13. Manager of the institution
14. Food Standards Authority
15. Water company
16. Occupational health physician

Figure 2G.6.1 Investigation of an outbreak

UK Note also that references to notifiable disease status in Tables 2G.7.1–2G.7.5 pertain to England and Wales
(see Section 2G.8).

246

Public [Link] 246 3/6/08 [Link]

 Table 2G.7.1 Food-borne gastrointestinal diseases
Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control

Public [Link] 247

Campylobacter Ranges from Stool culture Gastrointestinal Animal to person Preventive measures:
Asymptomatic to (high sensitivity tract of birds (water or food
Chlorination of drinking water supplies
severe diarrhoea with same-day (particularly contaminated with
(~50% cases bloody sample) poultry) and faeces) Milk pasteurisation
stools) animals, cattle
Microscopy Person to person Adequate hygiene, domestic and
and domestic pets
(sensitivity lower (direct contact with commercial
than culture) faeces of index case,
Kitchens
e.g. person changing
soiled nappies) Adequate cooking of poultry
Raw or undercooked Hand hygiene
meat (especially
Advice to travellers abroad
poultry), non-
pasteurised milk Control of patient, contacts and
immediate environment
Exclusion of symptomatic cases

247
Food hygiene and hand hygiene
Avoid non-pasteurised milk and
untreated water
Table contd overleaf
2G
Com mun icable Diseases

3/6/08 [Link]
 Table 2G.7.1 contd
2G

Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control

Public [Link] 248

Cholera Watery diarrhoea Determine if Untreated/ Consumption of Notifiable disease
toxin producing polluted water untreated water,
(toxin-producing Vomiting Secondary spread is rare if hygiene is
contaminated
Vibrio cholerae 01) Stool culture good
50% case fatality shellfish and foods
− severe untreated Direct eaten raw or washed Cases should be excluded until 48 h after
cases microscopy of in contaminated first normal stool
C o m mu n ic abl e D iseases

stools water. Person-to-

Fatalities rare in the UK: ensure oral
person spread (by
PCR rehydration and appropriate antibiotics
faecal−oral route) is
likely to be a threat Food advice to travellers
only when hygiene
Prevention:
is very poor and
sanitary facilities are Predominantly education, and adequate
inadequate hygiene and sanitation, especially for
travellers
Cholera rarely occurs
in the UK – most Advice for travellers to countries with

248
cases are imported; epidemic cholera: A simple rule of thumb
therefore, ask about is, ‘Boil it, cook it, peel it or forget it’
travel history in
Vaccination: killed whole-cell vaccine
week before onset
leads to poor short-lasting cover and is
of little value. Not available in UK and is
no longer a requirement for travel to any
country
Control:
Safe drinking water supplies

3/6/08 [Link]
 Cryptosporidiosis Healthy individuals – Stool microscopy Gastrointestinal Person to person Prevention:
(Cryptosporidium self-limiting tract of humans

Public [Link] 249

Intestinal biopsy Animal to person Hand hygiene
parvum) and animals
Immunocompromised
Serology (particularly Swimming pool Adequate water treatment
– severe illness may
farm and outbreaks
lead to death Genotyping Monitoring water quality
domesticated);
techniques
Diarrhoea – may be water Immunocompromised − avoid contact
bloody contaminated with farm animals, drink boiled water,
with faeces avoid contact with infected cases
Disinfectants, e.g. hydrogen peroxide
Control:
Exclusion of cases until 48 h after first
normal stool
Cases should avoid swimming for 2 weeks
Contact tracing
History of raw water consumption

249
Good practice guideline – nursery/farm/
swimming pool/hospitals
Table contd overleaf
2G
Com mun icable Diseases

3/6/08 [Link]
 Table 2G.7.1 contd
2G

Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control

Public [Link] 250

Shigellae S. sonnei – mild Isolation of Humans Person to person Prevention:
diarrhoea organism from
(Shigella sonnei – Contaminated food/ Hand hygiene
stools
common in the UK Others: water
Adequate cleaning of toilet area
and mild) Serotyping
Watery diarrhoea, Faecal−oral route
Treatment of drinking and swimming
(S dysentriae, S. vomiting Phage typing
C o m mu n ic abl e D iseases

water
flexneri, S. boydii
50% bloody stools
– imported and Advice to travellers
severe) S. dysenteriae 1 –
Control:
toxic megacolon/
(S. dysenteriae
haemolytic – uraemic S. sonnei: exclusion of case for 48 h
type 1– produce
disease and death after first normal stool
exotoxin – severe
illness) Other Shigella species: exclusion of case
for 48 h after first normal stool (unless
the case is in a risk group – when
clearance is needed, i.e. two negative

250
stools taken at least 48 h apart; contacts
in risk groups also need microbiological
clearance)
Contact tracing
Reinforce hygiene measures

3/6/08 [Link]
 E. coli E. coli O157 – Food, Gastrointestinal Contaminated food/ Prevention:
(vero cytotoxin asymptomatic/ environmental, tract of water

Public [Link] 251

Hand hygiene
producing) diarrhoeal illness/ animal samples cattle (and
Animal to person
haemolytic–uraemic possibly other Adequate cleaning (kitchen, toilet)
Most common Stool culture
syndrome (HUS domesticated Person to person
serotype in UK is E. Precautions during farm visits
− particularly in Biochemical animals)
coli O157.H7
children)/death and serological Well-cooked beef, lamb, venison products
testing –
Good practice − food processing and food
isolates
service industries
Reference labs
Control:
for VTEC
Hygiene advice cases/contacts/food
service industry
Cases not in risk groups are excluded
for 48 h after first normal stool.
Cases in risk groups are excluded
until microbiological clearance − two
consecutive stool samples 2 days apart

251
Household contacts in risk groups are
screened microbiologically
Salmonellae Diarrhoea Stool culture or Gastrointestinal Animal to person – Prevention:
rectal swab tracts of wild and contaminated food
(S. enteritidis PT4 Rare complication Vaccination of poultry flock
domestic animals,
– associated with is abscess formation Blood culture Person to person –
birds (especially Food processing industry systems to
eggs and poultry) and septicaemia faecal−oral
Reference poultry), ‘exotic’ identify, control and monitor potential
(S. typhimurium laboratory: pets (terrapins hazards
DT104 – increased and iguana) and
Serotyping Personal/food hygiene measures (home,
antibiotic occasionally
institutions)
2G

resistance) Phage typing humans

Control:
Hand/food hygiene advice
Exclude cases for 48 h after first normal
stool
Table contd overleaf
Com mun icable Diseases

3/6/08 [Link]
 Table 2G.7.1 contd
2G

Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control

Public [Link] 252

Paratyphoid fever Gastroenteritis Blood, urine, Humans Food borne Prevention:
faeces, bone
(Salmonella Early disease may Person to person Sanitation
marrow aspirate
paratyphi A (80% involve constipation in poor hygiene
culture Clean water
in UK), B (20% in – later – diarrhoea, conditions
UK) and C) vomiting Personal hygiene
C o m mu n ic abl e D iseases

Spots No effective vaccine

5% relapse Control:
Notifiable
Isolation in hospital is advisable
Screen all household contacts
Hygiene advice
Exclude food handlers (2 weeks after
antibiotic therapy, require six clear

252
consecutive samples – 2 weeks apart).
Other risk groups: exclude until three
clear consecutive samples taken 2 weeks
apart
Exclude cases not in risk groups until
well and have normal stools

3/6/08 [Link]
 Typhoid Similar to Blood, urine, Humans Food borne Prevention
(Salmonella typhi) paratyphoid – but faeces, bone

Public [Link] 253

Person to person Sanitation
more severe marrow aspirate
with poor hygiene
culture Clean water
Rose spots
(undetectable in Phage typing Personal hygiene
pigmented skin) − unexplained
Vaccination for travellers to endemic
clusters
Spleen may enlarge countries
Abdominal Control:
haemorrhage
Notifiable
(ulceration of Peyer’s
patches) Exclude food handlers (2 weeks after
antibiotic therapy require six clear
Intestinal perforation
consecutive samples – 2 weeks apart)
Renal failure
Exclude food handlers (2 weeks after
antibiotic therapy, require six clear
consecutive samples – 2 weeks apart).
Other risk groups: exclude until three

253
clear consecutive samples taken 2 weeks
apart. Exclude cases not in risk groups
until well and have normal stools
Contact tracing of household/close
contacts
2G
Com mun icable Diseases

3/6/08 [Link]
 Table 2G.7.2 Viral hepatitides
2G

Public [Link] 254

Disease Clinical features Diagnosis Reservoir Transmission Response/control
(organism)
Hepatitis A Ranges from Salivary IgM and IgG Human Person to person Notifiable disease
(HAv) asymptomatic to – useful in outbreak gastrointestinal spread – via faecal
Use of immunoglobulin (Ig) or
fulminant hepatitis investigations tract – oral route and
vaccination for close, household and
contaminated food
C o m mu n ic abl e D iseases

Young children Detection of IgM (serum/ sexual contacts should be offered

are commonly saliva) = acute infection Consumption of food via GP
asymptomatic. Adults grown or washed in
IgG persists for life Risk groups and all cases should be
are more likely to contaminated water
excluded for 7 days after onset of
have symptoms Persistent IgG – past (e.g. shellfish, fruit
jaundice and/or symptoms
– 75% of adults infection/immunisation and vegetables)
develop jaundice Travel advice for those to areas of
endemicity
Infectious from
2 weeks before
jaundice develops
Hepatitis B Non-specific See Table 2G.7.6 (HBV Humans Person to person by Notifiable

254
(HBv) prodromal illness marker interpretation) blood-borne routes.
Vaccination of high-risk group −
Intravenous drug
Jaundice (often after all health-care workers should be
users, sex, bites,
fever) immunised against hepatitis B
scratches
infection and should be shown to
Can lead to long-
Perinatal most have made a serological response to
term carriage
common in high the vaccine
(carriage rate in UK
prevalence countries
population ~0.5%) Boosters in poor or non-responders
Cirrhosis UK schedule 0, 1 and 6 months
Hepatocellular Accelerated schedule 0, 1, 2 and 12
carcinoma months
Co-infection with If susceptible, vaccinate household
hepatitis subviral and sexual contacts
satellite D
Screening of blood supply in UK
Antenatal screening and
immunisation of babies at risk

3/6/08 [Link]
 Hepatitis C Asymptomatic Enzyme immunoassay – Humans Historical: blood Hepatitis C National Register was
(HCv) detect HCAb products until established in 1998

Public [Link] 255

Mild infection
screened since 1991
Recombinant immunoblot Aim of the register is to inform the
Jaundice – unusual
assay (RIBA) confirms Now intravenous drug natural history of HCV infection in
80% carriers HCAb users: 80% the UK
80% chronic HCAb +ve – previous Sharing razors or Majority of these cases are
hepatitis exposure toothbrushes transfusion recipients who were
traced during the national HCV look-
10−20% cirrhosis PCR − detect infection Body piercing
back programme
(HCV RNA) (like tattooing or
1% liver cancer
acupuncture) DH released hepatitis C strategy
for England (2002) and hepatitis C:
Vertical transmission
Action plan for England (2004) to
rare
implement strategy
Risk is highest in HIV
No vaccine
mothers and those
who have high viral Interferon and ribovirin treatment
load for chronic infection with HCV
Insufficient evidence

255
to assess the risk
of transmission via
breast milk
Hepatitis D Exists only in Serology Humans As for HBV As for HBV
(HDv) conjunction with
HBV. Known also as
the ‘delta agent’
Increases risk of
cirrhosis
2G

Hepatitis E Illness similar to Serology Faecal−oral Provision of safe water supplies

(HEv) HAV without chronic
Specific IgM testing Person to person − Pregnant and older people, those
sequelae or carriage
low spread with weakened immune systems and
Most cases − young/ people with chronic liver disease
middle-aged adults might need closer observation for
deterioration in liver function
High case fatality
in third trimester of No vaccination
Com mun icable Diseases

pregnancy

3/6/08 [Link]
 Table 2G.7.3 Vaccine-preventable diseases
2G

Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control

Public [Link] 256

Diphtheria Pharyngitis, enlarged Nasal, throat, Reservoir of Animal to Statutory notifications
lymph nodes and skin ulcer C. ulcerans person
(diphtheria Rare in England and Wales following the
‘bull-neck’ appearance swabs – identify is cattle
toxin produced No direct introduction of mass immunisation in 1942
may cause respiratory C. diphtheriae
by toxigenic evidence of
obstruction, paralysis (toxigenic) Travel, and close contact with cattle or other farm
Corynebacterium person to
C o m mu n ic abl e D iseases

and cardiac failure – animals, are potential risk factors for infection
diphtheriae or by Occasionally person – but
fatal if untreated
C. ulcerans) toxigenic C. possible Contact in the previous 7 days with a case of
Many cases ulcerans infection caused by toxigenic C. diphtheriae or C.
vaccinated, so rarely ulcerans should be considered at risk
Confirmation
recognised on clinical
of toxigenicity Cases barrier nursed, antibiotic treatment,
grounds
from reference antitoxin, booster or primary vaccination
laboratory − can
Contacts – swabbed, food handlers/unvaccinated
be obtained within
children excluded, antibiotic, booster/primary
few hours by PCR
vaccination
Primary vaccine coverage (three doses) for

256
children aged 2 has been 94% since 2001, just
below the WHO target of 95%
No public health action reqiured if non-toxigenic
C. diphtheriae

3/6/08 [Link]
 Pertussis Cough, cold, fever Culture nasal Human Droplet spread Prevention:
(whooping cough) progressing to swab – but low

Public [Link] 257

Acellular pertussis vaccine is given in the primary
paraxysmal cough and sensitivity and
(Bordetella course with diphtheria, tetanus, polio and Hib,
bouts of coughing high specificity
pertussis) as DTaP/IPV/Hib, at 2, 3 and 4 months of age. A
ending with a whoop
PCR further booster dose with acellular pertussis, given
or vomiting
as dTaP/IPV, is given with the preschool boosters
EIA
Last 2−3 months between the ages of 3 and 5. See: [Link].
uk/infections/topics_az/vaccination/vac_sced.htm
<6 months at risk
Control:
Adults – milder
symptoms – Cases treated with antibiotic, isolated, vaccinated
recognised as a cause (if not vaccinated). If a vulnerable contact is
for chronic cough present in a household, all offered antibiotic
prophylaxis and vaccinate those under 10 and
unimmunised. If no vulnerable contacts, no
prophylaxis required.
Outbreaks – community-wide vaccination if
coverage low/case finding/antibiotic treatment

257
Tetanus Painful muscular Infrequently Animals, Dirty wounds Prevention:
contractions − obtained humans
(Clostridium Abdominal Vaccine − in UK schedule = three doses at 2, 3
especially of neck and
tetani) surgery and 4 months, boosters at 3−5 years and 13−18
jaw
years
Often history of
Case – vaccinated primary or booster − if 10 years
tetanus-prone wound
or more since last vaccine or if acquired tetanus-
prone wound
Control of outbreak:
Look for source, e.g. surgery, intravenous drug
2G

users
Table contd overleaf
Com mun icable Diseases

3/6/08 [Link]
 Table 2G.7.3 contd
2G

Public [Link] 258

Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control
Haemophilus Invasive disease Blood/CSF culture Humans Droplet Prevention:
influenzae type b spread/direct
Commonest PCR Vaccine in UK given at 2, 3 and 4 months with
(Hib) contact
presentation is diphtheria/tetanus/pertusis and polio vaccines
Reference
meningitis; other Unvaccinated
laboratory for Vaccination prevents carriage
− pneumonia, – higher
C o m mu n ic abl e D iseases

confirmation and
epiglottitis, bone and carriage rate Control:
typing
joint infection, facial – common in
Notifiable
cellulitis young children
Unvaccinated household children contacts
vaccinated and adults given chemoprophylaxis,
including case
Vaccination programme in cluster if coverage low
Meningococcal Non-specific early Blood/CSF culture Humans Direct or Prevention:
disease (Neisseria phase indirect person
PCR Vaccines against serogroups A, C, W135 and Y –
meningitidis) to person
Babies − floppy, fever, short-lived vaccine

258
Reference spread
vomiting
laboratory for Men C vaccine given in the UK at 2, 3 and 4
Photophobia, neck confirmation and months (children >1 year single dose)
stiffness typing
Control:
Petechial rash,
Chemoprophylaxis – close contacts
septicaemia, death
Vaccinate if vaccine-preventable strain
Complications:
deafness/convulsions/
limb amputation/
mental impairment

3/6/08 [Link]
 Mycobacterium Long incubation Chest radiograph Animals, Direct spread Prevention:
tuberculosis (MTB) period produces humans from infected
Sputum smear BCG vaccine: UK government now recommends
chronic disease with case

Public [Link] 259

(occasionally Sputum culture that the following risk groups be offered BCG
risk of reactivation
M. bovis and M. Bovine TB vaccination:
(particularly with Microscopy –
africanum) from ingesting • All infants living in areas where the incidence
age) and fatal sputum raw milk from of TB is 40 per 100 000 or greater
without treatment
Sensitivity testing infected cows • Infants whose parents or grandparents were
Symptoms can for multi-drug- born in a country with a TB incidence of 40
include: resistant TB per 100 000 or higher
Cough (MDRTB) • Previously unvaccinated new immigrants from
high prevalence countries for TB
Blood in sputum Molecular typing
for identifying Control:
Weight loss
clusters Notifiable
Night sweats
UK − enhanced TB surveillance since 1999
More common in
immigrant ethnic Cases are followed up by chest clinics to ensure
groups that adequate treatment is given and that
contacts are identified, screened and given
Mortality decreased prophylaxis where necessary

259
rapidly after
introduction Measures to maximise compliance such as directly
of effective observed therapy (DOTS)
chemotherapy Port health
Mumps Tenderness and Saliva, CSF, urine Humans Direct contact Prevention:
(paramyxovirus) parotid swelling culture with saliva
Vaccine in the UK given at 12−15 months and 3−5
or droplets
Meningitis Serology years (in combination with measles and rubella)
of saliva of
(commonest cause of
an infected Control:
viral meningitis pre-
person Notifiable
vaccine era)
2G

Orchitis As vaccination rates drop, resurgence and possible

outbreaks of these diseases are increasingly likely
Pancreatitis
Exclusion
Check vaccination status for case
Consider community-wide programme if coverage
low in outbreaks
Com mun icable Diseases

Table contd overleaf

3/6/08 [Link]
 Table 2G.7.3 contd
2G

Public [Link] 260

Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control
Measles Prodromal flu-like Salivary test kit for Human Direct contact See mumps
(paramyxovirus) symptoms measles IgM
Person to Human normal immunoglobulins (HNIG) for:
Koplik’s spots inside Serology person
Pregnant contacts who are measles IgG negative
the mouth
Babies less than 6 months old – maternal IgG
C o m mu n ic abl e D iseases

Rash starts on days

negative
3−4, over face, trunk
and limbs Exclusion
Not itchy
Complications
include:
• Pneumonitis
• Acute otitis
media
• Pneumonia
• Encephalitis

260
Rubella Moderately infectious Serum or saliva Human Direct person See mumps
detection of IgM to person
(rubella virus Pre-vaccination era Pregnant women in contact with case – tested; if
– a member of Viral culture from susceptible offer vaccination post partum
Affecting primary
Togaviridae) urine or serum
school-aged children,
susceptible pregnant
women – congenital
rubella syndrome
Pharyngitis
Conjunctivitis
Fever
Rash

3/6/08 [Link]
 Table 2G.7.4 Nosocomial infections

Public [Link] 261

Nosocomial (health care acquired) infections
Health care-associated infections (HCAIs) are infections acquired as a result of health care. Examples include meticillin-resistant Staphylococcus
aureus (MRSA), Clostrdium difficile, glycopeptide-resistant enterococci (GRE) and Acinetobacter species. Higher rates of HCAIs are often found in
specialist hospitals such as orthopaedic centres. However, with the move towards performing invasive procedures such as minor surgery in the
community, primary care HCAIs may become more common in future
HCAI risk factors
• More susceptible patients being treated (elderly patients or patients with severe or chronic diseases)
• Invasive procedures (e.g. indwelling lines into veins, access for dialysis or artificial ventilation)
• Immunosuppression (e.g. chemotherapy and anti-rejection drugs used in transplant surgery)
• Increased patient movement between wards or hospitals due to pressures on hospital beds
• Wider use of antibiotics and emergence of antibiotic-resistant microorganisms
Impact of HCAI

Effects on patients Effects on the health service

Severe or chronic illness Extended lengths of stay

261
Pain, anxiety, depression Costs of diagnosis and treatment of the infections and their complications
Longer stay in hospital Costs of specific infection-control measures – cleaning, disinfection, cohort nursing, etc.
Reduced quality of life Bed and ward closures and postponed admissions
Loss of earnings Provision of isolation facility/rooms
Death Antibiotic costs may be further increased if the infection is also due to a resistant microorganism
Control of HCAIs
• Hand washing is the most important prevention activity. In the UK the ‘Clean your hands’ campaign raised awareness among staff, patients
and the public
2G

• Prudent antibiotic prescribing

• Surveillance – in the UK surveillance systems for MRSA, surgical site infection
• Isolation, cohort nursing
• Local and national policies
Table contd overleaf
Com mun icable Diseases

3/6/08 [Link]
 Table 2G.7.4 contd
2G

Public [Link] 262

Disease Clinical features Diagnosis Reservoir Transmission Response/control
(organism)
MRSA Range from skin Gram stain, Humans, Direct contact Prevention:
infection and culture and rarely
(meticillin- Good personal hygiene − hand washing
conjunctivitis, to sensitivity testing animals
resistant
pneumonia and on appropriate Compliance with control measures
Staphylococcus
C o m mu n ic abl e D iseases

life-threatening specimen
aureus) Hand washing/aseptic techniques/handling waste/
septicaemia
16 strains waste disposal/ward or equipment cleaning
30% of the general
EMRSA-15 and Control:
population are
EMRSA-16 are the
colonised by S. Mandatory surveillance schemes in the UK:
dominant UK stains
aureus. In hospitals • Mandatory S. aureus bacteraemia surveillance
the percentage is • Mandatory MRSA bacteraemia-enhanced
higher because of surveillance scheme
more likely contact
Infection control policies – central and local
with infected cases
Central government initiatives (DH) – Towards

262
Cleaner Hospitals and Lower Rates of Infection,
Saving Lives Delivery Programme to reduce HCAIs,
including MRSA
Antibiotic prescribing policy

3/6/08 [Link]
 Table 2G.7.5 Sexually transmitted infections (STIs)
Prevention involves:

Public [Link] 263

• Health and sex education
• Early detection and prompt effective treatment
• Contact tracing and treating contacts
• Opportunistic or routine screening
• Surveillance
• In the UK KC60 returns from genitourinary medicine (GUM) clinics
Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control
Chlamydia Commonest bacterial STI in PCR or culture of Human Sexually Prevention:
trachomatis the UK urethral, cervical
National Chlamydia Screening Programme
or urine samples
Highest rates in young people commenced in 2002 in England
(especially under 24 years)
Antibiotic treatment
Majority asymptomatic in men
Contact tracing and treating partners
Untreated may lead to pelvic
Surveillance in England and Wales KC60
inflammatory disease (PID),
returns – GUM clinics

263
ectopic pregnancy and
ophthalmia neonatorum
Gonorrhoea Neisseria gonorrhoeae is the Microscopy Human Sexually Prevention:
(Neisseria second most common bacterial or culture of
Antibiotic treatment – but many strains
gonorrhoeae) STI in the UK urethral, cervical
resistant to commonly used antibiotics
swabs
Men more likely to have
Contact tracing and treating partners
symptoms than women
Outbreak control
Complications:
• PID, ectopic pregnancies Surveillance (KC60)
• Septic arthritis
2G

GRASP (gonococcal resistance to

• Ophthalmia neonatorum
antimicrobial surveillance programme)
Table contd overleaf
Com mun icable Diseases

3/6/08 [Link]
 Table 2G.7.5 contd
2G

Public [Link] 264

Disease (organism) Clinical features Diagnosis Reservoir Transmission Response/control
Syphilis Primary ulcer: third of cases Microscopy in Human Sexually Prevention:
develop secondary eruption. early syphilis
(Treponema Mother to baby In the UK Enhanced Surveillance − mainly
Late lesions of skin, bone,
pallidum) Serological due to outbreaks seen in Manchester,
central nervous system, heart Blood transfusion
tests* – London, Bristol and Brighton among gay
treponemal (e.g. men and heterosexual men and women
C o m mu n ic abl e D iseases

TPHA) and non-

KC60 return
treponemal (e.g.
VDRL) test Routine antenatal screening
Antibiotic treatment
Contact tracing
Outbreak control
Late syphilis – test for HIV also
Human HIV continues to be one of the HIV antibody Human Person to person Prevention:
immunodeficiency most important communicable test

264
Sexually Surveillance
virus (HIv) diseases in the UK. It is an
P24 antigen −
infection associated with Blood transfusion Routine antenatal screening
for early tests
serious morbidity, high
and screening Sharing needles Education
costs of treatment and care,
blood
significant mortality and high Vertical Contact tracing
number of potential years of Viral load/CD4 transmission
Antiretroviral treatment (no vaccine)
life lost count
Post-exposure prophylaxis
*TPHA, Treponema pallidum haemagglutination assay; VDRL Veneral Disease Reference Laboratory.

3/6/08 [Link]
 2G Com mun icable Diseases

HEPATITIS B SEROLOGY
Laboratory reports for hepatitis B contain details of a number of markers, the interpretation of which is outlined in
Table 2G.7.6.

Table 2G.7.6 Interpretation of hepatitis B markers

Laboratory abbreviation Serological marker Description Implication
HBsAg Hepatitis B surface Serological marker on surface Person is infectious. Presence
antigen of HBV that is present in for >6 months implies chronic
serum during acute or chronic carrier status
infection
Anti-HBs Hepatitis B surface Antibody to surface antigen Person is immune (either
antibody that is usually produced as due to recovery from
part of the normal immune prior infection or due to
response vaccination)
Total anti-HBc Total hepatitis B core Antibodies (of all classes) to Previous or ongoing infection
antibody a component of HBV
IgM anti-HBc IgM hepatitis B core IgM class of antibody Acute or recent infection
antibody that persists for 6 months
following exposure
HBeAg Hepatitis B e-antigen Marker present soon after High infectivity
exposure, then absent within
3 months
Anti-HBeAb Hepatitis B e-antibody Develops after HBeAg Low infectivity

NEW AND EMERGING INFECTIONS

Emerging infectious diseases are commonly defined as those that have newly appeared in a population or have
existed but are rapidly increasing in incidence or geographical range.

DRIvING FORCES BEHIND GLOBAL EMERGING INFECTIONS:

The pattern of communicable disease occurrence is in constant flux. Current influences include:
• Global travel
• Climate change
• Global trade and importation
• Urbanisation
• Population displacement
• Animal movements
• Changes in agriculture
• Emerging zoonoses
• Deforestation
• Bird migration
• Human conflict
• Antimicrobial resistance
• Genetic mutation/recombination
• Deliberate release.

265

Public [Link] 265 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

CURRENT UK AND EUROPEAN CONCERNS:

Diseases attracting particular research attention and interest in control policy include:
• Smallpox
• Anthrax
• SARS
• Avian influenza and pandemic influenza
• West Nile virus
• Changes in vector distribution
• Pet travel scheme
• Babesiosis
• Leishmaniasis
• Hantavirus
• MRSA, vancomycin-resistant enterococci (VRE).

NZ NEW ZEALAND – IMPORTANT INFECTIONS

In addition to the lists above, infectious diseases of particular importance in NZ include those shown in Box 2G.7.2.

Box 2G.7.2
Leptospirosis This zoonotic infection follows exposure to urine from infected animals and urine-
contaminated surface water. It is an important occupational zoonotic disease for farmers and
abattoir workers in New Zealand. Individual cases and outbreaks are common in developing
countries, particularly following flooding
Rheumatic fever Acute rheumatic fever may occur following streptococcal throat infection and may result
in serious damage to heart valves, leading to chronic rheumatic heart disease (CRHD).
Acute rheumatic fever and CRHD remain an important cause of morbidity and mortality for
indigenous people in New Zealand (Maori), Australia (Aborigines) and the Pacific Islands, as
well as people living in many developing countries

PANDEMIC INFLUENZA vIRUS IN HUMANS (H5N1)

Outbreaks of influenza A/H5N1 (a highly pathogenic form of avian influenza) are thought to represent a serious
threat to public health. Although A/H5N1 has appeared before, it has caused concurrent outbreaks in several
countries and is proving difficult to eliminate.
Box 2G.7.3 summarises some terminology.

Box 2G.7.3
Seasonal influenza RNA virus of the Orthomyxoviridae family. It rapidly spreads around the world in a
seasonal pattern
Avian influenza Form of influenza that is virulent in birds
Influenza A/H5N1 Highly pathogenic form of avian influenza
Pandemic influenza Worldwide epidemics of human influenza such as those that occurred in 1918 and 1957

It is feared, for several reasons, that A/H5N1 might trigger the next pandemic:
• Ability to infect humans and cause severe disease
• Ability to mutate and to acquire or exchange genes from other viruses
• Ongoing spread in birds.

266

Public [Link] 266 3/6/08 [Link]

 2G Com mun icable Diseases

The likelihood of a pandemic increases as more people become infected over time, especially if they are concurrently
infected with human influenza. In these circumstances a novel viral subtype may emerge that can be transmitted
readily person to person. The likelihood of such a mutation occurring is difficult to predict.

EMERGENCY PLANNING
The key steps in dealing with SARS, pandemic influenza and other emerging communicable disease threats are:
• Assess
• Prevent (vaccinate)
• Prepare (plan for surge capacity, stocks of facemasks)
• Respond
• Recover (from the event including psychological care).
Note that these are the same principles as for other disasters such as flood, fire or terrorist incidents.

2G.8 ORGANISATION OF INFECTION CONTROL

Organisation is at several levels.

LOCAL GOvERNMENT
Local authorities have the power to take action to control notifiable diseases within their boundaries. They are
required to appoint a ‘proper officer’ − usually a Consultant in Communicable Disease Control (CCDC) or a Designated
Medical Officer in Scotland. Environmental health services have a duty to register, inspect and investigate food
premises, and have legal powers of enforcement and prosecution.

HEALTH SERvICES
As part of their remit to promote health and prevent disease, health authorities/boards are responsible for the
surveillance of disease, identifying problems and establishing planning measures.

HOSPITALS
Each hospital is required to have in place an infection control team (ICT) consisting of an infection control
physician or microbiologist, and an infection control nurse (ICN) and in the UK one of them will act as Director of
Infection Prevention and Control (DIPC). The activities of the ICT are listed in Table 2G.8.1.

Table 2G.8.1 Activities of the ICT

Planning • Developing policies and procedures
• Accommodation
• Purchasing equipment, etc.
• Clinical waste
Education • Education of staff
• Audit
• Handwashing
Surveillance • Antibiotic use
Outbreak • Advise on outbreaks
• Use of isolation facilities

The ICT reports to the hospital infection control committee, the members of which may include the Chief Executive
(or director level deputy) and a CCDC. The committee meets at regular intervals to review infection control.

267

Public [Link] 267 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

COMMUNITY
Community ICNs (Infection Control Nurses) encourage collaboration among community staff, ICTs (Infection Control
Teams) and CCDC (Consultants in Communicable Disease Control) as well as with care homes, prisons, nurseries and
schools.

UK NATIONAL CENTRES: HEALTH PROTECTION AGENCY − CENTRE FOR INFECTIONS

These are responsible for:
• Infectious disease surveillance
• Provide specialist and reference microbiology and microbial epidemiology
• Coordinate investigation and cause of national and uncommon outbreaks
• Helping advise government on the risks posed by various infections
• Respond to international health alerts.

Eng DEPARTMENT OF HEALTH

Central government has overall responsibility for national policy matters relating to infection control. The Chief
Medical Officer and Chief Nursing Officer develop policies, guidance and tools for the NHS. They seek advice from
experts within and outside the DH to advise government on infection control matters, including:
• Aseptic technique
• Cleaning
• Decontamination
• Handwashing
• Invasive devices (installing catheters, etc.)
• Isolation
• Laboratory specimen guidance
• Laundry and linen handling/management
• Handling of sharps
• Waste disposal/management.
Examples of recent initiatives are given in Box 2G.8.1.

Box 2G.8.1
Examples: recent initiatives from the DH related to infection control
• Health Bill (2006) – includes Hygiene Code of Practice for the Prevention
and Control of Healthcare Associated Infections
• Saving Lives − a delivery programme to reduce health-care-associated
infections, including MRSA (2005)
• A Matron’s Charter − an action plan to cleaner hospitals (2004)
• Getting Ahead of the Curve − a strategy for combating infectious diseases
(including other aspects of health protection) (2002)

NOTIFIABLE DISEASES
Eng The diseases in Table 2G.8.2 are statutorily notifiable: all doctors working in England and Wales are
Wal
required to report to the proper officer of the local authority of any suspected cases. The proper officer is required
to provide anonymous details to the Health Protection Agency every fortnight.

268

Public [Link] 268 3/6/08 [Link]

 2G Com mun icable Diseases

Table 2G.8.2 Notifiable diseases in England and Wales

Anthrax Meningitis Scarlet fever
Cholera Meningococcal septicaemia Smallpox
Diphtheria Mumps Tetanus
Dysentery Ophthalmia neonatorum Tuberculosis
Encephalitis (acute) Paratyphoid fever Typhoid fever
Food poisoning Plague Typhus fever
Leptospirosis Poliomyelitis (acute) Viral haemorrhagic fever
Leprosy* Rabies Viral hepatitis
Malaria Relapsing fever Whooping cough
Measles Rubella Yellow fever
*Leprosy should be reported directly to the Centre for Infections at the Health Protection Agency.

Scot All doctors working in Scotland are required to report any suspected cases to the designated medical officer
of the local authority. The designated medical officer is required to provide anonymous details to Health Protection
Scotland every fortnight.
Note that, in Scotland, chickenpox and Legionnaires’ disease are notifiable diseases (in addition to the main UK
list above).

2G.9 MICROBIOLOGICAL TECHNIQUES

Basic understanding of the biological basis, strengths and weaknesses of routine and reference microbiological
techniques
There are two main methods of microbiological analysis − the traditional method involves growing a culture of
the specimen in order to isolate and identify the microorganism (bacteria, fungi, viruses and parasites). The
alternative is a range of modern molecular methods that involve the identification of specific DNA or RNA (e.g. RNA
transcriptase) within the specimen. Tuberculosis, for example, once took 12 weeks to diagnose; now using molecular
methods it takes 24 h.

MAIN CATEGORIES OF METHODS USED IN MICROBIOLOGY LABORATORIES

• Microscopy (including immunofluorescence)
• Culture
• Identification (e.g. typing of bacterial strain)
• Isolation of virus
• Drug sensitivity
• Serology (including immunoassay for antigen and antibody).

ROUTINE MICROBIOLOGICAL TECHNIQUES

Local hospital laboratories have tended to use traditional techniques, the strengths and weaknesses of which are
listed in Box 2G.9.1.

269

Public [Link] 269 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

Box 2G.9.1
Strengths Weaknesses
Relatively low cost Limited ability of laboratories to provide doctors with
timely and clinically relevant information
Can provide definitive diagnosis
Low sensitivity, e.g. samples taken after antibiotic has
been given may test negative
Samples taken after onset of illness may result in
difficulty isolating pathogen, e.g. viruses
Limited range of tests available – may not be able to
provide full identification, e.g. toxin-producing strains

REFERENCE MICROBIOLOGICAL TECHNIQUES

Molecular biological techniques form the basis of detecting and characterising an ever-increasing range of viruses,
bacteria, fungi and protozoa.
Nucleic acid probes are commercially available for cytomegalovirus, human papillomavirus, hepatitis B virus,
hepatitis C virus, Chlamydia trachomatis, Neisseria gonorrhoeae, Streptococcus pyogenes and mycobacteria, among
others.
Nucleic acid amplification systems are available for the direct detection in clinical specimens of hepatitis C virus,
HIV, M. tuberculosis, C. trachomatis and N. gonorrhoeae.
Strengths and weaknesses of these techniques are listed in Box 2G.9.2.

Box 2G.9.2
Strengths Weaknesses
Increased speed Need for specialised equipment
Increased sensitivity and specificity Segregated rooms in laboratories
Identify organisms that do not grow (or grow only slowly) in Currently detect only microorganisms
culture
Turnaround times for existing tests are much
Identify genes that result in resistance to antibiotics longer than can potentially be achieved using
molecular methods
‘Fingerprint’ individual isolates for epidemiological tracking
Recognition of newly emerging infectious diseases
Control of antibiotic resistance in Streptococcus pneumoniae,
Haemophilus influenzae, Staphylococcus aureus and common
Gram-negative bacilli

IMMUNOASSAYS
These are used in the detection of microbial antigens and offer the potential for rapid diagnosis. Examples include
enzyme-linked immunoassays and direct immunofluorescence antibody assays. Box 2G.9.3 lists strengths and
weaknesses of immunoassays.

270

Public [Link] 270 3/6/08 [Link]

 2G Com mun icable Diseases

Box 2G.9.3
Strengths Weaknesses
Technical simplicity Poor sensitivity
Rapidity Low negative predictive value
Specificity Low positive predictive value
Cost-effectiveness

AUTOMATED AND SEMI-AUTOMATED SYSTEMS

These fall into two main groups:
• Identification and susceptibility testing (some can provide results within a single working day); and
• Blood culture systems (most true positive results are detected within 24−36 h).
Some blood culture systems have been adapted for automated or semi-automated culture (e.g. for M. tuberculosis and
other mycobacteria). These enable the identification and susceptibility results to be processed from large numbers of
blood culture samples.
Box 2G.9.4 lists the strengths and weaknesses of automated and semi-automated systems.

Box 2G.9.4
Strengths Weaknesses
Reduce the traditional dependence on biochemical Organisms may be incorrectly identified, e.g. database
reactions to identify organisms does not include the correct identification
Avoid the many labour-intensive steps between isolating Bacteria with heteroresistance to b-lactam drugs,
and reporting clinically significant bacteria inducible resistance mechanisms or susceptibility gene
mutation may be misclassified
Provide rapid results
May miss resistance of an organism to antibiotic, e.g.
Perform tests more reproducibly
enterococci to glycopeptides; use supplemental testing
with manual methods for problematic combinations of
organisms and drugs

Molecular methods undoubtedly have enormous potential in diagnosing infectious diseases. New molecular methods
will be widely accepted and implemented routinely within the next decade.

2G.10 INTERNATIONAL ASPECTS OF COMMUNICABLE DISEASE CONTROL

International aspects of communicable disease control, including port health
Globalisation has increased the risk of international spread of infectious diseases. In the past, the most concrete
measures to stop importation of infectious diseases were thought to be quarantine and trade embargoes.

INTERNATIONAL OBLIGATIONS
INTERNATIONAL HEALTH REGULATIONS
This is a multilateral initiative by countries to develop a global tool for the surveillance of cross-border transmission
of diseases. It balances the protection of public health with the avoidance of unnecessary disruption to trade and
travel.

271

Public [Link] 271 3/6/08 [Link]

 2G C o m mu n ic abl e D iseases

CORE OBLIGATIONS FOR WHO MEMBER STATES

Countries are obliged to notify the WHO of public health emergencies of international concern. They must also:
• Respond to requests for verification of information regarding urgent national risks
• Control urgent national public health risks that threaten to transmit disease to other member states
• Provide routine port inspection and control activities to prevent international disease transmission
• Apply the measures recommended by the WHO during public health emergencies.

CORE OBLIGATIONS OF THE WHO

The WHO has a duty to respond to the needs of member states regarding the interpretation and implementation of
its regulations. It must update these regulations (and their supporting guides) so that they remain scientifically
valid. In addition, the WHO must publish recommendations for use by member states during public health
emergencies of international concern.

PORT HEALTH
In the UK, the regulations for ships, aircraft and international trains give local authorities and port health
authorities the power to appoint medical and non-medical port health officers who can prevent the entry of
communicable diseases into the country.

CIRCUMSTANCES REQUIRING THE INTERvENTION OF PORT HEALTH STAFF

• Outbreak of food- or water-borne disease on the vessel
• Contamination of aircraft by faeces or vomit
• Pests (rodents or insects) on board
• Passengers or crew who are suspected of being infected with viral haemorrhagic fever, yellow fever, plague,
cholera, diphtheria or TB.
The Port Medical Inspector advises immigration officers on matters of health protection. Immigration officers may
refer passengers who are emigrating to the UK, long-stay visitors and those visiting for health reasons. Some such
passengers may be required to have a chest X-ray, with the findings being passed to the CCDC.

272

Public [Link] 272 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

2H
Principles and Practice of Health Promotion

2H.1 Responsibility for health 273 2H.9 Communication in health education 293
2H.2 Determinants of health 275 2H.10 Legislation and health promotion 296
2H.3 Policy dilemmas 280 2H.11 Programmes of health promotion 298
2H.4 The prevention paradox 283 2H.12 Community development 300
2H.5 Health education 284 2H.13 Partnerships 301
2H.6 Settings for health promotion 284 2H.14 Evaluation 304
2H.7 Models of health promotion 287 2H.15 International initiatives 306
2H.8 Risk behaviour 292 2H.16 International health promotion initiatives 307

This chapter is about encouraging people to adopt actions that reduce the risk of developing disease. It is
increasingly recognised that health promotion is at its most effective when it focuses on enabling people to
increase control over their own health. For this reason it is often helpful to regard the discipline as a sociopolitical
process.

2H.1 RESPONSIBILITY FOR HEALTH

Collective and individual responsibilities for health: both physical and mental
There are different views about the extent to which health is a collective or an individual responsibility. These
perspectives determine how societies organise themselves to improve health.
• Social responsibility is a doctrine that holds that an entity (be it state, government, corporation, organisation
or individual) has a responsibility to society as a whole. This responsibility can be negative (i.e. a responsibility
to refrain from acting) or positive (i.e. a responsibility to act).
• In contrast, individualism is a moral, political and social philosophy, which emphasises the importance of the
individual. Its central tenets are individual liberty, the ‘virtues of self-reliance’ and personal independence.
Proponents of public initiatives and social responsibility argue that their policies are beneficial to the individual,
and that excessive individualism may actually be detrimental to the individuals themselves. In contrast,

273

Public [Link] 273 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

individualists hold that public initiatives may have unintended consequences beyond the issues that they are
intended to address. Many commentators find the ‘beneficial to the individual’ argument condescending and argue
that individualism is not about individual benefit so much as individual choice.
It generally falls to politicians to decide which paradigm dominates health policy in any one country at a given
time. For example, the absence of a universal health service in the USA has its roots in the political belief that
individuals, rather than society, have responsibility for health care.

COLLECTIVE RESPONSIBILITIES
Approaches that emphasise collective responsibilities for health encompass population-wide measures. They include
those shown in Box 2H.1.1 for the UK.

UK Box 2H.1.1
Policy Example
Legislation Drink−driving laws exist not just to protect the individual but to ensure that
the individual does not put others at risk
Regulation Health and safety legislation and regulations enable external bodies (e.g.
the Health and Safety Executive) to inspect and ensure that businesses are
working to protect their employees
Population-wide measures Fluoridation of the water supply
Progressive health service Universal taxation where high earners in society provide a larger
funding systems contribution for NHS costs, despite the fact that those earning least may
actually use the service more

INDIVIDUAL RESPONSIBILITIES
Approaches based on individual responsibility focus on initiatives to enable individuals to make an informed
choice. They include those shown in Box 2H.1.2.

Box 2H1.2
Policy Example
Information provision Providing safe drinking limits allows individuals to choose how much they drink given
knowledge of the health consequences
De-regulation Relaxation of licensing to allow pubs and shops to sell alcohol 24 h a day relies on
individual choice regarding when and how much to drink
Choice in health care Private health-care insurance enables individuals to choose where they receive health
care and whether or not to insure their health

MIXED APPROACHES
Eng England’s White Paper Choosing Health (DH 2004) contains many proposals both for collective action and for
encouraging people to assume individual responsibility. This paper set out the future direction for population health
improvement in the country. It covered six main themes, including a mixture of legislation and measures to help
people make positive, informed choices about risk factors associated with their future health.
The six main areas prioritised for action were:
1. Reducing the numbers of people who smoke

274

Public [Link] 274 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

2. Reducing obesity and improving diet and nutrition

3. Increasing exercise
4. Encouraging and supporting sensible drinking
5. Improving sexual health
6. Improving mental health.
At a collective level, the White Paper resulted in legislation to ban smoking in enclosed public places. Other
mechanisms for supporting people to make healthy choices included:
1. Marketing health
2. Improving information about health for the public
3. Tackling health inequalities
4. Partnership with industry (e.g. proposals for voluntary agreements on food marketing)
5. Promotion of healthy food for children
6. Further restrictions on tobacco advertising.

2H.2 DETERMINANTS OF HEALTH

Interaction between genetics and the environment (including social, political, economic, physical and personal factors)
as determinants of health, including mental health
See also Section 2I.2.
The factors that have the most significant influence on health are termed the determinants of health. While
health care and Social Services focus largely on dealing with the consequences of poor health, most of the key
determinants of health as a positive attribute lie outside the direct influence of these services. They are influenced
more by factors such as education, employment, housing and environmental policy.
Genetic and epidemiological studies offer understanding of the relative contributions of such factors to health and
illness. This is useful when developing health promotion interventions, in particular:
• Whether to use targeted or universal programmes
• How to allocate resources
• Predicting susceptibility
• Predicting uptake of health promotion in the population.
Several theories exist for contemplating the range of influences on health. This is potentially confusing, but it
reflects the rapid advances in this important field over the last half century, and the changing political and cultural
perspectives from which the theories arise. The four frameworks on the determinants of health are summarised in
Table 2H.2.1, and will then be discussed in turn.

Table 2H.2.1 Determinants of health frameworks

Framework Author(s) Year Summary
Health field concept Lalonde 1974 Health care is not the sole determinant of health
Fields are biology, lifestyle, environment and health care
Policy rainbow Dahlgren and 1991 Determinants of health exist as interrelated layers of
Whitehead influence
Health field model Evans and Stoddart 1990 Health is not only the absence of disease, but also takes
into account functional status and wellbeing
Social determinants Diderichsen and 1998 Social conditions affect individuals’ social situations, which
model Hallqvist in turn determine their health risks

275

Public [Link] 275 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

LALONDE HEALTH FIELD CONCEPT (1974)

Marc Lalonde, Health Minister for Canada 1972−1977, proposed the health field concept in 1974 in a seminal report:
A New Perspective on the Health of Canadians. Building on ideas of Thomas McKeown, it used evidence of mortality
and morbidity in Canadians to argue that health-care services were not the most important determinant of health.
He identified four fields – biology, lifestyle, environment and health care − as determinants of health (see Figure
2H.2.1) and started a new direction in Canadian health policy:
‘The Government of Canada now intends to give to human biology, the environment and lifestyle as much attention
as it has to the financing of the health care organisation so that all four avenues to improved health are pursued
with equal vigour.’
Reproduced from Lalonde (1974)

Health-care
systems

Human
biology Health Environment

Lifestyle

Figure 2H.2.1 Lalonde health field concept

The model has undergone refinements following criticism that it focused too much on lifestyle and too little
on environment. However, it was hugely influential throughout the world and was pivotal in the growth of the
discipline of health promotion.

DAHLGREN AND WHITEHEAD (1991)

Dahlgren and Whitehead’s health ‘rainbow’ identifies a range of determinants of health (see Figure 2H.2.2). It
recognises that some determinants (e.g. age and sex) are not modifiable, but that many can be altered. It builds
on previous models by giving an indication of the different levels at which health is affected: see Box 2H.2.1. The
model makes no attempt to explain the relationship between the different tiers, nor between factors in the same tier
– but instead aims to stimulate discussion about the interrelationship of different layers and the relative importance
of each to health. Furthermore, the model aims to promote debate regarding potential interventions for improving
health and reducing inequalities in each of the four layers. Box 2H.2.2 illustrates how the rainbow has been used to
develop policy.

276

Public [Link] 276 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Box 2H.2.1
Individuals Although some individual factors (e.g. age) are fixed, others (e.g. lifestyle and behaviour)
can be influenced. Influences include information and education, and also influences on the
‘distal’ determinants in the other layers
Communities Strengthening communities through action to improve the local environment and living
conditions, or through ‘bottom-up’ action led by local community groups
Access Improving access to services such as health care, leisure, transport in terms of location, cost,
appropriateness
Macroeconomics Engendering macroeconomic or cultural change at national or global levels, possibly through
legislative changes

The relative importance of factors in each layer depends on the health issue or population under discussion.

cultural or environm
n om ic , ent
ec o al
io co
oc nd
l s Living and itio
a n
er working
conditions

s
n
Ge

Work Unemployment
environment

community inf
l and lue
nc Water and
c ia e sanitation
o es
ual lif tyle fa
id cto
S

Education iv
nd r s
I

Age, sex and

Health care
constitutional
services
factors

Agriculture and
food production
Housing

Figure 2H.2.2 Dahlgren and Whitehead’s health rainbow. Reproduced from Dahlgren and Whitehead (1991)

277

Public [Link] 277 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Eng Box 2H.2.2

Example: London Health Commission reports
The work of the London Health Commission is informed by Dahlgren and Whitehead’s policy rainbow. The
Commission produces an annual report on the health of Londoners, focusing on variations on ten indicators
including:
1. Health outcomes: life-expectancy, infant mortality and self-reported health status
2. Determinants of health related to:
• Social and community influences – levels of crime
• Living and working conditions − employment, education and housing
• Environmental factors − road safety, air pollution
The Commission’s analysis links individual constitutional factors with the determinants in higher levels. For
example, there are clear physical and biological reasons why very young and very old people are more susceptible
to disease and injury than adults of working age. The report also considers the changes in people’s living
conditions as they age, in particular:
• Housing: young households (where the oldest member is 16−24) are most likely to live in poor housing.
Households with residents over 75 are also likely to live in poor housing
• Employment: 16−19 year olds have the highest unemployment rates
• Crime: young households are most likely to be burgled
Poor socioeconomic conditions are likely to affect old and young people in different ways: while young people
may be physically robust enough to withstand the health risks of poor living conditions, older people are more
susceptible.
Reproduced from the London Health Commission (2002).

EVANS AND STODDART (1990): THE HEALTH FIELD MODEL

In this health field model, health is explicitly conceptualised as more than the presence or absence of disease to
include functional status and wellbeing. By setting out a relationship between determinants, the health field
model helps practitioners to understand how the determinants are themselves influenced, and therefore how they
might be modified (see Figure 2H.2.3).
Evans and Stoddart’s framework of health fields encompasses a range of factors, including those shown in Box
2H.2.3.

Box 2H.2.3
Social environment Education, employment, family, poverty
Physical Poor housing, proximity to hazards, waste and conflict
environment
Genetics Genetic factors that interact with environmental conditions
Behaviour Viewed as an ‘intermediate’ determinant (i.e. not simply a voluntary act), behaviour is
shaped by a range of determinants, including education, access to facilities and financial
considerations
Health care Another ‘intermediate’ determinant, this encompasses access and quality of health care

278

Public [Link] 278 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Social Physical Genetic

environment environment endowment
Individual
response
• Behaviour
• Biology
Health and
Disease Health care
function

Wellbeing Prosperity

Figure 2H.2.3 The health field model. Reproduced with permission from Evans and Stoddart (1990)

SOCIAL DETERMINANTS FRAMEWORKS

The links between social conditions and health have been thoroughly explored. Research has strongly linked health
(considered as premature mortality, vulnerability to illness and injury, self-reported health and wellbeing) to
education and employment. For example, the Whitehall study found that those employed in lower grades of the
British civil service were more likely to die prematurely than those employed in higher grades. There are several
models that place social conditions as the main determinant of health and health inequalities.
Diderichsen and Hallqvist (1998) devised one of the most commonly cited ‘social determinants’ frameworks. The
model identifies four broad conceptual mechanisms:
1. Social stratification (social conditions, such as education and employment, will determine people’s social
situation) leading to
2. Differential exposure and
3. Differential vulnerability which together result in
4. Differential consequences.
These mechanisms work synergistically to generate health inequities. For each mechanism, the possible entry points
for policy interventions are identified (see Figure 2H.2.4).
Other models focusing on social determinants of health include explicit recognition that:
• Health influences social position as well as vice versa
• There are interacting effects between social and biological pathways
• Access to and quality of health care are related to social determinants.
Recently the focus for action has been on the social inequalities of health, which Dahlgren (quoted in an interview
with Koller, 2006) defines as:
‘Systematic differences in health between socioeconomic groups [which] are socially produced, modifiable and
unfair.’
This reflects growing disparity in life-expectancy and health status between socioeconomic groups, while overall the
average health and life-expectancy continue to improve.

279

Public [Link] 279 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

INDIVIDUAL

Social Influencing social Social

context stratification position

Decreasing exposures Differential Differential

exposure vulnerability

Specific
exposure

Decreasing vulnerability

Disease or Differential
injury consequences
Mechanisms that
play a role in
stratifying health
outcomes Preventing unequal consequences

Policy
Policy entry
context
points
Further social stratification

Figure 2H.2.4 Diderichsen and Hallqvist’s social determinants framework. Adapted from Diderichsen and Hallqvist
(1998)

2H.3 POLICY DILEMMAS

Ideological dilemmas and policy assumptions underlying different approaches to health promotion
While the aims of health-care treatment are usually relatively clear, the aims of health promotion may be
contentious, and they determine the approaches used to improve health.
Health promotion can reach beyond the traditional boundaries of health into policy-making, personal choice and
community development. In some contexts, the reach of health promotion into these spheres is more acceptable
than others. The acceptable scope for health promotion will depend on a number of factors including:
• Prevailing policy: two contrasting health policy statements – Health for All (1986) and Choosing Health (2004)
– exemplify how approaches to health promotion can be shaped by views regarding personal autonomy and the
role of communities.
• Socioeconomic circumstances affect both the community’s health needs and the resources available to
promote health. Naidoo and Wills (1998) describe the dilemma of health promotion for people whose primary
problem is economic poverty.
• The characteristics of the particular health issue, e.g. the epidemiology and perceptions of HIV and AIDS have
dictated the focus of health promotion.

280

Public [Link] 280 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

EFFECT OF POLICY: BALANCE OF INDIVIDUAL AUTONOMY, COMMUNITY INVOLVEMENT AND STATE

INTERVENTION
The Ottawa Charter (1986) applied the WHO ‘Health for All’ policy to the field of health promotion. The aims of the
Charter implicitly assume that circumstances beyond the control of the individual are necessary to support health
promotion, i.e. community involvement and state intervention are necessary. The Charter advocates that people
involved in health promotion should:
• Create supportive environments
• Enable community participation
• Develop personal skills for health
• Reorient health-care services towards prevention and health promotion
• Build wide-ranging public policy that protects the environment and promotes health.
Reproduced from [Link]/ph-sp/phdd/docs/charter/[Link] (The Ottawa Charter 1986).
Eng Choosing Health (DH 2004): the Department of Health Public Health White Paper focuses on enabling
individuals to make decisions about their own health. Here, the role of health promotion is to provide or facilitate
the provision of information. Health promotion through legislation or community involvement is largely outside its
scope (see Section 4C.3).

INDIvIDUAL ROLES
Personal roles have different requirements in terms of information, skills and involvement: see Box 2H.3.1.

Box 2H.3.1
Role Needs and entitlements
Patients Receive interventions only; do not need further input in the way
that programmes are delivered
Consumers Need sufficient information to make an informed choice about
whether they accept or reject health-promoting behaviours
Empowered Need adequate skills and resources to enable them to take part,
participants and have a responsibility to fulfil their role in shaping and
delivering health promotion

DILEMMA 1: TACKLE POVERTY OR ADDRESS NARROWER DETERMINANTS OF HEALTH?

An ideological dilemma of health promotion is whether its aim is to reduce inequalities in access to health services
or to reduce inequalities in health (which would require wider actions outside the health remit).
The Black Report (Black et al 1980) argued that poverty was one of four possible explanations for the widening
health inequalities seen in Britain in the twentieth century. See Sections 4A.8 and 4C.10.

POvERTY
Poverty (relative or absolute) is a major determinant of ill health. Low pay, inadequate benefits or unemployment
lead to various types of poverty, namely lack of food and fuel, poor housing and transport, and social isolation.
These in turn lead to:
• Physical problems (e.g. low-birth-weight babies, respiratory disease)
• Psychological problems (e.g. stress, depression, anxiety)
• Behavioural changes (e.g. smoking, low exercise, poor diet).

281

Public [Link] 281 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

BARRIERS
Barriers to effective health promotion in poverty include:
• Intrinsic ‘victim blaming’ culture
• Focusing on knowledge, attitudes and behaviour ignores the constraints on choice of healthy lifestyles that is
associated with poverty
• Focusing in medicine on one-to-one interventions can ignore wider social influences, and underestimates the
effect of poverty
• The lack of a common approach to poverty in health promotion
• Initiatives that may alleviate poverty (e.g. food cooperatives) are outside the health remit.

POLICIES
Policies to reduce poverty and improve ill health include:
• Macro-level changes (e.g. minimum wage)
• Collection of data on the wider social and economic determinants of health
• Multisectoral action at all levels (e.g. organisational: employment of benefits advisors in health clinics)
• Community development rather than simply health advice (e.g. food cooperatives to support healthy eating
messages).

DILEMMA 2: TARGETED OR UNIVERSAL HEALTH PROMOTION?

Targeting offers an opportunity to prioritise resources and to tailor messages and activities to the particular
characteristics of the people that need them the most. Targeting can be according to several factors. However, there
are disadvantages associated with targeting health promotion, outlined in Table 2H.3.1. The problems associated
with both targeting and universal health promotion approaches are illustrated with respect to HIV/AIDS in Box
2H.3.2.

Table 2H.3.1 Disadvantages of targeting health promotion

Targeted focus Examples Disadvantages
Behaviour Smoking, eating, This risks widening health inequalities by appealing to those with the
car driving resources and circumstances available to change their behaviour. For example,
advice to eat more fruit is easier to follow by people who can afford fruit and
who live in an area where fruit can readily be purchased
Group Children, older Assumes that groups are homogeneous (e.g. all gay men indulge in
people, gay men promiscuous, unprotected sex)
Can lead to culture blaming (i.e. ascribing the risk to an inherent aspect
of culture rather than a problem to be tackled jointly by the community and
health). For example, the Asian rickets campaign focused on health education
messages about increasing vitamin D in the diet, thereby placing the blame
for vitamin D deficiency on Asian families. In contrast, when vitamin D
deficiencies were first recognised in the 1950s as a general health problem,
the response was to fortify core foods
Groups that are deemed to be at risk of a certain conditions (e.g. sickle cell
disease in Africans) may be neglected with regard to broader health problems
(e.g. coronary heart disease)

282

Public [Link] 282 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Targeted focus Examples Disadvantages

Condition Hypertension, As part of the prevention paradox, many healthy people will be
diabetes ‘pathologised’ (e.g. hypertension will be diagnosed and treated in them,
thereby causing unnecessary worry and side effects of treatment in people
who will never become ill as a result of elevated blood pressure)
In contrast, focusing only on people at high risk can miss many people who
are at risk. For example, larger numbers of people who are normotensive die
of coronary heart disease than do people who are hypertensive. Furthermore,
concentrating on high-risk attributes while ignoring broader social influences
may be less successful than whole-population initiatives
The alternative strategy is to attempt to reduce risk across the whole population (see Section 1C.16). Where a
risk factor is widespread in society, universal approaches can have a larger effect than targeted interventions. For
example, universal approaches to reducing dietary salt (e.g. lobbying manufacturers to reduce levels of salt in their
products) can reduce average blood pressure and thereby reduce the burden of coronary heart disease. In this case,
however, while the population’s risk of disease may be substantially reduced, the average individual risk of disease
is essentially unaffected. This is known as the prevention paradox and its recognition can cause the public to lose
credibility in universal approaches (see Section 2H.4).

UK USA Box 2H.3.2

Example: Changing approaches with the changing perceptions and epidemiology of HIv/AIDS
When the HIV/AIDS epidemic began in the 1980s in the UK and the USA, health promotion messages were
initially targeted at gay men. This led to:
• An illusion that heterosexual people were immune to HIV/AIDS
• Increased homophobia
• Difficulties in attracting funding and resources for what became seen as a ‘marginal’ illness
By the mid-1980s, the epidemic had changed and HIV was recognised as an infection that could potentially
affect anyone. As a result, a whole-population approach for HIV/AIDS health promotion was adopted. However,
this caused:
• A shift in resources from gay projects to professionally led mainstream interventions
• Explicit, detailed guidance on safer sexual behaviour that would previously have been unacceptable to the
general public
• Change in messages from minimal behaviour change (sex, but safe sex) towards more conservative ones
(monogamy)
• Ignored the fact that gay men were still disproportionately at risk
• Downgraded the strong activist/support networks that had been built up
The debate still continues with HIV about whether to focus resources on the groups primarily at risk, i.e. people
from sub-Saharan Africa and men who have sex with men – or on the whole population.
Reproduced from Naidoo and Wills (1998).

2H.4 THE PREvENTION PARADOx

As described in Section 2H.3, public health practitioners often face a dilemma over whether to target people at
greatest risk of a disease, or to lower the risk across the whole population.

283

Public [Link] 283 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

The prevention paradox describes the effect whereby actions to reduce the risk of disease across the population
successfully reduce the population’s overall risk, but affects the outcome only for a minority. For example, the
mandatory wearing of seat belts is a policy that affects all car users but prevents death only in the small minority
involved in a road traffic crash.
‘A preventive measure that brings large benefits to the community offers few benefits to each participating
individual.’ (Rose 1981)

IMPLICATIONS FOR HEALTH PROMOTION

Because only a minority of people benefit directly from a population approach, the alternative approach of targeting
health promotional activities at high-risk people may seem more attractive (see Dilemma 2, Section 2H.3). In other
words, the prevention paradox can cause a loss of credibility in population-wide health promotion materials. For
example, health promotion advice recommends reducing fat in the diet in order to minimise the risk of coronary
heart disease. If it is evident that some individuals have a high-fat diet and do not develop heart disease, there
is less incentive for others in the community to reduce their dietary fat. Hunt and Emslie (2001) describe what
happens when health promotion materials do not take into account the prevention paradox:
‘The failure to acknowledge the prevention paradox more directly in health education material thus can lead, at best,
to greater mistrust among the general public of the messages contained, and at worst to their outright rejection.’

2H.5 HEALTH EDUCATION

Health education and other methods of influencing personal lifestyles that affect health
Health education is sometimes erroneously taken to be synonymous with health promotion – the flawed assumption
being that if people are sufficiently well informed about health then they will make healthy choices. Frameworks of
the determinants of health illustrate that health is actually affected by more than simply an individual’s personal
choices (see Section 2H.2). This reality is echoed in models of health promotion, in particular Ewles and Simnett’s
(2003) five approaches to health promotion explicitly include other activities, such as social change, within the
remit of health promotion (see Section 2H.7).
Education is a major component of health promotion: formal health education programmes usually aim to influence
beliefs, attitudes and behaviours by enabling individuals to make an informed decision about those aspects of their
life that affect their health.
Individuals do not receive their health education solely from planned, conventional, health promotion practitioners
and materials: there is a range of other routes, such as friends, family, television and magazines, as illustrated by
the example of children’s knowledge of drugs and alcohol (see Box 2H.5.1).

2H.6 SETTINGS FOR HEALTH PROMOTION

Appropriate settings for health promotion (e.g. schools, the workplace)
Key locations for conducting health promotion include:
• Workplaces
• Schools
• Primary care.
As shown in Table 2H6.1, these settings offer the structure (e.g. physical building), resources (e.g. staff) and
access to people (who spend considerable time in these settings) for health promotion to improve health and
productivity. These benefits may improve the performance of the organisation as a whole.

284

Public [Link] 284 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Eng Box 2H.5.1

Example: Schoolchildren’s knowledge and use of drugs and alcohol
In England, the National Curriculum requires schools to teach children about the effects of drugs and alcohol.
This occurs at different stages of their education and in both science classes and in personal, social and health
education (PSHE) lessons.
Regular surveys are conducted in England to explore secondary schoolchildren’s smoking habits and their
knowledge and use of drugs and alcohol. The 2004 survey included nearly 10 000 pupils aged 11−15.
Sources of information
The 2004 survey indicated that informal sources play as great, or a greater, role in where pupils receive their
information about drugs and alcohol than does formal education. It concluded that:
‘Pupils were most likely to say that they had received useful information on smoking, alcohol and drugs from
the television (82%), their parents (75%) or teachers (72%). Newspapers or magazines (65%) and friends
(52%) were also important sources. Pupils were least likely to have received useful information from the
government’s information and advice campaign, FRANK (15%) or from helplines (15%).’
Effect of social deprivation
The survey also looked at reported rates of smoking and drug use according to two measures of social deprivation
(namely receipt of free school meals, and few or no books in the home). Children living in higher social
deprivation (according to either measure) were more likely to smoke or have taken drugs. This indicates that
efforts to reduce smoking or drug taking among school-aged children need to take account of social deprivation
as well as education and information.
Reproduced from the National Centre for Social Research and the National Foundation for Educational Research
(2005).

Table 2H.6.1 Settings for health promotion

Schools Workplaces Primary care
Target Children and adolescents Employees Patients
Others Parents Employees’ relatives and friends Customers (e.g. in
pharmacies), patients’
relatives and friends
Potential Physical: Biological: Workplace hazards apply to
hazards at the • Exposure to unhealthy • Exposure to toxic chemicals staff
setting food and fumes
Psychological: Physical:
• Peer pressure • Musculoskeletal symptoms
• Stress from assessments lifting and handling
and examinations • Sedentary lifestyle, e.g.
• Bullying office work
Psychological:
• Stress, bullying and
harassment
Table contd overleaf

285

Public [Link] 285 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Table 2H.6.1 contd

Schools Workplaces Primary care
Potential Performance: Performance: Performance:
benefits to the • Better educational • Recruitment, morale, • Overall cost savings
organisation results retention • Reduced consultation rate
and individual • Reduced absenteeism • Reduced referral rates
Wellbeing:
• Reduced emergency care
• Stronger links between Wellbeing:
school and home • Exposure to healthy social
• Prevention or delay in norms (e.g. no-smoking
risk-taking behaviours workplaces)
(e.g. drugs) • Reduce hazards
Resources Staff: Employees: Staff:
available to • Teachers • Occupational health • GP
support health • School nurses • Health and safety officers • Practice nurses
promotion • Managers • Receptionists
Facilities:
• Workers • Community nurses
• Playing fields (if any)
• Pharmacist
• Classrooms Partners:
• Dentist
• Unions
• Optometrist
Facilities:
Facilities:
• Office space, financial
• Practice/shop space
Examples of National Healthy Schools Health and Safety Executive General practice:
major policies Programme: • Vaccinations
COSHH (see 2F)
or activities • Personal, social and • Lifestyle advice
health education Policies: • Screening
• Healthy eating • Harassment • Disease prevention
• Physical activity • Smoking treatment
• Emotional health and
Training: Pharmacies:
wellbeing
• Health and safety • Display leaflets
Ofsted: • Manual handling • Provide smoking cessation
• Inspections cover • Monitor blood pressure,
Subsidised or free:
‘Healthy Schools’ cholesterol
• Counselling
implementation as well
• Bike loans Dentists and optometrists:
as quality of teaching
• Canteens with healthy • Smoking advice
menus
• Gym membership
• Vaccinations for at-risk
workers (e.g. hepatitis B and
influenza offered to health-
care workers)

286

Public [Link] 286 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Advantages Most children in school Protect health Patients seeking primary care
may be more receptive to
‘Captive audience’ Access to healthy workers (who
health messages
otherwise may not have access
Children are in school for
to health messages) Most people have contact
several years so changes can
with primary care at some
be tracked over time
point
Opportunities for progressive
People are more receptive to
programme
health messages from senior
health professionals
Limitations Work pressures may affect Work pressures may affect Short consultations mean
implementation implementation insufficient time to discuss
health promotion
What about children Priorities for employees may
excluded or away from differ from health at work (e.g. If the intervention is not in
school for other reasons? earning money) the contract/not prioritised,
then it may not take place
Staff (particularly non- Priorities for employers may
health professionals) need conflict with health priorities, Some staff not convinced of
to develop skills and particularly in low-paid or benefits or trained to deliver
confidence to deliver health illegal work health promotion
messages
Unemployed people are not Work pressures may affect
covered implementation

2H.7 MODELS OF HEALTH PROMOTION

Models of health promotion have been developed to:
• Define the scope and aims of the discipline (e.g. Tannahill 1985)
• Enable health promotion practitioners to understand what motivates individuals and/or communities to adopt
health-seeking/harming behaviours.
• Inform the development of health promotion programmes to influence health behaviours.
The principal health promotion models are described in Table 2H.7.1.

TANNAHILL MODEL
Andrew Tannahill (1985) considered health promotion to be defined by ‘three overlapping spheres of activity’ (see
Figure 2H.7.1):
• Health education
• Protection against harm and enhancing wellbeing
• Prevention of disease, disability and injury.

287

Public [Link] 287 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Table 2H.7.1 Health promotion models

Name Author Year Brief description
Health belief Hochbaum et al 1958 Individuals will adopt health-related actions if they
model believe that they are faced with risk and have the
potential to reduce the risk
Social learning Bandura 1977 Behaviour is influenced by social norms,
theory expectations, observations and perceived ability to
control behaviour
Stages of change Prochaska and 1984 Individuals go through several stages to change
DiClemente behaviour
Modern theories Tannahill 1985 Overlapping spheres of protection, prevention and
education
Beattie 1991 Four approaches covering a range of levels of
authority and individuality
Ewles and Simnet 2003 Multidisciplinary perspective, from biomedicine to
sociopolitical change

Examples of the contents of areas of the overlapping spheres are shown

in Box 2H.7.1.
Health
Box 2H.7.1 education
Prevention Organised preventive programmes of health
care, e.g. immunisation
Health education Health education to prevent disease onset,
e.g. smoking cessation advice and information Health
Prevention protection
Health protection Health protection (legislation) to prevent
illness or injury, e.g. fluoridation of water

The strengths and weaknesses of Tannahill’s model are summarised in

Table 2H.7.2.
Figure 2H.7.1 Tannahill model of
health promotion. Reproduced from
Tannahill (1985)
Table 2H.7.2 Strengths and weaknesses of Tannahill’s health
promotion model
Strengths Weaknesses
Simple to understand Distinction between protection and prevention
is arbitrary at times
Widely adopted for defining what constitutes
health promotion, and for informing health
promotion practitioners how to plan and conduct
their work
Encompasses not simply the absence of disease but
also the positive enhancement of wellbeing

288

Public [Link] 288 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

BEATTIE MODEL
Beattie’s model of health promotion (summarised in Naidoo and Wills 2000) considers not just the activities
involved in health promotion but also how they are delivered, i.e. from the top down or from the bottom up. It is a
useful tool for critically evaluating programmes, particularly regarding the balance of authoritative and negotiated
approaches.
Beattie outlined four approaches to health promotion:
1. Health persuasion
2. Personal counselling for health
3. Legislative action
4. Community development.
Which approach should be used in a particular circumstance depends on the mode of the programme – ranging from
authoritative (top down) to negotiated (bottom up) – and its focus (individual or collective), as shown in Figure
2H.7.2.

Individual

Health persuasion Personal counselling

GP advises patient to stop Setting up a helpline for
smoking patients to quit

Authoritarian Negotiated

Community
Legislative action development
Smoke-free workplaces Quit – smoking support group
run by ex-smokers

Collective

Figure 2H.7.2 Beattie’s model of health promotion

EWLES AND SIMNETT

The model proposed by Ewles and Simnett (described in Ewles and Simnett 2003) considers health promotion from
a multidisciplinary perspective. It explicitly incorporates biomedical approaches and activities in order to achieve
policy and social change within the broader scope of the discipline. This model considers five approaches – the
most appropriate combination of these depends on the starting point of the health promotion initiative (see Table
2H.7.3).

289

Public [Link] 289 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Table 2H.7.3 Ewles and Simnett’s health promotion approaches

Approach Description Example
Medical Focused on disease and biomedical explanations of health Immunisation
Narrow concept of disease (ignores social and environmental dimensions) Screening
Behavioural Encourages individuals to adopt healthy behaviours Healthy cooking
classes
Educational Provision of knowledge and information and assists development of skills for Schools
individuals to make informed decisions
Empowerment Helps individuals to identify their own concerns and needs Community
development
Health educator as facilitator
work
Social change Focus on socioeconomic environment in determining health Lobbying
Involves lobbying, policy planning, negotiating Policy planning
Negotiating

HEALTH BELIEF MODEL (1958)

The health belief model (HBM) was first developed by American social psychologists Hochbaum, Rosenstock and
Kegels in response to the failure of a free tuberculosis health-screening programme.
The HBM is based on the understanding that a person will take a health-related action if that person believes all
three of the things shown in Box 2H.7.2.

Box 2H.7.2
Necessary beliefs
1. That a negative health condition can be avoided
2. That by taking a recommended action, they will avoid a negative health condition
3. Believe that they can successfully take a recommended health action

Moreover, the health belief model predicts that behaviour change requires individuals to believe all five of the things
shown in Box 2H.7.3.

Box 2H.7.3
Susceptibility They are susceptible to the condition or problem
Consequences It could have potentially serious consequences
Course of action A course of action is available to reduce the risks
Benefits outweigh costs The benefits of the action outweigh the costs or barriers
Ability* The individual perceives that they have the ability to carry out the action (‘self-
efficacy’)

*Note that the final parameter, self-efficacy, was added later by Rosenstock et al (1988). This addition is thought to
improve the way that the HBM meets the challenges of changing habitual unhealthy behaviours (e.g. being sedentary,
smoking or overeating).

290

Public [Link] 290 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

The strengths and weaknesses of the HBM are listed in Box 2H.7.4.

Box 2H.7.4
Strengths Weaknesses
Evidence supports the usefulness of the model in Less useful for complex, long-term behaviours (e.g.
predicting behaviour or improving the effectiveness of alcohol dependence)
interventions
Most useful for traditional preventive behaviours (e.g. It does not account for other forces aside from
immunisations, health checks) individuals’ beliefs that influence behaviour (e.g.
socioeconomic circumstances or access to health care)

SOCIAL LEARNING THEORY

Also known as the social cognitive theory (developed by Bandura in the 1970s and 1980s – Bandura 1977), this
model focuses on three influences on behaviour: see Box 2H.7.5.

Box 2H.7.5
Reciprocal determinism The continuous, subtle and complex interactions between
people’s behaviour and their environment
Social norms The effect of social and cultural conventions on behaviour
Cognitive factors These encompass observational learning, expectations and
self-efficacy

Observational learning is the concept that humans learn not just by doing (participation) but also by watching
other people’s behaviour and the rewards that others receive from their behaviours.
The term ‘expectations’ is used to describe the capacity of a person to anticipate and value the outcomes of a
behaviour. These vary between individuals, underlining the importance of exploring personal attitudes and beliefs
when looking to change behaviour. For example, young women who believe that smoking helps with weight loss are
more likely to be persuaded to give up if they are given information about other ways to control weight, rather than
by warning them about the risks of smoking and lung disease.
Self-efficacy is a person’s perceived ability to control their own behaviour. This is both person specific and
environment specific. For example, someone may be very confident of their ability to avoid alcohol at home, but less
so if in a social situation.
Strengths and weaknesses of social learning theory are listed in Box 2H.7.6.

Box 2H.7.6
Strengths Weaknesses
Realistically complex solutions to health problems (i.e. not too simplistic) Can be difficult to implement
because of its broad scope
Unlike the stages of change and the health belief models, it explicitly recognises
and complexity
the impact of other factors: environmental, social and behavioural
Widens the role of health promotion beyond individual persuasion about a
discrete behaviour. Instead covers the entire social environment and wider
personal beliefs

291

Public [Link] 291 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

STAGES OF CHANGE MODEL

This was developed by Prochaska and DiClemente (1984) over a number of years and is also known as the trans-
theoretical model. It describes behaviour change as a process (not a one-off event) and it predicts that all
individuals who change their behaviour will go through the stages shown in Box 2H.7.7.

Box 2H.7.7
Stages of change
• Pre-contemplation
• Contemplation
• Determination
• Action
• Maintenance
• (Termination)
People can enter or exit at any point – and they can stall at any stage. The model can be applied to people
who initiate change themselves, as well to people in organised programmes. Programmes based on this model
require initial understanding of the stage or stages at which people may enter the programme.
Strengths and weaknesses of this model are listed in Box 2H.7.8.

Box 2H.7.8
Strengths Weaknesses
Useful for long-term, complex behaviour changes Less useful for programmes aimed at whole
(e.g. giving up smoking, weight management) communities
Useful for practitioners who wish to tailor their
counselling (and their expectations for change)
according to the stage of the model in which the
individual is currently located
Useful for programme planning, to organise
interventions sequentially, and to match
interventions to stages of the population

2H.8 RISK BEHAvIOUR

Risk behaviour in health and the effect of interventions in influencing health-related behaviour in professionals,
patients and the public
A person’s aversion or predilection to risky behaviour is influenced by several factors, including:
• Familiarity with the outcome of the risky behaviour
• Degree of personal control over the risk factor: in contrast to environmental risks, individuals tend to downplay
personal risks. This is due to beliefs of personal invulnerability and that other people are at greater risk (‘It
won’t happen to me’)
• Demographics (age, gender and ethnicity) – young people are more likely to take risks (partly due to greater
peer pressure) and women are more likely to be risk averse.

HEALTH RISK FACTORS

Factors known to be risks to health include those shown in Box 2H.8.1.

292

Public [Link] 292 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Box 2H.8.1
Harm Smoking, cannabis use
Harm and benefits Alcohol or food
Harm to others Unprotected sexual intercourse or driving while intoxicated

RISK INTERVENTIONS
Interventions can be implemented at different levels:
• Professionals – to reorient services
• Patients – to receive preventive treatment, make lifestyle changes
• The public – to make healthy choices and protect or promote the health of society.
See also Section 2F.2.

2H.9 COMMUNICATION IN HEALTH EDUCATION

Theory and practice of communication with regard to health education
There are several requirements for health education to be effective; it needs:
• To be received
• To be understood
• To stimulate a change in attitude
• To provoke behavioural change.
The success of health education messages depends not simply on what is said, but on how it is said (including
which media are used to communicate it).

HEALTH MESSAGES
As with all communication, there are four components of health messages: source, message, channel and receiver
(see Figure 2H.9.1).

Channel

Source Message Receiver

Figure 2H.9.1 Components of health messages

SOURCE
This is the person or organisation that generates the message. The credibility of a source depends on several factors,
including:

293

Public [Link] 293 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

• Source’s position in society

• Training and qualifications of the source
• Shared characteristics with the recipient (e.g. age, culture)
• Perceived conflict of interests.
For example, a health advice message from a health minister may not be as credible as one from a doctor.

MESSAGE
A message may be verbal (written or spoken words) or non-verbal (images or sounds), and it may be horizontal or
vertical (Box 2H.9.1)

Box 2H.9.1
Horizontal General lifestyle improvement messages (e.g. eating a healthy diet)
vertical Specific issue messages (e.g. advice not to binge-drink)

Health promotion messages can be designed to persuade or to empower.

CHANNEL
This is the medium or media through which a message is conveyed. Channels include:
• One to one (e.g. midwife−patient consultation)
• Small groups (e.g. antenatal classes)
• Drama, storytelling or songs
• Mass media (broadcast, internet, newspapers, leaflets).
The most appropriate setting for the communication to take place will depend on the message. Examples include:
• Home
• Schools
• Community centres
• Workplaces.
Integrated marketing communication (IMC) achieves greater efficacy with:
• A communication mix of multiple channels
• Use of public relations, advertising and promotion.

RECEIvER
The target audience should be the prime consideration of any communication.
Messages may be targeted, for example, at:
• Individuals, families or communities
• Adults, adolescents or children
• Men or women.

SUCCESSFUL COMMUNICATION
In order for a communication to improve health it has to fulfil several requirements: see Table 2H.9.1.

294

Public [Link] 294 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Table 2H.9.1 Requirements for successful communication

Requirement Description Example
Be seen or heard Will the target audience see/hear the A poster for young people is displayed in places
message? where young people go such as schools and
colleges
Attract attention Will the target audience notice the A newspaper insert is sufficiently enticing for
message? readers to notice it rather than throw it away
unread
Be understood Is the language intelligible to the The wording and images on a leaflet are pre-
target audience? tested on a sample of the intended audience to
ensure that it is readable and unambiguous
Be accepted Does the message reinforce current Stop-smoking advertising is often effective in
attitudes and beliefs? influencing smokers who have already decided
that they wish to quit. It is less effective on
those who do not want to give up
Change behaviour Are all the factors that could prevent In order to change their diet, people on low
behaviour change being addressed? incomes may require messages about healthy
nutrition to be supported by food subsidies

MEDIA
Different channels of communication may be suited to different aims and different settings. Often, however, they
will be most effective when used in combination. Table 2H.9.2 compares some features of mass communication
methods with small group methods.

Table 2H.9.2 Features of mass communication

Mass media Small group
Methods Broadcast: TV, radio, cinema Face to face: consultations, classes, groups
Print: newspapers, magazines, posters and At a distance: telephone, email, chat room
leaflets
Electronic: websites, email lists
Scope Many recipients Few recipients
Flexibility Low: one style produced for all recipients High (message can be tailored to the individual
(electronic media excepted) or group)
Feedback Low High − integral part of the process
Strengths Useful for reinforcing attitudes and behaviours Can be used to challenge current attitudes and
behaviours
Simple, unambiguous messages (e.g. stop
smoking) Useful for conveying complex messages (e.g.
benefits and risks of alcohol)
Ensures that same message reaches all
recipients

Table contd overleaf

295

Public [Link] 295 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Table 2H.9.2 contd

Mass media Small group
Weaknesses Weak link between mass media production Can reach only small numbers, so it is less well
and the receipt of information (e.g. the suited to messages that need to reach whole
information broadcast on TV may not be populations
watched, and even if watched, it may not be
Low profile without the mass media
listened to or understood)
Hard to control how information is disseminated
Limited scope to tailor information so that it
by individual practitioners
is more suitable for particular individuals or
groups
Expensive: mass media may be unaffordable
for local campaigns

2H.10 LEGISLATION AND HEALTH PROMOTION

Role of legislative, fiscal and other social policy measures in the promotion of health
Health promotion can make use of a range of levers at national and community level to create a supportive
environment that encourages health-seeking behaviour. These include those shown in Box 2H.10.1.

Box 2H.10.1
Legislation Description Example
Social policy Local, national or international health cultures and Free-market or regulatory
policies approaches to the economy
Restrictions and Discourage behaviours known to have a damaging effect Ban on children purchasing
bans on health cigarettes
Fiscal measures Systems of taxation to discourage certain behaviours, and Alchohol duty
subsidies to encourage others

SOCIAL POLICY
Social policy measures encompass a wide range of arrangements and structures designed to increase harmonisation
in society – be it at an international, national or local level (Bunton 2002). While some policies will specifically be
targeted at promoting health, other policies on a wide range of other issues will also influence health (see Section
2I.7).
A society’s dominant ideology will influence its policy development, coupled with the degree to which state
intervention is acceptable in that society. For example, cultures that prize the freedoms of the individual may be
less inclined to promote policies of collective actions, such as banning smoking in enclosed public places.

BANS AND RESTRICTIONS

Restrictions or bans on activities or goods can take several forms, including the following.

AvAILABILITY
Eng Cigarettes to young people: in England the Children and Young Persons (Protection from Tobacco) Act 1991
makes it illegal to sell cigarettes to under 16s. This age limit was increased to 18 years in 2007 and is enforcable

296

Public [Link] 296 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

by local authorities [[Link]/ACTS/acts1991/Ukpga_19910023_en_1.htm Children and Young Persons

(Protection from Tobacco) Act 1991 (c. 23)].
Eng Alcohol at particular times of day: in England, the Licensing Act 2003 has liberalised when alcohol can be
sold so establishments can, potentially, sell alcohol 24 h a day
UK Drugs/medicines: the Medicines Act 1968 controls the availability of medicines – ranging from drugs that are
on general sale to those that are available only on prescription.

USAGE
UK Smoke free public places. In Ireland this became law in 2004. Similar legislation was enacted in Scotland in
2006, Wales, Northern Ireland and England in 2007 ([Link]).

SALES
UK Illicit drugs. The Misuse of Drugs Act 1971 designates controlled drugs (both ‘medicinal’ drugs and those with
no known therapeutic benefit) into three classes – A, B or C – with corresponding restrictions on their availability
and penalties for selling or being in possession. The most severe sentences are for class A drugs and the least severe
for class C.

ADvERTISING AND INFORMATION

Eng Tobacco warnings. Compulsory labelling of cigarette packets with health warnings is stipulated in the Tobacco
Advertising and Promotion (Point of Sale) Regulations 2004, which also bans tobacco advertising except at the
point of sale ([Link]/si/si2004/[Link]).
These measures often require legislation, although voluntary industry codes of practice also exist (e.g. major soft
drinks retailers have voluntarily pledged not to target adverts at children). Restrictions and bans are only successful
if they are supported by favourable public opinion or by enforcement.

FISCAL MEASURES
Fiscal measures generally take place at a national level, and can involve either taxes or subsidies. Such measures
alter the price of goods to reflect the externalities associated with the particular action or goods. An externality is
a cost or benefit from consumption that falls on someone other than the consumer (e.g. herd immunity is a positive
externality of immunisation).

TAxATION
Taxation can serve two purposes:
1. To raise revenue
2. To decrease demand by increasing the price of the good to the consumer (or sometimes to the producer, e.g.
the polluter pays principle).
In the UK, taxation to influence health is largely limited to alcohol and tobacco. Its success depends on the price
elasticity of the goods (i.e. by how much the demand for a product is affected by its price). The price elasticity of
goods will be different for different groups within society. So, for example, UK tax rises on tobacco products have
had the greatest effects on:
• Young people
• Those on lower incomes.
Specific fees for particular activities can also be used to increase the price of producing or consuming goods (e.g.
the congestion charge levied for driving into central London during busy times). The arguments for and against
taxation as a health promotion measure are illustrated with respect to fatty foods in Box 2H.10.2.

297

Public [Link] 297 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

SUBSIDIES
Subsidies decrease the price of consumption of goods to the consumer in order to adjust for the positive (i.e.
beneficial) externalities associated with consuming them. Examples in health policy in England include the Healthy
Start Scheme (previously known as the Welfare Food Scheme) where families with young children receive vouchers
for liquid milk, infant formula milk or fresh fruit and vegetables so as to encourage the consumption of these foods
among children.
Note that there may be unintentional health effects associated with subsidies. For example, the Common
Agricultural Policy has subsidised tobacco farmers across Europe in a way that has encouraged tobacco production.
This subsidy is being gradually phased out, and in the short term it is being revised so that subsidies are no longer
linked to the amount of tobacco that is produced.

UK Box 2H.10.2
Example: Should fatty foods be taxed?
In January 2000, the British Medical Journal featured an article exploring the merits of imposing value-added tax
(VAT) on a range of common foods high in saturated fat. The revenue generated would be used to compensate
lower income families who would be hardest hit by the policy. The arguments for taxing high fat foods include:
• There are precedents for taxation to influence health, e.g. increasing tobacco duties
• Diet is partly responsible for ischaemic heart disease
• The costs of consumption therefore are partly borne by health services
• Increasing price could decrease the purchase and consumption of high-fat foods (assuming that these goods
are price elastic)
• The effects would be greatest on those on low income, who would be most sensitive to price changes and
who are at higher risk of heart disease
In contrast to the taxation of tobacco, however, the argument for foods was less persuasive because the
relationship between fats and heart disease is complex. Some fats are necessary and beneficial, and people
are affected by fat to a greater or lesser extent due to their genetic make-up. Since the article was published,
the prevalence of obesity has increased in the UK and it has become a key public health problem. As a result,
taxation of foods that could increase the risk of obesity is being reconsidered.
Adapted from Marshall (2000) and Marshall et al (2000).

2H.11 PROGRAMMES OF HEALTH PROMOTION

Methods of development and implementation of health promotion programmes
Health promotion programmes can be developed at local, national and international levels and therefore vary greatly
in their scope and aims. However, all health promotion programmes should be based on the aspects shown in Box
2H.11.1.

Box 2H.11.1
Evidence Available evidence of effectiveness: evidence-based practice (see Sections 1A.35 and 1A.34)
Theory Appropriate health promotion theory and other scientific disciplines
Need Local need (see Sections 1C.1 and 1C.2)
Resources Resources available locally, in terms of funding and human resources
Priorities Local and national priorities

298

Public [Link] 298 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

APPROACHES TO HEALTH PROMOTION

Programmes often include a combination of the following approaches:
• Changes in policy to shift culture and/or behaviour (see Section 2H.10)
• Distribution of resources to provide incentives or remove barriers to change
• Community development (see Section 2H.12)
• Information, communication and education to inform people of the risks and benefits associated with
behaviours, and to influence attitudes towards change (see Section 2H.9).

IMPLEMENTATION
Front-line clinical staff and members of the community conduct far more health promotion than do health promotion
or public health staff. Implementation therefore requires working with diverse practitioners to ensure that they are
equipped to carry out programmes. Some considerations are set out in Table 2H.11.1 and the issues in practice are
described with reference to smoking cessation in Box 2H.11.2.

Table 2H.11.1 Issues affecting implementation of health promotion

values Do those who are involved have the cultural and professional values to support the
programme’s aims? For example, if members of the community are opposed to sex outside of
marriage, then they will be unlikely to publicise or support community-based testing for sexually
transmitted infections in young people
Motivation What is the motivation for staff not connected with health improvement to change their
practice to improve health? This could involve providing incentives (e.g. financial bonuses
related to programme implementation)
Guidance Is it clear from the policies and guidance what exactly is required of staff and community
members? People may support a programme in principle but, if their role is unclear, then the
programme may not be implemented successfully
Skills Do those involved have the relevant skills and competencies to carry out their roles? If
staff are required to perform tasks that they have not conducted before, then they may need
additional training
Time Is there enough time to put changes in place? While consultations in general practice could be
an ideal opportunity to discuss behaviour changes with patients, it may not be feasible to do
this in a 10-min consultation (or shorter)

In order to ensure that a programme has been implemented effectively, it should be evaluated during or after the
programme’s completion (see Section 2H.14).

299

Public [Link] 299 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

UK Box 2H.11.2
Example: Reducing smoking among NHS staff
In an NHS trust, a programme to reduce smoking among staff involved:
• Responding to national policy and targets regarding smoking
• Staff surveys to find out the smoking prevalence and the attitudes of smokers towards giving up
• Research into guidelines, evidence of effectiveness and case studies of other organisations’ experience
• Application of relevant health promotion theory (e.g. stages of change model)
• Consideration of the resources available to support the programme
The implementation comprised the following steps:
• Creation of a policy regarding smoking at work, with staff involvement to ensure that the policy was workable
and that staff felt some ownership of it
• Dissemination of the smoking policy to ensure that staff were aware of whether they could smoke at work,
and the support available to help them stop
• This was followed by monitoring and enforcement of the policy, e.g. recording occasions where the policy
was not adhered to, and disciplinary action for staff smoking in no-smoking areas
• Ensuring a strong partnership with the local NHS stop-smoking service, with time being made available to
enable smokers to attend support groups
• Sufficient incentives for staff to give up (e.g. free or subsidised nicotine replacement therapy, stop-smoking
support groups and one-to-one help)
• Recruitment and training of sufficient smoking cessation advisers at the trust to support staff giving up

2H.12 COMMUNITY DEvELOPMENT

Building stronger communities is a key strategy for health improvement and for reducing inequalities. It is
fundamental to building healthy environments and providing individuals with the social support to adopt and
maintain health-seeking behaviours. Community development encompasses a range of activities to generate social
networks and to, empower people to shape their local services and have an input into their community. It can
stimulate innovative and creative solutions to problems that are not amenable to conventional health promotion
programmes. Its challenges lie in evaluating success and in ensuring that those in greatest need of stronger
communities actually benefit.

DEFINING THE COMMUNITY

A community is not a static entity. Rather, it consists of groups of people with a common characteristic at a
particular time, most commonly:
• Geographical (e.g. housing estates, villages)
• Social (e.g. workers’ groups, student unions, lesbian/gay communities)
• Cultural (e.g. religious, ethnic).
Terms such as community development, community participation and community renewal are often used
interchangeably. In contrast to social planning and other initiatives aimed at changing communities, community
development is a ‘bottom-up’ approach. Social planning would be described as ‘top down’.

METHODS
Community development uses a combination of activities. Smithies and Adams (1990) suggest that there are five
core activities of community development, described in Table 2H.12.1.

300

Public [Link] 300 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Table 2H.12.1 Activities of community development

Activity Description Examples
Formal Formal participation in decision-making Focus groups
participation
Consultation days
Community Priorities are developed by community groups Lobbying
action
Self-help activities
Facilitation Health service employees promote community activities Provision of meeting rooms and
refreshments
Interface Statutory services working closely with communities and Consultation with local imams
community leaders
Strategy Strategic support from national initiatives Neighbourhood renewal funds
Local strategic partnerships

Projects may progress from one type of activity to another as they mature, as illustrated in the example in Box
2H.12.1.

Box 2H.12.1
Example: cooking skills in the community
Top-down target Tackling obesity
 More user led

Formal A decision by the local strategic partnership to fund training in cooking skills for
participation local people following consultation with mothers who attend the local children’s
centre in a deprived area
Facilitation Local mothers attend cooking skills course funded by the PCT and local authority
Community action Trained mothers organise and provide training for other local mothers

ADVANTAGES AND CHALLENGES OF COMMUNITY DEVELOPMENT

The advantages and disadvantages of community development are outlined in Table 2H.12.2.

2H.13 PARTNERSHIPS
Working with people outside the field of public health is essential for delivering effective health promotion.

PARTNERSHIPS IN PUBLIC HEALTH

In the UK, Public Health is currently part of the NHS, itself both a provider and a commissioner of health services.
However, health is influenced by factors outside the remit of health care. According to Lalonde’s (1974) health field
concept, health is determined by:
• Genetics
• Environment
• Lifestyle/behaviour
• Health care.

301

Public [Link] 301 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Table 2H.12.2 Advantages and disadvantages of community development

Advantages Challenges
Initiating projects User led: can achieve better community Resource intensive
support if based on community priorities
Time-consuming
Can be difficult to secure funding
(especially given the unknown outputs)
Goals Can focus on root causes of ill health Long timescales: it may take years for
rather than simply lifestyle choice (e.g. health outcomes to appear and for
organising a food cooperative rather than communities to change
just providing dietary advice)
Evaluation/outputs The process of enabling communities to Results are often intangible and
participate is an end in itself unquantifiable
Enhances self-esteem, confidence and
control
Communities involved Can reach disadvantaged or excluded There may be a conflict of accountability
groups that conventional interventions for community development workers: do
would miss (see Section 2H.6) they work for the community or for the
statutory service?

Therefore, working with those actors (individuals and groups) who have influence over factors other than health care
(e.g. local councils and businesses) can have an impact on health.
Other reasons for partnership work include:
• Avoiding duplication
• Pooling resources – funding, experience, contacts, information, skills
• Political imperatives – duty to work together (e.g. health and social care provided to patients discharged from
hospital).

PARTNERS
Public health practitioners need to work with partners from a range of disciplines and organisations: see Box
2H.13.1.

Box 2H.13.1
Within organisation Primary care directorate, modernisation team, clinical governance officers
and community services
Within health Acute hospitals, mental health hospitals, neighbouring health authorities
Within the community Local authorities, education, police, voluntary sector, local businesses

TYPES OF PARTNERSHIP WORKING

Public health practitioners work in partnership in a number of ways: see Box 2H.13.2.

302

Public [Link] 302 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Box 2H.13.2
Type of partnership Example
Statutory committees Local strategic partnerships
Shared targets and Local area agreements
monitoring
Community strategies
Joint projects Production of joint reports (e.g. Health in London was jointly produced by the
London Health Observatory, the Health Development Agency and the Greater
London Authority
Specifically resourced SureStart
initiatives
HealthySchools
Teenage Pregnancy Partnerships
Shared posts Eng Director of Public Health employed jointly by a PCT and a local authority
Shared budgets Eng A cluster of PCTs may choose to implement one chlamydia screening
programme across all partners’ areas, with pooled budgets and pooled resources

BARRIERS AND CHALLENGES TO SUCCESSFUL PARTNERSHIP WORKING

Partnership working can be challenging and it can be easy to blame problems on the other members of the
partnership. However, many of the difficulties involved in partnership working are due to a mismatch in the ways
that the organisations or individuals within them work. Table 2H.13.1 outlines some of these factors and some
questions that might help to identify why partnerships are not working effectively.

Table 2H.13.1 Challenges to partnership working

Aspirations Do all partners want the same outcomes from a partnership?
Are partners working to different (even conflicting) performance drivers? (e.g. NHS is driven
by health outcomes whereas private industry is driven by profit)
Are all partners clear about what the objectives of the partnership are?
Processes Organisational cultures: do different organisations function in the same ways? Consider
Handy’s work (see Sections 5B.1 and 5B.3)
Are there compatible systems for making decisions? Do some members need to take decisions
back to in-house committees or boards? Can representatives authorise spending or actions at
the meeting on behalf of their organisation?
Team working: how do different members of the partnership work in a team? Consider Belbin’s
team roles (see Sections 5A.1)
Commitment The opportunity to build up a relationship can be lost if there is no continuity of people
attending meetings
Who attends the meetings?
Are they the right people?
Do they attend regularly?
Do the same people attend each meeting?
Table contd overleaf

303

Public [Link] 303 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Table 2H.13.1 contd

Influence Are people from equal levels of seniority attending the meeting?
As there may be no direct leverage over members of the partnership through line-
management, how can it be ensured that group members deliver on what they promise?
Power Big fish/little fish situations (e.g. statutory agencies versus user groups in the community)
How is the agenda developed?
Are all parties given plenty of time to contribute?
Where are meetings held: always in the PCT? Or out in the community?

MEASURING SUCCESS
Partnerships, like other aspects of health promotion, should be evaluated to ensure that they are meeting their
aims. There are a number of tools that could be used to do this but Donabedian’s evaluation model (see Section
1C.9) is a simple approach for considering what a partnership has achieved. Examples of successful indicators of
partnership under Donabedian’s three categories are shown in Box 2H.13.3.

Box 2H.13.3
Indicator Examples
Structure Joint funding
Joint posts
Process Internal and partnership plans are aligned
Meetings well attended
Outcome Objectives and milestones are achieved
Outputs (e.g. reports) are used and valued by the partners and by the local community

2H.14 EvALUATION
Evaluation of health promotion, public health or public policy interventions
The WHO Working Group on Evaluation in Health Promotion (2001) identifies the eight steps listed in Box 2H.14.1.

Box 2H.14.1
1. Describe
2. Identify
3. Design
4. Collect
5. Analyse
6. Recommend
7. Disseminate
8. Use

DESCRIBE THE PROGRAMME

Clarify the initiative’s mandate, timeframe, aims and objectives. Create a logical model showing the activities,
outputs, impacts, effects, objectives and goals.

304

Public [Link] 304 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Assemble an evaluation group

• Engage the stakeholders
• Clarify the purpose of the evaluation
• Identify key questions
• Identify evaluation resources.

IDENTIFY THE ISSUES AND QUESTIONS FOR THE EvALUATION TO ADDRESS

Consider the purpose of the evaluation:
• Formative (for an ongoing programme): to maximise the effectiveness of a programme. Provides ongoing
feedback and improvement
• Summative (once a programme has finished or reached a particular point): to decide to what extent the
programme met its aims
• Persuasive: if the aim of the evaluation was to evaluate a particular outcome.

DESIGN THE DATA COLLECTION

Choose measurement methods, considering which indicators will be used as criteria of success. Both qualitative and
quantitative data may need to be collected.

COLLECT THE DATA

Consider: confidentiality, validity, reliability of sources and information produced.

ANALYSE AND INTERPRET THE DATA

Consider:
• Comparisons with similar programmes elsewhere
• Differences between expected and observed results
• Strengths and limitations of the data (e.g. statistical issues such as chance and biases)
• Different interpretations are possible for any set of data
• Involve stakeholders to consider the implications of the findings.

MAKE RECOMMENDATIONS
Identify the costs and benefits of implementing recommendations and the costs of ignoring them. Involving
stakeholders in generating recommendations means that they will ‘already be committed to acting on the findings
and receptive to the results’.

DISSEMINATE THE FINDINGS

Key audiences include:
• Funders
• Practitioners considering implementation of a programme elsewhere
• Stakeholders of the original programme.

ACTION: USE THE FINDINGS AND RECOMMENDATIONS

Change ongoing programmes in light of the recommendations. Where changes are required in existing programmes,
change-management techniques should be used (see Section 5C.2).

305

Public [Link] 305 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

2H.15 INTERNATIONAL INITIATIvES

International initiatives in health promotion
The principal actors in international health promotion have changed since the middle of the twentieth century,
when there were relatively few agencies working across countries. The WHO (a subsidiary of the United Nations) had
the greatest role and was largely uncontested in its activities. As outlined by Walt (2001) and Lincoln and Nutbeam
(2005), today international health promotion is characterised by:
• A smaller, more contested role for the WHO
• A range of other UN organisations with a remit for health (such as UNICEF, the UN Population Fund and the
World Bank – which is currently the largest donor for health projects)
• Increasing activity from non-governmental organisations (NGOs), particularly from the private sector
• Bilateral activity between two countries’ governments (e.g. host country and the UK Department for
International Development, DFID)
• Resurgent interest in ‘vertical’ programmes that focus on discrete diseases (e.g. the Gates Foundation’s activity
focuses on particular diseases).

TIMELINE
Some key health promotion landmarks are summarised in Table 2H.15.1. Box 2H.15.1 provides examples of the
application of other major global health policy developments.

Table 2H.15.1 Key health promotion landmarks

Name Setting Year Description
Lalonde Report Canada 1974 Commissioned by the then Canadian Health Minister, Marc Lalonde,
and generally regarded as a key turning point in health promotion.
Established the health field concept of health (see Section 2H.2)
Health for All: Alma-Ata 1978 WHO stated the aim of providing universally accessible primary care,
Declaration of (Kazakhstan) which included health education
Alma-Ata
Health for All Canada 1986 WHO established the core principles of health promotion
2000: Ottawa internationally
Charter
Millennium UN 2000 Eight targets aimed at improving the lives of the world’s poorest
Development peoples
Goals
Global Fund International 2000 Global partnership among governments, civil society, the private
sector and affected communities to combat AIDS, tuberculosis and
malaria
Global Strategy WHO 2004 Global Strategy on diet, physical activity and health provides
member states with a range of global policy options to address the
growing problems of unhealthy diet and physical inactivity
Bangkok Charter Thailand 2005 WHO recognised the growing burden of communicable and non-
communicable diseases and called for greater coherence across
governments, international organisations and civil society

306

Public [Link] 306 3/6/08 [Link]

 2H Principles and Practice of Health Promotion

Box 2H.15.1
Examples:
Healthy Cities (1988)
Launched by WHO Europe to support cities in prioritising health improvement, and to provide resources and
guidance for cities to improve health. Over 1000 European cities now participate, linked through national and
international networks. The programme sets priorities for 5-year periods. A recent phase had three core themes:
healthy ageing, healthy urban planning and health impact assessment. In addition, all participating cities
focus on the topic of physical activity.
Reproduced from Healthy Cities and urban governance (2006).
Framework Convention on Tobacco Control (2005)
This framework was the WHO’s first treaty. It is a binding international legal instrument with broad commitments
and governance for national governments regarding tobacco control. It sets international standards on a range
of tobacco-related issues, including tobacco price and tax, tobacco advertising and sponsorship, labelling, illicit
trade and second-hand tobacco smoke.
Reproduced from the WHO Framework convention on tobacco Control (2005).

2H.16 INTERNATIONAL HEALTH PROMOTION INITIATIvES

Opportunities for learning from international experience
International agencies such as the WHO are central for disseminating learning on key issues, by means of the
methods shown in Box 2H.16.1.

Box 2H.16.1
Guidance and tool kits Where practice and evidence exists
Consensus statements Where evidence is not conclusive but advice is helpful
Expert networks Where opinion at an international level is required (see Section 1C.13)

Other systems for sharing experience include international scholarships, study visits, conferences and the internet.
Note that there are limitations to learning from international experience that may preclude an approach that was
successful in one country from being implemented or successful in another. These include:
• Different cultures
• Different health systems and funding of health care
• Demographic structures
• Varying disease epidemiology and health needs in different regions.
The example described in Box 2H.16.2 describes some of the problems in transferring lessons learnt about HIV
health promotion from western Europe to central and eastern Europe.

307

Public [Link] 307 3/6/08 [Link]

 2H P r i nc i p l e s and Practice of Heal th Promotion

Box 2H.16.2
Example: Transferring lessons learnt regarding HIv/AIDS prevention from western Europe to central and
eastern Europe
The HIV epidemic has been relatively minor so far in central and eastern European countries, providing an
opportunity for instituting preventive measures and for learning lessons from HIV prevention in western Europe.
Wright (2005), however, highlights the fact that there are diversities within Europe that may affect the extent to
which western European approaches can be used. One key issue is the cultural difference between some groups
of men who have sex with men and the gay identity.
In western Europe, the central role of the gay community in leading health promotion to prevent the spread of
HIV/AIDS was one of the ‘success stories’. In eastern European countries, it cannot be assumed that there is an
established gay culture among men who have sex with men. For example, legal restrictions on homosexuality are
severe in parts of central and eastern Europe
Wright argues that the focus should be on:
‘Assisting each country to adapt basic principles of HIV prevention to their current political and social situation.’
Reproduced from Wright (2005).

308

Public [Link] 308 3/6/08 [Link]

 2I Disease Prevention and Models of Beh aviour Ch ange

2I
Disease Prevention and Models of
Behaviour Change

2I.1 Prevention in the early years 309 2I.6 Community development 319
2I.2 Pre-determinants of health 313 2I.7 Health impact assessment 319
2I.3 Social marketing 314 2I.8 Strategic partnerships 319
2I.4 Involving the public 315 2I.9 Setting targets 319
2I.5 Deprivation and its effect on health 318

Following on closely from the theories explored in Section 2H, this chapter tackles the challenges, approaches and
priorities for improving health in practice. Preventive actions are arguably most important in children and families,
whose health can be particularly vulnerable to the effects of living in deprivation. However, as this chapter
discusses, robust evaluation and research of effective strategies is often lacking.
The principal tools used to improve health and to prevent disease are discussed. These range from social marketing
(a relatively new technique that is becoming increasingly widespread) to target setting (which has become integral
to the delivery of modern health services).

2I.1 PREvENTION IN THE EARLY YEARS

Evaluation of preventive actions, including the evidence base for early interventions on children and families, support
and emotional development
The goal of most public health interventions is to prevent disease and to maximise health. Interventions aimed at
pregnant women, babies and young children are particularly attractive to public health practitioners because of
their potential to influence health and wellbeing throughout life. Working with these groups holds some particular
challenges, not least because the outcomes are often realised only decades into the future.

EVALUATION OF PREVENTIVE ACTIONS

‘There is currently limited evidence about what works in terms of preventive and
public health interventions, how effectively to implement them, and even less
evidence on their impact on inequalities and the cost-effectiveness of these
interventions.’ (Wanless 2004 p 110)

309

Public [Link] 309 3/6/08 [Link]

 2I D i s e a s e P revention and Model s of B ehaviour Change

In commenting on the evaluation of preventive action, the Wanless report emphasised the need for sustained
support and funding, and observed that well-controlled studies are rarely performed in this field.
Evaluation of public health interventions may be complicated by a number of factors: see Box 2I.1.1.

Box 2I.1.1
Study design It has been argued that randomised controlled trials can be difficult to justify ethically for
preventive studies. However, without a well-designed study, biases and confounders cannot be
excluded. It is therefore difficult to identify whether the intervention was directly responsible for
the averted disease
Lack of It is not always possible to predict or model the course of events if a preventive action was not
control group put into practice. For example, it is impossible to tell what the disease occurrence would have
been had a vaccine programme not been delivered
Lead times Long time delays between an intervention and the manifestation of its effects can require lengthy
(and therefore costly and complex) studies, e.g. effect of early years’ education on achievement
at 18 years of age, or effect of smoking cessation on lifetime risk of cancer. Where effects are not
observed in the short term, it can be difficult to ensure continued support for preventive actions
See also Sections 1C.9, 1C.16, 2H.4 and 2H.14.

THE EVIDENCE BASE

Studies that investigate the impact of early interventions on children and families may be grouped into five areas:
1. Education
2. Health and nutrition
3. Socioeconomic benefits
4. Emotional and social support
5. Combined programmes.

EDUCATION OF THE CHILD AND PARENTS

Pre-school education can improve children’s social and intellectual development and long-term outcomes, as
illustrated in the examples in Boxes 2I.1.2 and 2I.1.3. Education for parents can include parenting skills but may
also involve ensuring that young parents are still able to access their own education.

HEALTH AND NUTRITION

Health promotion in the early years starts before birth with interventions to reduce the risk of low-birthweight
babies. Low birthweight is associated with higher infant mortality but also long-term effects, such as a higher
risk of chronic conditions in adult life (e.g. diabetes, heart disease). It is more common for parents of lower
socioeconomic background to have babies with low birthweight. A Health Development Agency systematic review
of published evidence identified two major modifiable factors that influence low birthweight and explored the
outcomes of intervention to reduce these risks, namely:
• Poor maternal nutrition at conception and during pregnancy can led to low-birthweight babies. While there is
evidence that calcium supplements and folate can be effective, there is little evidence about the effectiveness
of other supplements.
• Smoking during pregnancy doubles the risk of having a low-birthweight baby. Formal smoking cessation
programmes with nicotine replacement therapy can enable some pregnant women to give up smoking. However,
other factors (e.g. partner’s smoking status and mother’s socioeconomic group) affect the success of such
programmes.

310

Public [Link] 310 3/6/08 [Link]

 2I Disease Prevention and Models of Beh aviour Ch ange

EU Box 2.I1.2
Example: the Effective Provision of Pre-School Education (EPPE) Project
A cohort study of 3000 children across Europe considered the effects at age 6−7 of different forms of pre-school
education, and the circumstances of the home learning environment, upon children’s literacy, numeracy and
social development (e.g. anxiety, antisocial behaviour and positive social behaviours), together with the impact
on social inequalities. The findings included:
• Children who had attended pre-school (full or part time) showed higher educational and social attainment
than those who had no pre-school experience – even when home and social circumstances were taken into
account
• The quality of pre-school education influenced the educational levels achieved
• The home environment was important, but parents’ socioeconomic status did not affect children’s benefit
• The report concluded that ‘what parents do is more important than who they are …. Children whose parents
read to them, taught them letters and numbers, songs and nursery rhymes, and took them to the library had
better outcomes at 6 and 7 years’.
Reproduced from Sylva et al (2004).

The evidence indicates that breastfeeding can improve the health and wellbeing of infants and mothers.
Breastfeeding is less common in lower socioeconomic groups and is becoming less widespread among certain
minority ethnic groups that traditionally breastfed their babies. Prenatal support and education of both mothers and
health-care staff have been shown to improve breastfeeding rates.

USA Box 2I.1.3

Example: Head Start – long-term effects of comprehensive child development in the early years
The Head Start programme is a pre-school education scheme run for disadvantaged families and children in the
USA to reduce social, educational and health inequalities between children from disadvantaged backgrounds and
their peers. As well as pre-school education, Head Start provides a range of other services, including:
• Facilitating use of medical care for children (e.g. immunisations and dental health)
• Provision of healthy food and snacks
• Encouragement of parents’ involvement in their children’s education
Since Head Start was started in the 1960s, it has been possible to evaluate its long-term outcomes. A large-
scale survey of social, health and economic behaviours (called the Panel Survey of Income Dynamics) has been
conducted in a cohort of 8000 families since 1968. In 1995, adults aged 18−30 were asked as part of the survey
whether they participated in Head Start or other pre-schools as a child. The survey found that adults who had
attended Head Start were more likely to complete high school and attend college than their siblings who
attended other pre-schools.
The effects of the Head Start programme are not just restricted to educational achievement, but encompass
crime and economic benefits: Head Start graduates from African−American origins were less likely to have been
later charged or convicted of a crime than were their siblings who attended other pre-school programmes. Also,
white graduates in their 20s who had attended Head Start earned more than comparator groups.
Reproduced from Administration for Children and Families (2006) Head Start General Information, available online
at: [Link]/programs/hsb/, Garces ET (2002), Fight Crime: Invest in Kids (2006) Head Start reduces crime
and improves achievement, available online at: [Link]/reports/[Link].

311

Public [Link] 311 3/6/08 [Link]

 2I D i s e a s e P revention and Model s of B ehaviour Change

In the early years, interventions to promote maternal and child nutrition include:
• Education (e.g. nutritional content of foods, cooking skills)
• Subsidies (e.g. free school meals, free fruit in schools)
• Supplements (e.g. fluoridation of water supplies, nutritional supplements to food staples, such as bread).

SOCIOECONOMIC BENEFITS
In the UK, the Acheson Report (an official enquiry commissioned by the government in 1997) reviewed evidence
from a range of stakeholders and concluded that families with young children were at increased risk of poverty. Many
of these families found themselves in a ‘benefit-dependent poverty trap’, i.e. they were unable to seek work because
affordable childcare was unavailable. The enquiry recommended that poverty in young families should be reduced
through:
• Provision of accessible and affordable childcare
• Increased benefits (and the uptake of benefits) to pregnant women and to families with young children.

EMOTIONAL/SOCIAL SUPPORT
Family support programmes can be based at:
• Community level (i.e. addressing poverty, social isolation and lack of community resources)
• Individual or small group level (e.g. home visiting during pregnancy and after birth).
The aims of such support are to:
• Provide parents with respite, problem-solving skills, capacity and wellbeing (e.g. decrease incidence of postnatal
depression)
• Ensure physical, emotional and cognitive development in children
• Prevent child abuse.
Few studies have evaluated the long-term outcomes of family support. However, in 2005, the European Early
Promotion Project (an ongoing cohort study of approximately 1000 families across Europe) identified that training
health-care workers to support early parent−infant relationships leads to fewer psychosocial problems in young
children.

COMBINED PROGRAMMES
These programmes include elements of the above four types of intervention and are often run by multi-agency
teams. An example in England that developed in the late 1990s was the SureStart programme, now developed into
children’s centres.
An evaluation of this programme (Belsky et al 2006) reported that differences between areas with the Sure Start
Local Programme (SSLP) and comparison areas were limited, small and varied in degree of social deprivation. SSLPs
had beneficial effects on non-teenage mothers (better parenting, better social functioning in children) and adverse
effects on children of teenage mothers (poorer social functioning) and children of single parents or parents who did
not work (lower verbal ability). SSLPs led by health services were slightly more effective than other SSLPs.
They concluded that SSLPs seemed to benefit relatively less socially deprived parents and their children, but to have
an adverse effect on the most disadvantaged children. Programmes led by health services seem to be more effective
than programmes led by other agencies.

312

Public [Link] 312 3/6/08 [Link]

 2I Disease Prevention and Models of Beh aviour Ch ange

2I.2 PRE-DETERMINANTS OF HEALTH

Understanding of pre-determinants of health, including the effect of social cohesion on health outcomes
Pre-determinants of health are those factors that portend the determinants of health. However, the distinction
between pre-determinants and determinants is variable, e.g. income can be seen both as a determinant of health or
as a pre-determinant of a determinant such as housing.
See Section 2H.2.
Pre-determinants can be grouped in terms of individual and community material goods, policies and societal factors:
see Box 2I.2.1.

Box 2I.2.1
Material Policies Society
Sufficient and healthy food Minimum wage Social cohesion: the extent to which
a society is mutually supportive and
Pure water Health at work
minimises inequalities
Clean air Maternal services
Values and attitudes, e.g. the
Income Childcare balance between competitive and
cooperative approaches
Housing Benefits
Ethnic diversity and the tolerance of
Green spaces General education
different cultures
Languages

IMPACT OF PRE-DETERMINANTS OF HEALTH

Pre-determinants mediate changes in health through various routes, as described by Kahan (2005/2006), and shown
in a visual form in Figure [Link] is also illustrated in the example of social cohesion in improving health in
Europe: see Box 2I.2.2.

EU Box 2I.2.2
Example: Social cohesion as the driver for improving pre-determinants of health is itself affected by health
policies
The Council of Europe directs member states to consider social cohesion as ‘an essential condition for democratic
security and sustainable development’. As such, its policies are designed to reduce inequalities and enhance active
participation in the community.
The Council recognises not only that social cohesion influences the incidence of disease and death, but also that
it can be influenced by health policy. For example, the funding of health care through private insurance systems
may cause more inequalities than social insurance or tax-based systems. Hence, private insurance may adversely
affect social cohesion.
Reproduced from Council of Europe (2006).

313

Public [Link] 313 3/6/08 [Link]

 2I D i s e a s e P revention and Model s of B ehaviour Change

Respiratory
Health characteristic
health

Housing
Determinant
conditions

Income: Employment Tolerance of

Enabling practices: different
individuals to Enabling cultures:
afford individuals to Affecting
Pre-determinants adequate earn enough employment
housing to support opportunities
themselves and areas in
and their which
family people live

Figure 2I.2.1 Impact of pre-determinants of health

2I.3 SOCIAL MARKETING

See [Link].
The term social marketing was first used by Kotler and Zaltman in 1971. It describes the use of techniques of
commercial marketing to ‘sell’ a health message in order to benefit individuals and society. Social marketing
approaches involve the six steps described in Table 2I.3.1.

Table 2I.3.1 The social marketing process

Identify the Health approaches have traditionally segmented groups by risk/disease categories, e.g.
target group diabetic patients, HIV+ve or sociodemographic characteristics (e.g. age group, income
bracket). Marketing approaches use other tactics, e.g. psychographic, lifestyle characteristics
and consumption patterns
Research target Surveys and focus groups are used to assess:
group • Attitudes and beliefs
• Habits and lifestyles
• Needs
Set objectives Clear objectives are needed for the campaign, e.g. raise awareness or change behaviour
Develop the The message needs to be thoroughly pre-tested with target group to ensure that it is credible
message

314

Public [Link] 314 3/6/08 [Link]

 2I Disease Prevention and Models of Beh aviour Ch ange

Sell the message Sell the message through a mixture of considerations (4 Ps, see Box 2I.3.1)
Evaluate The campaign should be evaluated so as to monitor success and to guide refinements during
the campaign
Selling a health promotion message, like the sale of any commodity, involves consideration of the product, its price,
placement and promotion: see Box 2I.3.1.

Box 2I.3.1
Product It must be clear what exactly is being ‘sold’. For example, in a campaign to boost the proportion
of children receiving MMR, the product could be the procedure (delivery of vaccine), the service
(visit to nurse) or the outcome (immunity from measles). Each of these will have different appeals
for different groups
Price This is the relationship between the costs and benefits of the programme to the behaviour change.
For example, in the case of vaccination, some parents do not see measles as a serious disease and
therefore may not value the benefits of immunity. The price should be considered in economic
terms, i.e. not simply the financial cost but also the opportunity cost
Place The channel through which the message is communicated will affect who has exposure to it, and
therefore levels of awareness of the message among the audience. The type of message will also
affect which channels should be used
Promotion This can be achieved through various media and advertising
Marketing campaigns can include emails, mail-outs, text messaging, events, merchandising (e.g. red
ribbon for AIDS) and partnerships with commercial companies and third sector organisations

Social marketing has a number of strengths and weaknesses, as shown in Box 2I.3.2.

Box 2I.3.2
Strengths Weaknesses
Based on a clear understanding of the target group, Assumes that the individual is fully able to choose to
not on the health promoter’s perceptions of the group change behaviour, i.e. that health is an individual
choice and that socioeconomic barriers are not a factor
Clear objectives are integral to the approach
in health choices
Makes use of techniques that have been successfully
As in commercial marketing, there is a danger of
applied commercially
portraying only partial information in an effort to
change behaviour, e.g. ‘Just Say No’ is catchier than a
rounded picture of positives and harms of drug taking
Danger of reinforcing the same stereotypes and
attributes used by commercial marketing to sell
products such as equating health with physical
characteristics (e.g. youth, attractiveness, health, being
thin) or with moral attributes (e.g. being in control)

2I.4 INvOLvING THE PUBLIC

Involvement of the general public in health programmes and their effect on health care
Reasons for including the public in health programmes are shown in Box 2I.4.1.

315

Public [Link] 315 3/6/08 [Link]

 2I D i s e a s e P revention and Model s of B ehaviour Change

Box 2I.4.1
Improved treatment Patients who are not informed or involved in treatment decisions are less likely to be
outcomes concordant with treatment
Empowerment The process of participation can empower individuals and communities to understand
their own situations and to assume increased control over the factors affecting their
lives. This process can, in turn, enhance people’s sense of wellbeing and quality of life
Democracy Community participation in decision-making, planning and action is a human right
Integrated approaches Communities that are not restricted in their thinking by organisational boundaries can
help to develop integrated, holistic and cross-cutting approaches to address complex
issues
Better decisions Involving people can result in more responsive, effective and appropriate services
Ownership and Community participation is essential if interventions and programmes aimed at
sustainability promoting health, wellbeing, quality of life and environmental protection are to be
widely owned and sustainable

LEVELS OF INVOLVEMENT
The general public can be involved as patients or community members: see Box 2I.4.2.

Box 2I.4.2
Patients Individual patients discussing their own treatment decisions with
a practitioner
Community Geographical areas (e.g. housing estate, village)
members
Age group (e.g. children)
Condition-related groups (e.g. diabetes, mental health, self-help)

Brager and Specht (1973) described a health ladder that specifies levels of community involvement: see Table
2I.4.1.
A criticism of the health ladder approach is that it implies that organisations should be striving to reach the top of
the ladder with the community. However, full involvement of the community may not always be feasible or desirable:
often the appropriate level of involvement will be lower down the ladder.

PUBLIC INVOLVEMENT
Ways to involve the public in health-care programmes include working with individuals and groups. See also Section
2I.6.

WORKING WITH INDIvIDUALS

Individuals may be co-opted to work alongside professionals in health and social care partnerships.
Eng Local strategic partnerships (LSPs) are multi-agency bodies that function over an area of local government
or social care. They provide a forum where local chief executives can meet to discuss interagency working, and to
hold partnership structures in an area to account. Many partnership structures (e.g. prescribing committees, mental
health fora, coronary heart disease networks, cancer networks) contain individuals who will represent themselves
and their experiences to the professionals running the groups. Such people may not be typical of people with the

316

Public [Link] 316 3/6/08 [Link]

 2I Disease Prevention and Models of Beh aviour Ch ange

Table 2I.4.1 Brager and Specht’s ladder of involvement

Control Participants’ Examples
action
High Has control Organisation asks a community to identify the problems and make all key
decisions on goals and means; it is willing to help the community at each step to
accomplish goals
Has delegated Organisation identifies and presents a problem to the community; it defines limits
authority and asks the community to make a series of decisions that can be embodied in a
plan that it will accept
Plans jointly Organisation presents tentative plans that are open to change by those who
are affected; it expects to change its plans at least slightly and perhaps more
subsequently
Advises Organisation presents a plan and invites questions; it is willing to change plans
only if absolutely necessary
Is consulted Organisation tries to promote a plan; it seeks to develop support to facilitate
acceptance
Receives Organisation makes a plan and announces it; the community is informed and
information compliance is expected
Low None Community told nothing

same conditions and members will often need support from the chair of such groups to participate fully, but the
perspective that service users offer is invaluable. The impact of children’s involvement in the construction of Evelina
Children’s Hospital in London is described in Box 2I.4.3. The Expert Patient Programme in England also illustrates
ways to involve patients in their own care: see Box 2I.4.4.

Eng Box 2I.4.3

Example: Evelina Children’s Hospital
A children’s board composed of children treated at the existing hospital and living nearby helped
to design the new Evelina Hospital building. There is clear evidence of how children’s views
influenced the building design: children said that they did not want long straight corridors so
instead the building has a curvy ‘snake’ floor plan.
Reproduced from the Evelina Children’s Hospital Appeal, available online at [Link]/
hospital/[Link].

WORKING WITH GROUPS

Groups run by voluntary services, and community groups such as self-help groups, can be used to facilitate
participation in:
• Advocacy activities
• Social networking through organising face-to-face consultation events or attending events held in the
community.

317

Public [Link] 317 3/6/08 [Link]

 2I D i s e a s e P revention and Model s of B ehaviour Change

Eng Box 2I.4.4

Example: Expert Patient Programme
People with chronic illness receive training from lay volunteers who themselves have
chronic illness. The training helps them to understand their illness, develop coping
skills and minimise the impact of symptoms.

2I.5 DEPRIvATION AND ITS EFFECT ON HEALTH

Concepts of deprivation and its effect on health of children and adults
Deprivation can be material or social in nature and can encompass all aspects of life. It is associated with:
• Lower life-expectancy
• Higher risk of tobacco, alcohol and drug dependence
• Higher chance of developing a long-term illness.
See Section 1C.8 for measures of deprivation.

CONCEPTS OF DEPRIVATION
Deprivation manifests itself in a number of ways (Box 2I.5.1), and the longer that people are exposed to
deprivation, the greater its effects.

Box 2I.5.1
Area Issue
Housing Temporary accommodation, damp, overcrowded, poorly maintained housing
Environment High levels of crime; poor access to facilities and transport
Income Low income
Employment Low status posts, hazardous work, job insecurity
Education Lack of education during childhood and adolescence
Social exclusion Poor social support, isolation or abusive relationships

Poverty can be absolute or relative: see Box 2I.5.2.

Box 2I.5.2
Type of poverty Definition
Absolute poverty Lacks the basic material necessities for life
Relative poverty Lives on under 60% of the median national income

FACTORS THAT REINFORCE DEPRIVATION

A number of phenomena are known to result from and exacerbate the effects of deprivation: see Table 2I.5.1.

318

Public [Link] 318 3/6/08 [Link]

 2I Disease Prevention and Models of Beh aviour Ch ange

Table 2I.5.1 Factors that reinforce the effects of deprivation

Social exclusion Relative poverty leads to exclusion from society, e.g. from decent housing
Discrimination Racism, homophobia, ageism and discrimination against those who have received
psychiatric treatment or been in prison or care all serve to exclude people from accessing
services, and act as barriers to the opportunities open to other people (such as jobs,
housing and social networks)
Employment Unemployment and job insecurity lead to stress and health effects
The degree of control and level of demand in a job both influence health
There is a gradient across high-, middle- and low-ranking staff even when employed by
the same employer (as shown by, for example, the Whitehall study)
Stress The effects of ‘continuing anxiety, insecurity, low self-esteem, social isolation and lack
of control’ have both mental and physiological effects, such as increased risk of heart
disease
Antenatal effects Stress and lifestyle (e.g. smoking or ill health) during pregnancy have an effect on the
fetus, which may manifest as fetal abnormalities and low birthweight
Low birthweight is associated with increased risk in adult life of diseases such as
diabetes and coronary heart disease (see also Sections 2A.1, 2B.1.5 and 2B.1.11)

2.I6 COMMUNITY DEvELOPMENT

Benefits and means of community development, including the roles and cultures of partner organisations such as local
authorities
See Section 2H.12.

2I.7 HEALTH IMPACT ASSESSMENT

Health impact assessment of social and other policies
See Section 1C.17 for details of HIA. Box 2I.7.1 provides an example of the use of HIA on policies in London.

2I.8 STRATEGIC PARTNERSHIPS

Role of strategic partnerships and the added value of organisations working together
See Section 2H.13.

2I.9 SETTING TARGETS

Role of target setting, e.g. public service agreements, local authority agreements
See also Section 4D.4.
Reasons for setting targets in health care include:
• Adoption of management practices and culture in health services
• Improvement of performance and accountability
• Ensuring consistency across services.

319

Public [Link] 319 3/6/08 [Link]

 2I D i s e a s e P revention and Model s of B ehaviour Change

UK Box 2I.7.1
Example: Health impact assessment of policies in London
The London Health Commission conducted rapid health impact assessments of the Mayor of London’s draft
strategies that aimed to improve the lives of Londoners. These strategies included policies on social issues (such
as culture, children and young people). The health impact assessments took a largely pragmatic approach to
meet time and resource constraints. They involved workshops with key stakeholders, written submissions, and the
synthesis and review of available evidence.
An evaluation noted that the health impact assessments succeeded in:
• Influencing strategy: those drafting the strategy considered health because they knew that it would be
subject to an HIA and they later made revisions as a result of HIA recommendations
• Involving a wider group of stakeholders than would otherwise have been involved in the policies
• Providing the evidence base for decision-makers to make choices
• Raising the profile of HIAs
Challenges included:
• Short timescales
• No established quality standards for HIAs
• Gaps in research evidence
• Involving the right stakeholders
• Relative priority of different types of research
Reproduced from HIA Gateway, available online at: [Link], London Health Commission, available online
at: [Link]/[Link]#Top.

Targets can relate to each component of Donabedian’s framework (i.e. structure, process, output or outcome − see
Section 1C.6): see Box 2I.9.1.

Box 2I.9.1
Indicator Example
Structure Director of Public Health in post at 85% of PCTs
Process 80% of GP practices maintain a register of their diabetic patients
Output 98% of patients seen within 4 h of arrival in A&E
Outcome Cancer deaths reduced by 20% by 2010 relative to 1990 baseline

Targets can be set at the micro, meso or macro level: see Box 2I.9.2.

CHARACTERISTICS OF ‘GOOD’ TARGETS:

‘SMART’ targets comply with the points in the mnemonic of Box 2I.9.3.

320

Public [Link] 320 3/6/08 [Link]

 2I Disease Prevention and Models of Beh aviour Ch ange

Box 2I.9.2
Level Example
Organisational Personal development plan (PDP) agreed with line manager
Local Local area agreement (LAA) negotiated at council level and agreed with central government to
ensure that local targets reflect the local situation
National Public service agreement: in return for investment, the government sets minimum standards
to be delivered across a range of public services, including health, education and crime
prevention (HM Treasury 2007)
International WHO’s ‘Health For All’ targets

Box 2I.9.3
S Specific Relate to what they want to achieve
M Measurable Defined indicators (quantitative if possible) to show if the target is met
A Achievable Achievable but challenging, in order to improve performance
R Relevant Relevant to current performance. For example, lung cancer deaths are mainly due to
smoking practices 30 years ago rather than current health service interventions
T Timescales These must be defined in advance for each target

Rewards (or punishments) attached to meeting (or failing to meet) targets can lead to particular areas of health
care receiving particular managerial interest. For example, smoking cessation and tobacco control are now discussed
at board level in English PCTs as a result of the quitters’ target. However, the high stakes involved (e.g. risk of job
loss or attainment of foundation trust status) can lead to gaming and distortion of priorities.

ADVANTAGES AND DISADVANTAGES OF TARGETS

Strengths and weaknesses of targets are described in Table 2I.9.1.

Table 2I.9.4 Targets – strengths and weaknesses

Strengths Weaknesses
• Provide a focus to performance improvement • Often, information is unavailable to measure
meaningful health outcomes
• Priorities are set explicitly
• Distortions of practice or gaming. For example,
• Create a level playing field for the organisations
some hospitals attempted to reclassify their A&E
under comparison (e.g. hospitals, local authorities)
trolleys as beds in order to meet the waiting
• Number and domains of targets can be set to time target. Elsewhere, ambulance services were
reflect organisational priorities and goals de-prioritising people in rural areas because they
knew that they had no chance of meeting transport
• Opportunity for sharing good practice and learning
time target
• Distort priorities – those areas of health care that
are not amenable to targets are paid insufficient
attention
• Disengagement of clinicians if targets are externally
set, or set from the top down

321

Public [Link] 321 3/6/08 [Link]

Public [Link] 322 3/6/08 [Link]
 Section 3
HEALTH INFORMATION

Health information is essential for health service planning and evaluation. Without it, services would be
unresponsive to changes in circumstances – and it would be impossible to make predictions or to increase efficiency.
There is a vast array of information available on populations, sickness and health. However, the quality and
appropriateness of this information vary widely too.
Section 3 provides practitioners with an appreciation of what information is available, how data are collected, and
the advantages and disadvantages of their use in public health.

323

Public [Link] 323 3/6/08 [Link]

Public [Link] 324 3/6/08 [Link]
 3A Populations

3A
Populations

3A.1 Conduct of censuses 325 3A.8 Effect of population structure on fertility,

3A.2 Collection of routine and ad hoc data 328 mortality and migration 338
3A.3 Demography 329 3A.9 Historical changes in population size and
structure, and factors underlying them 339
3A.4 Major demographic differences 330
3A.10 Effects of demographic change on health
3A.5 Methods of population estimation and
care 340
projection 334
3A.11 Policies to address population growth
3A.6 Life-tables 336
nationally and internationally 340
3A.7 Population projections 337

Public health practitioners require an understanding of the population, e.g. calculating rates and risk ratios requires
a viable denominator as well as a count of disease occurrences. The population size, demography and social
characteristics in most developed countries are now enumerated by regular censuses. This chapter discusses the
methods by which census information is obtained and, where this information is not available or sufficient, the
methods by which populations are estimated. It also summarises trends in population structures across time,
comparisons between different regions and approaches to address the health consequences of population changes.

3A.1 CONDUCT OF CENSUSES

A census is a snapshot enumeration survey of all the people in a country; as such, it provides the most complete
set of population data. Its findings are of use within the health, housing, education and transport sectors for:
• Making comparisons between regions
• Targeting resources
• Planning future resource allocations.
Most countries hold a census every 5–10 years, often enshrined in the constitution. They can be conducted either by
interviewing or by self-enumeration (with a post-enumeration survey to assess under-enumeration).

325

Public [Link] 325 3/6/08 [Link]

 3A Po p u l a t io ns

Questionnaires can be delivered and collected by the following means:

• Traditional ‘drop-off’ and ‘pick-up’
• Post-out and post-back
• Internet.

UK UK CENSUSES
The first UK census was held in 1801. They have been held decennially since then, except during wartime. The 2001
census cost ª £250m. The Census Act makes completion of the census form compulsory: non-responders face a
£1000 fine.

UK 2001 CENSUS
See Table 3A.1.1.

Table 3A.1.1 A summary of the conduct of the 2001 census

Preparation Planning began with a consultation exercise
There was then a parliamentary debate on the choice of questions, followed by a publicity exercise
Census forms were designed for self-completion by the so-called ‘head of the household’
Delivery Census forms were delivered to each household by a ‘field force’
The forms requested information about ‘persons present’ and ‘usual residents temporarily absent’
88% of forms were posted back to temporary local offices and 6% handed back at the doorstep,
with 4% collected by the field force
Content Accommodation
Relationships between people living in the household
Demographics (age, sex, marital status)
Migration
Ethnicity (see Section 3A.4)
Self-reported health
Qualifications, employment, journey to work
In Wales there was also a question about the Welsh language
Analysis Completed forms were first scanned
Scanned forms were then coded using automated and manual systems (a process that took
approximately 1 year to complete)
Paper forms were destroyed (with the scanned images being stored to be made public 100 years
later)
Data were quality assured: 2001 was branded the ‘one number census’ since missing responses were
estimated in order to adjust for under-enumeration
Results were then tabulated into databases according to output areas and super-output areas (see
Section 1A.18 for details), ready for analysis
A Census Output Prospectus was published

326

Public [Link] 326 3/6/08 [Link]

 3A Populations

UK CENSUS SMALL AREAS

Historically, census information was collected by enumeration district, and local data were generally analysed at
the level of the electoral ward. However, enumeration districts and electoral wards varied greatly in size (the latter
ranging from 100 residents to 30 000+) and were subject to frequent boundary changes – causing problems for
longitudinal study.
For the 2001 census, output areas (OAs) were designed (using a geographical information system) that are:
• Co-terminous with postcodes
• Uniform in population size
• Compact in shape
• Socially homogeneous.
A new hierarchy of super-output areas (SOAs) groups OAs into units that are similar in population size and are
highly stable. There are three levels of SOA, the lower SOA representing four to six OAs: see Box 3A.1.1.

Box 3A.1.1
Unit Approximate
population size
Output area 300
Lower SOA 1500
Middle SOA 7000
Upper SOA 25 000

COLLECTION IN OTHER COUNTRIES

HK In Hong Kong, censuses are conducted every 10 years, with by-censuses in the middle of the intercensal
periods. The 2006 by-census involved 1 in 10 households. Data were collected by thorough interviews. The 2001
population census used two types of questionnaires:
• A short form (in six of seven) households on basic characteristics. Self-enumeration forms were mailed to the
householders and collected by enumerators.
• A long form administered to one of seven households collecting data on a broad range of socioeconomic
characteristics through face-to-face interviews.
The thematic household survey (THS), a further data collection tool, is used in Hong Kong. This provides information
on health status, patterns of health service utilisation and health-care service expenditure profiles, and was
performed in 1999, 2001, 2002 and 2005.
NZ Aus In New Zealand and Australia, censuses are administered using a similar approach to the UK:
• The Australian Bureau of Statistics conducts a nationwide census every 5 years
• Statistics NZ conducts a census every 5 years. The 2006 census allowed completion via the internet for the first
time in this country.

HARD-TO-COUNT GROUPS
The following groups are typically difficult to enumerate:
• Young, inner-city men
• Multiple-occupancy buildings and student houses

327

Public [Link] 327 3/6/08 [Link]

 3A Po p u l a t io ns

• Those who do not speak the official language(s) of the country

• Babies
• Very elderly people
• Military personnel.

ALTERNATIVES TO TRADITIONAL CENSUSES

Between census years, planners must compromise between using either out-of-date results from the most recent
census or more up-to-date results from less robust sources. The subsequent census often leads to dramatic revisions
in statistics. For these reasons, some continental European countries are switching to:
• Rolling censuses
• Population registers with sample surveys (advantages and disadvantages of which are listed in Box 3A.1.2)
• Population projections.
Box 3A.1.2
Advantages of population registers Disadvantages of population registers
More up-to-date results Not a snapshot, so complicated to compare regions
Linked statistical database allows multivariate analysis Loss of ‘brand’ ➔ lower response rates
Improved planning, provision and monitoring of surveys ➔ results have less impact
More consistent statistics Depends on quality and availability of administrative
data
Supports evidence-based policy
Actual risk of confidentiality breaches
Improved efficiency and quality from permanent systems
Perceived risk of confidentiality breaches

3A.2 COLLECTION OF ROUTINE AND AD HOC DATA

The raw data used by researchers and public health practitioners either can come from routine sources (see Section
1A.1) or may be specifically commissioned.

ROUTINE DATA
Databases of routinely collected data are a plentiful source of health information – often covering entire populations
and spanning many years. Their advantages and disadvantages are listed in Box 3A.2.1.

Box 3A.2.1
Advantages Disadvantages
Cheap Limited to what is actually collected
Can be complete (e.g. register of births) Difficult to assess quality control
Large numbers of subjects Access sometimes restricted
Prospective, therefore avoid recall bias Potential delays in publication
Particularly useful when different data sources are Data linkage complex
linked

328

Public [Link] 328 3/6/08 [Link]

 3A Populations

AD HOC DATA
These data can be obtained either by commissioning a specific data collection exercise (e.g. a patient survey), or by
requesting ad hoc extracts from a routine data source (e.g. from a cancer registry). Advantages and disadvantages of
such data are listed in Box 3A.2.2.

Box 3A.2.2
Advantages Disadvantages
Can specify exactly what data are to be collected Sampling frame may be unknown
Can target data collection to the subgroup of interest Potentially costly
Can collect qualitative data May be difficult to link to routine data sources
Rapid data collection sometimes possible Typically the number of subjects is small
Quality can be readily assessed Data linkage is complex

3A.3 DEMOGRAPHY
This is the study of the characteristics and dynamics of human populations. Population change may arise because of
any of the following four factors:
• Births
• Deaths
• Migration
• Ageing.
Table 3A.3.1 provides a summary of important demographic concepts.

Table 3A.3.1 Important demographic concepts

Concept Description
Total population Measured at each census, this represents the historical trend in population growth or
size shrinkage
Age structure Ages are divided into a series of bands (e.g. 0–4 years, 5–9 years), and the population is
described according to the number of people in each band
Fertility This is the number of offspring per female by age band
Overall fertility rate = number of live offspring per 1000 per year in women aged 15–49
(Note: fecundity = number of offspring biologically possible per female)
Mortality This is the count of the number of deaths per year by age bands. It is the calculated
probability of a person in that age band dying each year:

= (Number that died per year) ¥ (Length of age band)

(Total number of individuals in each age band)
Survival Probability of survival = (1 – Probability of dying)

Once these parameters have been calculated, then population projections can be made (see Section 3A.7).

329

Public [Link] 329 3/6/08 [Link]

 3A Po p u l a t io ns

3A.4 MAJOR DEMOGRAPHIC DIFFERENCES

Important regional and international differences in populations, in respect of age, sex, occupation, social class,
ethnicity and other characteristics
Some of the differences in demographics between different parts of the UK are shown in Table 3A.4.1.

UK Table 3A.4.1 Differences in demographics between different parts of the UK

Constituent Number of people Percentage of Percentage Population Percentage of
country at 2001 census UK population population growth density population
1993–2003 (number/km2) non-white
(58 789 194)
England 49 138 831 84 +3.6 383 9

(London) (7 172 036) (12) (+5.4) (4700) (29)

Northern Ireland 1 685 267 3 +4.1 125 1
Scotland 5 062 011 8 –0.7 65 2
Wales 2 903 085 5 +1.9 142 2
Reproduced from Office for National Statistics (2001 Census).

AGE
Age distributions can be represented graphically as population pyramids: see Figures 3A.4.1–3A.4.5.

75–79
70–74
65–69
60–64
55–59
50–54
45–49
40–44
35–39
30–34
25–29
20–24
15–19
5–15
0–4

3 2 1 1 2 3
Figure 3A.4.1
Millions Males Females Population pyramid –
England 2005
330

Public [Link] 330 3/6/08 [Link]

 3A Populations

75–79
70–74
65–69
60–64
55–59
50–54
45–49
40–44
35–39
30–34
25–29
20–24
15–19
5–15
0–4

3 2 1 1 2 3
Figure 3A.4.2
Bars represent 100 000s Males Females Population pyramid –
Ireland 2005

75–79
70–74
65–69
60–64
55–59
50–54
45–49
40–44
35–39
30–34
25–29
20–24
15–19
5–15
0–4

6 5 4 3 2 1 1 2 3 4 5 6
Figure 3A.4.3
Bars represent 100 000s Males Females Population pyramid –
Australia 2005
331

Public [Link] 331 3/6/08 [Link]

 3A Po p u l a t io ns

75–79
70–74
65–69
60–64
55–59
50–54
45–49
40–44
35–39
30–34
25–29
20–24
15–19
5–15
0–4

4 3 2 1 1 2 3 4
Figure 3A.4.4
Bars represent 100 000s Males Females Population pyramid –
Hong Kong 2005

80+
75–79
70–74
65–69
60–64
55–59
50–54
45–49
40–44
35–39
30–34
25–29
20–24
15–19
5–15
0–4

5 4 3 2 1 1 2 3 4 5
Figure 3A.4.5
Bars represent 100 000s Males Females Population pyramid –
Sierra Leone 2005
332

Public [Link] 332 3/6/08 [Link]

 3A Populations

GENDER
UK In the UK, slightly more boys are born each year than girls (330 600 boys were born in England and Wales in
2005, compared with 315 235 girls), but overall there are fewer males than females in the UK population (29 271 000
males in the UK in mid-2004, compared with 30 563 000 females). This is because, from age 22 upwards, there
are more women in each age group due to higher female immigration and lower female deaths from accidents and
suicide. The gap narrows for those in their 40s (more immigration of men in this age group) and then widens (due to
longer female life-expectancy and the effect of World War II).
There are differences in the male-to-female ratio across ethnic groups: see Box 3A.4.1.

Box 3A.4.1
More men More women
Pakistani White
Bangladeshi Black
Chinese Indian

ETHNICITY
UK People whose ethnicity is non-white tend to live more commonly in England than in other parts of the UK,
especially in London and the West Midlands. Black, Bangladeshi and Irish populations are particularly concentrated
in London. See Box 3A.4.2.

Box 3A.4.2
England London Wales Scotland Northern Ireland
White (%) 91.0 71.2 97.9 98.0 99.3
Mixed (%) 1.4 3.2 0.6 0.3 0.2
Asian (%) 4.6 12.1 0.9 1.1 0.2
Black (%) 2.3 10.9 0.2 0.2 0.0
Chinese and other (%) 0.8 2.7 0.4 0.4 0.2
Reproduced from ONS, Northern Ireland Statistics and Scottish Executive Statistics, with permission from the Controller
of HMSO.

ETHNICITY AND AGE

UK Non-white groups have a younger age structure than the white population, as a result of past immigration and
fertility patterns. The Irish population is the oldest, and the mixed group has the youngest structure. Because of
differences in mortality rates, women aged 65+ outnumber men. However, the migration patterns of certain groups
mean that there are fewer elderly women in the Pakistani and Bangladeshi populations.

MEASURING ETHNICITY
UK The 2001 Census question on ethnicity asked, ‘What is your ethnic group?’ and allowed respondents to choose
from one of the 16 options shown in Table 3A.4.2.
As different ethnic populations mix and inter-marry, the fastest growing ethnic group is now people of mixed
ethnicity. The 2001 UK census was the first census to ask about the backgrounds of this ethnic group (Table 3.4.2)
and, although very useful, this has made comparisons between censuses difficult.

333

Public [Link] 333 3/6/08 [Link]

 3A Po p u l a t io ns

Table 3A.4.2 Classification of ethnicity according to the 2001 census

A White 1. British
2. Irish
3. Any other white (write in)
B Mixed 4. White and black Caribbean
5. White and black African
6. White and Asian
7. Any other mixed (write in)
C Asian or Asian British 8. Indian
9. Pakistani
10. Bangladeshi
11. Any other Asian (write in)
D Black or black British 12. Caribbean
13. African
14. Any other black (write in)
E Chinese or other 15. Chinese
16. Any other (write in)
Other related questions were, ‘What is your country of birth?’ and, ‘What is your religion?’ The latter was a voluntary
question.

3A.5 METHODS OF POPULATION ESTIMATION AND PROJECTION

UK The UK’s Office of National Statistics produces mid-year population estimates on 30 June each year. These are
estimates of the resident population and are calculated using the cohort component method.

COHORT COMPONENT METHOD

See Table 3A.5.1.

Table 3A.5.1 Cohort component method for mid-year population estimates

Steps Data sources
1. Take previous mid-year estimate Previous mid-year estimate (or the census findings in
census years*)
2. Increase the population’s age by 1 year
3. Add births Registrar General’s office
4. Subtract deaths Registrar General’s office
5. Adjust for external migration International Passenger Survey
Irish National Household Survey (for migration
between the UK and the Republic of Ireland)
6. Adjust for internal migration GP registration data (linked by NHS number)
7. Quality control (check for consistency and against
previous estimates)
*Population estimates are more reliable in census years: the cohort component method is still used, but the census
population needs to be adjusted only by the number of weeks between the census and 30 June.

334

Public [Link] 334 3/6/08 [Link]

 3A Populations

SPECIAL GROUPS
The groups listed in Box 3A.5.1 are not included in the general population calculation, but are estimated separately.

Box 3A.5.1
Special group Information source
Boarding school pupils DfES/Welsh Assembly Government
Prisoners (only if >6 months in prison) Home Office
Home Armed Forces Defence Analytical Services Agency
Foreign Armed Forces US Forces

ESTIMATES
UK Population estimates are produced for the populations listed in Box Box 3A.5.2
3A.5.2, according to age, sex and marital status.
• UK as whole
PROJECTIONS • Constituent countries
UK In the UK, population projections are produced by the Government • Government office regions
Actuary’s Department (GAD) and are used for planning across government
• Local authorities
sectors. The core function of the department is to produce, every 2 years,
25-year projections of the population of the UK and its constituent nations, • Health authority areas
according to age, sex and marital status. Less accurate 70-year projections
• Primary care areas
are also made.

REPLACEMENT FERTILITY
In the absence of migration, the growth or decline of a population depends on sustained patterns in replacement
fertility. Changes in replacement fertility are slow to take effect because of population momentum, i.e. large cohorts
of the population in childbearing years will continue to have high numbers of births even if fertility falls. However,
in the long run, replacement fertility depends upon three factors:
• Fertility
• Birth sex ratio (males to females)
• Female mortality before the end of reproductive age.
These factors are usually predicted on the bases shown in Box 3A.5.3.

Box 3A.5.3
Factor Basis of forecast
Birth sex ratio Stable at 1.05 males per female
Mortality Extrapolate historical trends
Fertility Judgement
Migration Judgement

The uncertainty is greatest regarding fertility and migration. Traditionally, all factors have been forecast along
smooth paths, but there is now a move towards stochastic predictions, where the probabilities of random
fluctuations are incorporated. Likewise, the uncertainty regarding population forecasts has traditionally been

335

Public [Link] 335 3/6/08 [Link]

 3A Po p u l a t io ns

expressed as a range (lower to upper), but probabilistic population forecasting (with 95% confidence intervals) is
now superseding this.

3A.6 LIFE-TABLES
Life-tables and their demographic applications
Life-tables, also known as mortality-tables, list the probabilities (according to age and sex) that a person will die
before their next birthday. In the UK these tables are generated by the Government Actuary’s Department for the
country as a whole and for its constituent nations.
Life-tables are of two principal types: period and cohort.

PERIOD LIFE-TABLES
These are calculated using the age-specific mortality rates for a given historical period (either a single year, or a run
of years), with no allowance made for any later actual or projected changes in mortality.
A period life-table displays the life-expectancy of people of a given age in a given year if they experienced that
year’s age-specific mortality rates for the rest of their lives: see Table 3A.6.1.

Table 3A.6.1 Example of a period life-table of people aged 70 in 2005, calculated using mortality rates
Age (x) Central rate Probability that Number of people who Number of people Average life-
of mortality* someone will die survive to age x who die aged x expectancy at
before their next age x
birthday
lx+1 = lx – (1 – qx) dx = lx – lx+1
(x + 1)

Age x mx qx lx dx ex

70 0.027274 0.026907 100 000 2690.7 12.86

71 0.030869 0.030400 97 309.3 2958.2 12.20
72 0.034271 0.033694 94351.1 3179.1 11.56

For period life-tables, use the mortalities:

• For age 70 in 2005
• For age 71 in 2005
• For age 72 in 2005, etc.

*Number of deaths in people aged x over the past 3 years divided by the average population at that age in the same
period.
Adapted from GAD website: [Link].

Official life-tables that relate to past years are generally period life-tables.

336

Public [Link] 336 3/6/08 [Link]

 3A Populations

COHORT LIFE-TABLES
In contrast, cohort life-tables are calculated using age-specific mortality rates which do allow for known or
projected changes in mortality in later years.
A cohort life-table displays the average life-expectancy of a group of people of a given age in a given year if they
experienced projected future age-specific mortality rates from the series of future years in which they will actually
reach each succeeding age if they survive: see Table 3A.6.2.

Table 3A.6.2 Example of a cohort life-table of people aged 70 in 2005, calculated using projected future
mortality rates
Age Central rate Probability that Number of people who Number of people Average life-
(x) of mortality* someone will survive to age x who die aged x expectancy at
die before their age x
next birthday
lx+1 = lx – (1 – qx) = lx – px dx = lx – lx+1
(x + 1)
Agex mx qx lx dx ex
70 0.027274 0.026907 100 000 2690.7 12.86
71
72

Cohort life-tables would use the

predicted mortalities of:
• Age 70 in 2005
• Age 71 in 2006
• Age 72 in 2007, etc.

*Number of deaths in people aged x over the past 3 years divided by the average population at that age in the same
period
So, if mortality rates are projected to fall in the future, then the cohort life-expectancy at a given age will be longer
than the period life-expectancy for that age.

APPLICATIONS OF LIFE-TABLES
Life-tables are used for planning in all sectors of government, but particularly with regard to pensions and social
insurance. Several values can be derived from life-tables, including:
• Proportion of people born in different years who are still alive
• Remaining life-expectancy of people at a particular age
• Probability of surviving to a particular age.
As well as separate life-tables for men and women, it is also possible to distinguish other factors that affect
mortality, including ethnicity, social class and smoking status.

3A.7 POPULATION PROJECTIONS

See Sections 3A.5 and 3A.6.
UK Note that, in making official predictions, the Office of National Statistics assumes that there will be
improvements in mortality and increases in net immigration.

337

Public [Link] 337 3/6/08 [Link]

 3A Po p u l a t io ns

3A.8 EFFECT ON POPULATION STRUCTURE ON FERTILITY, MORTALITY AND MIGRATION

If the effects of migration are ignored then, for the population size to remain stable, the average female needs to
have one female child who survives long enough for one female grandchild to be born. This level of fertility – called
the replacement fertility – needs to be higher than two children per female in order to compensate for:
• Mortality (not all daughters will survive to the end of reproductive age, especially in developing countries)
• Unbalanced sex ratio at birth (in the UK the ratio of male births to female births is approximately 105:100).
Replacement fertility values vary between countries (see Box 3A.8.1), and the commonly quoted figure of 2.1
children applies only to the developed world. International variation is mostly due to mortality differences –
particularly with respect to HIV/AIDS.

Box 3A.8.1
World region or country Replacement fertility
Reunion 2.05
Europe 2.10
Africa 2.70
Sierra Leone 3.43

UK In the UK, replacement fertility has fallen because of decreasing mortality of the young. Again ignoring
immigration, if period fertility drops below replacement fertility, then the size of the population will eventually fall.
However, population decline may not be observed immediately because of the buffering effects of other factors,
including:
• Age structure of the population
• Changes in age-specific mortality rates
• Childbearing postponement.
The last phenomenon has occurred in several developed countries in recent decades. It has the effect of stretching
out the population into the future so that there are fewer people alive at any moment in time.

POPULATION TIME BOMB

Population decline in
the UK would lead to
problems often referred Low fertility
to as the population time
bomb: see Figure 3A.8.1. ‘Population time bomb’
Falling birth rates and Working population needs
low mortality in older age to support more
groups both result in an Ageing population dependants
ageing population. As the Implications for pensions,
number of people who productivity, economy,
are economically active housing, carers
relative to those requiring
health and social care Low mortality
at older ages
starts to fall, the resulting
effects may have profound
implications for society. Figure 3A.8.1 Population time bomb

338

Public [Link] 338 3/6/08 [Link]

 3A Populations

In economic terms, a country achieves its optimum population when productivity per capita is highest.
• Under-populated countries are those that can increase their productivity by increasing their population.
• Over-populated countries are those that can increase productivity by reducing their population.

IMMIGRATION
See 4C11.
Net migration is the difference between the number of people emigrating and immigrating. Both immigration and
emigration occur most commonly among young adults, with slightly more males than females migrating each year
overall.
UK In the UK, both immigration and emigration have increased in recent years. Immigration into the UK has been
predominantly of citizens from the 10 accession countries that joined the European Union in 2004. Emigration has
been mostly in people from the 25–44 age group, and has been mainly to other EU countries, and to Australia and
New Zealand.

3A.9 HISTORICAL CHANGES IN POPULATION SIZE AND STRUCTURE, AND FACTORS UNDERLYING
THEM
Over the course of centuries, large Box 3A.9.1
populations have changed with regard
to their age structure and geographical Historical phase Population Fertility and Age
distribution. The populations of many growth mortality
countries can be seen to have fitted into Pre-industrial Slow High Young
three phases: pre-industrial, industrial and
Industrial Fast Intermediate Intermediate
post-industrial: see Box 3A.9.1
Post-industrial Slow Low Old
A more detailed ‘demographic transition
model’ describes five phases rather than
three:
1. Pre-industrial
2. Developing Stage 1 2 3 4 5
3. Urbanisation
4. Developed Birth rate
5. De-industrialised (i.e. switch 25 40
from manufacturing to service-
based economy).
Death rate per 1000 population

Birth rate per 1000 population

20
See Figure 3A.9.1. 30

15 Death rate
20
10

10
Figure 3A.9.1 Illustration 5
Total population
of birth rates, death rates and
population sizes for the five phases.
Reproduced from [Link] 0 0
org/wiki/Demographic_transition. Time

339

Public [Link] 339 3/6/08 [Link]

 3A Po p u l a t io ns

UK UNITED KINGDOM
The UK population of around 60 million (2001 census) has increased by 18% since 1951 due to births exceeding
deaths. There were more births than deaths in the UK in every year since 1901 (with the exception of 1976). Net
immigration has also been a factor since the mid-1990s. The population was projected to peak at 67 million in 2005
and then gradually fall.
Underlying factors affecting population size include changes in fertility, migration and increasing urbanisation:
see Box 3A.9.2.

Box 3A.9.2
Fertility There are fewer under-16s and more over-65s
Birth rates rose after both World Wars, ‘baby boom’ in 1960s, steadily falling until reaching a
trough in 1970s, some increases in 1980s and 1990s but lowest levels early 2000s
The average age at which women give birth to their first child is increasing. More women remain
childless (1 in 5 now compared with 1 in 10 for women born in the mid-1940s). See also Section 3B.4
Migration Net immigration into the UK is an increasingly important factor in population size
Urbanisation England is one of the most crowded countries in the world. Over 90% of inhabitants live in
urban areas – covering just 8% of the land area

WORLD
The United Nations Population Division expects the absolute number of the world’s infants to begin falling in 2015,
and the number of children under 15 to begin falling by 2025. Other forecasters have calculated that the world
population will peak at 9 billion in 2070, with the average age of the population steadily rising.

3A.10 EFFECTS OF DEMOGRAPHIC CHANGE ON HEALTH CARE

Significance of demographic changes for the health of the population and its need for health and related services
See Table 3A.10.1.

3A.11 POLICIES TO ADDRESS POPULATION GROWTH NATIONALLY AND INTERNATIONALLY

Policies in a particular country are shaped by the nature of demographic challenges faced by that country. In many
developed countries, relatively low birth rates require policy-makers to address issues associated with an ageing
population. Worldwide, there is a need for policies to restrict or limit population growth but also to recognise that
there will be an increasingly ageing population.

UK UNITED KINGDOM
The ageing population of the UK means that there will be a relative shortage of working-age adults compared
with the demand from the total population. This is particularly the case regarding people with key professional
qualifications (e.g. clinical staff).
Policy-makers can increase the number of working-age adults through encouraging:
• More people to have larger families by means of family friendly policies (e.g. subsidised paternity and
maternity leave, tax allowances for parents)
• People of working age to move to the UK through managed migration strategies.

340

Public [Link] 340 3/6/08 [Link]

 3A Populations

Table 3A.10.1 Changes in population size and structure and effects on the need for health care
Demographic changes Effect on health and need for health and related services
Population structure Age An ageing population will place greater demand for geriatric,
intermediate and social/personal care
Ethnicity Greater ethnic diversity will lead to different risk factors
for disease, different patterns of disease and demand for
different models of provision (e.g. bilingual health-care
workers, culturally specific services such as women-only
group sessions)
Access to health care should be monitored to ensure that
there is no discrimination due to language, cultural or
knowledge barriers
Population mobility Short term (travel) Increased global spread of infectious disease and pandemics
(e.g. SARS)
Longer term (migration) International spread of emerging diseases (e.g. TB)
Urbanisation Increased spread of infectious diseases
Housing People living longer Demand for new homes has increased over the past 30 years.
More people living alone Housing shortages in south-east England are acute, leading
to a shortage of workers. The situation is due to get worse if
current trends continue
Lower building rates Risk of health problems associated with overcrowding/poor
housing, e.g. respiratory illnesses

Demand for affordable housing is increasing, and can be met by:

• Increasing the supply and affordability of housing in areas of high demand (e.g. Thames Gateway)
• Keyworker policies (providing subsidised housing for staff, e.g. science teachers)
• Neighbourhood renewal to tackle problems in declining areas
• Sustainable development.

INTERNATIONAL
According to circumstances, strategies may be aimed at:
• Restricting population growth (e.g. China’s one-child policy)
• Limiting rate of growth (e.g. developing countries encouraging birth control through providing subsidised
contraception or sterilisation)
• Targeting resources into the research of diseases associated with ageing (e.g. Alzheimer’s disease)
• International agreements to tackle environmental problems and the demand for natural resources.

341

Public [Link] 341 3/6/08 [Link]

Public [Link] 342 3/6/08 [Link]
 3B Sickness and Health

3B
Sickness and Health

3B.1 Routine mortality and morbidity data 343 3B.4 Measurements of health status 352
3B.2 Biases and artefacts in population data 350 3B.5 Prescribing data and pharmacovigilance 354
3B.3 The International Classification of Diseases 350 3B.6 Data linkage 357

Information regarding sickness and health can be derived from a variety of sources, not just from health service
data. Effective use of this information requires a familiarity with the major resources, an understanding of the
limitations to data validity, and an appreciation of methods for relating one set of data to another by means of data
linkage.

3B.1 ROUTINE MORTALITY AND MORBIDITY DATA

Sources of routine mortality and morbidity data, including primary care data, and how they are collected and published
at international, national, regional and district levels.
Routine information is that which is regularly collected and made partially or fully available. It provides information
on mortality or morbidity in a standardised format (see Section 3A.2). The sources described below relate mainly
to national systems in England and Wales, although similar systems exist in Scotland and certain other countries.
Additional sources may be available locally and regionally.

MORTALITY DATA
Sources of mortality data are listed in Table 3B.1.1.

343

Public [Link] 343 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

UK Table 3B.1.1 Sources of mortality data

ONS – death certificates and death registration
Information Cause of death (inserted by medical practitioner on the certificate)
Additional information collected when the death is registered includes:
• Date and place of death
• Name and surname of the deceased, date and place of birth, occupation, etc.
• Details of any spouse or civil partner
• Usual address
Collection, coding and Deaths must be registered within 5 days by a relative or friend of the deceased to the
analysis local registrar (8 days in Scotland)
Statutory duty for death to be registered before funeral can occur
Register information is stored at district level registry office
Uses Legal requirement
Analysis of trends, comparisons between areas
Health needs analysis for serious conditions
Strengths Complete and timely
Relatively accurate
Weakness Clinical code less accurate for older patients with several co-morbidities
Coding accuracy varies
Problems comparing different years if different ICD classification used (e.g. shift from
ICD-9 to ICD-10 in January 2001 – see Section 3B.3)
Risk of bias in social class measures due to occupational advancement (see Section
3B.2)

Additional information is collected from deaths that result from a road traffic accident: see Table 3B.1.2.

UK Table 3B1.2 Police road traffic accidents – ‘STATS19’

Information Injury/death due to road traffic accidents
Collection, coding an All reported road accidents* involving personal injury are recorded at the time of the
analysis accident on the STATS19 form
Police process the STATS19 data and send the information to the Department for
Transport, which adds the information to the National Road Casualty database
Both the police and the Department for Transport validate the information received
Uses Indications of mortality from incidents occurring on roads
Can link with A&E data on deaths from road accidents
Strengths More detailed information about each incident and the types of vehicles involved than
recorded by A&E
May include accidents not seen by health services

344

Public [Link] 344 3/6/08 [Link]

 3B Sickness and Health

Weakness Not all accidents are reported to the police

Morbidity is rated differently by the police and the health service
*Note that, in Australia, public health practitioners encourage use of the term ‘road traffic collisions’ rather than
‘accidents’, since the latter carries a connotation of non-preventability.

MORBIDITY DATA
Information on morbidity in the UK comes from:
• Condition-specific registers and datasets (see Tables 3B.1.3 and 3B.1.4)
• Individual-level secondary care databases (inpatient, see Table 3B.1.5, and outpatient, see Table 3B.1.6)
• Aggregate-level community records (see Table 3B.1.7)
• Primary care: limited information (not described here) is collected on optometry, pharmacy and dental services,
mainly for payment purposes. More information is available on general practice and prescribing (see Tables
3B.1.8 and 3B.1.9).
International data sources are described in Table 3B.1.10.

UK SOURCES OF PUBLIC HEALTH INFORMATION

Table 3B.1.3 Condition-specific information: regional cancer registers on every new diagnosis of cancer
Information Personal identifiers (needed in order to eliminate duplicates)
Socioeconomic characteristics
Disease status (cancer type, stage)
Treatment
Outcomes
Collection, coding and Sources for collection of information include cancer centres, treatment centres, hospices,
analysis private hospitals, cancer screening programmes, other cancer registers, general practices,
nursing homes and death certificates
Cancers are coded using a system common to all the registries in the UK
The Office for National Statistics collates, analyses and publishes the registers’ data
Uses Monitor trends for incidence and survival of cancer
Compare epidemiology and performance in different areas
Strengths Complete
Useful for continuing conditions with reliable diagnosis
Weakness Time lag between collection of data and available information
Time and labour are intensive to set up and maintain
Risk of inaccuracies if diagnostic criteria change

345

Public [Link] 345 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

Table 3B.1.4 Condition-specific information: minimum datasets specified in areas such as mental health and
cancer
Information Standardised collections of information about specific conditions, including:
• Personal: ethnicity
• Clinical: treatment received, inpatient/outpatient, re-admissions
• Administrative
Collection, coding and Information collected by clinicians at the point of service and collated at regional and
analysis national levels
May be fed to other databases, e.g. cancer registers
Uses Monitor trends
Performance management, e.g. regarding meeting the milestones set out in National
Service Frameworks
Allow health and social care professionals to measure and compare the care that they
provide
Strengths Potential to standardise care and reduce inequities in service provision
Weakness Many still in development

Table 3B.1.5 Inpatient and day-case treatment: hospital episode statistics (HES)
Information Consultant episodes
Personal: name, NHS number, date of birth, ethnicity (address added but stripped out if
NHS number present)
Administrative: start and end dates of the stay, hospital, ward, specialty code, waiting
time
Clinical: ICD-10 code, ONS Classification of Surgical Operations and Procedures
Collection, coding and Hospital episode coded by hospital coders not clinicians
analysis
Collected as part of monthly mandatory information submission by hospitals
Sent to Secondary Uses Service (SUS)
HES provided as nationally available extract
Trusted organisations (e.g. public health observatories) have full database access
Uses Payment from commissioner to provider
Analysis of hospital usage, waiting times
Assessment of quality and outcomes of care (also external performance management by
Department of Health, and inspection by Healthcare Commission)
Estimation of health need for conditions routinely managed in secondary care
Strengths Generally complete – hospitals must submit information if they are to be paid
Timely – information collected routinely; available quarterly and yearly from HES
Mortality data can be linked to HES databases to produce statistics linking episode to
outcome and to individual patients

346

Public [Link] 346 3/6/08 [Link]

 3B Sickness and Health

Weaknesses Accuracy dependent on quality of NHS coders (very variable)

Variable completeness
Relates to episodes, not patients (therefore may overestimate need if one patient has
several episodes)
Only useful for conditions that are generally admitted to hospital

Table 3B.1.6 Individual outpatient activity: secondary uses service (SUS)

Information Personal: NHS number, date of birth, postcode, not ethnicity
Administrative: referral and appointment dates, attendance/DNA (do not attend),
outcome
Clinical: name of clinic recorded but no information on diagnosis
Collection, coding and Collected as part of monthly mandatory submission by hospitals
analysis
Uses Not always used at PCT level
Strengths Measure of outpatient service demand
Useful if clinic name gives indication of need (e.g. HIV clinic, diabetes clinic)
Weakness Missing useful information, e.g. ethnicity, clinical data

Table 3B.1.7 Community level aggregated data: Körner data*

Information Sexually transmitted infections: KC60 GUM clinic data on attendances, demographic
characteristics
Immunisations: KC50, e.g. MMR, flu
Adult screening programmes: KC53, 63
Information Anonymous information collected at clinic; community level information required for
national returns to DH
Collection, coding and Performance management, e.g. uptake of routine immunisations, screening programmes
analysis?
Indication of use of services, level of infections, e.g. GUM
Strengths Confidential
Weaknesses For KC60, patients are not obliged to give truthful demographic details; therefore
difficult to draw conclusions from data analysis
Not straightforward link to geographical area: people can self-refer and may choose to
go far from home in order to protect confidentiality

*Data-set devised by Dame Edith Körner’s working party to record NHS service activity.

347

Public [Link] 347 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

Table 3B.1.8 Primary care data at aggregate level: QMAS (Quality Management and Analysis System)*
Information Disease registers – see Section 3B.3
146 evidence-based indicators forming the quality and outcomes framework (QOF) over
four domains:
1. Clinical, based on 11 conditions including hypertension, asthma, diabetes,
coronary heart disease, mental health
2. Organisational, including records, practice management issues
3. Patient experience (assessed through surveys and consultation length)
4. Additional services
Collection, coding and Monthly extract from a general practice sent to QMAS either automatically or manually
analysis
Payments are linked to evidence-based indicators through the Quality and Outcomes
Framework (QOF)
QMAS accessible online to PCTs
Uses Paying general practices according to the services delivered and the degree to which
certain milestones are met
Registers can indicate disease prevalence
QMAS helpful to plan primary care/referral services
Strengths Many conditions treated only in primary care
Incentive for improving the quality and comprehensiveness of information from general
practice (linked to payment)
QOF score gives an indicator of the quality of clinical care in a practice
Relatively complete – most people are registered with a general practice
Weaknesses Accuracy dependent on the coding and currency of the registers in general practice
QMAS primarily for payment, not performance management
Cannot use QOF scores to compare practices on performance: different list sizes and
different population characteristics may affect QOF reached
QOF is voluntary, though most practices have signed up to it
*The national system supporting the new GMS GP contract.

UK Table 3B.1.9 Primary care data prescribing information (see Section 3B.5): PACT, ePACT
Information By quantity – number of items prescribed or number of tablets/grams of active ingredient
By cost, e.g. net ingredient cost (NIC)
Collection, coding Information from prescriptions made (in large part in general practice) and sent to
and analysis Prescription Pricing Authority
Data on each practice sent back to PCTs for payment
Uses Monitoring of prescribing practice
Payment
General practice prescribing incentive scheme (being replaced by QOF in many places)

348

Public [Link] 348 3/6/08 [Link]

 3B Sickness and Health

Strengths Complete
Weaknesses Not linked to patients or conditions, so difficult to use for performance management

INTERNATIONAL COMPARISON
NZ The Ministry of Health collects, analyses and disseminates information on population health either directly or
through contracted service providers. Most of this work is conducted by two Ministry units, the New Zealand Health
Information Service (NZHIS) and Public Health Intelligence (PHI). The main systems are listed in Table 3B.1.10.

NZ Table 3B.1.10 New Zealand sources of morbidity information

Health outcomes Mortality data
Hospital discharges (public and private)
Cancer registrations and deaths
Mental health service use
Maternal and newborn data
Surveillance (see surveillance section)
Health status and behaviour NZ Health Survey
(regular surveys programme)
NZ Alcohol and Drug Use Survey
NZ Sexual and Reproductive Health Survey
NZ Tobacco Use Survey
NZ Adult and NZ Child Nutrition Surveys
Health-care services and Health workforce data
interventions
Surgical throughput and waiting lists data
Population screening data
Pharmaceutical prescribing data
Other information sources Census data
Social wellbeing (indicators collected by
Ministry of Social Development)
Environmental quality (information collected
by multiple government agencies, e.g. air
quality, water quality)

349

Public [Link] 349 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

3B.2 BIASES AND ARTEFACTS IN POPULATION DATA

Potential sources of bias and artefact should always be borne in mind when working with population data.

BIASES
A bias is a systematic error (see 1A.14). Biases can affect population data in a number of ways, including the
following.

UK LIST SIZE
GP practices sometimes overestimate their list of registered patients by failing to remove, or delaying the removal
of, people who have died or moved away from the practice. This leads to a systematic error with regard to the
estimated population size and structure. As a consequence, it can lead to an underestimate of service provision, e.g.
in terms of vaccine uptake.

UK CENSUS ERRORS
Census estimates of population size can be biased by differential response rates: in 2001 there was under-
enumeration of men in their 20s (the group least likely to respond to the survey). Other groups with low response
rates included: inner-city areas (particularly London) and areas with a high proportion of people with difficulties
filling in the census (e.g. language problems).

UK STATUS INFLATION
There is a tendency for people to inflate the socioeconomic status of the deceased when registering the death or
completing surveys. This ‘occupational advancement’ biases the population structure with regard to social class.

ARTEFACTS
Artefacts in population data are caused by spurious differences between an observed population characteristic and
the true underlying characteristic. Artefacts may hinder accurate comparisons between areas and trends over time.
Adjusting for or reducing artefacts requires an understanding of how data are collected and the potential sources of
inaccuracy, bias and confounding.

INTERNATIONAl ClASSIfICATION Of DISEASE

In January 2001, ICD-9 was replaced by ICD-10 (see Section 3B.3), thereby causing an artefact in mortality data
when comparing deaths before and after this date.

UK CENSUS CHANGES
Census ethnicity information codes changed between 1991 and 2001, when for the first time people were allowed to
identify their ethnicity as ‘mixed’. This complicates comparisons between the two censuses.

3B.3 THE INTERNATIONAl ClASSIfICATION Of DISEASES

The International Classification of Diseases and other methods of classification of disease and medical care
The International Classification of Diseases (ICD) belongs to the WHO Family of International Classifications (see Box
3B.3.1), the purpose of which is to provide a common language for health-related topics.

350

Public [Link] 350 3/6/08 [Link]

 3B Sickness and Health

Box 3B.3.1
WHO Family of International Classifications
• International Classification of Diseases (ICD)
• International Classification of Functioning, Disability and Health (ICF)
• International Classification of Health Interventions (ICHI)

Between them, these classifications enable the consistent collection, analysis and presentation of data, to enable
comparisons over time and between populations. In particular, they allow:
• Analysis of population health
• Monitoring of disease frequency
• Classification of death certificates and hospital records
• Mortality and morbidity statistics to be collated using a common framework.
ICD is primarily an international standard classification for mortality statistics and contains a standard format for
death certification. Its history dates back to the 1850s (International List of Causes of Death), and since 1948 it has
been administered by the WHO.
Diseases mentioned on the death certificate are translated into codes, ranging between A00 (‘Cholera’) and Z99.9
(‘Dependence on unspecified enabling machine and device’): see Table 3B.3.1. By applying coding rules contained in
the ICD (which prioritise and consolidate codes), a single cause of death will be selected.
The ICD is revised every 10–20 years. The WHO advises that it is problematical to translate between the codes of
one revision to another. Deaths during the bridging period should be dual-coded (according to the old and new
revisions) to enable comparisons to be made of mortality measures as collected under the two systems.
The 10th Revision (ICD-10) came into use in 1994 and contains twice the number of codes in ICD-9. A clinically
modified version of ICD-9 (ICD-9-CM) contains more precise detail for describing mortality as opposed to morbidity.
Although ICD-10-CM has now been written, at the time of writing ICD-9-CM remains the standard for reporting
morbidity.

ICD-10
ICD-10 was the first revision to adopt alphanumeric codes (and hence potential problems of confusion between zero
and letter O, and between I and 1). These codes range between three and six characters in length (with a decimal
after the third character if the code is four characters long or greater). Box 3B.3.2 shows examples.

Box 3B.3.2
M60 Myositis
J85.2 Abscess of lung without pneumonia
M11.9 Crystal arthropathy, unspecified

All codes for injuries (range S00–T98) must have a corresponding external cause code (range V01–Y98).
Neoplasms are coded by morphology and site codes. Code U is reserved for future use.

351

Public [Link] 351 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

Table 3B.3.1 ICD-10 disease codes

Code Contents
A00–B99 Certain infectious and parasitic diseases
C00–D48 Neoplasms
D50–D89 Diseases of the blood and blood-forming organs and certain disorders involving the immune
mechanism
E00–E90 Endocrine, nutritional and metabolic diseases
F00–F99 Mental and behavioural disorders
G00–G99 Diseases of the nervous system
H00–H59 Diseases of the eye and adnexa
H60–H95 Diseases of the ear and mastoid process
I00–I99 Diseases of the circulatory system
J00–J99 Diseases of the respiratory system
K00–K93 Diseases of the digestive system
L00–L99 Diseases of the skin and subcutaneous tissue
M00–M99 Diseases of the musculoskeletal system and connective tissue
N00–N99 Diseases of the genitourinary system
O00–O99 Pregnancy, childbirth and the puerperium
P00–P96 Certain conditions originating in the perinatal period
Q00–Q99 Congenital malformations, deformations and chromosomal abnormalities
S00–T98 Injury, poisoning and certain other consequences of external causes
Z00–Z99 Factors influencing health status and contact with health services
U00–U99 Codes for special purposes

DIAGNOSIS-RELATED GROUPS AND HEALTH-CARE RESOURCE GROUP

Diagnosis-related group (DRG) is an American system to classify hospital cases into one of approximately 500
groups that are expected to require similar hospital resource use. DRGs are based on ICD diagnoses, together with
procedures, age, sex and the presence of complications or co-morbidities.
UK In the UK, the analogous health-care resource group (HRG) is the unit of currency for commissioning health
services.

3B.4 MEASUREMENTS OF HEALTH STATUS

Rates and ratios used to measure health status, including geographical, occupational, social class and other
sociodemographic variations
See Section 1A.1.

UK HEALTH SURVEYS IN THE UK

Health surveys supplement data provided from patients’ interaction with the health-care system. Different models
are in use in different parts of the UK.

352

Public [Link] 352 3/6/08 [Link]

 3B Sickness and Health

Eng HEALTH SURvEY FOR ENGLAND

The Health Survey for England seeks to obtain a representative sample of people living in private households
each year and it includes both adults and children. The survey consists of a health interview and an examination
conducted by a nurse who takes a variety of measurements (e.g. height, weight and blood pressure). The Health
Survey for England contains a core that is repeated each year, and has one or more supplements: modules on
subjects of special interest.
The ‘core’ includes questions on:
• General health and psychosocial indicators
• Smoking
• Alcohol
• Demographic and socioeconomic indicators
• Use of health services and prescribed medicines.

Scot SCOTTISH HEALTH SURvEY

The Scottish Health Survey was undertaken in 1995, 1998 and 2003. As in England, the Scottish surveys also target
both adults and children, and utilise a health interview and an examination by a nurse. In addition to demographic
data, the survey in 2003 included questions on:
• General health and illness
• Cardiovascular disease
• Respiratory disease (including asthma)
• Dental health
• Health-related behaviours
• Use of health services
• Height and weight.
The nurses measured blood pressure and lung function, and they collected saliva, blood and urine samples. Some
subjects were asked to have an ECG recording.

Wal WELSH HEALTH SURvEY

The Welsh Health Survey (WHS) was first carried out in 1995 and repeated in 1998 and 2003–5. This survey does not
include a nurse examination, and it was targeted at adults only during the first two iterations. It is a postal survey
designed to provide a picture of health of the people of Wales, the way the NHS is used and areas where services
could be improved. The 2003–05 survey included a 15-minute face-to-face questionnaire. Results from the 2003–05
survey are not comparable with those from the previous surveys because of differences in the questionnaires and the
ways in which the survey was designed and conducted.
In addition to demographic data, the content of the 2003–05 survey included:
• General health status (SF-36)
• A range of reported illnesses and other conditions (such as eyesight or hearing difficulty)
• Reported lifestyle behaviours, including smoking, drinking, fruit and vegetable consumption, physical activity
and body mass index
• Reported use of a range of health services.

NI NORTHERN IRELAND HEALTH AND SOCIAL WELLBEING SURvEY

The Northern Ireland Health and Social Wellbeing Survey was conducted in 1997, 2001 and 2006. It is designed to
yield a representative sample of all adults aged 16 and over living in Northern Ireland.

353

Public [Link] 353 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

The questionnaire consists of a household interview followed by an individual interview with each person in the
household aged 16 and above. The individual interview consists of core modules and modules that will recur on
a regular cycle. Core items include accommodation, tenure, employment status, educational qualifications, family
information, smoking and drinking, and health and ill health. Non-core items include physical activity and sexual
health.
In 1997 the survey included physical measures for one respondent selected at random from each household.
Qualified nurses recorded details of prescribed medication, and measured height, waist, hip, weight and blood
pressure. A blood sample was also taken to measure the level of cholesterol (non-fasting).
In addition to demographic details, the questionnaire covers:
• Lifestyle habits, including sexual health
• General health, long-standing illness
• Common chronic diseases and conditions
• Stressful life event and possible mental health problems (SF-36 was also used in 1997)
• Informal care and the lifestyle of carers
• Characteristics of the people whom carers are looking after.

3B.5 PRESCRIBING DATA AND PHARMACOvIGILANCE

Prescribing data (information about the volumes, costs and types of drugs prescribed and dispensed) can be useful
for both clinical and managerial reasons, as well as for identifying adverse drug effects.
Reasons for monitoring prescribing include:
• Cost containment – prescription costs are rising; money spent on prescriptions means that less is available for
other areas of health care
• Monitoring adherence to guidelines (e.g. NICE)
• Detecting aberrant or inappropriate clinical performance (e.g. controlled drug prescriptions, antibiotics)
• Addressing local priorities (e.g. reduction of heart disease through prescribing drugs such as statins)
• Performance-related incentive payments (e.g. prescribing elements of QOF – see Section 3B.1).

CHALLENGES
While prescribing data are accessible and useful for measuring costs, it is more difficult to measure the quality
of prescribing. Information on why a drug was prescribed or for whom it was prescribed is not collated, and
information currently can be obtained only by the laborious process of auditing individual patient records.

SOURCES OF DATA
PACT
Prescribing Analysis and CosT (PACT) data provide GPs and other prescribers with reliable and regular information on
their NHS prescribing habits and costs.

EPACT
This is an electronic system for pharmaceutical and prescribing advisors. It allows real-time on-line analysis of the
previous 5 years’ prescribing data held on the NHS Prescribing Database. The data available include:
• Budgets and expenditure forecasts
• Costs and volumes of prescribing
• Prescribing totals by prescriber at all British National Formulary (BNF) levels
• Prescribing from the Nurse and Extended Nurse Formularies

354

Public [Link] 354 3/6/08 [Link]

 3B Sickness and Health

• Working environment for nurses and supplementary prescribers (i.e. community or practice)
• Patient list sizes
• Low Income Scheme Index scores for practices (released in May 2004)
• Average daily quantities and defined daily doses.

UNITS OF MEASURING PRESCRIBING

ITEM
This is the number of prescription items listed on a prescription form. While easy to measure, caution is needed
regarding repeat prescriptions. For example, a GP who writes monthly prescriptions will appear to prescribe more
than a GP who prescribes quarterly – even though the total amount of drug is identical.

QUANTITY
• Number of tablets or millilitres, milligrams, etc.
• Strength of active ingredients: stronger active ingredients may require fewer milligrams for the same number of
tablets
• Dosing schedules
Note that caution is needed regarding potency (e.g. fewer milligrams of one type of statin are needed to achieve a
particular drop in blood cholesterol compared with another).

NET INGREDIENT COST

This is the basic price of a drug. It can be used to measure the volume of similarly priced groups of drugs at
equivalent doses. However, where there is a large price difference, it is not an accurate measure of use.

ACTUAL COST
This is calculated by taking the basic price of the prescription items, deducting the National Average Discount, and
then adding an allowance for the container. Actual cost is used in Prescribing Monitoring Documents (PMDs).

DEFINED DAILY DOSES

These are based on maintenance doses and are not suitable for one-off doses. Note that the defined daily dose
(DDD) is not the recommended dose, nor is it necessarily a dose that a patient could practically receive. For
example, simvastatin has a DDD of 15 mg but is available only in 10 or 20 mg tablets.

AvERAGE DAILY QUANTITIES

This is an England-specific system developed by the Prescribing Support Unit (PSU) and is equivalent to the DDD.

PRESCRIBING UNITS
Prescribing costs can vary across different organisations because of different prescribing practices, but also because
of the features of the local population. As a result, it is not valid simply to use an average cost per patient as a
measure of prescribing spend. In England, prescribing units (PUs) have been developed to take account of the fact
that older patients have a greater need of medication. This allows comparisons of prescribing costs across areas with
different age structures in their populations. Since 1983, prescribing units have been further refined into:
• ASTRO-PUs (age, sex and temporary resident originated prescribing units): take account of a wider range of
demographic factors other than age in comparisons of prescribing or resource allocation decisions
• STAR-PUs (specific therapeutic group age–sex weightings related prescribing units): units specific to different
therapeutic areas which also take into account demographic factors.

355

Public [Link] 355 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

PHARMACOvIGILANCE
Pharmacovigilance aims to prevent or reduce harm resulting from medicines through processes described in Table
3B.5.1.

Table 3B.5.1 Processes involved in pharmacovigilance

Monitoring Monitoring the use of medicines in everyday practice to identify previously
unrecognised adverse effects or changes in the patterns of adverse effects
Risk assessment Assessing the risks and benefits of medicines in order to determine what action, if any,
is needed to improve their safety
Information provision Providing information to health-care professionals and patients to optimise safe and
effective use of medicines
Measuring Assessing the impact of any action taken

INFORMATION SOURCES USED FOR PHARMACOVIGILANCE

A range of different national and international systems is used:
• Spontaneous adverse drug reaction (ADR) reporting schemes, such as the Yellowcard and black triangle systems
(see below)
• Clinical and epidemiological studies in the worldwide medical literature
• Information from pharmaceutical companies
• Information from worldwide regulatory authorities
• Morbidity and mortality databases.
Information from any of these sources may identify unexpected side effects or indicate that certain side effects
occur more commonly than was previously believed. They may also reveal that certain patient groups are more
susceptible to particular problems than others. Such findings can lead to changes in the marketing authorisation of
the medicine, through:
• Restrictions in use
• Changes in the dose of the medicine
• Introduction of specific warnings of side effects in product information.

UK THE YELLOW CARD SCHEME

The Medicines and Healthcare Products Regulatory Authority (MHRA) and the Committee on Safety of Medicines
(CSM) run the UK’s spontaneous adverse drug reaction reporting scheme, called the Yellow Card Scheme. Yellow Cards
are distributed to health-care professionals, including as an appendix to the BNF. This receives reports of suspected
adverse drug reactions from health professionals and patients.

UK BLACK TRIANGLE SCHEME

New medicines and vaccines are labelled with a black triangle symbol (▼) in the BNF and on all product information
and advertisements. Health professionals are urged to report any suspected adverse reactions that might involve one
of these products. For established medicines that are not marked with a black triangle, health professionals should
report only serious or unusual suspected adverse reactions.

RISK MINIMISATION
Occasionally, where the risks of a medicine are found to outweigh the benefits, the drug (or even an entire drug
class) may be removed from the market. More commonly, the risk of a side effect may be avoided or reduced by:

356

Public [Link] 356 3/6/08 [Link]

 3B Sickness and Health

• Including warnings in the product information or on the package labelling

• Restricting the indications for use of a medicine
• Changing the legal status of a medicine (e.g. by switching from pharmacy to prescription only).

COMMUNICATION WITH HEALTH-CARE PROFESSIONALS AND PATIENTS

UK The MHRA communicates with health-care professionals and patients to warn about adverse effects and to
provide feedback of information. It does this through:
• Patient Information Leaflets (PILs) and Summaries of Product Characteristics (SPCs) for medicines that are
updated when new safety issues are identified
• Urgent warnings about drug hazards via letters to all doctors and pharmacists
• MHRA and CSM regular drug safety bulletin, ‘Current Problems in Pharmacovigilance’, sent to doctors and
pharmacists
• Fact sheets of major safety issues for both health-care professionals and patients
• Safety alerts on the MHRA website.

3B.6 DATA LINKAGE

Data linkage is the process of matching information in one data source with that in another. Linking two databases
can provide useful extra information and enable different analyses to be carried out. For example, linking hospital
episode statistics (HES) databases to mortality data can provide information about the outcomes of hospital
activity. One of the earliest record linkage studies was the Oxford Record Linkage Study (ORLS), which started in
1963. The ORLS consists of computerised abstracts of records of all types of hospital inpatient care, and records of
births and deaths in the Oxford region. When data collection ceased in 1999 all patient identifiers were removed
from the database. The data-set includes 10 million records relating to over 5 million people.

Eng Box 3B.6.1

Example: Linkage in the NHS – Connecting for Health
Connecting for Health is a national programme for improving the information technology of the NHS to ensure
that the appropriate information is available to any clinician regardless of geography. It should mean that
patients will no longer be asked the same information every time they see a different health professional.
Furthermore, it should reduce the chances of information (e.g. X-rays, patient notes) becoming lost. Connecting
for Health involves the development of several information systems containing patient information, which will be
linked to provide a complete record.
• NHS Care Records Service: patients will have a single overarching electronic care record, which will hold all
of their health information in one place
• Choose and Book: GP systems will connect to secondary care providers to enable electronic booking of
hospital appointments
• Electronic transmission of prescriptions system: this will mean that patients can have a repeat prescription
sent electronically to a nominated pharmacy without having to visit the GP practice
• A new national broadband IT network for the NHS
• Picture Archiving and Communications System (PACS): for the electronic storage and communication of
X-rays, scans, etc.
• IT supporting GP payments: including QMAS (Quality Management and Analysis System)
• Email and directory service

Eng Data linkage requires a unique identifier across both data sets. In England, the ideal unique identifier in health
systems is the NHS number (see Box 3B.6.1). An alternative is the National Insurance number (now issued at birth).

357

Public [Link] 357 3/6/08 [Link]

 3B S ic k ne s s a nd Heal th

However, where a unique identifier is not available, other information such as patient name, date of birth and
postcode can be used (see Section 3A.2 and Boxes 3B.6.2 and 3B.6.3).

Eng Box 3B.6.2

Example: ‘Combined model’ predictive risk algorithm
This model, developed for the NHS, links together several routinely collected data sources (inpatient, outpatient,
A&E, GP and Social Services) to make predictions of future emergency hospital admissions. See Section 3C.4.

Aus Box 3B.6.3

Example: Data linkage in Western Australia
Since the 1980s Western Australia (WA) has had a population health data linkage system that is unique in
Australia, and one of only five similar systems worldwide. Since its inception more links have been added as its
value, particularly to health researchers, becomes increasingly apparent.
The data linkage initiative is a collaborative venture of the WA Department of Health, the Centre for Health
Services Research at the University of WA, the Institute for Child Health Research and the Health Science
Division at Curtin University. The Data Linkage Unit, part of Information Collection and Management in the
Statewide Health Support Service (WA Department of Health), is responsible for creating and maintaining the
links within and between the core population health data collections, including births, and mental health,
hospital inpatient, emergency, cancer and death records.
Very stringent ethical requirements and strict protocols are applied to the use of linkable data and, to the extent
possible, de-identified data files are used. Links to ambulance services and home nursing clients have been added
and special arrangements with the Australian government have allowed links to aged care and Medicare records
to be created.
[Link]/icam.

358

Public [Link] 358 3/6/08 [Link]

 3C Applications of Health Services I n formation

3C
Applications of Health Services Information

3C.1 Use of information for health service 3C.4 Uses of mathematical modelling in health
planning and evaluation 359 service planning 364
3C.2 Specification and uses of information 3C.5 Indices of needs for and outcome of services 366
systems 362 3C.6 Issues with routine health information 368
3C.3 Common measures of health service 3C.7 Information technology and health-care
provision and usage 362 provision 370

Health service information is used in a number of ways to improve health. These include: informing, planning and
commissioning; providing key data for epidemiological studies; and identifying unusual or substandard practice
through audit. This chapter illustrates how information from Section 1 and other parts of Section 3 can usefully
be linked. It contains examples of recent practice in order to demonstrate how the collection and collation of
this information may be achieved in practice. The strengths and limitations of health service data are considered,
together with the technologies used for their manipulation.

3C.1 USE OF INFORMATION FOR HEALTH SERvICE PLANNING AND EvALUATION

Health information is essential for health service planning and evaluation. Without it, services would be
unresponsive to changes in circumstances, and it would be impossible to make predictions or to increase efficiency.
Eng In England, the responsibility for local health service provision lies with PCTs, which plan health services over
a 3-year period: see Table 3C.1.1.
Wal Similar arrangements apply in Wales, where responsibility for securing health services lies with local health
boards (LHBs). The 22 LHBs are co-terminous with local authorities. This facilitates joint working across the health
and social care agenda. However, it also means that arrangements for commissioning secondary care are complex,
with a number of commissioning partnerships in place for each hospital.
Scot In Scotland, the health boards, and acute and primary care organisations all work within a single system.
Planning and evaluation are led by the health boards with all organisations working together to develop improved
systems and solutions.

359

Public [Link] 359 3/6/08 [Link]

 3C A p p l ic a t io ns of Heal th Services Information

Table 3C.1.1 Steps in health service planning and evaluation

Step Description
Priorities Identifying the national and local priorities and the key targets for delivery over the next 3
years
Capacity Agreeing the capacity needed to deliver them
Responsibilities Determining the specific responsibilities of each health and social care organisation
Plans Creating robust plans which show systematically how improvements will be made and which
are based on the involvement of staff and the public
Monitoring Establishing sound local arrangements for monitoring progress and NHS performance
management which link into national arrangements
Accountability Improving communications and accountability to the public locally so as to demonstrate
progress and the value added year on year

INFORMATION FOR HEALTH-CARE PLANNING

Data can be used to assess changes in health and health care over time, thereby enabling demand for services to be
forecast: see Box 3C.1.1.

Box 3C.1.1
Health-care The identification of health needs and priorities involves epidemiological, qualitative and
needs comparative methods to describe the health problems of a population. These are assessed in
terms of inequalities in health and in degrees of access to health services. Population trends
(e.g. age distribution, relative deprivation) will affect need and these can also be monitored
from routine data
Health-care The next step is to determine priorities for the most effective use of resources. This is achieved
priorities by reviewing routinely collected national and local data, the literature (both published and the
grey literature) and best practice (as defined in national guidance and guidelines). Workforce
trends should also be taken into account
Health-care This begins with a description of the existing service, in terms of utilisation and distribution. An
review analysis is then undertaken of the differences between needs and existing services (deficient and
superfluous services)

ROUTINE DATA SOURCES

See also Section 1A.1.
Eng Many population measures can be obtained from routinely collected data. Several of these are published
annually by the Department of Health, with the statistics made available at national, regional, health authority and
local authority levels: see Box 3C.1.2.

PERFORMANCE MANAGEMENT
Performance management indicators are used to identify whether services are delivering against a dimension agreed
to be of importance for that service. They are used to highlight inadequate performance or problems with service
delivery. Where problems are identified, more detailed work may be required to elucidate the underlying causes so
that they can be addressed.

360

Public [Link] 360 3/6/08 [Link]

 3C Applications of Health Services I n formation

Eng Box 3C.1.2

Measure Example
Mortality Office for National Statistics
Serious morbidity Hospital episode statistics data
Cancer registries
Minor morbidity GP consultation rates
General health General Household Survey (self-reported long-standing illness)
Deprivation Jarman score
Demographics Census (age, sex, ethnicity)

HEALTH SERVICE EVALUATION

Health service activity may be assessed in terms of process or outcome measures. Evaluations may be formative or
summative: see Box 3C.1.3.

Box.3C.1.3
Formative Assesses whether a problem is occurring while the activities are being developed. For
example, in a pilot scheme, continuous feedback is obtained from service users and service
providers to revise the original plans
Summative This focuses on the impact and the effectiveness of an established programme. For
example, data may be collected over an extended period to assess the impact of a service
on a community

Health service evaluations typically consider a number of dimensions. The impact and cost of services are almost
invariably evaluated. However, service quality can also be assessed through a consideration of acceptability to
service users, access to care and the impact of a service on health inequalities.
The nature of a health service evaluation often depends on the level of confidence held in the effectiveness of the
service in question. If there is high confidence then the evaluation may well be limited to performance management
using routine data. If there is low confidence, then the service may be subjected to more rigorous evaluation, e.g.
by means of a randomised controlled trial. There are levels of confidence between these extremes. See Box 3C.1.4.

Wal Box 3C.1.4

Example: Evaluation of CHD Secondary Prevention Programme
The literature suggests that not smoking and maintaining appropriate levels of body mass index, blood pressure
and blood cholesterol can all help to prevent secondary cardiac events. Evaluation of people re-admitted to
hospitals for cardiac events in the borough of Rhondda Cynon Tâf showed that, although people had been
supported to bring these parameters back to normal at time of discharge home, they rapidly relapsed once
discharged home. A new community-based pilot scheme has recently been developed to support people after
discharge from hospital. Initially this service will be evaluated to see whether it can help people to maintain
healthy blood pressure and cholesterol levels in the longer term (at 12 months post-discharge). If the pilot is
shown to be effective and cost-effective, and secures ongoing funding, then performance management will be
used to monitor for inequalities in access to the service and drop-out rates by social class.

361

Public [Link] 361 3/6/08 [Link]

 3C A p p l ic a t io ns of Heal th Services Information

BENEFITS OF EVALUATION
The prime aim of an evaluation is to determine whether the objectives of a programme are being met. The
evaluation of the project should therefore reflect those objectives, but it will be affected by what can readily be
measured and by the budget for the evaluation.
Depending upon the findings, an evaluation may justify continuation of a programme or else question whether
it should continue as currently provided. If an evaluation is of sufficient breadth and depth, then it may identify
changes that could be made to improve the programme, or it may highlight problems requiring more detailed
investigation.
A balance must be struck between the benefits and risks associated with the programme, and the costs of the
programme and its evaluation. However, it is important to evaluate work – otherwise limited resources will be
unlikely to achieve the maximum potential impact. A high-quality evaluation of a programme in one location can
often provide sufficient evidence to support rolling out that programme elsewhere – with further monitoring limited
to process outcomes.
Evaluation can assess performance with regard to:
• Effectiveness
• Cost-effectiveness
• Efficiency
• Quality outcomes
• Accountability
• Impact on the community’s health
• Inequality and other adverse impacts
• Access.

3C.2 SPECIFICATION AND USES OF INFORMATION SYSTEMS

An information system is a process in which raw data are transformed into meaningful information: see Figure
3C.2.1.

Input Process Output

Figure 3C.2.1 Simple diagram of an information system

SPECIFICATION
The specification of an information system is the set of requirements agreed between the user and the producer of
the system. In modern information systems, information creation and processing require a high level of engagement
of all users during the developmental stages.

USES
See Table 3C.2.1.

3C.3 COMMON MEASURES OF HEALTH SERvICE PROvISION AND USAGE

Health service provision refers to the supply of health care (buildings, staff, services), while access is affected by
opening hours, demand, capacity, travel times, language and cultural barriers. Health service usage is a complex

362

Public [Link] 362 3/6/08 [Link]

 3C Applications of Health Services I n formation

Table 3C.2.1 Uses of information systems in health care

Clinical information To ensure that all health-care professionals have access to the relevant clinical
information necessary to support patients under their care
Clinical guidance To provide health-care professionals with on-line access to up-to-date guidance and
evidence on effective treatment, both local and national, and to the information
required to evaluate the effectiveness of their work, and to support professional
education, development, research and the planning of services
Standardisation To provide the basis for uniformity of clinical systems, so that comparable recording
of information on activity and quality will be available to assist in the maintenance of
the highest possible standards of practice
Aggregation To analyse aggregate data for monitoring of quality (including outcomes), planning
new services and supporting research activity
Security To ensure that sensitive or critical electronic information and systems are not lost,
destroyed, misappropriated or corrupted
Exchange To exchange information securely and automatically and to protect data
Linkage To manage, link and process the different types of data electronically
Analysis To analyse, display, report and map accumulated data and share data and technologies
for analysis and visualisation with other professionals
Regulation To conform to national standards recommended by the government or professional
bodies, including those for a common clinical terminology and headings for
communication from the record

product of provision and access issues and is affected by health service need. The concepts of use and need are more
fully explored in relation to equity in Sections 1C.10 and 4C.1.
Measures of health service use can be valuable in assessing the quality and appropriateness of health care provided.
Indicators of health service utilisation are also described in Section 1C.3, but those particularly useful in primary
and secondary care are shown in Table 3C.3.1.

Table 3C.3.1 Primary and secondary care indicators of service utilisation

Primary care Secondary care
Consultations per patient Hospital episodes:
• Emergency or scheduled/elective
Secondary/tertiary care referral rates
• Outpatient appointments or inpatient stays
• Re-admission rates
Time spent with practitioner Length of stay
General practice list sizes Bed occupancy
Prescriptions: No comparable data
• Number of items
• Cost
• Type of medication
Preventive health services: No comparable data
• Screening uptake rates
• Scheduled immunisation uptake rates

363

Public [Link] 363 3/6/08 [Link]

 3C A p p l ic a t io ns of Heal th Services Information

Examples of service usage information in England are shown in Box 3C.3.1.

Eng Box 3C.3.1

Practice profiles: service usage information to improve the quality of primary care
Several PCTs in England provide general practices in their area with summarised information on their performance
in a number of domains. In Croydon, the PCT compiles practice profiles, a comprehensive summary for each
practice which shows their performance compared with all other practices in the PCT.
Practices receive information on four domains, to enable them to link activity with health service need,
behaviour and outcomes:
• Characteristics of patients that may affect their use of primary care services
• Indicators that may be affected by GP clinical behaviour
• Patient survey results
• Disease-specific areas
Each indicator is coloured to give an instant indication of how well the practice is performing. For services with
a proven health benefit, e.g. uptake of screening, high activity is marked green, low activity red, with gradations
in between. For services with little evidence of clinical benefit, e.g. X-rays for knee problems, high activity is
represented by red, with low activity as green.
Excerpt from standard Croydon PCT practice profile

3C.4 USES OF MATHEMATICAL MODELLING IN HEALTH SERvICE PLANNING

Modelling is the actvity of bridging the real world (where observations of phenomena or behaviours occur) with the
conceptual world (where understanding of real world observations occurs). In a model, the observations are analysed
using statistical or other analysis, then used to predict events or to provide solutions. An iterative or feedback
process allows the model to be refined. See Figure 3C.4.1.

Conceptual Mathematical
Real world Solution
world modelling

Figure 3C.4.1 Simple concept of mathematical modelling. Courtesy of Professor E Carson and Dr A Roudsari, Centre
for Health Informatics School of Informatics, City University, London

364

Public [Link] 364 3/6/08 [Link]

 3C Applications of Health Services I n formation

LIMITATIONS OF MODELS
The usefulness of a model may be constrained by:
• Availability of data
• Quality of data
• Incorrect assumptions in statistical analysis
• Application of an inappropriate or flawed techniques.

ADVANTAGES OF MATHEMATICAL MODELLING

Used properly, mathematical modelling can assist in health-care planning by means of:
• Aiding decision-making
• Dealing with complexity (both complex organisations and complex activities)
• Creating alternative scenarios
• Modelling at any level of detail
• Creating plans for a few months or several years into the future.

PREDICTIVE RISK MODELLING

Predictive risk modelling is a technique that has been used in the financial and banking sectors for many years,
and has recently been applied to health care. Predictive modelling involves building an algorithm by examining the
historical relationships between known data and a variable of interest (an outcome) contained within the data. A
model is built either by means of multiple-regression or by using neural network technology. Once the algorithm
has been built, it is used to process contemporary data in order to make predictions about the future. There are six
prerequisites to implementing a health intervention based on predictive risk modelling: see Table 3C.4.1.
An example is given in Box 3C.4.1.

Table 3C.4.1 Prerequisites to implementing a health intervention based on predictive risk modelling
Prerequisites
Sufficient historical data must be available to build the algorithm.
The data sources must be routinely available and frequently updated
It must be possible to link these data sources Algorithm
There must be a meaningful outcome contained within one of the data
sources
An effective intervention must be available to prevent the selected endpoint
The intervention must be economical, considering the specificity of the Intervention
algorithm for the chosen outcome at a particular level of risk

365

Public [Link] 365 3/6/08 [Link]

 3C A p p l ic a t io ns of Heal th Services Information

UK Box 3C.4.1
Example: King’s Fund/New York University/Health Dialog algorithms
The NHS has commissioned the King’s Fund and partners to develop a series of algorithms that predict which
people in a population are at highest risk of future multiple emergency hospitalisations. Advance predictions are
needed because of the rapid turnover of the group of patients who are frequently admitted to hospital: within 18
months they will have regressed to the mean admission rate for the population.
Several years’ worth of historical data were used to build the algorithm. The data sources included inpatient,
outpatient, A&E and GP practice data. These were linked using NHS number and the outcome variable chosen
was multiple emergency admissions (information that was contained within the inpatient data records). The
algorithms allow PCTs to rank their population in terms of likelihood of needing multiple emergency admissions,
and an ‘upstream’ intervention (e.g. case management by a community matron, or admission to a virtual ward) is
offered to the patients at highest risk.

3C.5 INDICES OF NEEDS FOR AND OUTCOME OF SERvICES

Service indices used by health care are listed in Table 3C.5.1.

Table 3C.5.1 Service indices used by health care

Output index How much of each service is being produced (e.g. number of patients treated)
Welfare index Value to final users (e.g. degree of pain reduction)
Performance management index How the services are being produced (e.g. extent to which doctors use
appropriate treatment)
Composite index This includes elements of the above three indices. With multiple services,
weights are added to each service, often combining indicators that are
measured in non-comparable units

OUTPUT INDICES
The outputs of a health service may be weighted in a number of ways.

IDEAL vALUE-WEIGHTED OUTPUT INDEx

This index has two fundamental features:
• A value attached to each output reflects its relative contribution to health outcomes; and
• The values of other important characteristics of health care (such as the process of care delivery) are also
incorporated.
A lack of health outcome data makes calculation of this index unfeasible.

COST-WEIGHTED ACTIvITY/OUTPUT INDEx

This weights separate activities or outputs using the costs of providing them. The implication of this is that
decision-makers are equating costs and benefits of the service provided. There are many arguments for and against
this assumption: some might feel that doctors do or should do what is best for the patient whatever the cost.
Decisions about capacity may use cost–benefit analysis in a more explicit way, which will affect the decisions made
by doctors.

366

Public [Link] 366 3/6/08 [Link]

 3C Applications of Health Services I n formation

COMPOSITE INDICES
Composite indices of health-care performance are aggregations of several underlying individual performance
measures. They are used worldwide to rank health-care organisations or systems: see Box 3C.5.1. They are designed
to be easy to interpret and present the ‘big picture’. However, there is a need to pay attention to methodological
issues in their construction, otherwise misleading conclusions may be drawn. For example, some hospitals may
be promoted up the league table of performance as a result of subtle changes in the methods of creating the
composite.

Box 3C.5.1
Example: The World Health Report 2000 – Health Systems: Improving Performance
In this report the WHO ranked the overall health system performance world’s health-care systems according to
an index composed of five dimensions:
1. Overall population health
2. Health inequalities
3. Health system responsiveness
4. Distribution of responsiveness
5. Distribution of financial burden for the health system
France was ranked top and Sierra Leone bottom. The UK was ranked 18th, Ireland 19th, Australia came 32nd, the
USA 37th, New Zealand 41st and South Africa 175th.
Reproduced from [Link]/whr/en.

PROPERTIES OF PERFORMANCE MANAGEMENT INDICES

Features of an ideal performance index are listed in Table 3C.5.2 and an example is given in Box 3C.5.2.

Table 3C.5.2 Features of an ideal performance index

Fit for purpose Which index to use will depend on the question being addressed
Consumer’s wellbeing Welfare index is required (of interest to organisations providing the service as well
as central government)
Local information Departments, as providers of services, require performance management indices

367

Public [Link] 367 3/6/08 [Link]

 3C A p p l ic a t io ns of Heal th Services Information

UK Box 3C.5.2
Example: Star rating system
Originally introduced in 2000 by the Department of Health, the following key targets were used in the star
ratings for PCTs in 2004–05:
• Access to a GP
• Financial management
• 4-week smoking quitters
• A larger set of ‘balanced scorecard’ indicators
The performance indicators used in star ratings for PCTs in 2004–05 included:
• Risk management
• Quality of ethnicity data
• Workforce indicators
In 2005–06, the star ratings were replaced by the Healthcare Commission’s ‘Annual Health Check’ which includes
a small set of ‘key targets’ and ‘performance indicators’.

3C.6 ISSUES WITH ROUTINE HEALTH INFORMATION

Strengths, uses, interpretation and limitations of routine health information
Routine data are derived from automated, ongoing, data collection systems. They include large local and national
databases associated with health and social services. Data are collected irrespective of the procedure or outcome.
Examples include HES, deaths, births, statutory notifications of infectious diseases and cancer registry data. See
Figure 3C.6.1.

STRENGTHS AND WEAKNESSES OF ROUTINE DATA

See Box 3C.6.1.

368

Public [Link] 368 3/6/08 [Link]

 3C Applications of Health Services I n formation

Population health assessments

• Needs assessment
• Establish baseline characteristics
regarding the health status of the
community

Research
Clinical audit • Examine risk factors, control of
• Highlight variations confounding
in care
Use of routinely • Generate hypotheses based on
• Identify events that collected data sex, age, cohort or geographical
should not normally variation
occur
• Identify areas for further research

Health-care planning and commissioning

• Understand current health service use
• Reveal trends over time
• Predict future health requirements

Figure 3C.6.1 Uses of routine data

Box 3C.6.1
Strengths Weaknesses
Readily available Incompleteness:
• Statutory returns do not guarantee completeness,
Low cost
e.g. even meningococcal septicaemia is only 70%
Up to date notified
• Poor levels of ethnicity coding
Large population coverage
Bias:
Collection usually spans a significant time period
• If those who participate in providing data are
Breadth and diversity to explore unexpected avenues significantly different from those who do not, the
collated data on health disease may be biased
Useful for initial assessment: provides baseline data on
• Political interference in what data are collected
expected levels of health/disease
(or not collected), and how they are presented
Simple statistical analysis often sufficient due to size and
Poor collation, presentation and analysis limit value
completeness of database
in informing practitioners and providing policy-makers
with usable information
More information on process rather than health status
and outcomes
Lack of uniformity in data structures, coding systems
and definitions

369

Public [Link] 369 3/6/08 [Link]

 3C A p p l ic a t io ns of Heal th Services Information

IMPROVING ROUTINE DATA

Improving the reliability, validity and completeness of routine data is important to avoid waste and to maximise the
use of resources. There should therefore be a good reason to begin or to stop collecting each item of data. Ways of
improving data quality are listed in Table 3.6.1.

Table 3C.6.1 Ways of improving the quality of data

Computerised data collation and Improves the accuracy and timeliness of the preparation and dissemination
analysis of information
Feedback Improving feedback of collated data to providers of primary care is essential
if their interest is to be maintained and their attention to providing quality
data is to be sustained
Presentation Data should be presented in a variety of ways which are meaningful to
policy-makers, media, professionals and the lay public
Training Training the coders and those responsible for data entry in the use of
standard definitions, terminology, etc.

3C.7 INFORMATION TECHNOLOGY AND HEALTH-CARE PROvISION

Use of information technology in the processing and analysis of health service information and in support of the
provision of health care
Health systems in many countries are implementing large-scale information and communication technology (ICT)
projects to transform the delivery of services. Some important features of electronic communication include:
• Faster and more varied methods for communicating (e.g. email, videoconferencing and mobile
telecommunications)
• Near-instant access to vast amounts of information (e.g. through the internet)
• Emergence of new forms of inequality (the so-called ‘Digital Divide’).
Consequently, for public health, ICT presents both developmental opportunities and new challenges to equity.
Information underpins:
• Assessment of health needs
• Development of health strategies
• Monitoring of progress.
Communication provides essential links for consultation, discussion and dissemination of knowledge between all
those individuals and organisations with a role to play in improving the health of the public. New technologies can
offer public health practitioners rapid access to:
• Key data from international down to local levels
• Networks of professionals in health and related disciplines (e.g. managed public health networks)
• The public’s views on health service development
• Electronic libraries of evidence, peer-reviewed research and practice guidance (e.g. Medline, Cochrane, NICE).
Sensitive data can be securely transferred between organisations by encrypted email.

370

Public [Link] 370 3/6/08 [Link]

 3C Applications of Health Services I n formation

HEALTH INFORMATICS
The field of health informatics is concerned with the application of information technology to the acquisition,
processing, interpreting, storage, transmission, and retrieval of health and health care-related data. Moreover it uses
this knowledge to facilitate health-care delivery, education, management and research.
Health informatics tools include computers, clinical guidelines, diagnostic and monitoring equipment, and
information and communication systems.
In practice, health informatics may be broadly divided into: public health informatics, clinical informatics, nursing
informatics, dental informatics, bioinformatics and pharmacoinformatics.

APPLICATIONS
The uses of informatics within health care are expanding rapidly. Current applications include:
• Patient monitoring
• Clinical care
• Geographical information systems in public health surveillance
• Integrating data sources for improved decision-making
• Electronic health records (especially in general practice)
• Remote consultations (especially in dermatology).

UK NATIONAL IT PROGRAMME
The Connecting for Health project is reportedly the world’s largest ever IT investment. The programme has four
goals:
1. Electronic appointment booking
2. Electronic care records service
3. Electronic and fast transmission of prescriptions
4. A fast, reliable, interconnected IT infrastructure.

371

Public [Link] 371 3/6/08 [Link]

Public [Link] 372 3/6/08 [Link]
 Section 4
MEDICAL SOCIOLOGY, SOCIAL POLICY AND
HEALTH ECONOMICS

Public health is concerned with generating high-quality evidence and credible health advice. However, both lay and
professional behaviours often diverge from scientific evidence or professional guidance.
Two disciplines help explain what people actually do and why. Sociology describes the rules and processes of groups
– communities, cultures and organisations. Economics provides insight into how decisions are made in a world with
scarce resources and infinite needs.
Section 4 explores different concepts of health, and the factors that underlie the ways that people seek and use
health care, in order to influence behaviours. The distinctions between equity and equality are discussed, together
with their impacts on health policy. Finally, the section on health economics provides the background and core
principles of the discipline to equip practitioners with the insight to lobby more effectively for equitable and cost-
effective approaches to health-care provision.

373

Public [Link] 373 3/6/08 [Link]

Public [Link] 374 3/6/08 [Link]
 4A Health and I llness

4A
Health and Illness

4A.1 Human behaviour 375 4A.6 Social and structural iatrogenesis 383
4A.2 Illness as a social role 376 4A.7 Role of medicine in society 383
4A.3 Concepts of primary and secondary 4A.8 Social patterns of illness 384
deviance 379 4A.9 Social factors in the aetiology of disease 386
4A.4 Stigma and how to tackle it 379 4A.10 Social capital and social epidemiology 387
4A.5 Disability and handicap 382

Reaction to illness has certain common features across cultures and some important differences. This chapter
explores the social role of illness, together with the norms and behaviours that are tacitly expected of those who
are ill. In disability and handicap and stigma the chapter touches on the preconceptions surrounding health and
ill health. Finally, it considers how personal characteristics and position in society affect how, and whether, people
seek help for their symptoms.

4A.1 HUMAN BEHAvIOUR

Theoretical perspectives and methods of enquiry of the sciences concerned with human behaviour
The social sciences aim to understand the attitudes, motivations and behaviours of human social behaviour and
why these change over time. Society is a group of interacting people who share a geographical region, a sense of
common identity and a common culture. As such it is more than an aggregate of individuals. The social sciences
encompass the fields of study in Box 4A.1.1.

Box 4A.1.1
Psychology Study of individuals’ mental processes and behaviour
Sociology Study of social processes and interactions in societies, groups and institutions. Sociology
recognises that people in societies may behave in ways that differ from the behaviour of
individuals
Anthropology Study of human cultures
History Recording and interpretation of past events

375

Public [Link] 375 3/6/08 [Link]

 4A He a l t h a nd Il l ness

These disciplines are of importance to public health insofar as they can help explain:
• Individual behaviour patterns
• Behaviour of groups within a population
• Behaviour of health-care organisations.
Data from social research may be quantitative (numerical data) or qualitative (textual data), although in practice
most research considers data of both types.

QUANTITATIVE METHODS
See also Sections 1A.25–1A.29.
This asks questions such as ‘How many?’ and ‘What proportion?’ Examples include:
• Questionnaires
• Surveys (face-to-face or telephone)
• Routine data sources (e.g. mortality data).
As with all quantitative research, three potential causes of error that should always be considered are:
• Chance
• Bias
• Confounding.

QUALITATIVE RESEARCH
See also Section 1D.
This asks questions about ‘How?’ and ‘Why?’ Methods include ethnography, interviews, focus groups and case studies:
see Box 4A.1.2.

Box 4A.1.2
Ethnography This is an anthropological research method in which the investigator studies a group’s behaviour
in great detail, often by living among them for a protracted period of time
Interviews This can either be a semi-structured interview (loose set of questions) or an in-depth interview
(respondent guides the conversation)
Focus groups The researcher brings together a small group of up to 15 people. The researcher uses in-depth
interview techniques and is interested in the group’s opinions and deliberations
Case studies Multiple data-collection methods are used to generate a rounded picture of a ‘bounded system’,
i.e. an organisation fixed in place and time. An example might be a particular GP surgery in
2007

TERMINOLOGY
The theoretical terminology shown in Box 4A.1.3 may be encountered in the social sciences literature.

4A.2 ILLNESS AS A SOCIAL ROLE

The concept of illness as a social role introduces the notion that people who feel ill, and those who care for and
treat them, behave in ways that are related to society’s implicit ideas of what it means to be sick. The American
sociologist, Talcott Parsons, described this as ‘the sick role’.

376

Public [Link] 376 3/6/08 [Link]

 4A Health and I llness

Box 4A.1.3
Epistemology This is the study of knowledge: its origin, nature, methods and limits. Epistemology is
concerned with what it means to be ‘true’ or ‘false’, what constitutes valid ‘information’, and
whether information is absolute or relative
Ontology Ontology is the study of being. It considers whether facts are constructed in people’s minds or
whether they exist in an external world
Positivism Social scientists who advocate positivism value the scientific method. They believe that the
social world can be studied in the same way as the material world: hypotheses can be tested
according to observable facts. Positivists often employ a quantitative approach
Constructivism This philosophy is based on the premise that our understanding of the world is constructed by
reflecting on our experiences. Each of us generates our own ‘rules’ and ‘mental models’ which
we use to make sense of our experiences
Reflexivity This acknowledges that, through the process of observing, researchers always affect the
environment that they are studying

THE SICK ROLE

Parsons (1951) wrote that people who are ill have rights and responsibilities that work together in the interest of
society: see Box 4A.2.1.

Box 4A.2.1
Rights of the sick Responsibilities of the sick
Exemption from blame for having their illness Duty to seek medical assistance
Exemption from normal responsibilities such as Duty to want to get better
work

These rights and responsibilities are all both temporary and universal.
Strengths and weaknesses of this approach are listed in Box 4A.2.2. An example of its use is presented in Box
4A.2.3.

Box 4A.2.2
Strengths Weaknesses
Applies well to acute infections, e.g. cold, flu Applies less well to chronic conditions (or some
acute illnesses) where:
• Individuals can be ‘blamed’ for their ‘illness’, e.g.
obesity, sexually transmitted infections
• There is no need for individuals to be exempted
from normal duties, e.g. controlled diabetes,
asthma
• Medical assistance not always perceived as
‘helpful’, e.g. Huntington’s disease, inoperable
cancers
• A duty to ‘want to get better’ is part of the ‘sick
role’, e.g. one of the symptoms of schizophrenia
is a lack of insight into the condition

377

Public [Link] 377 3/6/08 [Link]

 4A He a l t h a nd Il l ness

Box 4A.2.3
Example: The sick role in a New Guinea village
Gilbert Lewis’s ethnographic descriptions of sickness in a New Guinea society highlight variations in the way that
different cultures treat those with sickness. In his account, both western medical approaches and local rituals
are used to attempt to cure a sick man. The source of illness is sought in his own previous behaviour (fights,
disputes with others); the spiritual cures attempted include crucifixes round the bed and a Malyi ceremony. Many
of these behaviours contrast with the Western ideas of the sick role but also emphasise that a ‘sick role’ is not
confined to Western cultures.
Reproduced from Lewis (2000).

THE DOCTOR–PATIENT ROLE

Doctors often face a conflict between acting in their patients’ best interests and serving the wider interests of
society. For example, if a doctor saw a patient who worked as a lorry driver and the patient reported having had
a blackout, then the doctor would be obliged to inform the Driver and Vehicle Licensing Authority (DVLA), thereby
jeopardising the lorry driver’s livelihood.
Scambler (1997) describes the traditional Box 4A.2.4
doctor–patient role as ‘paternalistically doctor
centred’ but in recent years there has been a Patient centred Doctor centred
shift in some countries towards more patient- Consultation style: Consultation style:
centred care. See Box 4A.2.4 and Figure 4A.2.1. • Open questions • Closed questions
It is increasingly recognised that patients • Disease centred
Focused on:
and professionals each have their own area of • Patients’ Focused on:
knowledge and expertise, and that both parties experience of • Reaching a diagnosis
benefit from working together. illness • Patient compliance
• Reaching
concordance
UK The NHS is promoting such cooperation
by means of the Expert Patients Programme
([Link]), which uses lay facilitators
to empower patients to make the most of their contact
More doctor centred
time with professionals and to engage in shared decision-
making.
See Box 4A.2.5.

Default Paternalistic
More patient centred

Consumerist Mutuality

Figure 4A.2.1 The doctor–patient role

378

Public [Link] 378 3/6/08 [Link]

 4A Health and I llness

Eng Box 4A.2.5

Example: Combating drug taking and deviant behaviour through Drug Testing and Treatment Orders
Use of illicit drugs is a prime example of deviant behaviour in Western societies, and it is often linked to a range
of other deviant behaviours (e.g. stealing, prostitution). However, it is also recognised in many Western cultures
as an addiction – and, as such, an illness. In England, the programme of Drug Testing and Treatment Orders
(DTTOs) for drug misusing offenders explicitly links this deviant behaviour with the sick role. Those arrested who
have a history of taking illicit drugs are required to attend intensive treatment and rehabilitation for their drug
use in an attempt to break the cycle of drugs and crime. The DTTOs are a community sentence and are provided
as an alternative to prison (the traditional way of dealing with deviant behaviour).
For more information on DTTOs and their effectiveness, see: Comptroller and Auditor General (2004) The Drug
Treatment and Testing Order: early lessons. London: National Audit Office. Available online at: [Link]/
publications/nao_reports/03-04/[Link].

4A.3 CONCEPTS OF PRIMARY AND SECONDARY DEvIANCE

Unacceptable behaviour within a particular culture is known as deviance. Behaviour seen as perfectly acceptable in
one culture may be regarded as unacceptable in another (Scambler 1997). On being recognised, deviant behaviour is
subject to sanctions, punishment, ‘correction’ or ‘treatment’.
In medicine, deviance has labelling implications with regard to organic and psychiatric disease. Parsons (1951)
considered illness as a form of deviance where the doctor is an agent of social control (i.e. the doctor ’restricts’
access to the sick role by determining who is sick and who is healthy).

PRIMARY AND SECONDARY DEVIANCE

Lemert (1967) differentiated between primary and secondary deviance: see Box 4A.3.1.

Box 4A.3.1
Type of deviance Description Example
Primary Relates to the deviant behaviour Rape
Secondary Relates to the deviant status Rapist

It is important to be aware of secondary deviation because society’s reaction to labelling can sometimes hamper
treatment and thereby reinforce the deviant behaviour.

4A.4 STIGMA AND HOW TO TACKLE IT

A stigma is a mark of disgrace or infamy. The American sociologist Erving Goffman (1963) defined stigma as:
‘An attribute that is deeply discrediting within a particular social interaction.’
This underlines two features of stigma: that it is an undesirable characteristic in a particular context (i.e. in a
particular time and society). For example, in the 1950s in England, knowledge that a person was born outside of
marriage would have been considered shameful, but in twenty-first-century England, it is considered normal.
In the medical literature, stigma is used to describe diseases or conditions that lead to exclusion from society: see
Box 4A.4.1.

379

Public [Link] 379 3/6/08 [Link]

 4A He a l t h a nd Il l ness

Box 4A.4.1
Examples of stigmatised diseases
• Leprosy
• Psychiatric illnesses
• Epilepsy
• HIV/AIDS
• Sexually transmitted infections (STIs)

CAUSES OF STIGMA
Stigma is rooted in ingrained cultural norms. Stigma thrives on inequalities, fear and misinformation: see Box
4A.4.2.

Box 4A.4.2
Inequalities Women (e.g. HIV-positive women in Africa)
Marginalised groups (e.g. homosexual men, transgender people, prostitutes)
Fear Fear among the public of having to deal with a person having a fit is often a cause of the
stigma attached to epilepsy
Misinformation There is a popular misconception that the term ‘schizophrenia’ means dual personality,
whereas in fact schizophrenia is characterised by impaired social functioning, distorted
thought and hallucinations

The media and religious groups may perpetuate these stigmas, and in so doing will bolster those who are not
currently in a stigmatised group.

CONSEQUENCES OF STIGMA
Stigma affects individuals and society, and is manifested as either expressed stigma or as enacted stigma: see
Box 4A.4.3.

Box 4A.4.3
Felt stigma Enacted stigma
Shame and guilt Loss of job
Self-stigmatisation Compulsory testing
Depression Violence
Unwillingness to speak up Quarantine
Withdrawal from society Denial of health services

380

Public [Link] 380 3/6/08 [Link]

 4A Health and I llness

TACKLING STIGMA
Tackling stigma benefits stigmatised individuals and society as a whole. For example, reducing the stigma of STIs
removes barriers to diagnosis for people with genitourinary symptoms. Early treatment of such symptoms benefits
the individual (e.g. reduces the risk of infertility) and reduces the spread of the infection among the population.
Tackling stigma requires changes in the attitudes and behaviours of both the stigmatised and society at large. Ways
of tackling stigma include those shown in Box 4A.4.4.

Box 4A.4.4
Measure Description Example
Education Public education by means of challenging World AIDS day initiatives
negative stereotypes and raising awareness
of illness
Language Challenging the language that is used to Promoting the term ‘people with
describe illness schizophrenia’ instead of ‘schizophrenic’
because the illness does not define a
person’s entire identity
Public Public acknowledgement of illness by In 1985, Rock Hudson publicly declared
acknowledgement celebrities that he was homosexual and that he was
dying from AIDS
Treatment Advances in the management of illness Newer antipsychotic therapies that do
not produce parkinsonian-like symptoms
reduce the visible marks of illnesses such as
schizophrenia
Legislation Certain manifestations of stigma can be The UK’s Disability Discrimination Act
outlawed provides a legal framework for promoting
the rights of disabled people

A ‘virtuous circle’ may be established where positive challenges to stigma can change attitudes and thereby reduce
felt stigma. This in turn can reduce enacted stigma: see Figure 4A.4.1.

Stigmatised
group

Reduces
Reduces felt
enacted
stigma
stigma

Figure 4.4.1 A virtuous circle to challenge stigma

381

Public [Link] 381 3/6/08 [Link]

 4A He a l t h a nd Il l ness

4A.5 DISABILITY AND HANDICAP

See also Section 1C.
In 1980 the WHO defined impairment, disability and handicap as distinct but interrelated concepts: see Box 4A.5.1.
These have since been updated, see Table 1C.7.1.

Box 4A.5.1
Concept Impairment Disability Handicap
Definition A loss or abnormality of a An inability or restricted ability A disadvantage due to
body function (anatomical, to perform an activity (in the impairment or disability that
physiological or psychological) normal human range) limits the role of an individual
Description Malfunctioning body parts or Activities a person cannot do Social sequelae of impairment
systems or disability
Example Bilateral above-knee Unable to walk: wheelchair Unable to mingle while
amputations bound socialising in bars

MEASURING DISABILITY
In hospital settings, the Barthel ‘Index of Activities of Daily Living’ (ADL) score is generally used to assess disability.
Patients’ independence is assessed in 10 domains: see Box 4A.5.2.

Box 4A.5.2
Domains of the Barthel index
Bowels Transfer
Bladder Toilet use
Feeding Walking
Grooming Stairs
Dressing Bathing

The index is scored out of 20 and describes the level of support that will typically be required (e.g. a patient scoring
10 will require a maximal package of home care; patients with scores >10 will need residential or nursing home
care).

MEASURING HANDICAP
Various tools exist for assessing handicap, including:
• Rankin scale (used in stroke research)
• Hearing Handicap Inventory
• London Handicap Scale (domains include mobility, physical independence, occupation, social functioning and
economic self-sufficiency).

THE SOCIAL MODEL OF DISABILITY

In contrast to the above indices, the social model of disability considers how society disables those with physical
impairments by means of a wide range of barriers. These barriers are both environmental (e.g. no wheel ramps at
entrances to buildings) and cultural (e.g. patronising attitudes towards people with impairments).

382

Public [Link] 382 3/6/08 [Link]

 4A Health and I llness

Eng In England, the Disability Rights Commission (now replaced by the Equality and Human Rights Commission)
describes how the:
‘Arrangement of transport, leisure facilities, public services and work excludes disabled people and how people’s
attitudes also demean and isolate.’

4A.6 SOCIAL AND STRUCTURAL IATROGENESIS

In his book Medical Nemesis, Ivan Illich (1975) introduced the concept of iatrogenesis, i.e. disease caused by
medicine. He described three ways in which medicine can cause illness: clinical, social and structural.

CLINICAL IATROGENESIS
Medical treatment sometimes worsens the original illness or creates a new illness. Examples include:
• Adverse drug reactions
• Diabetogenic drugs (e.g. steroids, certain combinations of antihypertensive drugs)
• Hospital-acquired (‘nosocomial’) infections.

SOCIAL IATROGENESIS
This describes the way in which medicine invades normal life: see Box 4A.6.1.

Box 4A.6.1
Aspect of life Medicalisation
Normal childbirth Caesarean sections on demand
Ageing Cosmetic surgery
Unruly children Attention-deficit hyperactivity disorder

This is reflected by the growing proportion of the GDP that is spent by many countries on health care.

STRUCTURAL IATROGENESIS
This is the impact that the medical profession has upon a population. As a result of increasing reliance upon
medicine, the public has lost its traditional ways of coping with illness, death, pain and misfortune.
‘The so-called health professionals have an even deeper, structurally health-denying effect insofar as they destroy the
potential of people to deal with their human weakness, vulnerability and uniqueness in a personal and autonomous
way.’ (Illich 1975, p26).

4A.7 ROLE OF MEDICINE IN SOCIETY

The role of medicine has been described as being:
‘To cure sometimes, to heal often and to comfort always.’ (Variously attributed to Hippocrates, to a fifteenth-
century French proverb, and to Sir William Osler)
In particular, the role of biomedicine (i.e. branch of medical science that applies biological and physiological
principles to clinical practice) has become an increasingly dominant – though not always unchallenged – part of
western society.

383

Public [Link] 383 3/6/08 [Link]

 4A He a l t h a nd Il l ness

EXPANDING BOUNDARIES OF MEDICINE

Health-care professionals and professional organisations are now routinely involved in areas of life that were
previously outside the scope of medicine. Key examples of the role of medicine in areas previously not the preserve
of health care are:
• Childbirth
• Euthanasia
• Abortion.

CHALLENGES TO THE ROLE OF MEDICINE

The place of medicine in Western society is not undisputed. Some of the areas where it is frequently challenged are
shown in Box 4A.7.1.

Box 4A.7.1
Clinical iatrogenesis Well-publicised accounts of side effects and poor outcomes as a result of
medical care have led some commentators to question the dominant role of
medicine in society
See Section 4A.6
Anti-psychiatry More than other medical disciplines, psychiatry has been challenged. Many
commentators, but especially RD Laing in the 1960s, have questioned the
whole notion of mental illness. In western societies, however, psychiatry
remains the dominant, if contested, model of care for mental ill health
Health literacy Growth in consumer access to health information (e.g. via the internet)
is removing the absolute dependence of the public on the opinions of
doctors. Medical opinion becomes only one of many sources of information
Complementary and Other approaches to treating illness are available, and many such
alternative health care treatments are growing in popularity

4A.8 SOCIAL PATTERNS OF ILLNESS

Explanations for various social patterns and experiences of illness (including differences of gender, ethnicity,
employment status, age and social stratification)
Subjective (i.e. experienced) health differs markedly from objective health. According to the Health Survey for
England, approximately 1% of the population regards itself as being in ‘very bad health’, with the proportion
being correlated to age. A study by Scambler 1997 found that a typical person experiences symptoms of illness
approximately one day in three. However, only on about 6% of occasions will the person consult a health
professional. Whether the person seeks medical advice is determined by a number of triggers, including:
• Presence of a concomitant crisis
• Sanctioning by others
• Interference with normal activity
• Temporising deadlines.
Some explanations for the varying patterns of illness seen between different social groups include:
• Biology and genetics
• Behaviour (differences in health seeking or risk taking between groups)

384

Public [Link] 384 3/6/08 [Link]

 4A Health and I llness

• Social circumstances (e.g. disproportionate effects of poverty on health)

• Artefact (the categories used to distinguish social groups are in themselves problematic).

GENDER
Clearly, certain diseases affect only men (e.g. prostate cancer) or women (e.g. endometriosis). Others are commoner
in men (e.g. renal stones) or in women (e.g. gallstones). Differing behavioural patterns also affect morbidity and
mortality, e.g. smoking, dangerous driving.
Although life-expectancy is lower for men, women report more ill health. This is thought to be partly due to
different reporting behaviours between the sexes.

ETHNICITY
Certain diseases are commoner in some ethnic groups, e.g. sickle cell disease in black people. Reasons include
genetic differences, consanguineous marriage, as well as:
• Poverty: in the UK, minority ethnic groups are mostly less affluent
• Migration: loss of social capital (see Section 4A.10)
• Behavioural: differing patterns of smoking, diet, etc.
• Racism (direct and indirect) affects health through increased stress and social isolation
• Access to health services.

EMPLOYMENT
Fulfilling and secure employment provides not only material resources for individuals and their families, but also
psychological and social support. A summary by the Health Development Agency (now part of NICE) underlined the
effects of unemployment on:
• Physical health: unemployment is associated with mortality (greater suicide rates and cardiovascular mortality)
• Mental health: those who are unemployed or in insecure employment are more prone to common mental
disorders, e.g. depression.
The relationship between unemployment and health is complex. In some circumstances, health problems may be a
factor in losing a job, while, in others, the loss of work may precipitate health problems.

AGE
There are clearly different patterns of mortality at different ages, but experiences of health services and illness also
vary with age. For example, surveys indicate that older people are generally more positive about the standard of care
that they have received than are younger people.

SOCIAL CLASS
Social class gradients in health occur at every age (e.g. low birthweight) and for all major causes of death.
UK In the UK, the Black Report (1980) found that, despite general improvements in health and prosperity in
the UK, there were still pronounced, and possibly increasing, disparities in health and illness across the five social
classes. The report was the work of the Department of Health Research Working Group on Inequalities in Health and
was led by Sir Douglas Black. It suggested four possible explanations for differences in health: see Box 4A.8.1. The
response to the report is discussed in Section 4C.10.

385

Public [Link] 385 3/6/08 [Link]

 4A He a l t h a nd Il l ness

Box 4A.8.1
Artefact The association between social class and health is an artefact of the way in which
these concepts are measured
Social selection Health determines social class through the process of health-related social mobility
Behavioural and cultural Social class determines health through social class differences in health-damaging
or health-promoting behaviours
Materialistic Social class determines health through differences in the material circumstances of
life (e.g. Whitehall studies)
This may manifest directly (e.g. accidents), indirectly (e.g. social capital) or
through psychosocial mechanisms

An example of the relationship between occupation and health is shown in Box 4A.8.2.

Box 4A.8.2
Example: Occupation and health – Whitehall studies I and II
The first Whitehall cohort study examined mortality rates over 10 years among male British civil servants aged
20–64 in the 1960s and 1970s. This revealed differences in health between different employment grades:
• Men in the lowest grade (messengers, doorkeepers, etc.) had a threefold higher mortality rate than men in
the highest grade (administrators)
• Blood pressure at work was associated with ‘job stress’, including ‘lack of skill utilisation’, ‘tension’ and ‘lack
of clarity’ in tasks. The rise in blood pressure from the lowest to the highest job-stress score was much larger
among low-grade men than among upper-grade men. Blood pressure at home, on the other hand, was not
related to job-stress level
A second longitudinal study of British civil servants (Whitehall II) started in the 1980s and focused on
occupational effects on health and disease. The study involved around 10 000 men and women aged 35–55 in the
London offices of 20 civil service departments, and many people in the cohort are still being followed up. The
study found that employment grade was strongly associated with work control. Lack of control in the job was
related to long spells of absence and an increased risk of cardiovascular disease.
Reproduced from [Link]/projects/[Link].

4A.9 SOCIAL FACTORS IN THE AETIOLOGY OF DISEASE

Role of social, cultural, psychological and family relationship factors in the aetiology of illness and disease
See also Section 4A.8.
These factors are listed in Box 4A.9.1.

386

Public [Link] 386 3/6/08 [Link]

 4A Health and I llness

Box 4A.9.1
Social Inequalities, environment, economic
Cultural Lay health beliefs, explanatory models
Psychological Health beliefs (perceived costs, risks and benefits of health behaviour)
Locus of control – internal, external orientations
Family relationships Benefits to health of stable relationships
Lay referrals (where family members sanction each other to seek professional advice)

4A.10 SOCIAL CAPITAL AND SOCIAL EPIDEMIOLOGY

Although social capital and social epidemiology were both first described in the mid-twentieth century, academic
interest in the two concepts only began in earnest in the early 1990s. The concepts are now widely regarded as
being important and meaningful across the social sciences.

SOCIAL CAPITAL
Social capital gives a value to the social networks in which individuals live. These networks provide norms (i.e.
defined limits of acceptable behaviour) and sanctions when these bounds are crossed (e.g. social exclusion, gossip,
stigma). Social capital functions in two ways:
• Bonding: social capital strengthens the links between members of families and tight-knit communities and
provides social support
• Bridging: social capital strengthens the links with members outside the group, i.e. ‘networking’.
Social capital can be considered at the micro- (individual), meso- (community) and macro- (national) levels. At all
three levels it is demonstrably correlated with economic affluence, low crime, educational attainment and health.
The WHO regards social networks as a determinant of health, and it advocates the provision of social support to
improve health outcomes through increased social capital.

SOCIAL EPIDEMIOLOGY
Social epidemiology is the study of the social determinants of the Box 4A.10.1
distribution of disease within a population. Multi-level analysis is
used to determine to what extent an individual’s health is shaped Social determinants of health
by micro (individual), meso (household/small area) and macro
• Social gradient
(large area) characteristics.
• Stress
Macro-level epidemiological factors may be compositional (e.g. • Early life
prevalence of childhood poverty) or contextual (e.g. population • Social exclusion
density) – with the latter being irreducible to the individual. • Work
• Unemployment
Social determinants of health are listed in Box 4A.10.1.
• Social support
• Addiction
• Food
• Transport
Reproduced from Wilkinson and Marmot
(2003).

387

Public [Link] 387 3/6/08 [Link]

Public [Link] 388 3/6/08 [Link]
 4B Health Care

4B
Health Care

4B.1 Alternative sources of health care 389 4B.5 Illness behaviour 396
4B.2 Hospitals as social institutions 393 4B.6 Psychology of decision-making in health
4B.3 Professions 394 behaviour 397
4B.4 Clinical autonomy 395

Most of the discourse concerning health care focuses on formal health systems and the role of professionals
as health-care providers. In different approaches to health care, the topic is widened to encompass other
kinds of care, including self-help and complementary practices. This chapter also considers the social roles and
characteristics of health-care providers. In hospitals as social institutions, the archetypal providers of health care
are considered in terms of their functions in society other than health care and their potential to constrain the
actions of individuals. Professions looks at how professional status was created and is maintained, and examines
the situations where conflicts due to professional roles can arise.

4B.1 ALTERNATIvE SOURCES OF HEALTH CARE

Different approaches to health care (including self-care, family care, community care, self-help groups)
The decisions of whether or not to access health care and which type of health care to access are influenced by a
range of factors including:
• Characteristics of the person (age, gender, ethnicity, previous experiences with health care)
• Nature and duration of symptoms
• Accessibility of formal health care (cost, convenience, welcoming attitude).
The term clinical iceberg was used by Last (1963) to describe the finding that professional health services treat
only a small fraction of the total burden of ill health. It has been estimated that a typical adult experiences some
sort of somatic symptom once every 3–6 days (Barsky and Borus 1995). Assuming that some of these symptoms
will require health care, it is probable that the bulk of health care occurs in the so-called informal sector because
patients in Britain visit their GP only four or five times a year (see [Link]).

389

Public [Link] 389 3/6/08 [Link]

 4B He a l t h C a re

SELF-CARE
Self-care (taking care of oneself without professional assistance or oversight) is the commonest form of health care.
For a new symptom, such as a cough, a person may typically instigate a management plan such as waiting to see
what happens, unless a:
• Symptom changes (e.g. the cough becomes productive) or
• Symptom persists beyond a time limit (e.g. the end of the week).
In these circumstances, the person may choose to consult another person (e.g. a family member or a GP) or else try
an over-the-counter remedy (e.g. cough mixture).
Self-care remedies can be orthodox or complementary/alternative. Orthodox medicines in the UK can be accessed
without consulting a clinician for a prescription if they are listed on the pharmacy (P) or general sales list (GSL).
Restrictions apply both to which items may be sold and to the quantities that may be purchased (e.g. 32 tablets of
paracetamol on the P list and 16 on the GSL list). See Table 4B.1.1.

UK Table 4B.1.1 Classes of medications and their restrictions

Class of medication Abbreviation Restriction Example
Controlled drug CD Can be dispensed only with a detailed prescription. Diamorphine
Stored in locked cupboard; closely monitored
Prescription-only POM Can be dispensed only with a prescription Flucloxacillin
medication
Pharmacy list P Can be sold only under the supervision of a Ranitidine
pharmacist
General sales list GSL Available on open shelves, e.g. in supermarkets Ibuprofen

LONG-TERM CONDITIONS
Self-care accounts for the vast majority of treatment of any long-term medical condition. For example,
the Department of Health has calculated that a typical patient with diabetes has 3 h of contact time
with professionals per year, and will therefore self-care for the remaining 8757 h (see [Link]/
assetRoot/04/10/17/02/[Link]).
The population of people with a long-term condition is often represented as a triangle (called the Kaiser pyramid)
(Figure 4B.1.1), which depicts the large number of patients at the base of the triangle who have straightforward
conditions, rising up through the triangle to the small numbers at the top with complex disease.
The proportion of self-care undertaken by patients varies with the complexity of their condition: see Figure 4B.1.2.

COMPLEMENTARY OR ALTERNATIVE THERAPY

Complementary/alternative treatments can be delivered as self-care or by practitioners. Increasingly, complementary
practitioners share many of the attributes of health-care professionals (see Section 4B.3). Complementary therapies
range in complexity and the degree to which they can be self-administered: see Box 4B.1.1.
The use of complementary therapies can differ somewhat from conventional medicine. Their uses are not restricted
to the treatment of specific health complaints, but are also employed by the healthy for the maintenance and
promotion of wellbeing.

390

Public [Link] 390 3/6/08 [Link]

 4B Health Care

Level 3
Case
Patients with highly complex long-term conditions (LTCs) (�1% LTC population)
management

Level 2
Disease
Higher-risk patients (20–30% of LTC population)
management

Self- Level 1
management Lower-risk patients (70–80% of LTC population)

Figure 4B.1.1 Kaiser pyramid. Reproduced from [Link]/cms/336

Complex cases with

co-morbidities
High percentage of
professional
care
Pr
of

Equally shared
es

High-risk cases
sio

care
na
lc
S

ar
e
e
lf
-c

Figure 4B.1.2 Pyramid showing

a

proportion of patients who self-care,

re

under professional care and mixed

High
according to complexity of their
percentage condition. Reproduced from
of self-care Low-risk [Link]/assetRoot/04/10/17/
cases 02/[Link]

Eng The Health Survey for England asks about the use of complementary and alternative medicines among adults
(see [Link]). Data from 2004 indicate that:
• 40% of the general population have used some form of complementary and alternative medicine, though women
(51%) are more likely to use them than men (21%)
• the most commonly used forms were osteopathy (11%), massage (10%) and acupuncture (8%)
• over 90% of osteopathy and acupuncture, and 70% of massage therapies, were delivered by a practitioner. In
contrast, less than half of all those surveyed who used herbal medicine, aromatherapy and meditation visited a
practitioner for this therapy.

391

Public [Link] 391 3/6/08 [Link]

 4B He a l t h C a re

Box 4B.1.1
Examples: Complementary therapies
Homemade:
• Honey and lemon in hot water for a cold
• Avocado pulp as a skin moisturiser
Commercial:
• Ginseng for energy
• St John’s wort to combat depression
Self-administered practices:
• Meditation
Therapies delivered by a practitioner:
• Massage
• Acupuncture

FAMILY CARE
Friedson (1959) described the way in which people tend to discuss medical issues with family, friends or colleagues
before seeking professional advice. This lay referral system is used for:
• Interpretation of symptoms
• Reassurance
• Advice about a remedy
• Advice about referral to another lay person or professional.
Where lay culture and professional culture differ, a ladder of consultations begins with the nuclear family, through
progressively more distant and authoritative lay people, until the professional is reached. In contrast, where lay and
professional cultures are more alike, patients typically take a great deal of time trying to treat themselves, and then
go directly from self-treatment to consulting a doctor.

COMMUNITY CARE
As well as forming an important part of the lay referral system referred to above, the term community care has taken
specific meanings in certain countries:
• UK In Britain, the National Health Service and Community Care Act 1990 led to a large-scale relocation of
people with mental illness from large psychiatric hospitals into small local hostels or sheltered housing.
• Aus In Australia, the term relates to a joint Commonwealth, State and Territory initiative in which frail
older people and people with disabilities are given funding to support themselves in continuing to live in the
community.
• SA In South Africa, the term is typically applied to mean home-based care and support for people living with
HIV/AIDS.
• Ire In Ireland, the term includes public health nursing, home help, and out-of-hospital physiotherapy,
occupational therapy, chiropody service, day care, respite care service, etc.

SELF-HELP GROUPS
Self-help groups now exist for almost every conceivable medical condition – literally from achondroplasia to XXY
syndrome. Typically these societies exist to:

392

Public [Link] 392 3/6/08 [Link]

 4B Health Care

• Provide information to patients and their carers through leaflets, helplines and websites
• Put people in touch with others who are affected with that condition
• Raise funds in order to commission research into the condition
• Lobby government and clinicians.
The rise in self-help groups in recent times is attributed to a general desire by patients to take more responsibility
for their own health. As treatment options become more complex, and time with clinicians more limited, self-help
groups have filled the gap between professionals’ availability and the demand for disease-related information and
support. Their expansion has been facilitated by the rise of the internet and email.

4B.2 HOSPITALS AS SOCIAL INSTITUTIONS

There are various ways that hospitals can act as a social institution, shaping cultures and practices in their
community. Hospitals are not just a centre for treatment. They also function positively in many other ways in
society, as:
• Employers: for clinical staff; and non-clinical, including technicians, caterers, porters, builders, secretaries.
Health services are often one of the largest employers in a local area
• Purchasers: health care-related products, such as drugs and medical equipment, and other goods, e.g. food and
drink for patients, visitors and staff
• Community resource: hospital resources can be important for the local community, where rooms in the hospital
building are used for public meetings, hospital sports, catering, arts facilities.
There are negative ways in which a hospital can function, too:
• Polluter: travel to and from hospital by staff, patients and visitors adds significantly to congestion; hospital
waste makes an impact on carbon emissions
• Means of isolation or exclusion: at its extreme, there is an intentional element to separating patients with
infectious diseases from others. However, in general hospitals serve as a means of separating people from the
rest of society until they are healthy.

Box 4B.2.1
Example: Eng NHS as a corporate citizen
The Sustainable Development Commission is working with the NHS to improve its performance as a corporate
citizen. It highlights the market power of the NHS as a purchaser and its potential to exercise significant
leverage to influence practice. As a consumer the NHS spends around £17 billion a year. Every year it buys:
• 1.3m chicken legs
• 12.3m loaves of bread
• 13.5m kg of potatoes
• 250 000 l of orange juice
Reproduced from [Link].

As a responsible corporate citizen (see Box 4B.2.1), the NHS can engage in a range of processes with wide-ranging
potential benefits, covering transport, procurement, facilities management, employment and skills, community
engagement and new buildings.
Goffman (1963) defined total institutions as places where people are isolated from society over a period
of time and lead life in an enclosed and formally administered way. He went on to consider the effects of
institutionalisation on social relationships in the outside world, the ways in which people adapt to and become
attached to the institution, and the complicit role of medicine in the process. See Box 4B.2.2.

393

Public [Link] 393 3/6/08 [Link]

 4B He a l t h C a re

Box 4B.2.2
Erving Goffman: asylums
Goffman used participant observation to study the inner workings of an American psychiatric hospital. Working
as a hospital porter, he observed the effects of institutionalisation – notably the staid behaviour patterns of both
staff and patients.
Goffman found that institutionalised patients were apathetic and became progressively less able to make decisions
and care for themselves. Ways have subsequently been found for avoiding these negative effects, including:
• Providing information to patients prior to treatment
• Promoting mobility and self-care while in hospital
• Pre-discharge planning and education
• Reducing length of stay
• Community care

4B.3 PROFESSIONS
Professions, professionalisation and professional conflicts
See also Section 5A.11.
The three original professions – clergy, medicine and law – have been characterised by the following traits:
• Specialist area of knowledge
• A professional association
• Ethical code
• Control over certification or licensing.
Medicine is often used as a model for the study of professions. In contrast, the status of nursing as a profession has
been controversial and taken much longer to establish.

PROFESSIONS
Talcott Parsons in the 1950s described various aspects of the medical profession and its effects on the relationship
with patients:
• A clearly defined knowledge base that is highly developed, theoretical and specialised
• Patients expected to defer to doctors’ authority
• Self-governing and self-policing
• Potential to exploit power over patients for financial gain
• A commitment to public service and ethics
• Protection for patients against exploitation.
In the 1960s, nursing was often described as a ’semi-profession’ because it lacked the powers of self-regulation and
a specialised body of knowledge.

PROFESSIONALISATION
Sociologists have studied the emergence of the professional status in medicine. Professionalisation tends to involve
establishing:
• Acceptable qualifications (e.g. in medicine, Larson [1977] argues that the introduction of a university-based
medical degree improved the credibility of the profession)

394

Public [Link] 394 3/6/08 [Link]

 4B Health Care

• A professional body (e.g. General Medical Council) to oversee the conduct of members of the profession
• Occupational closure (where there is no entry from outsiders, amateurs or the unqualified).
Successful professionalisation protects members of an occupation from external control, and acquires specialised
knowledge, monopoly and autonomy for the profession which is guided by a code of professional ethics.

PROFESSIONAL CONFLICTS
Parsons’ functionalist account of how medicine works as a profession has since been challenged. Sociologists started
to question whether the espoused principles, e.g. of altruism, coincided with what doctors actually did. Feminist
critiques of the profession focused on the gendered nature of the profession:
• In the 1970s, professions (such as medicine) were largely populated by men
• Professions seem to espouse ‘traditional masculine’ values (technical expertise, rationality)
• Entrance to a profession was geared more towards the opportunities offered to men rather than women.
The medical profession is currently prone to conflicts in a number of areas:
• With nursing and allied health professions regarding disputes over professional boundaries
• With complementary practitioners over recognition
• With management regarding issues of professional autonomy and self-control
• With patient groups regarding issues of consumerism and paternalism
• With government over terms and conditions of employment, and health service policy.

4B.4 CLINICAL AUTONOMY

Role of clinical autonomy in the provision of health care
As discussed in Section 4B.3, autonomy (control over terms and conditions of work – clinical, financial,
organisational) is a key attribute of professions. Clinical autonomy refers specifically to the control over content and
delivery of health care.
In the 1960s, 1970s and 1980s, sociology focused largely on how the medical profession had achieved its autonomy.
Since the 1990s, the focus has shifted towards threats to the medical profession‘s autonomy. Threats to clinical
autonomy include:
• Increasing role of management in health care, in setting clinical priorities, monitoring standards of care
• Cost containment: since all health-care decisions are spending decisions, cost-containment is necessary –
explicitly or implicitly – to safeguard resources. Explicit cost containment rules can limit clinicians’ capacity to
make decisions based on the characteristics of the individual
• Guidelines and protocols: some clinicians argue that the existence of guidelines and protocols removed
the role of clinical judgement from medicine. Alternatively, David Armstrong (2002) views evidence-based
medicine as the medical profession’s response to falling public trust in clinical practice
• Both external assessment and revalidation open doctors’ practice to greater scrutiny and to scrutiny from
those outside the profession (see also 5A.11)
• Market forces: in a private health-care system, in particular, consumer satisfaction is required in order to
remain a viable business.
The structure of health-care systems is often balanced so there is a trade-off between clinical autonomy, financial
autonomy and control over terms and conditions of work: see Box 4B.4.1.

395

Public [Link] 395 3/6/08 [Link]

 4B He a l t h C a re

Eng Box 4B.4.1

Example: A new contract for GPs
Since its inception in 1948, the NHS has not employed GPs. Instead, they work as small businesses contracted to
provide medical services, and hence have had a large degree of autonomy.
In 2003, the majority of GPs voted to accept a new contract with the NHS for their medical services. The
contract in general led to a substantial pay rise for GPs, but the autonomy of the GP has, to some extent, been
traded for lesser clinical autonomy and a stronger role for all employed by the practice.

Earlier GP contract 2003 contract

Clinical Contract held between individual GP and Contract held between practice (not just
autonomy Secretary of State individual GP) and local NHS organisation
Clinical Contract not dependent on clinical practices Quality and Outcomes Framework rewards
Practice practices for certain clinical behaviours
Terms and GP responsible for patients 24 h a day GP can opt out of out-of-hours care
conditions
Financial GP pay not linked to service provision Pay related to services provided
autonomy
Reproduced from [Link]; [Link].

4B.5 ILLNESS BEHAvIOUR

Behaviour in response to illness and treatments
Definitions of ‘health’ and ‘illness’ vary across cultures, communities and households. Sociologists study illness
behaviour to learn why people seek or decline professional help.
Mechanic (1968) identified 10 variables that influence illness behaviour:
• Visibility of symptoms and signs
• Perceived seriousness (by the patient) of the symptoms, for present and future probabilities of danger
• Amount of disruption caused by the symptom to work, family, etc.
• Frequency and persistence or recurrence of signs or symptoms
• Tolerance threshold of person exposed to symptoms
• Knowledge, information and assumptions of the evaluator
• Basic needs leading to denial
• Needs leading to competition with illness
• Competing interpretations assigned to symptoms once recognised
• Availability of treatment: access, cost (not only money but also emotional, e.g. stigma).

CULTURAL vARIATION
Pilowski and Spence (1977) noted marked cultural differences between Anglo-Saxon (stoical, withdrawn) and
Mediterranean groups (experience) in their interpretation of a response to symptoms and signs. Zborowski (1952)
found that Americans of Irish origin had a matter of fact attitude towards pain, whereas people with an Italian or
Jewish background were more demanding and dependent on medical help.

396

Public [Link] 396 3/6/08 [Link]

 4B Health Care

PHENOMENOLOGY OF SYMPTOMS
Diseases that present with striking symptoms (e.g. severe pain, jaundice) are more likely to receive prompt medical
attention than those that are less dramatic.

LAY REFERRAL AND INTERvENTION

See Section 4B.1.1, Family care, a lay person may also intervene to initiate medical consultation, e.g. on behalf of a
child, or by calling an ambulance for someone who is having an epileptic fit or chest pain.

4B.6 PSYCHOLOGY OF DECISION-MAKING IN HEALTH BEHAvIOUR

The success of medical practice can be linked to patient behaviour. Understanding the issues that influence patients’
decisions when they choose not to follow preventive or therapeutic recommendations is instrumental to improving
concordance and, ultimately, to improving health outcomes.
There are a number of models and theories that attempt to offer explanations for why people behave as they do with
regard to health: see Section 2H.7.

397

Public [Link] 397 3/6/08 [Link]

Public [Link] 398 3/6/08 [Link]
 4C Equality, Equity and Policy

4C
Equality, Equity and Policy

4C.1 Need and social justice 399 4C.9 Power, interests and ideology 411
4C.2 Priorities and rationing 402 4C.10 Health inequalities 413
4C.3 Balancing equity and efficiency 405 4C.11 Migration and health effects of
4C.4 Consumerism and community participation 406 international trade 414
4C.5 Public access to information 408 4C.12 International influences on health and
social policy 417
4C.6 User and carer involvement 409
4C.13 Investment in health improvement 418
4C.7 Problems of policy implementation 409
4C.8 Formulating policy 410

The NHS in the UK was established to provide equal access for equal need. At first glance, this does not appear to be
a contentious or ambiguous aim. As will be seen in the beginning of this chapter, however, the concepts of access,
need, equality and equity can be complex and they may conflict with other priorities such as efficiency. The rest of
the chapter covers the extent to which equity and equality are present in society and policy, and how they fare in
competition with other priorities.

4C.1 NEED AND SOCIAL JUSTICE

Concepts of need and social justice
The concept of need relates to the capacity to benefit.
Social justice is characterised by the availability of equal rights, opportunities, obligations and benefits available
to all members of the population.

NEED
Given that the resources available to any health service are finite, the issue of determining the relative needs
of patients will always be crucially important. In terms of allocating health-care resources, Bradshaw (1972)
differentiated need into four categories: see Table 4C.1.1.

399

Public [Link] 399 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

Table 4C.1.1 Bradshaw’s categories of need

Type of need Description Measurement
Felt need This is also known as a need for health and relates to an individual’s Felt need can be
subjective experience of feeling unwell. Felt need does not necessarily measured through
relate to health service use. For example, someone with a headache may surveys, e.g. in the
well report that she is in pain but may seek no health service intervention UK, the census asks
about experience of
illness
Expressed This is synonymous with demand. Expressed need occurs when a patient Waiting lists can
need seeks health care for a felt need be used as a proxy
measure of expressed
need
Normative This is also known as a need for health care and relates to a Health needs
need professional’s judgement of an individual’s health state. A professional’s assessment
assessment of whether an individual has a normative need depends not
only on the experience of symptoms but also on a range of other factors,
such as whether an effective treatment exists, whether the treatment is
available and whether the patient is in a suitable state to benefit from the
treatment. Normative need is therefore dependent upon the health system
in place and the technology available. In public health, practitioners may
identify a normative need where individuals may have no felt need (e.g.
recommendation for weight loss in someone with a high BMI who has no
wish to change their size and lifestyle)
Comparative This is also known as relative need and describes the process of Deprivation and
need comparing the services available for areas with similar prevalence of mortality measures
disease and epidemiology. If one has a greater service provision, the other may be useful for
area could be said to be in relative need of health services. Comparator indicating a need for
areas could be neighbouring boroughs, or the country as a whole health but may not
reflect a need for
health care

In certain circumstances these needs may overlap, as is illustrated in the Venn diagram in Figure 4C.1.1.

SOCIAL JUSTICE
Humans come together as societies for their mutual benefit. However, for societies to function, they must curtail the
freedoms of each individual for the purpose of the greater good. Theories of social justice address:
• The extent to which individuals’ freedoms are (or should be) restricted
• How decisions are made in a society (and by whom)
• What constitutes ‘the greater good’.
Three major theories of social justice are utilitarianism, distributive justice and procedural justice.

UTILITARIANISM
Utilitarian philosophy as expounded by the eighteenth-century philosopher, Jeremy Bentham, stipulates that
individuals and societies should make their choices with the aim of achieving the greatest good for the greatest
number. Utilitarianism forms the basis for many economic and efficiency arguments within public health – but it
dismisses the principles of distributive justice and equity.

400

Public [Link] 400 3/6/08 [Link]

 4C Equality, Equity and Policy

Normative need
and expressed Felt and expressed
need, but no felt EXPRESSED need, but no
need (e.g. NEED normative need
immunisations, (e.g. minor
health promotion cosmetic surgery)
interventions)

NORMATIVE
FELT NEED
NEED

Felt need and

normative need,
but no expressed
need (e.g. where
stigma exists, such
as treatment for
sexually transmitted
infections)

Figure 4C.1.1 Venn diagram illustrating Bradshaw’s overlapping categories of need. Adapted from
Bradshaw (1972)

DISTRIBUTIvE JUSTICE
The twentieth-century philosopher, John Rawls, set out a theory of ‘Justice as Fairness’ that outlined two principles
for achieving a fair society: see Box 4C.1.1.

Box 4C.1.1
Basic liberties are a Restricting the individual liberties of some members of society is not justified even if
right for everyone it could lead to the greater good of society
Difference principle Resources need not be distributed equally, but social and economic inequalities
should benefit those who are most disadvantaged. Society can agree to pay a doctor,
for example, more than a cleaner because the doctor has the potential to save
people’s lives

Rawls proposed a model for societal decision-making named the veil of ignorance. In this model, those who make
resource allocations should do so from a stance of ignorance regarding their personal future position in society.
A health-care system supported by progressive taxes is an example of justice as fairness because, in general,
individuals pay tax with no knowledge of what health-care resources they will need in the future. On this basis,
society has deemed that those currently earning more should pay more tax in order to provide resources for those
who are currently unable to pay.

401

Public [Link] 401 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

PROCEDURAL JUSTICE
While distributive justice is concerned with the outcomes of society in distributing rights and resources, procedural
justice also examines the ways in which those outcomes are reached. Three models of procedural justice are based
on outcomes, balancing costs and outcomes, or participation: see Box 4C.1.2.

Box 4C.1.2
Outcomes Society should adopt procedures that produce the fairest outcomes
Balancing Society should adopt procedures that balance the costs and the benefits of the procedure.
For example, an insurance system that provides rapid settlement of claims but does not seek
to investigate the validity of those claims may pay out for invalid cases. However, it will
save resources through not carrying out laborious investigations
Participation Society should ask those who will be affected by a decision to choose which procedures to
adopt. This model holds that those who will be affected by a decision should be involved in
determining its outcome, whatever that outcome may be. This model is particularly relevant
to health systems that enshrine public participation as a central part of their service
development

4C.2 PRIORITIES AND RATIONING

Rationing in the context of health care can be defined as the process of allocating finite resources. The term is
often used synonymously with prioritisation.

RATIONING
‘There are two certainties in life: death and scarcity.’ (Maynard 2001)
In economics, the term scarcity refers to finite resources, which may be monetary, but also include materials,
equipment and human resources. Since in the field of health and health care there will always be finite resources
but infinite demand, a system of decision-making is required that will decide which services should be provided and
which should not.
Although explicit rationing systems (where transparent, consistent criteria are used to make entitlement decisions)
are contentious, rationing decisions are still made every day and at every level of health care. These decisions range
from:
• A GP receptionist’s judgement of whether a patient should be offered an emergency appointment (based on an
implicit decision of whether this is justified by the severity of symptoms), through to
• National guidance recommending or limiting the use of a new health-care technology (based on an evaluation
of its cost-effectiveness).

NATIONAL, ORGANISATIONAL AND INDIvIDUAL RATIONING DECISIONS

See Table 4C.2.1.

402

Public [Link] 402 3/6/08 [Link]

 4C Equality, Equity and Policy

Table 4C.2.1 Levels of rationing

Level Agency Advantages Disadvantages
Macro Government or state level (e.g. NICE Saves duplication of work No scope for local decision-making
guidance for health technologies in at local level or autonomy
the UK)
Meso Health-care organisation (e.g. local Can respond to local Little scope to address individual
exceptional treatments panel to circumstances circumstances
decide whether interventions sought
Risk of unnecessary duplication
by patients should be funded by a
if same activity happening in
health authority)
neighbouring areas
Micro Individual clinician (e.g. consultation Can respond to individual Vulnerable to inconsistencies and
duration and frequency; intensity and circumstances lack of accountability in decision-
scope of interventions provided) making

STAKEHOLDER INVOLVEMENT IN RATIONING

Rationing decisions are greatly affected by the views and experiences of decision-makers: see Table 4C.2.2.
Therefore, a wide range of stakeholders is to be encouraged. These may include:
• Voters
• Clinicians
• Patients
• Health-care management
• Pharmaceutical industry.

Table 4C.2.2 Ways in which the views and experiences of decision-makers can influence rationing
Perception of outcomes For example, if large breasts were seen as a necessity rather than an aesthetic
choice, then plastic surgery for breast augmentation would be more likely to
be funded
Acquisition of opinions Willingness-to-pay methods for valuing benefits will reach different conclusions
according to:
• An individual’s ability to pay
• An individual’s likelihood of requiring the service
• An individual’s likelihood of being required to pay for the service
Position of groups in society For example, attitudes towards older people may influence the degree of
funding provided for social care and dementia treatments, etc.
Social norms For example, given four hypothetical candidates for a heart transplant which
would be chosen from the following?
• A person who smokes
• A person who drinks excessive alcohol
• A person who is overweight
• A person who is in prison

403

Public [Link] 403 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

RATIONING CRITERIA
Rationing decisions in health care generally take into account the three criteria shown in Box 4C.2.1.

Box 4C.2.1
Necessity for an intervention This can apply at two levels:
• Individual (see Section 4C.1) (e.g. someone with severe atherosclerosis
is more likely to have a myocardial infarction than another person with
less severe atherosclerosis)
• Societal (e.g. societies with a declining birth rates may be more likely
to invest in IVF treatment for infertility than those with a rapidly
growing population)
Cost-effectiveness By definition, this includes only effective treatments. NICE uses cost per
QALY as a measure of the incremental cost-effectiveness ratio – ICER (see
Section 4D.8)
Fairness Individuals in similar circumstances should have equal access to care,
but the process for allocating resources should also be fair (i.e. it should
consider concepts of procedural justice – see Section 4C.1)

However, the relative weight applied to each criterion will vary. As a result, even if the criteria for allocating
resources are agreed, there can still be disagreement regarding the resultant decision. See Box 4.2.2.

USA Box 4C.2.2

Example: Setting health-care priorities in Oregon
In the late 1980s, the American state of Oregon chose to expand its publicly funded medical system – Medicaid –
so as to increase the proportion of the population that was eligible for state-funded treatment. The ambition was
for Medicaid to cover people on low incomes as well as those in absolute poverty. However, in order to remain
within budget, the range of services funded by Medicaid would need to be reduced.
The rationing process
Oregon undertook an extensive research programme to assess the cost-effectiveness of a range of treatments. The
outcome was a list of interventions ranked according to cost-effectiveness. The methods used initially produced
some counterintuitive rankings, e.g. tooth capping was rated similar to appendicectomy.
The list went through several iterations to ensure that it was both politically acceptable and methodologically
sound; then it was used to select which interventions Medicaid would fund.
‘Priority lists’ versus ‘rule of rescue’
While Oregon’s revised system provided access for more people on low incomes to certain interventions, those
individuals in the lowest income bracket initially saw their access to health care narrowed. For example, Oregon’s
initial priority list resulted in denial of bone marrow transplantation to a child with leukaemia who subsequently
died. The ensuing public outrage in Oregon typifies the inherent controversy of rationing processes. Application
of the rules leads to cases where a named individual, who intuitively should receive treatment but does not
fit the criteria, is denied a potentially life-saving intervention. This difficulty for society is termed the rule of
rescue, i.e. society feels obliged to ‘rescue’ the named individual. The process of rationing denies services to
some people and not others, and explicit rationing processes make this fact uncomfortably apparent.
Reproduced from Hadorn (1991).

404

Public [Link] 404 3/6/08 [Link]

 4C Equality, Equity and Policy

4C.3 BALANCING EQUITY AND EFFICIENCY

It is generally accepted in public health that the allocation of health-care resources should be guided by the
principles of efficiency and equity. However, the nature of these concepts depends on the exact definitions used.
Furthermore, a system that distributes resources equitably requires some sacrifice of efficiency and vice versa. See
Box 4C.3.1.

Box 4C.3.1
Concept Definition
Equity Fairness
Equality Same for all
Efficiency Greatest benefit achievable from a
given resource

CONCEPTS OF EFFICIENCY
Definitions of efficiency typically relate to a utilitarian philosophical position, i.e. to achieve the greatest
aggregated good across the greatest number in the whole community. For example, Donaldson et al (2004) defines
efficiency in health care as securing the:
‘… greatest improvements in wellbeing from available resources.’
Therefore, the concept of efficiency within health care depends on the definition of wellbeing. This is problematic
because an improvement in health status may not necessarily improve wellbeing. For example, offering people a
choice of how they wish to be treated may actually provide more wellbeing than would treating all individuals with
the same treatment known to have the best evidence of success.
A state of ultimate efficiency is described by economists as being a Pareto optimal state (not to be confused
with the Pareto principle [also known as the 80:20 rule] which states that, for many phenomena, 80% of the
consequences result from 20% of the causes). In these circumstances, efficiency has reached the point where no
further improvements can be made in one part of the system without disadvantaging others. In a health-care system
that is not Pareto optimal, providing more resources for one part of a health system may not disadvantage others,
and so this change is regarded as being an efficient choice.

CONCEPTS OF EQUITY
The two principal types of equity, vertical and horizontal, are described in Section 1C.10.

CONFLICT BETWEEN EFFICIENCY AND EQUITY

An intervention is efficient if it produces the greatest net health gains in a population for a given budget, whereas
to be equitable it should be distributed fairly within the population.
Overall improvements in health status may mask widening differences in access to health care between constituent
groups of the population. Even when access to services is similar, some groups within the population will require
extra resources in order to achieve the same health gain.

ALIGNMENT BETWEEN EQUITY AND EFFICIENCY

Conflicts between equity and efficiency can sometimes be reconciled by adjusting the definition of wellbeing, e.g.
by incorporating the concepts of distributive justice or externalities.

405

Public [Link] 405 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

DISTRIBUTIvE JUSTICE
Distributive justice states that, for a system to be just, its benefits should be shared fairly. Justice then depends
not only on the absolute amount of benefit received by individuals but also its distribution among them. A Pareto
efficient system which is inequitable (i.e. the system benefits everyone but some people receive unfair amounts
of benefit) is distributively unjust. By defining the wellbeing of individuals as being dependent both on their own
health status and on distributive justice, the intervention will be seen as inefficient.

ExTERNALITIES
An externality is a by-product of the production or consumption of goods that are enjoyed by society in general.
Equitable care provides positive (i.e. beneficial) externalities, and therefore improves the wellbeing of society. See
box 4C.3.2.

Box 4C.3.2
Example: A hypothetical equity/efficiency alignment in the provision of universal immunisations
In order to achieve herd immunity to measles, 95% of the population need to be vaccinated. Uptake of
vaccinations using a standard call–recall system with injections given at general practices is relatively cheap and
easy to use, and achieves 80% coverage of the population.
For a variety of reasons, the remaining 20% do not attend for immunisations offered in the standard way.
Additional resources will be required to encourage these groups to be vaccinated (e.g. mobile vaccination clinics,
targeted health promotion, individual calls and visits). Although the unit cost of vaccinating these hard-to-
reach people is much higher than the standard unit cost, it may actually be more efficient to spend this money
because the 95% coverage achieved will offer herd immunity to the whole population.

4C.4 CONSUMERISM AND COMMUNITY PARTICIPATION

In health and health care, consumerism transforms patients from being passive recipients of care into customers
who have the right and the capacity to choose whether and where to seek health or health care.

CONSUMER RIGHTS IN HEALTH CARE

In the UK health system, the concepts of the ‘patient as a consumer’ and ‘health care as a product’ were introduced
in the 1980s. In 1991, the Department of Health produced ‘The Patients’ Charter’ which set out a list of rights that
individuals could expect from health-care services. In common with many private companies, the NHS now uses
satisfaction surveys as a key tool to assess the quality of its provision.

PATIENT AND CONSUMER CHOICE

Consumers of products have the choice of whether to consume services and, if so, where. In England, several policies
to modernise the NHS have focused on providing more choice for patients. They are intended to:
• Introduce competition between health-care providers
• Improve the responsiveness of the health-care system to consumers’ preferences
• Enhance the efficiency of health-care services.
None the less, choice within health care can be constrained: see Table 4C.4.1 and an example illustrates this in Box
4C.4.1.

406

Public [Link] 406 3/6/08 [Link]

 4C Equality, Equity and Policy

Table 4C.4.1 Constraints in choice within health care

Health-care gatekeeping Where patients are obliged to see a GP (or other gatekeeper) in order to
access specialist health-care services
Urgency of treatment In circumstances of acute health-care need (e.g. following a road traffic
accident), individuals do not have the choice of whether to accept health care
and may lack the capacity to decide where to seek services
Disempowerment Individuals who are unwell in hospital may not feel as empowered to
question decisions that they would have challenged when well or on familiar
territory. In a small minority of circumstances (e.g. under a section of the
Mental Health Act), this freedom to challenge decisions is suspended
Information asymmetry Patients may be dependent upon clinicians to advise them about the
complexities of the management of their conditions. While growing access
to the internet has meant that some consumers have an abundance of
information available to them, this potentially widens inequalities. One
reason for this is that older people and people on low incomes are the least
likely to have internet access

Eng Box 4C.4.1

Example: Choosing health: making healthy choices easier
The language and policy recommendations in the government’s public health White Paper for England (DH, 2004)
typify the high profile of choice in the country’s health policy in the 2000s. The strategy explicitly recognises
individuals as consumers: e.g. its chapter on health in a consumer society notes that, in a market economy, it is
not for ‘Government to dictate … what [people] can and cannot consume.’
Recommendations focus on enabling individuals to choose health by means of:
• The provision of information to make an informed choice about whether to follow advice to promote or
protect health. Therefore, many of the recommendations in the paper centre on the provision of effective,
accurate and clear information. For example, there are recommendations for working with food manufacturers,
on providing information on the fat, salt and sugar content of products
• Availability of health products for consumers to choose, e.g. through increasing the availability of foods
with lower fat, salt and sugar content or in smaller portion sizes
• Creating demand for health, through the use of social marketing of health
Reproduced from Department of Health (2004).

CONSUMERISM AND COMMUNITY PARTICIPATION

Health care has traditionally had connotations of a passive role for patients, who were cared for by expert health
professionals. However, other models of health-care provision involving self-help and community involvement have
developed – both formally and informally. Formal examples include expert patient programmes and the provision
of health care through not-for-profit organisations and social enterprises: see Sections 2I.4, 3 and 4B.1, and Box
4C.4.2.

407

Public [Link] 407 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

Eng Box 4C.4.2

Example: NHS foundation hospitals – community participation in running health care
Foundation trusts have different regulatory and financial arrangements from other NHS trusts, and were first
introduced in England in April 2004. It is of note that they were established as public benefit corporations,
meaning that they have a board of governors (comprising patients, staff and members of the public) who have
responsibilities regarding the appointment of the trust’s chair, non-executive directors and chief executive,
and the right to be consulted on the trust’s strategic direction. Foundation trusts are also required to recruit
members, i.e. people in their local community who want to be informed about the management of the trust. In
this respect, they have the potential to bring true community participation to the hospital system.
In its early evaluation of foundation trusts, the King’s Fund found that they had attracted a large membership
(with over half a million people being members of the first wave of foundation trusts), but it was unclear
to what extent these member were representative of local populations. With a few notable exceptions, the
evaluation concluded that, ‘so far there is little compelling evidence that members or governors have made a
significant impact on the management of foundation trusts’.
Reproduced from Lewis et al (2006) and Foundation Trust Network (2005).

4C.5 PUBLIC ACCESS TO INFORMATION

Providing the public with access to health-care information has the potential to:
• Improve individuals’ understanding of health and disease, so as to inform their decisions about whether to seek
care and how to prevent disease
• Make services more accountable, through the provision of information on patients’ rights, the availability and
quality of services. This enables patients to decide where to seek care if they are given a choice
• Keep patients informed about their own health care. In recent years, health services have made moves to
improve access to information for patients about their own health. For example, in the NHS, patients now have
a right to see information in their medical records

UNDERSTANDING OF HEALTH AND DISEASE

The internet has transformed access to health information for many people, and according to the National Consumer
Council it has:
‘The potential to result in a much more active patient and more balanced relationship with health professionals.’
(Sihota and Lennard 2004)
However, alongside this explosion of information, the following issues should be borne in mind:
• Variable quality of information available
• Applicability of information to the individual patient
• Health literacy regarding technical terminology
• Potential to worsen health inequalities because of varying access to the internet.

FREEDOM OF INFORMATION ACT

UK The Freedom of Information Act 2000 applies in England, Wales and Northern Ireland, and there is a similar act
in Scotland. The Act applies to all ‘public authorities’, including:
• Central and local government
• NHS

408

Public [Link] 408 3/6/08 [Link]

 4C Equality, Equity and Policy

• Schools, colleges and universities

• Police
• Many other non-departmental public bodies, committees and advisory bodies.
The Act requires public authorities to specify the kinds of information that they publish, how it is made available,
and whether it is available free of charge or upon payment. In addition, the Act gives any person the legal right
to ask for, and be given, any information held by a public authority – although certain exceptions apply. Public
authorities must provide the information requested within 20 working days. Information can be withheld to protect
various interests which are allowed for by the Act, in which case an explanation that the information was withheld
and the reasons why should be provided.
Those requesting information can be asked to pay a small amount for making photocopies or postage. If the public
authority thinks that it will cost it more than a set amount (around £500 in 2007) to find the information and
prepare it for release, then it can turn down the request.
When people ask for information that a public body holds about them, the request is handled under the Data
Protection Act rather than under the Freedom of Information Act. There are slightly different rights under the two
Acts, with different fees and timeframes being applicable.

NZ Under the Official Information Act 1982, any person can request government agencies (ministers,
departments, local authorities) to release information which must be made available unless there is a good
reason for withholding it.

4C.6 USER AND CARER INvOLvEMENT

User and carer involvement in service planning
Since the 1980s, governments throughout western Europe and North America have encouraged patients to contribute
to the planning and development of health services. In England and Wales the involvement of patients is a key
component of current strategies to improve the quality of health care. Underlying these changes is the belief that
involving patients improves their health and quality of life. In particular, it leads to more accessible and acceptable
health services.

POLICY INITIATIVES TO PROMOTE INVOLVEMENT

UK The UK has a long history of patient and public involvement in policy initiatives (Department of Health 1999)
– in both health and social care. Key developments in this history include:
• Creation of community health councils that provided an early forum for public voices
• Development of a needs-based approach to the planning of health and social services
• Raising of expectations about patient rights and responsibilities through the Patients’ Charter
• Recognition of the skills of patients and carers in managing long-term conditions
• Focus on the patient experience as the driver of NHS modernisation
• Pursuit of closer partnerships among health services, local authorities and the third sector
• Establishment of public involvement as a statutory duty of NHS organisations
• Patient Advice and Liaison Services (PALS).

4C.7 PROBLEMS OF POLICY IMPLEMENTATION

Policy implementation involves three principal activities: see Box 4C.7.1.

409

Public [Link] 409 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

Box 4C.7.1
Activity Description
Interpretation The intent of policy-makers is discerned, and details are added
Organisation A strategy for accomplishing policy goals is decided
Administrative units are defined
Methods of service delivery are chosen
Application Policy is put into action

Typical barriers to implementation include:

• Overambitious time scales
• Lack of skills or training
• Ill-defined roles and responsibilities
• Poor project management
• Inadequate contingency planning.
The UK’s National Audit Office lists the following six adverse consequences that can result from poor implementation
of policy:
• Users’ expectations not met
• Poor quality public services
• Adverse effects on economic competitiveness
• Adverse social or environmental consequences
• Little or no benefit delivered or not sustainable in the longer term
• Sections of society excluded from benefits.

4C.8 FORMULATING POLICY

Principal approaches to policy formation
Policy formulation is subject to a number of disparate influences: see Table 4C.8.1. The relative strengths of these
influences will vary according to the nature of the policy and the organisation in which it is made.

Table 4C.8.1 Types of influences involved in policy formation

Type of influence Explanation
Incremental Policy is influenced by chance events, learning from mistakes, experimentation and a
range of other influences
Pluralist Many actors and interest groups influence the policy process. There may be a range of
mechanisms by which these different voices are heard
Policy narratives Different stories evolve to describe events. Some gain more authority and have more
influence on policy decisions than others
Actor networks Certain individuals or institutions spread narratives through chains of persuasion and
influence
Political Both personal politics and party politics influence policy decisions
Practitioners Front-line staff have a strong influence on policy
Reproduced from Anderson and Sotir Hussey (2001).

410

Public [Link] 410 3/6/08 [Link]

 4C Equality, Equity and Policy

MODELS OF POLICY FORMATION

Traditionally, two models of decision-making have been proposed: rationalist and incrementalist. A third model –
that of mixed scanning – has recently been added. See Table 4C.8.2.

Table 4C.8.2 Models of decision-making in policy formation

Model Proponent Type Description Criticisms
Rationalism Simon Prescriptive The aims of the policy are stated In reality, decisions do not begin
(1947) explicitly and are analysed in the with explicit goals, they do not
light of existing and potential involve consecutive stages, and
problems decision-makers have restricted
time and resources available to
A list of available options is
them. In practice, policy-making
drawn up, and each option is
is actually a process of ‘muddling
evaluated in terms of costs and
through’
benefits, and the option that
maximises benefits relative to
costs is chosen
Incrementalism Lindblom Descriptive This model describes the process This model is criticised as
(1959) of ‘muddling through’ in which: being too conservative in that
• Explicit goals are not set it neglects the potential for
• Only marginal changes are innovation
made to existing policy
• Decision-making is iterative
• Policies adopted are
those enjoying the widest
agreement
Mixed scanning Etzioni Prescriptive This model divides policy • Which decisions count as
(1967) decisions into fundamental and ‘fundamental’ and which as
regular decisions: ‘regular’?
• Fundamental decisions set • How much time and
new policy directions and resources should be invested
are to be attained using a into making fundamental
rational approach decisions?
• Once the parameters • How long should regular
of policy are set, an incremental changes be
incremental approach is used continued? Many small
for regular decisions changes may alter the
fundamental nature of policy

4C.9 POWER, INTERESTS AND IDEOLOGY

Appreciation of concepts of power, interests and ideology
Policy changes are heavily influenced by power relationships within and between groups of people.

POWER
Power is the capacity to make something happen and often involves making others do things that they would not
otherwise do. French and Raven (1960) identified several different types of power: see Box 4C.9.1.

411

Public [Link] 411 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

Box 4C.9.1
Resource Also known as reward power, the person with this power has control of resources (e.g. budgets,
people) and has the power to reward people with promotion or funds
Position By virtue of holding a particular job within an organisation, the post holder is entitled to the
rights and privileges of that role
Coercive The type of power that comes from the ability to punish (e.g. through the withdrawal of privileges
or the imposition of penalties)
Personal Also known as charisma. This power resides in the personality
Expert This is power vested in someone because of their acknowledged expertise. A public health
practitioner within an organisation has a degree of expert power
Negative This power is the capacity to stop things happening or from even being discussed

Identifying which individuals have which types of power can be helpful for identifying which people need to be
influenced in order to secure support for a new policy. It can also be used to identify and counter negative sources
of power. Finally, it can help identify which sources of power an individual is under-using.

INTERESTS
See Section 5B.2.

IDEOLOGY
A political ideology describes a belief of how power should be allocated and to what ends it should be used. It can
be a construct of political thought, and it often defines political parties and their policies.
Box 4C.9.2 attempts to group common ideologies into themes, but note that one ideology may belong to several
groups and that related ideologies often overlap. For this reason, modern political parties often subscribe to a
combination of ideologies. Note also that that the meanings of political labels differ between countries.

Box 4C.9.2
Type of ideology
Class struggle Collectivity Ethnicity Religion
Socialism Socialism Nationalism Christian-based ideologies
Communism Religious socialism Fascism Christian anarchism
Marxism Christian socialism Nazism Hindu-based ideologies
Leninism Democratic socialism Neo-Nazism Hindu nationalism
Stalinism Communism Racism, racialism Islam-based ideologies
Neo-Marxism Religious communism Islamism, Muslim fundamentalism
Marxism Jewish-based ideologies
Religious Zionism

412

Public [Link] 412 3/6/08 [Link]

 4C Equality, Equity and Policy

4C.10 HEALTH INEQUALITIES

Inequalities in the distribution of health and health care and its access, including inequalities relating to social class,
gender, culture and ethnicity, and their causes
See Section 4A.8.
Health inequalities are differences in health that are attributable, among other things, to a range of factors such as
social class, age, gender, ethnicity and geography.

SOCIAL CLASS
Social class is a strong predictor of health outcomes.

UK BLACK REPORT (1980)

This key report was commissioned by the Labour government in the late 1970s, but was published after the
Conservative party came to power in 1979. The incoming government attempted to suppress the publication of the
report (although the endeavour backfired) and did not support its findings. Black and colleagues noted a social
class gradient in both morbidity and mortality that was:
• Present at every age
• Present for all major diseases
• Increasing over time.
The authors listed four possible explanations for the gradient that they observed (see also Box 4A.8.1, p 385), namely:
• Statistical artefact – the way that social class is measured
• Social selection – the ill become poorer
• Behavioural factors – direct impact of nutrition, etc.
• Indirect impact of deprivation – effect on physical health is mediated by psychology.
The implication of the Black Report is that income fundamentally influences health, and that this influence lies
outside the scope of the health-care system. A number of studies have been published since the Black Report that
investigated the issue in more detail and provided potential solutions: see Table 4C.10.1.

Table 4C.10.1 Solutions to inequality proposed by key studies

Benzeval (1995) Acheson Inquiry (1998) Saving Lives: Our Wanless reports
Healthier Nation (1999)
Improve physical Evaluate government Prominence given to health 2002 – made the economic
environment policies regarding health
inequalities and four case for public health
inequalities priority areas:
Address socioeconomic 2004 – securing good
• Cardiovascular disease
factors: income and Focus on policies affecting health for the whole
• Accidents
employment health of families and population
• Cancer
children
Promote healthier • Mental health 2006 – funding of social
lifestyles Reduce inequalities in care for older people
income and improve
Improve access to services
housing standards

AGE
Older people report more illness, and use health services more frequently than younger adults. This is due to a
combination of material, biological and social factors:

413

Public [Link] 413 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

• Avoidable illness and death linked to poverty (e.g. cold houses)

• Biological factors, which may be more common in certain ethnic groups (e.g. hypertension, type 2 diabetes)
• Cultural treatment of older people (e.g. disengagement/stigma) and social isolation.

GENDER
Demographic statistics and surveys show gender differences in health-related behaviour, illness and mentality.
Risk-taking is greater in men – they are more likely than women to smoke and drink heavily, while women are more
likely to eat fresh fruit and vegetables. Women are more likely to be diagnosed with anxiety or depression, although
suicide rates are greater in men. Life-expectancy is greater for women than men at all ages. In childhood boys are
more likely to die from accidents, poisoning and injury, and in adult life, men die earlier for a range of common
causes such as cardiovascular disease.

UK ETHNICITY
There are differences in prevalence of illness and health seeking between ethnic groups:
• Cardiovascular disease, type 2 diabetes and hypertension are more prevalent in people of South asian and
African–Caribbean ethnicities
• Mental illness is commoner in the African–Caribbean population
• Lowest levels of health service usage are seen in Chinese population.

GEOGRAPHICAL AREA
Geography may have an independent effect on health through the following mechanisms:
• Physical environment
• Home and work environments
• Local services
• Sociocultural factors
• Reputation (psychological factors).

4C.11 MIGRATION AND HEALTH EFFECTS OF INTERNATIONAL TRADE

While migration that is properly managed can bring huge financial and cultural benefits to both individuals and
countries, mismanaged migration and human trafficking represent significant health risks.
The UN’s International Organization for Migration estimates that 3% of the world’s population are migrants (the
highest proportion in world history). It expects migration to continue growing because of the factors listed in Box
4C.11.1.

Box 4C.11.1
Economic Increases in trade and globalisation lead to increased demands for skilled labour –
liberalisation especially in the IT, financial and hospitality sectors
Economic decline Counter to what might be expected, temporary economic recessions tend not to lead to a
downturn in migration
Demographic Most developed countries have populations that are expected to shrink and to become
changes older over the course of the coming decades. The young, growing populations of
developing countries may serve to counter this ‘population time bomb’

414

Public [Link] 414 3/6/08 [Link]

 4C Equality, Equity and Policy

Transnational These are migrants who shuttle between multiple homes and maintain links with more
migration than one country. Such people are said to live in ‘transnational migration space’. This
growing phenomenon is facilitated by dual citizenship and multiple properties and voting
rights
Conflict There is no end in sight for wars and political upheaval

A distinction is made between voluntary and forced migrants: see Box 4C.11.2.

Box 4C.11.2
Reasons for voluntary migration Reasons for forced migration
Employment War
Study Industrial, environmental or natural disasters
Rejoin family members Famine
Retirement Development projects, e.g. dams

CLASSIFICATION OF MIGRANTS
Migration is the permanent relocation of people between one country and another: see Box 4C.11.3.

Box 4C.11.3
Term Alternative Definition Rate
name
Emigration Out-migration Process of people leaving one country Number of people departing from a
on a permanent or semi-permanent country per 1000 of its population per
basis annum
Immigration In-migration Process of people entering a country Number of people arriving per 1000
to take up residence of the population of the receiving
country per annum

An asylum seeker is a person who requests sanctuary in a destination country on grounds of having escaped
persecution in the country of origin. A convention refugee is a person recognised by the destination country as
having a ‘well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular
social group, or political opinion, is outside the country of his nationality, and is unable to, or owing to such fear, is
unwilling to avail himself of the protection of that country’ and is therefore accorded the full rights of the 1951 UN
Convention on Refugees. In contrast, a quota refugee is a person who is granted limited refugee status by the
destination country before leaving the country of origin, usually as part of an agreement by which the destination
country agrees to take a finite group of refugees over a short period of time.
Migrants may be classified as documented or undocumented, depending on whether the state authorities in
the host or transit country have authorised residence and employment. Undocumented migrants (sometimes
inappropriately referred to as ‘illegal immigrants’) are people who have either entered a host country without legal
authorisation or overstayed their period of temporary, authorised entry.
UK A Home Office study (Robinson and Segrott 2002) in which asylum seekers were asked why they came to the
UK found that the main reason was to seek a place of safety. Those who were in a position to choose a destination
country, selected the UK for the following reasons:

415

Public [Link] 415 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

• Relatives or friends already in the UK

• Their belief that it is a safe, tolerant and democratic country
• Established links between the country of origin and the UK (including colonialism)
• Ability to speak English or desire to learn the language.

MIGRATION AND HEALTH

Health implications can be divided into those on the country of origin, the migrants themselves and on the
destination country: see Table 4C.11.1.

Table 4C.11.1 Migration and health

Country of The WHO is particularly concerned about the drain of health-care workers from developing
origin countries to developed countries because it can cause serious deficiencies in local provision of
services.
Most migrants are highly skilled, and the loss of new graduates places financial and
human resource strains on the countries of origin. This is by no means compensated for by
remittances (the term for the portion of migrants’ wages that is sent back to the country of
origin), so it contributes to the widening gap in income between rich and poor countries
Migrants During conflicts, health-care workers may be displaced from the populations that they serve
Poor sanitation and overcrowding in refugee camps are ideal conditions for epidemics to
develop
Sexual violence in conflict situations and refugee camps is a cause of sexually transmitted
infections
Post-traumatic stress disorder is highly prevalent among asylum seekers and refugees
People being trafficked across international borders suffer death in transit (e.g. asphyxiation
of a group of Chinese migrants in a cargo container while crossing the English Channel in
2000)
On arrival in the destination country the health outcomes of migrants are affected by poor
living conditions, a lack of social integration, stigma and open xenophobic hostility
Destination Screening of migrants at the point of entry is a politically sensitive issue
country
In many countries, temporary migrants (most fall under this category for a time) are only
entitled to free emergency health care
Undocumented migrants avoid health-care workers for fear that they have links to the
immigration authorities
Moreover, migrants experience impaired access to health care because of unfamiliarity with
local health-care systems and conventions, as well as linguistic or cultural difficulties in
communicating their symptoms
Adapted from WHO International Migration, Health and Human Rights (2003b).

First-generation migrants often retain patterns of disease from their country of origin. For example, stomach cancer
(which has a high prevalence in Japan and China due to diets that are rich in smoked, salted and pickled foods) is
common among first-generation Japanese and Chinese immigrants to the USA. However, the incidence is much lower
among descendants who adopt the lifestyle practices of the destination country.

416

Public [Link] 416 3/6/08 [Link]

 4C Equality, Equity and Policy

INTERNATIONAL TRADE
The effects may be direct (e.g. an infectious disease is transported within traded goods or by means of an infected
tradesperson) or indirect (e.g. trade of health technologies). Equally, health regulations may have effects on trade
(e.g. transport of food) (WHO 2005). See Box 4C.11.4.

SA Box 4C.11.4
Example: spread of HIv/AIDS along trade routes
In Africa, international trade is implicated in the spread of HIV/AIDS. As a representative of the South African
Medical Research Council put it:
‘If you wanted to spread a sexually transmitted disease, you’d take thousands of men away from their families,
isolate them in single sex hostels and give them easy access to alcohol and commercial sex. Then to spread the
disease, you’d send them home every once in a while to their wives and girlfriends.’ (Mark Lurie, South African
Medical Research Council)

4.C12 INTERNATIONAL INFLUENCES ON HEALTH AND SOCIAL POLICY

When the United Nations was formed in 1945, member states agreed to work together to promote the ‘economic and
social advancement of all peoples’. More than 60 years later, the health of the world population is under threat from
environmental crises, and income and health inequalities continue to rise between the richest and poorest nations.

HEALTH PROMOTION
Cooperation at an international level to promote health is of particular value in the following respects:
• Exchange of ideas and mutual learning
• Pooling of resources
• Joint action.

SMOKING IN ENCLOSED PUBLIC PLACES

In recent years, one of the most significant steps taken by developed countries to improve public health has been to
prohibit smoking in enclosed public places. The rapid spread of smoke-free policies is, however, attributed more to
zeitgeist than concerted international policy. That said, the European Commission does promote cooperation between
EU Member States on public health (particularly regarding health protection) and it commissions Eurobarometer
surveys to investigate major public health issues such as physical activity, smoking and mental health.

SOCIAL POLICY
The field of international social policy compares the welfare provision found in different countries, ranging from the
Nordic welfare model of comprehensive provision at one extreme, through to the scantier provision available in
other countries.
UK In Britain, the Welfare State was established following the Beveridge Report in 1942, which identified
five ‘giant evils’ in society: squalor, ignorance, want, idleness and disease. By creating the welfare state (a series
of policies designed to support people with financial, health or social needs) the government acknowledged its
responsibility to care for the population ‘from the cradle to the grave’.
International social policy also addresses the potential impact of globalisation on the welfare state, and the
influential roles played by international actors such as the International Monetary Fund, the United Nations, the
World Bank and the World Trade Organization.

417

Public [Link] 417 3/6/08 [Link]

 4C E q ua l i t y, E q uity and Pol icy

4C.13 INvESTMENT IN HEALTH IMPROvEMENT

Critical analysis of investment in health improvement, and the part played by economic development and global
organisations
Although health promotion interventions have the potential to reduce future health-care costs by avoiding disease,
the evidence for this is equivocal. Even where a health promotion programme is demonstrably successful (e.g. stop-
smoking programmes), the overall cost to the state from a population with higher longevity may be more.
It is therefore important to remember that the objective of health promotion is not to save money but to reduce
morbidity and mortality. Health promotion programmes should therefore be assessed along with other health-care or
social care programmes, using the methods of economic evaluation and cost-effectiveness analysis (see Section 4D).
Hale (2000) identifies a number of potential reasons why the use of economic analysis is problematic in the context
of health promotion interventions, including:
• Health promotion programmes have multiple objectives
• Clients are essentially healthy at the time and the benefits are broader than simply gains in health status
• QALYs do not capture the full range of benefits from health promotion
• Randomised controlled trials are needed to confirm that utility benefits were caused by the intervention, but
these are difficult to conduct in the field of health promotion because of the timeframes involved
• The conclusion of the economic evaluations of health promotion interventions are highly dependent on the rates
of discounting applied.
An example from the UK is given in Box 4C.13.1. The roles in investment played by global organisations are outlined
in Table 4C.13.1.

Eng Box 4C.13.1

Example: NICE economic evaluation of physical activity interventions
In 2005, the Department of Health asked NICE to evaluate four health promotion interventions aimed at
increasing physical activity:
• Primary care brief interventions
• Exercise referral
• Pedometers
• Community walking and cycling programmes
The review found an absence of evidence for the cost-effectiveness of pedometers and of community walking
and cycling programmes. It found limited evidence that brief interventions in primary care were less cost-
effective than usual care. For exercise referral, it found some cost-effectiveness evidence to suggest that the
intervention is both more effective and more costly than usual care. The review was unable to determine the
ICER (incremental cost-effectiveness ratio), however, because of the different outcome measures employed in the
studies that it reviewed.
Other public health interventions being reviewed by NICE include:
• Smoking and tobacco control
• Obesity, diet and nutrition
• Alcohol
• Sexual health
• Mental health
• Drug misuse
• Promoting the health of children and young people
• Preventing accidental injury

418

Public [Link] 418 3/6/08 [Link]

 4C Equality, Equity and Policy

Table 4C.13.1 Roles played by global organisations in investment in health improvement

Public health Child health Emergency aid
World Health Organization UNICEF (UN Children’s Emergency World Food Programme
Fund)
World Bank International Committee of the Red
Cross/Crescent
Food and Agriculture Organization
Medécins Sans Frontières

419

Public [Link] 419 3/6/08 [Link]

Public [Link] 420 3/6/08 [Link]
 4D Health Econom ics

4D
Health Economics

4D.1 Health economics 421 4D.5 Economic appraisal 435

4D.2 Assessing performance 431 4D.6 Marginal analysis 444
4D.3 Financial resource allocation 432 4D.7 Decision analysis 444
4D.4 Health-care systems and incentives 433 4D.8 Economic evaluation and priority setting 446

Economics is based on the notion that we operate in a world of scarcity. In other words, there are infinite demands
but only finite resources. This chapter considers how these finite resources can meet the infinite demands for health
and health care through the practice of resource allocation. It provides a reference resource of techniques used in
health economics, in particular economic evaluation, together with a consideration of their role in policy-making
and public health.

4D.1 HEALTH ECONOMICS

Principles of health economics (including the notions of scarcity, supply and demand, marginal analysis, distinctions
between need and demand, opportunity cost, margins, efficiency and equity)
It is an axiom in health care that demand exceeds supply. This is because there is an infinite amount that could
potentially be spent on health care, so the actual resources available will always be relatively scarce. This means
that choices must be made regarding how to spend the health-care budget. Health economics is the science of
making these decisions, i.e. how best to employ scarce resources that have alternative uses.
The science is divided into positive and normative economics: see Box 4D.1.1.

Box 4D.1.1
Positive economics Study of how markets work and how interventions will affect outcomes
Normative economics Study of determining what should be produced, what resources to use and how to
distribute goods

SPECIAL FEATURES OF HEALTH CARE

Economics has become a core discipline within public health. However, health care is seen as being special in
economic terms for the reasons shown in Box 4D.1.2 (these will be explored later in this chapter).

421

Public [Link] 421 3/6/08 [Link]

 4D He a l t h E c onomics

Box 4D.1.2
Supply and demand Demand and supply are not truly independent in health care
Imperfect markets All health-care systems (but especially those that are publicly funded) are imperfect
markets
Immediacy Life and death decisions often need to be made with very short timescales
Agency The nature of health care that people need, especially when they are critically ill, is
largely specified by health-care providers
Uncertainty Illness is often unpredictable
Necessity Health care is an unavoidable commodity

SCARCITY
Health care can be regarded as a production process that uses a number of inputs in a process that produces
outputs. The inputs (or ‘factors of production’) are divided into four categories: see Box 4D.1.3.

Box 4D.1.3
Land Physical resources of the planet including mineral deposits
Capital Resources created by humans to aid production, such as tools, machinery and factories
Labour Human resources in the sense of people as workers
Enterprise Human resource of organising the other three factors to produce goods and services

It can be seen that none of these resources is infinite. The term ‘scarcity’ is used where more of a resource is wanted
than is available. In these circumstances, every choice made involves a sacrifice, since the same resource cannot
subsequently be used for something else. This sacrifice is called the opportunity cost (i.e. the benefits that are
forgone are effectively the value of the benefits that we enjoy) and this is a fundamental concept of health economics.

CONCEPTS OF SUPPLY AND DEMAND

Demand is what people request; supply is what is provided.
Note that demand is different from need (need is something that a person will benefit from receiving) because
people do not benefit from everything that they demand (e.g. antibiotics for a common cold).
Note also that demand is different from what is received because people do not receive treatments that they
demand when they do not meet the eligibility criteria for that treatment (e.g. breast enlargement surgery under the
NHS in England).

RISING DEMAND
In developed countries, demand for health care is rising for the following reasons
• Demographics (ageing population)
• Innovation (technology)
• Lifestyle (abuse)
• Information (educated consumer)
• Standards of living (quality-of-life expectations).
It is a principle of economics that – in a perfect market – supply and demand are determined independently, i.e.
producers determine supply and purchasers determine demand. The price of goods rises or falls until the amount
supplied equals the amount demanded, i.e. equilibrium is reached.

422

Public [Link] 422 3/6/08 [Link]

 4D Health Econom ics

Another fundamental principle of economics is that demand will equal supply in a perfect market. While health
and health care are not perfect markets, certain aspects of supply and demand do remain applicable.

SUPPLY
The supply of health care is the care that is made available; it is the capacity of services to meet need. Health-care
supply may be quantified in terms of:
• Staffing (e.g. whole-time-equivalent consultants, nurses, etc.)
• Beds
• Equipment
• Budget.

DEMAND
Demand is the expressed need. Note that there is a distinction between the demand for health (i.e. to feel good
and participate in all areas of life) and the demand for health care (i.e. a service required to achieve health), the
latter being a derived demand.
Demand for health care may be quantified in terms of:
• Bed occupancy
• Consultation rates
• Waiting lists.
According to economic theory, demand is determined by four factors: price, income, preferences and alternatives
(Table 4D.1.1).

Table 4D.1.1 Four factors that determine demand under economic theory
Price of health/health care When fees are introduced to health-care systems (e.g. a charge to see a doctor
or to obtain medicines), then demand drops (known as price elasticity). In
England certain people (e.g. children) are exempt from prescription charges
in order to ensure that their demand for health is not affected by the price of
health care
Individuals’ income Studies have shown that the introduction of user fees reduces demand
disproportionately in those on lower incomes (i.e. their demand is more income
elastic compared with people on higher incomes)
Tastes and preferences Different people place different values on different lifestyle factors. People also
place different values on the benefits of health care: some people are more
likely to seek health care for particular symptoms than others. For example, an
individual may choose to trade off the unhealthy effects of a takeaway meal
against the convenience of not having to cook
Price and availability of Substitutes for health care could include other types of health care (e.g.
complements and substitutes complementary medicine). Some individuals who use these services may choose
not to use conventional health care. Others (e.g. those who place a particularly
high priority on health) may use both types of health care

AGENCY
As will be seen in the section on markets below, one of the characteristics of a perfect market is that each consumer
has perfect knowledge about the products on offer. There are several reasons why health care is not a perfect

423

Public [Link] 423 3/6/08 [Link]

 4D He a l t h E c onomics

market: one is that consumers do not have perfect knowledge about the complexities of health care available.
Instead, they rely on agents such as doctors to inform them about what services they need. For example, patients
with angina only receive angioplasty if their cardiologist thinks that they will benefit from it.
Perfect agents (like perfect markets) do not exist, because a perfect agent would have to strike a perfect balance
between all of the following conflicting priorities:
• Health status of an individual patient
• Preferences of an individual patient
• Utility to society.
Moreover, agents may sometimes be motivated by other factors (e.g. self-interest).

SUPPLIER-INDUCED DEMAND
Supplier-induced demand occurs when agents act in their own interests and thereby recommend more health care
than is necessary, i.e. more than would a perfect agent. For example, a dentist might recommend a dental filling
that was not strictly necessary in order to gain the fee for performing the procedure.
The phenomenon is difficult to identify on a macroscopic scale because only price equilibrium points can be
observed (i.e. if costs of dentistry increase and demand decreases, then it is difficult to say if it has decreased to
the extent that would be expected if there were no supplier-induced demand. However, should the cost of a dental
consultation increase when more practitioners entered the market, then the market would be demonstrably abnormal
and supplier-induced demand would have been detected).

SUPPLIER-REDUCED DEMAND
In systems where the supply of health care is particularly scarce, supplier-reduced demand is seen. In this case an
agent may not recommend a particular health-care intervention, whereas a perfect agent would have done so in the
same circumstances. To an observer it would appear that there is less demand for a service than is in fact the case.
Rather than seeking to demonstrate and punish supplier-reduced demand, policy-makers can design contracts to
eradicate it (e.g. by linking payment to quality indicators).

LAWS OF SUPPLY AND DEMAND

The law of the downward sloping demand curve states that demand is affected by the following:
• Price
• Price of other products
• Income
• Consumer preferences.

SUPPLY AND DEMAND CURvES

See Figure 4D.1.1.
If a product sells at a low price, then producers will be disinclined to make large amounts of it. If the price rises,
then producers will make more of the product and therefore the supply curve (S) slopes upwards. In contrast,
if the price of a product is high, then consumers buy relatively little of it. If the price of the product drops, then
consumers will be willing to buy more of the product, and thus the demand curve (D) slopes downwards.
Price reaches equilibrium at the intersection of the supply curve and the demand curve. If the price is set higher
than this point (a move from price P1 to price P2), then producers will want to produce more of it (i.e. move from
quantity Q1 to Q2). Since there is now more of the product available, customers will not be willing to pay as much –
thereby returning the price back to the equilibrium point.

424

Public [Link] 424 3/6/08 [Link]

 4D Health Econom ics

P
D1 D2 S P � Price

Q � Quantity of goods

S � Supply

D � Demand

P2
P1

Figure 4D.1.1 Supply and

Q1 Q2 Q demand curves

SHIFTS AND MOvEMENTS IN THE DEMAND CURvE

A change in price (ceteris paribus – all other things being equal) leads to a movement along the curve, whereas a
change in other factors leads to a shift of the curve (i.e. the curve itself moves). See Box 4D.1.4.

Box 4D.1.4
Phenomenon Causes Change seen on the curve
Rise in price More product being produced Movement rightwards along the
demand curve
Fall in price Less product being produced Movement leftwards along the
demand curve
Purchasers not willing to pay as Falls in income Leftward shift of the demand curve
much
Purchasers willing to pay more Increased population Rightward shift of the demand curve
Increased income
Changes in taste (e.g. as a result of
advertising)
Producers find it easier to produce More land Rightward shift of the supply curve
the product
More labour
Technology available
Price drops

SUBSTITUTES AND COMPLEMENTS

Substitutes and complements are related goods where a change in price of one good affects the demand for the
other: see Box 4D.1.5.

425

Public [Link] 425 3/6/08 [Link]

 4D He a l t h E c onomics

Box 4D.1.5
Definition Example
Substitutes Products where an increase in price of one type of good Two different brands of the same
causes an increase in demand for the other vaccine
Complements Products where an increase in price of one type of good Needles and syringes
causes a decrease in demand for the other

PRICE ELASTICITY
The price elasticity of demand is a measure of how sensitive is the demand for a particular good to changes in price.

PRICE ELASTICITY OF DEMAND

Percentage change in quantity demanded
Price elasticity of demand (PED) =
Percentage change in price
See Box 4D.1.6.

Box 4D.1.6
Absolute value* of PED Definition Characteristic
|PED| >1 Price elastic Large response to price
|PED| = 1 Unit elasticity Proportionate to price
|PED| <1 Price inelastic Small response to price
|PED| = 0 Perfectly inelastic No response to price
*The absolute value of PED is the magnitude with the + or – sign removed. It is symbolised using two vertical straight
lines, e.g. |–0.5| = 0.5.

PRICE ELASTICITY OF INCOME

Percentage change in quantity demanded
Income elasticity of demand (IED) =
Percentage change in income
See Box 4D.1.7.

Box 4D.1.7
value of IED Definition Characteristic
IED >1 Luxury goods Disproportionately large amounts demanded as incomes rise
IED >0 Normal goods Amount of these goods demanded changes in line with
income as would be expected from demand curve
IED <0 Inferior goods Disproportionately large amounts demanded as incomes fall

MARKETS
A perfect market has the characteristics shown in Box 4D.1.8.

426

Public [Link] 426 3/6/08 [Link]

 4D Health Econom ics

Box 4D.1.8
Atomicity Many buyers and many sellers
Homogeneity Identical products
Free entry Sellers are free to join and leave the market
Equal access Production technology is available equally to all sellers
Perfect information All buyers and all sellers know the products and prices of all sellers
No externalities An externality is a benefit (or disbenefit) to someone other than the
purchaser, e.g. an externality of an immunisation programme is herd immunity

In a perfect market, the producers are price-takers, i.e. the market sets the price. In such circumstances, producers
produce at the lowest possible cost in the long run, and they earn only normal profits. A normal profit is the same
profit that could be achieved in the best alternative business (rather than an economic profit, which is revenue
additional to this).
The long run is defined as the timeframe in which firms can enter or exit the market, and in which they can change
their capital (e.g. build an extra operating theatre).
If producers do not operate in this way and, in particular, if they have a significant power to influence price or the
total quantity being produced, then the market will fail.

CAUSES OF MARKET FAILURE IN HEALTH CARE

A market may fail if any of the conditions listed in Box 4D.1.8 are not met. The principal causes of market failure are
listed in Box 4D.1.9.

Box 4D.1.9
Externality This is a side effect of the product that is not traded on the market (e.g. herd immunity
as a side effect of immunisation is an externality). Externalities may be beneficial (termed
positive externalities) or harmful (negative externalities)
Public goods These are extreme examples of an externality, such as a health promotion poster campaign,
and are characterised by:
• ‘Non-rivalness’ (when one person reads the poster, other people do not suffer)
• Non-excludability (it is impossible to stop a person from reading the poster)
Monopoly Monopoly (single producer)
Monopsony (single purchaser)
Oligopoly (few producers)
Imperfect Uncertainty (unable to predict demand, e.g. when trauma care needed)
knowledge
Moral hazard (no price to consumer at the time of use)
Adverse selection (those at low risk of illness opt out of health insurance)
Merit goods Belief that health services are in some way special

DISTINCTIONS BETWEEN NEED AND DEMAND

Economists define need as the capacity to benefit from health care (see Section 4C.1 for the differences between
expressed, felt and normative need). In contrast, demand is a request for health care.

427

Public [Link] 427 3/6/08 [Link]

 4D He a l t h E c onomics

The key distinction, then, between need and demand is that the latter must be expressed by people by one of the
following:
• Attending the place where the service is offered
• Waiting for the service
• Paying for the service.
Where need is not identified, it is not expressed, and it is therefore not a demand. Moreover, the health care that is
demanded does not represent all health-care needs.
See Figure 4D.1.2.

Screening Need New technology

programme for antibiotic resistance

Supply Demand

Homeopathy
Figure 4D.1.2 Need, supply and demand

MARGINS
Choices often involve demanding a little more of one product or a little less of another. These are known as marginal
changes. The margin is defined as the incremental variation in inputs that is required to have a corresponding
variation on outputs. See Box 4D.1.10.

Box 4D.1.10
Marginal cost Cost of producing one extra unit of service. This will reflect any stepped costs that are
encountered (such as having to open an additional operating theatre because the capacity of
the existing theatres has been exceeded)
Marginal Benefit derived from one extra produced unit
benefit

Marginal analysis examines the effect of small changes in the existing pattern of expenditure. It can identify:
• Where additional resources should be targeted

428

Public [Link] 428 3/6/08 [Link]

 4D Health Econom ics

• Where reductions should be made if expenditure must be cut

• How resources can be reallocated to achieve an overall gain in benefit with no overall change in expenditure.
See Box 4D.1.11.

Box 4D.1.11
Example: Targeting a screening programme
A small-scale screening programme targeted at the highest risk groups may show a low cost per positive case
detected. Continual expansion of the programme will entail screening progressively lower risk groups or screening
more frequently. The number of screens required to detect each additional positive case will rise, increasing the
cost per case detected.
Reproduced from Cohen (1994).

ECONOMIES AND DISECONOMIES OF SCALE

The average cost curve for the production of a good is U shaped: see Figure 4D.1.3.

Economies of
scale
Average cost

Diseconomies
of scale

Amount produced

Figure 4D.1.3 Cost curve for the production of a good

As more of a type of goods is produced, so the average cost falls due to more efficient use of inputs. However, a
point is reached where the diseconomies of scale begin to feature. These are caused by:
• Difficulties in managing an organisation that is so large that it is unwieldy
• Diseconomies of scope where it becomes increasingly costly to reach remote areas.

OPPORTUNITY COST
In health care, opportunity cost is quantified as the health benefits (life years saved, QALYs gained, etc) that could
have been achieved had the money been spent on the next best alternative intervention or health-care programme.
It can be calculated directly by means of cost-effectiveness or cost–utility studies (see Section 1C.14). Many studies
attempt to compare particular interventions with existing practice, which itself may not be well defined. Failure to
select an appropriate comparator can make the intervention appear more cost-effective than it should – leading to
incorrect estimates of the opportunity cost.

429

Public [Link] 429 3/6/08 [Link]

 4D He a l t h E c onomics

EFFICIENCY
See Section 4C.3 for a comparison of equity, equality and efficiency.
Efficiency reflects how much health-care benefit is being achieved from the available resources. It can be considered
in three ways: see Box 4D.1.12.

Box 4D.1.12
Technical efficiency Maximum output for given inputs or minimal inputs needed
for a given output
Economic efficiency Maximum output for a given expenditure or minimal costs
needed for a given output
Allocative efficiency Efficient budget to produce the goods according to demand
(i.e. produce what consumers value more than their cost)
or set the level of production such that marginal benefit >
marginal cost

TECHNICAL EFFICIENCY
Technical efficiency addresses the issue of using given resources to maximum advantage. An intervention is
technically efficient if the same (or greater) outcome could not be produced with less of one type of input. For
example, consider the treatment of osteoporosis using alendronate: if a 10 mg daily dose is as effective as a 20 mg
dose, then the lower dose is more technically efficient.

ECONOMIC EFFICIENCY
Economic efficiency refers to the maximisation of health outcomes for a given cost, or the minimisation of costs
for a given outcome, e.g. a policy of changing from maternal age screening to biochemical screening for Down’s
syndrome. If the sum of the costs of the new biochemical screening programme is smaller than or the same as the
maternal age programme and outcomes are equal or better, then the biochemical programme is economically efficient
in relation to the maternal age programme.

ALLOCATIvE EFFICIENCY
Allocative efficiency is the achievement of the best combination of health-care programmes to maximise the health
of society. The concept of allocative efficiency takes account not only of the productive efficiency with which
health-care resources are used to produce health outcomes, but also the efficiency with which these outcomes are
distributed among the community. Allocative efficiency is achieved when resources are allocated so as to maximise
the welfare of the community. Given that there will always be scarcity, a decision-making system is required that
determines how much of which kinds of health care is provided. There are three possibilities: the free market, the
command system and the mixed system (Box 4D.1.13).

Box 4D.1.13
Free market Health-care resources are allocated according to consumers’
purchasing behaviour
Command system Planning is used to allocate health care according to some pre-
determined criterion such as ‘need’
Mixed system Combines elements of the free market with elements of the command
model

430

Public [Link] 430 3/6/08 [Link]

 4D Health Econom ics

EQUITY
See Section 1C.10 for a comparison of vertical and horizontal equities and Section 4C.3 for a comparison of equity
and efficiency. An example is given in Box 4D.1.14.

Box 4D.1.14
Example: Equity versus efficiency in cervical screening
The NHS policy on cervical cancer screening has been primarily aimed at maximising coverage by using economic
incentives to GPs. However, there has been lower participation by high-risk women particularly those in
disadvantaged socioeconomic groups. It has been calculated that the programme could have achieved the same
cost-effectiveness in terms of cancers avoided with less extensive but more equitable coverage.
Reproduced from Sassi et al (2001).

DISCOUNTING
Discounting is a method used in economics to deal with the phenomenon of positive time preference, i.e. the human
nature of preferring benefits to be realised now and for costs to be borne at a later date.
The reason for positive time preference is that the future is uncertain, so it is logical to want benefits earlier and
costs later. However, many public health interventions (e.g. stop-smoking campaigns) are costly today but will not
realise their benefits (e.g. reduction in lung cancer deaths) until many years into the future. In order to allow fair
comparisons of costs and benefits to be made at different times, positive time preference can be compensated for by
means of discounting. The strength of the time preference is reflected in the discount rate (which need not be the
same for costs and benefits). Discount rates range widely between 0% and 6%, and the discount rate for benefits is
particularly controversial.
Present cost = (Cost in year n) ¥ (Discount factor)
where
1
Discount factor =
n(1 + r)
and r = discount rate.
Because of the large effect of discounting on public health interventions, sensitivity analyses should always include
a range of discount rates for both costs and benefits (see Section 4D.5).

4D.2 ASSESSING PERFORMANCE

See also Section 1C.9.
The WHO advises that performance assessment should encompass the aspects listed in Box 4D.2.1.

Box 4D.2.1
Social goals Measuring the health system’s contribution to socially desirable goals
Resource use Measuring the health system and non-health system resources used to achieve these outcomes
Efficiency Estimating the efficiency with which the resources are used to attain these outcomes
Review Evaluating the way the functions of the system influence observed levels of attainment and
efficiency
Feedback Designing and implementing policies to improve attainment and efficiency and monitoring the
effect

431

Public [Link] 431 3/6/08 [Link]

 4D He a l t h E c onomics

Economic analysis (see Section 4D.5) allows performance to be assessed in terms of the benefits derived per unit of
expenditure. The performance of an allocation system is assessed in terms of efficiency (including Pareto efficiency)
and equity (see Section 4C.3).

4D.3 FINANCIAL RESOURCE ALLOCATION

The objective of financial resource allocation is the transfer of funds appropriately from purchasers to providers in
order to meet health-care objectives. The allocation process can be used to promote:
• Equity
• Changes in activity
• Efficiency (by means of incentive mechanisms).
Expenditure by the health service can broadly be classified as recurrent and capital: see Box 4D.3.1.

Box 4D.3.1
Recurrent Staff, drugs, consumables
Capital Buildings, equipment

NHS FUNDING
Eng The majority of funding comes from general taxation and National Insurance contributions, with the remainder
coming from patient charges, e.g. prescription charges and dental charges. Recurrent revenue allocations to primary
care trusts (PCTs) cover hospital and community health services, prescribing, primary medical services and HIV/
AIDS. See Figure 4D.3.1.

Weighted Contracts/
Providers Deliver
Department Primary
(GPs, Population
of Health capitation care trusts Agreements services
hospitals)

Eng Figure 4D.3.1 The allocation of funds (DH 2006)

ALLOCATION METHODS
Eng The Department of Health (DH 2005) makes allocations to PCTs according to four factors, including the
national weighted capitation formula: see Box 4D.3.2. The DH is in the process of moving away from recurrent
baselines towards the target of allocations based purely on weighted capitation.
NI Scot Wal In Northern Ireland, Scotland and Wales, the Barnett formula is used by the UK government to
allocate central funding for services that are the responsibility of the devolved legislatures. This represents about
80% of public spending in these countries, and includes health care. The formula has been in use since the late
1970s although it has no statutory basis. The devolved assembly or parliament decides how to spend its financial
block, including the relative proportions to be spent on health care and other services.
Scot The Scottish Executive health department allocates funds within Scotland on the basis of the Arbuthnott
formula.
Wal The Welsh Assembly decided to move towards allocating resources to local health boards (LHBs) purely
according to the Welsh Townsend formula (see Section 1C.8), rather than according to previous allocations.
However, the pace of this transition is constrained by political considerations. LHBs are in turn expected to allocate
their resources to reflect local need rather than demand.

432

Public [Link] 432 3/6/08 [Link]

 4D Health Econom ics

Box 4D.3.2
Weighted capitation PCTs receive their share of resources calculated according to:
• Size and age distribution of the population
• Additional need
• Unavoidable geographical variations in the cost of providing services (called the
market forces factor)
Recurrent baseline This is the previous year’s actual allocation, plus any adjustments made in-year
Distance from target This is the difference between the weighted capitation and the recurrent baseline
Pace of change policy This determines the level of extra resources that are allocated to PCTs that are below
their weighted capitation target. The pace of change policy is decided by ministers for
each allocation funding round

RATIONING
See Section 4C.2.

4D.4 HEALTH-CARE SYSTEMS AND INCENTIvES

Systems of health and social care and the role of incentives to achieve desired endpoints
Health-care systems can be considered in terms of a revenue collector, a payer, a purchaser and a provider.
Eng In England these can be thought of as follows (although the reality is more complex with each element of the
financial system being affected by the others): see Box 4D.4.1.

Box 4D.4.1
Revenue collector HM Treasury
Payer Department of Health
Purchaser Primary care trust
Providers General practice, hospital trust

Social care covers a wide range of services that facilitate people to carry on in their daily lives, and particularly
focuses on the groups shown in Box 4D.4.2.

Box 4D.4.2
Elderly people Through residential care homes, nursing homes, home carers, meals on wheels, day centres,
lunch clubs
Disabilities People with physical disabilities or learning disabilities
Mental health Ranging from support for those with mild mental illness, up to exercising legal powers for
compulsory admission to psychiatric hospitals of potentially dangerous people
Ex-offenders Those leaving prison may need help with resettlement, especially those with drug or alcohol
problems
Families Particularly where children have special needs such as a disability
Child protection Including monitoring of children at risk
Children in care Through fostering, accommodation in children’s homes and adoption

433

Public [Link] 433 3/6/08 [Link]

 4D He a l t h E c onomics

Eng Around 70% of the Social Services budget for England is spent on adult community care services – in particular
older people, people with physical or learning disabilities and mentally ill people. Some of these people pay for their
own social care, with the remainder receiving financial assistance from the state (either through welfare benefits or
through Social Services funding). Some people who require social care are given direct payments that provide them
with the freedom to purchase their own care rather than having this arranged by the local authority.

INTEGRATED CARE
Integrated care is the provision of both health and social care services in combination, and it ensures that
individuals receive the medical treatment and social care that they need (which frequently overlap). Integrated care
requires health and social care organisations to work together to deliver flexible services that are tailored to allow
people to live independent lives.
Examples of integrated care include cross-organisational services for drug users who have a range of other
difficulties in their lives such as housing and education. Clients and staff decide together what medical and social
care support is needed.
For frontline staff, the provision of integrated care requires:
• Working with individuals to identify their whole range of needs
• Knowing what other services are available
• Working alongside other professional groups
• Taking responsibility for arranging the right care or service.

MONITORING PERFORMANCE
Eng The Commission for Social Care Inspection (CSCI) provides an independent assessment of the quality and
performance of the social care sector on behalf of the government and the public. It assesses how well councils are
providing Social Services against objectives set by ministers from the Department of Health, and supports councils
in improving their performance.
Eng Wal The performance of the NHS and independent health care is monitored by the Healthcare Commission
(HC), which publishes the Annual Healthcheck: an annual performance assessment for each health-care organisation.
The HC monitors the quality of the NHS and independent health care. It also evaluates the value of additional
investment in the NHS and thereby judges how performance is improving. Other roles of the HC are to identify how
the quality of health services may be improved, and whether appropriate arrangements are in place to promote the
public health needs of the local population.
Eng NHS foundation hospitals are regulated by Monitor rather than by the HC.

FINANCIAL INCENTIVES
Financial factors may be used to encourage or discourage the provision of particular services, and may also
encourage or inhibit the use of these services by patients. For example, there is evidence that prescription charges
are negatively associated with the uptake of prescription medicines.
UK As part of the UK government’s public service modernisation agenda, explicit incentives have been introduced
with the aim of improving the efficiency of the health service: see Table 4D.4.1.

434

Public [Link] 434 3/6/08 [Link]

 4D Health Econom ics

UK Table 4D.4.1 Incentives to improve the efficiency of health care

Quality and Outcomes In England, PCTs and general practices have increasing incentives to be cost-conscious
Framework (QOF) in their decision-making. The ‘Quality and Outcomes Framework’ is a voluntary system
of financial incentives aimed at improving quality within the General Medical Services
(GMS) contract for GPs
Delayed discharge Delivering the NHS Plan requires local authorities to use some of their additional
resources to reduce the number of people who remain in hospital after they have been
deemed fit to be discharged. Failure to make appropriate alternative provision available
to patients will result in hospitals charging Social Services departments for the costs
incurred in keeping older people in hospital unnecessarily
Readmissions Incentives also apply to NHS trusts, which are held accountable for the cost of
emergency hospital re-admissions following a recent discharge. This is aimed at
ensuring that patients are not transferred prematurely
vaccinations Fee for service is used to fund certain items of service provided to patients. In these
cases, providers have an additional incentive to provide the services as it attracts more
fees, e.g. fees received by GPs for administering vaccinations
Case mix Case-mix payments are made according to measures of illness severity. For example,
the UK health economy uses the Healthcare Resource Group (HRG) for units of charging
(e.g. inguinal hernia repair without complication). HRGs are based on the average
cost of a patient treated with that diagnosis; therefore, providers have an incentive to
deliver care costing no more than the fixed payment of that HRG

4D.5 ECONOMIC APPRAISAL

Techniques of economic appraisal including cost-effectiveness analysis and modelling, cost–utility analysis, option
appraisal and cost–benefit analysis, the measurement of health benefits in terms of QALYs and related measures
Economic evaluation is the comparative analysis of costs and consequences between two or more alternative
interventions. For example, it may be used to compare the costs and benefits of switching from one vaccination
programme to another. The aim of economic evaluation is to improve efficiency in the context of scarce resources,
while also considering the impact on equity. Three types of efficiency are considered in health economics:
technical, economic and allocative (see Section 4D.1).

COSTS
Evaluation of costs begins by defining:
1. The perspective (i.e. the viewpoint from which the costs and benefits are regarded). Commonly used
perspectives are those of the patient, the hospital, the health authority or wider society
2. The timeframe for the evaluation.
The costs of the alternative interventions are then calculated. Depending on the perspective and timeframe chosen,
different costs may or may not need to be included. For example:
• Costs to other government agencies (e.g. Social Services)
• Costs from loss of productivity
• Costs to the patient’s family
• Out-of-pocket costs to the patient
• Future costs.
Costs may be classified as shown in Box 4D.5.1.

435

Public [Link] 435 3/6/08 [Link]

 4D He a l t h E c onomics

Box 4D.5.1
Direct Salaries, drugs, diagnostic tests, patient transport, etc.
Indirect Costs associated with reduced productivity due to illness, disability and death
Tangible Costs that can readily be measured in currency terms and with certainty
Intangible Psychological costs associated with illness or treatment, e.g. pain and suffering
Fixed Costs that do not vary with the volume level of activity
variable Costs that vary in proportion to the quantity produced

Resource use is evaluated by using micro (bottom-up) or macro (top-down) costing, using either an RCT where
the interventions represent different arms of the trial, or using some form of economic modelling (see below).
The analysis should take account of opportunity costs, take a long-run economic perspective and be adjusted for
discounting: see Box 4D.5.2.

Box 4D.5.2
Opportunity costs It is the opportunity cost that should always be considered, i.e. the forgone value of the
next-best use of the resources
Long-run costs Costs should be assessed from the long-run perspective, i.e. where all inputs (including
capital inputs such as buildings) can be altered freely. In the long run, the average cost
is equal to the marginal cost
Discounted costs Costs should be adjusted for human time preference by means of discounting (see Section
4D.1)
Marginal cost Cost of producing one extra unit of service, reflecting any stepped costs that are
encountered

Finally, a sensitivity analysis should be performed. This takes account of uncertainties in the economic analysis by
establishing confidence intervals around the mean costs. One such uncertainty is the discounting rate that should
be applied; therefore, the sensitivity analysis will include adjustments for a range of plausible discount rates. Note
that because costs are often not normally distributed, bootstrapping (i.e. iterative) techniques may need to be
used.

BENEFITS
Benefits may be expressed in a number of ways: see Box 4D.5.3.

Box 4D.5.3
Type of benefit Definition Example
Fixed benefit Benefit is pre-determined such as a set number of 100 hip replacements
operations
Clinical benefit Benefit is a clinical endpoint 20 mmHg drop in blood
pressure
Utility benefit Benefit is expressed in terms of utility Gain of 2.5 QALYs
Cost benefit Benefits are translated into monetary values Gain of £60 000

Approaches for the monetary valuation of life are shown in Box 4D.5.4.

436

Public [Link] 436 3/6/08 [Link]

 4D Health Econom ics

Box 4D.5.4
Approach Advantages Limitations
Human capital The expected value of the • Easy to calculate • Value of children, elderly
individual’s productivity is and unemployed people
• Easy to define
calculated (both market appears less than that of
and household ‘non-market’ males of working age
productivity), then adjusted
• Does not reflect how
for the individual’s life-
much society is willing to
expectancy
pay for treatment
Contingent Surveys • Does not rely on markets • Requires large and costly
valuation or observed behaviour surveys
• Willingness to pay (WTP),
i.e. how much a person • Can be applied to any • Relies on hypothetical
is willing to pay for a good or service scenarios that may not
health benefit or to avoid reflect reality
harmful risks
• Susceptible to bias:
• Willingness to accept people may state
(WTA), i.e. the minimum different preferences from
amount a person is those that they actually
willing to accept as hold
compensation for a
• WTP and WTA can
loss, or a reduction in a
be affected by an
health-care service
individual’s income
Hedonic wage/ Individuals’ preferences • Based on actual consumer • Focused on immediate
revealed regarding the value of health choices (indicates actual accidental deaths as
preference risk or benefit gain are traded willingness to pay for opposed to deaths due to
against income (e.g. British items such as airbags, chronic exposures (e.g.
soldiers’ pay is greater than smoke alarms, etc.) asbestos)
other comparable public
• Gives insight into an • Biased towards males of
sector posts because soldiers
individual’s valuation of working age
face risk during their service
their own life
life) • Ignores imperfect
labour markets (i.e. no
knowledge of risks, may
not have plenty of job
choices)
• Limited generalisability:
people who undertake
risky jobs are unlikely to
be representative of the
general population

ECONOMIC ANALYSIS
Having considered costs and benefits, the outcome of an economic analysis combines the two into a unified
measure: see Box 4D.5.5 with further detail in Boxes 4D.5.6, 4D.5.7, 4D.5.8 and 4D.5.9.

437

Public [Link] 437 3/6/08 [Link]

 4D He a l t h E c onomics

Box 4D.5.5
Type of analysis Type of Example Advantages Disadvantages
benefit
Cost- Fixed £ per hip Simple to conduct Restricted to one technology
minimisation outcome replacement
Cost- Clinical £ per Straightforward outcome One-dimensional
effectiveness outcomes 20 mmHg measures (life gained, blood
Limited comparability
reduction in pressure reduced)
blood pressure
Cost-utility QALY £ per QALY Allows comparison across Practical problems in valuing
health-care field utility
Cost-benefit Currency £ Allows comparison across Places monetary value on life
(£/$, etc.) sectors (e.g. road building, which is considered priceless
schools)
Practical problems in valuing
health

COST-MINIMISATION ANALYSIS
This type of economic evaluation is used to find the least expensive method of achieving a single outcome: it
assumes that interventions have an equivalent effect and compares programmes solely on the criterion of cost.
For example, treatments A and B both prevent 100 strokes per year. Under cost-minimisation analysis, the cheaper
option would be chosen.
The main advantage and disadvantage of cost-minimisation analysis are shown in Box 4D.5.5. Additional advantages
and disadvantages of cost-minimisation analysis are listed in Box 4D.5.6.

Box 4D.5.6
Advantage Disadvantages
Simple (focuses on cost alone) Assumes that equivalence of benefits has been proved unambiguously: much
research effort would be needed to demonstrate this
Few health-care interventions produce identical benefits

COST-EFFECTIvENESS ANALYSIS
This compares alternative treatments where both the costs and the benefits vary. Benefits are measured in natural
units (e.g. years of life gained; fits of coughing prevented; mmHg drop in blood pressure; mmol/l drop in serum
cholesterol). Since costs and benefits are measured in non-comparable units, the results are often expressed as cost-
effectiveness ratios, i.e. the cost per unit of benefit for intervention (A) versus the cost per unit of benefit for
intervention (B). Relative cost-efficiency can then be assessed, with the preferred option being that with the lower
ratio.
If an intervention is both more expensive and more effective than an alternative, then the criterion for efficiency
becomes the ratio of the net increase in costs to the net increase in effectiveness, i.e. the incremental cost-
effectiveness ratio (ICER). The additional expense of the new intervention requires resources to be redirected
from elsewhere. An economic evaluation assesses whether or not the additional benefits generated by the new
intervention are greater than the loss of benefits from the reduction of other programmes (i.e. whether the
re-allocation is efficient).

438

Public [Link] 438 3/6/08 [Link]

 4D Health Econom ics

The main advantage and disadvantage of cost-effectiveness analysis are shown in Box 4D.5.5. Further advantages
and disadvantages are listed in Box 4D.5.7.

Box 4D.5.7
Advantage Disadvantage
Frequently incorporated within RCTs – hence cost- Inability to compare interventions with differing
effectiveness analyses often provide the least natural effects. For example, interventions aimed
biased estimates of effectiveness at increasing life-years gained cannot be directly
compared with those that improve physical
functioning. Cost-effectiveness analysis therefore
cannot directly address allocative efficiency

COST–UTILITY ANALYSIS
Cost–utility analysis is a form of economic evaluation that measures the effect of an intervention on both morbidity
and mortality. By using a utility-based unit, such as QALYs, to measure benefits, cost–utility analysis is able to
compare alternative health interventions that have completely different types of benefit.
Decision-makers can then be presented with league tables that rank the incremental cost–utility ratios of
different interventions in order that they may select those interventions with the lowest ratios (i.e. best value) until
the budget is expended.
The lower the incremental ratio for an intervention, the higher its priority should be in terms of maximising health
benefits derived from a given level of expenditure. The point at which resources are exhausted defines a maximum
price for a unit of effectiveness, e.g. £20 000 per QALY might be the upper limit of affordability within the budget.
Eliminating interventions with an incremental cost above this threshold price in favour of those below the threshold
is a means of improving allocative efficiency. As with cost-effectiveness analysis, relative efficiency is assessed
using an incremental ratio – here a cost utility ratio which takes as its units the cost per QALY:
Marginal cost (£)
Incremental cost–utility ratio =
Marginal effect (QALY)
An intervention is deemed economically efficient, relative to an alternative, if it results in equal or higher benefits
at lower cost.
For example, should £1m be spent on primary stroke prevention through antihypertensive treatment, or should
£1.5m be spent on expanding a ‘Children’s Hospital at Home’ service? A cost–utility analysis will deem that the
money should be spent on whichever intervention produces more health gain (i.e. QALYs) per £ spent.
The main advantage and disadvantage of cost–utility analysis are in Box 4D.5. Further advantages and
disadvantages are listed in Box 4D.5.8.

Box 4D.5.8
Advantage Disadvantages
The use of a single measure of health benefit Complex to conduct
enables diverse health-care interventions to be
Debatable comparability of utilities arising from
compared, so cost–utility analysis can address both
different measurement instruments, health problems
productive efficiency and allocative efficiency
and interventions

439

Public [Link] 439 3/6/08 [Link]

 4D He a l t h E c onomics

COST–BENEFIT ANALYSIS (CBA)

Finally, this type of economic evaluation involves measuring both costs and benefits in monetary terms. Benefits
may be converted into monetary terms using willingness-to-pay exercises (revealed preference, contingent
valuation). CBA compares two or more interventions from more than one sector (e.g. transport and education)
according to their net effects on welfare.
For example, should £1m be spent on primary stroke prevention or on buying more books for local primary schools?
CBA would quantify the monetary value of the health benefits and the educational benefits, and would recommend
the programme in which the monetary benefit is the greater.
The main advantage and disadvantages of CBA are in Box 4D.5.5. Further advantages and disadvantages are listed in
Box 4D.5.9.

Box 4D.5.9
Advantages Disadvantage
Comprehensive: all costs and all benefits are considered in monetary units Practical difficulty of assessing
benefits in monetary terms
Increased use of interventions with the greatest net gain will increase efficiency
By valuing all costs and benefits in the same units, cost–benefit analysis
compares diverse interventions using the net benefit criterion
Cost–benefit analysis thus simultaneously addresses issues of productive and
allocative efficiency

ECONOMIC MODELLING
In circumstances where a real experiment would be impractical or too time-consuming, economic models can be
used instead to simulate the trial. Economic evaluation is particularly useful in the evaluation of interventions with
benefits that will not be observed for many years (e.g. stop-smoking clinics).
While models are explicit and transparent, their validity remains questionable because they could easily be
manipulated by changing one parameter slightly. This problem of parameter uncertainty can be addressed by
conducting a sensitivity analysis in which parameters are varied across their range of plausible values to assess the
effect on the ICER. Parameters can be altered one at a time (one-way sensitivity analysis) or simultaneously (multi-
way sensitivity analysis and probabilistic sensitivity analysis).
See Figures 4D5.1–4D5.3.

OPTION APPRAISAL
This is the appraisal of various options chosen to achieve specific objectives. The relative advantages and
disadvantages of the options are examined before resources are committed.

QUALITY-ADJUSTED LIFE YEARS

Economists use utility as an expression of an individual’s preference for a particular health status or health outcome.
A commonly used unit of utility is the QALY, which combines the quality and duration of life gained from an
intervention as a single measure. For example, the UK’s National Institute for Health and Clinical Excellence (NICE)
collects evidence on the cost per QALY produced by the treatments that it appraises.
QALYs measure the years spent by an individual in different health states. Each year is weighted according to a scale
between 0 (dead) and 1 (perfect health). Negatively weighted values may also be used for intractable symptoms.

440

Public [Link] 440 3/6/08 [Link]

 4D Health Econom ics

Examples are shown in Box 4D.5.10, and advantages and disadvantages in Box 4D.5.11.

Decision trees The

Thefirst
firstdivision
divisionofofthe
the tree
tree represents
representsthe
theold
old
and
andthethenew
newtreatments.
[Link] Thenext
nextbranch shows
branch shows
the
the potential
potential outcomes
outcomesfor foreach
eachtreatment
treatment(either
unproblematic or adverse events).
(either unproblematical or adverse Subsequent
events).
Old treatment branches
Subsequent branches show potential of those side
show potential consequences
effects (such as of
consequences recovery or death)
those side effects (such as
Arecovery or death)
probability is assigned for each branching point
(except for the initial division into old and new
New treatment A probability is assigned for each branching
treatments)
point (except for the initial division into old and
Each final branch is assigned a cost (in £) and a
new treatments)
benefit (in QALY). For the old treatment, the sum of
the costs and benefits are added together, multiplied
Each final branch is assigned a cost (in £) and a
by the probability of each branch. The same is done
benefit (in QALY). For the old treatment, the sum
for all the branches of the new treatment. The ICER
of the costs and benefits are added together,
then represents the incremental costs and benefits of
multiplied
moving bythe
from theold
probability of each
to the new branch.
treatment
The same is done for all the branches of the new
Figure 4D.5.1 Decision trees treatment. The ICER then represents the
incremental costs and benefits of moving from
the old to the new treatment

Markov modelling Markov

Markovmodels areare
models useful forfor
useful analysing chronic
analysing chronic
diseases
diseases
Each circle represents a health state, e.g. healthy
Each
(H), circle disease
chronic represents
(C), adead
health
(D).state,
Usinge.g.
the best
H C
available evidence, each state is assigned(D).
healthy (H), chronic disease (C), dead Using
a quality-
the best available evidence,
of-life value and a cost each state is
assigned a quality of life value and a cost
Straight arrows represent the probability, by the
year-end, that a patient
Straight arrows moves
represent the from one state
probability, to
by the
another. A new treatment will affect the probabilities
year-end, that a patient moves from one state
D associated with these arrows
to another. A new treatment will affect the
Circular arrowsassociated
probabilities represent probability, at the year-end,
with these arrows
that a patient
Circular remains
arrows in theprobability,
represent same stateat the
Ayear-end, thatexperiment
hypothetical a patient remains in thetwo
is run where samecohorts
state
of, say, 1000 patients each are followed over 25
years. One cohortexperiment
A hypothetical receives the
is new treatment
run where two and the
other receives
cohorts standard
of, say, care. Patients
1000 patients move
each are between
followed
health
over 25states
[Link] One
stay in the same
cohort health
receives thestate
new
according
treatmenttoandthethe
probabilities defined
other receives above
standard
[Link]
Total Patients move
(£) and between
total health states
consequences (QALYs) or
stay in the same health state according to the
over the 25 years are calculated for both cohorts.
probabilitiesare
Adjustments defined
made above
to take account of discounting.
An ICERcosts
Total for the
(£) new
and treatment can then be
total consequences calculated
(QALYs)
over the 25 years are calculated for both
Figure 4D.5.2 Markov modelling cohorts. Adjustments are made to take account
of discounting. An ICER for the new treatment
441 can then be calculated

Public [Link] 441 3/6/08 [Link]

 4D He a l t h E c onomics

Reporting results Thedotted

The dotted line
line represents
representsthe
thecost-effectiveness
cost-effectiveness
i.e. the
plane, i.e.
plane, the maximum
maximumacceptable
acceptableICER.
ICER.
ForFor
the
the this
UK UK this is generally
is generally accepted
accepted as being
as being approximately

New treatment more

approximately
£30 000 per QALY £30 000 per QALY
expensive
Cost difference ∆C

Treatmentsbelow
Treatments belowand
andtotothe
theright
rightof
ofthe
theplane
planeare
likely to betoaccepted,
are likely thosethose
be accepted; aboveabove
and to thetoleft
and
rejected
the left rejected
New treatment more effective

Effect difference ∆E

Figure 4D.5.3 Reporting results: cost effectiveness plane

Box 4D.5.10
Examples: QALYs
5 years of perfect health = 5 QALYs
3 years of perfect health followed by 2 years in a state
calculated to be 0.8 of perfect health = 4.6 QALYs

Box 4D.5.11
Advantages Disadvantages
Single measure combining quality and quantity-of-life Theoretically problematical
measures
Difficulty of estimating consequences while healthy
Allow comparisons of health-care outcomes across
Moral questions about negatively weighted years (i.e
medical specialties
life not worth living)
Large studies using standard quality-of-life measuring
Apparent discrimination in favour of paediatric
instruments (e.g. SF-36 or EQ5D) lead to valid, reliable
treatments and against those for older people
results
Do not capture externalities (e.g. benefits to the
patient’s family and friends); therefore, are likely to
undervalue health care
Calculation is dependent upon who and how asked
No weighting for age (compare DALYs, below)
No discounting

442

Public [Link] 442 3/6/08 [Link]

 4D Health Econom ics

DERIVATION OF QALYs
QALY = value of preference in a particular state x Length of time in that state.
Two key issues in determining value preferences are:
• Who is consulted
• How they are asked.
See Box 4D.5.12.
Methods for calculating QALYs are shown in Table 4D.5.1.

Box 4D.5.12
Who There is evidence that the value of a QALY changes radically depending on who makes the value
judgement. However, it remains a moot point whether the people consulted should be health
professionals, the general public or patients who have experience of the particular medical condition or
treatment
How Five main approaches are used to derive quality-of-life weightings (see below). Note, however, that the
chronicity of the hypothetical illness influences valuations, as does the way in which questions are
asked. Since responses are given to imagined situations, they may not reflect real life accurately

Table 4D.5.1 Methods for calculating QALYs

Method Description Problem
Time trade-off Respondents are asked how many years of life with the disease they Influenced by
are willing to give up for one year of full health time preference
(compare
discounting)
Standard gamble Respondents imagine that they have the disease and are asked to Biased by
gamble on taking a hypothetical cure which will either fully cure the respondents’
disease or kill them. The treatment has a probability (p) of full cure attitude to risk
and a probability (1 – p) of death, and p is varied between 0 and 1
Rating scales Subjects are asked to attribute values between 0 and 1 to a series of Doubtful interval
health states. With visual analogue scales, respondents place a mark properties
on a 10 cm line that represents death at one end and full health at
the other
Person trade-off Subjects are asked to imagine groups of patients with a severe disease May be affected
(X) and a mild disease (Y). Imagine there is a cure for both diseases by factors other
and sufficient funds to pay for one patient with X. How many patients than the health
with disease Y would have to be cured for you to be indifferent? state alone
Subjects are asked whether they would cure one person in health state
X or (n) people in health state Y. The value of (n) is varied until the
subject is indifferent between the two alternatives
Quality-of-life Quality-of-life questionnaires (e.g. EuroQol, Nottingham Health Profile) Scales are not
questionnaires can be used to derive utility values for certain disease states appropriate for
some disease
states

443

Public [Link] 443 3/6/08 [Link]

 4D He a l t h E c onomics

DISABILITY-ADJUSTED LIFE YEARS (DALYs)

In contrast to the QALY, which measures the effect of disease on an individual, the DALY is a measure of the burden
of disease in a population. It quantifies the impact on the population of premature death and disability by
combining the two into a single measure.
DALY = Years of life lost to death (mortality) + Years of life lost to disability (injury and illness).
Disability is weighted between 1 and 0 (by using estimation methods similar to those for QALYs). DALYs are
commonly used in low-income country settings.
DALYs are compared with QALYs in Box 4D.5.13.

Box 4D.5.13
Similarities with QALYs Differences from QALYs
Combine morbidity and mortality into a single Measure disease burden (rather than quality of life)
dimension
Are age weighted, i.e. more weight is afforded to
Similar methods for calculating morbidity weightings productive years (childhood is valued less than early
(e.g. standard gamble) adult life)
Are discounted at 3%

SENSITIVITY ANALYSIS
A sensitivity analysis should be included in every economic evaluation. It involves varying the assumptions and
estimates that underlie the study, e.g. by changing the discount rate to 2% from 6%, or by including a range
of intangible costs. Since costs and other assumptions may not be normally distributed, iterative methods (e.g.
bootstrapping) are often used.
Sensitivity analyses test the robustness of economic evaluations by:
• Making explicit any uncertainty, imprecision or methodological controversy within the evaluation
• Highlighting any areas where extra data are needed
• Allowing decision-makers to gauge how much confidence they should place in the study results.

4D.6 MARGINAL ANALYSIS

See Section 4D.1.

4D.7 DECISION ANALYSIS

Decision analysis is a process that involves:
• Division of a problem into simpler, manageable components
• Detailed examination of each component
• Formation of components into a logical sequence so as to identify the best solution.
All options are identified and listed, together with the probability and utility of each possible outcome.
See Box 4D.7.1.

444

Public [Link] 444 3/6/08 [Link]

 4D Health Econom ics

Box 4D.7.1
Example: decision analysis
For a particular disease there is a standard treatment (that results in either cure or death) and a new treatment
(that may result in cure or death or disability). A decision tree is constructed to illiustrate these possibilities.
Note that the branches emanating from each node are mutually exclusive, and that the probabilities at each
branch must therefore add up to 1.

Utility

Complete cure (p � 0.95) 1.0

Standard treatment Chance

node

Death (p � 0.05) 0.0

Decision
node

Death (p � 0.01) 0.0

Chance
Scar (p � 0.25) 0.9
New treatment node

Chance
Cure (p � 0.99) node

Complete cure (p � 0.75) 1.0

Reproduced from Jefferson et al (2000).

The probabilities of each outcome in a chance node are included, and utilities are assigned to every possible
outcome using a common scale (a value between 0 and 1).
For each treatment, the sum of the utilities for each possible outcome multiplied by its probability is calculated.
Standard treatment = (0.95 ¥ 1.0) + (0.05 ¥ 0.0) = 0.95
New treatment = (0.01 ¥ 0.0) + (0.99 ¥ 0.25 ¥ 0.9) + (0.99 ¥ 0.75 ¥ 1.0) = 0.966
The option with the higher overall utility is chosen – in this case the new treatment. Finally, a sensitivity
analysis is conducted (in which each probability and utility varied within a confidence interval) so as to test the
robustness of conclusions.

445

Public [Link] 445 3/6/08 [Link]

 4D He a l t h E c onomics

4D.8 ECONOMIC EvALUATION AND PRIORITY SETTING

Role of economic evaluation and priority setting in health-care decision-making, including the cost-effectiveness of
public health, and public health interventions and involvement
Wanless’s first report (2002) argued that there would be financial benefits to the UK from investing more resources
in public health. His second report (2004) called for a single framework for evaluating the cost-effectiveness of
both health care and public health – the implication being that investment in the latter would represent markedly
better value for money.
Both cost–utility analyses and cost–benefit analyses allow interventions with disparate outcomes to be compared
on the same scale. These include public health interventions that tackle the risk factors for multiple diseases (e.g.
smoking and obesity) and the wider determinants of health. Since the benefits of public health interventions (health
promotion, screening and treatment) may not be realised for decades, so the principle of discounting becomes
particularly pertinent (see Section 4D.1).

ECONOMIC EVALUATION
UK In the UK, NICE provides priority-setting advice to governments based on economic evaluations of the
categories shown in Box 4D.8.1.

Box 4D.8.1
Health technologies Drugs, procedures and other treatments
Clinical practice Management of clinical conditions and diseases
Public health interventions Health promotion and preventive medicine

For each of these categories, the process of economic evaluation consists of identifying, then critically appraising
and synthesising the evidence, so that recommendations of cost-effectiveness (expressed as cost per QALY) can be
made: see Box 4D.8.2.

Box 4D.8.2
Identification of A search of the published and unpublished evidence is performed (using economic
evidence search filters). Both qualitative and quantitative studies are sought from the literature
and the grey literature
Critical appraisal The quality of individual studies is appraised, following the hierarchy of evidence (see
Section 1C.5), where RCTs (or cluster RCTs for community interventions) are particularly
valued
Synthesis of evidence Assessment of applicability, construction of evidence tables, meta-analysis and
summaries of the evidence (in the form of evidence statements)

PUBLIC HEALTH INTERVENTIONS

Public health activities may be direct or indirect: see Box 4D.8.3.

446

Public [Link] 446 3/6/08 [Link]

 4D Health Econom ics

Box 4D.8.3
Direct Stop-smoking services
Obesity clinics
Health promotion posters
Indirect Creation of parks and open spaces
Restrictions on advertising fast foods

When evaluating a public health intervention, NICE follows the process shown in Box 4D.8.4. NICE uses a discount
rate of 3.5% for costs and benefits in its analyses, and adopts the following economic perspectives: NHS, society
and the individual.

Box 4D.8.4
Background work Topics for evaluation are set by ministers
Stakeholders are identified who will scrutinise and validate the evaluation process
Quality assurance arrangements are put in place
Preliminary work The scope of the evaluation is determined (i.e. definition, settings, population, exclusion
criteria)
A preliminary literature search is conducted
Detailed work Key questions are set and an analytical framework is constructed
A review of the evidence is conducted (selection of studies, quality assessment, evidence of
implementation)
The evidence is synthesised in the form of evidence statements

For each question addressed by the evaluation, an evidence statement is presented that documents both the
strength of the evidence (i.e. its quality, quantity and appropriateness) and its applicability to the target
population. NICE rates the strength of evidence according to the scale shown in Section 1C.5.
For example, a hypothetical evidence statement might be:
‘A body of 2+ evidence of efficacy offers consistent findings about the impact of pogo sticks on weight loss. The
evidence is directly applicable to the target population in terms of ethnicity, age and gender.’

COST–CONSEQUENCE ANALYSIS
Because of the complexities of public health interventions, NICE sometimes uses a cost–consequence approach in
addition to cost-effectiveness analysis. This technique enables non-quantifiable outcomes, equity and distributive
justice to be considered in an explicit way (see Section 4C). It presents a table of the lifetime impact of the new
treatment on individuals or groups of individuals in terms of:
• Resource use (both health-care and productivity losses)
• Health outcomes (symptoms, life-expectancy and quality of life).

447

Public [Link] 447 3/6/08 [Link]

Public [Link] 448 3/6/08 [Link]
 Section 5
ORGANISATION AND MANAGEMENT OF HEALTH CARE
AND HEALTH-CARE PROGRAMMES

Management is not limited to managers. Every member of an organisation has some managerial role. At its simplest
level, this may require controlling one’s own time and resources; at its most complex, it could demand directing
projects, people and budgets.
Section 5 covers the major theories of management, from working with people to managing budgets. The challenges
and benefits of working in teams are covered, together with an overview of how to motivate individuals. To enable
team working to function optimally, an understanding of personalities, the ways that people think and how they
function when they work together is needed. At a higher level, organisational structure and culture can account for
the success or failure of a health system in meeting its goals. Understanding organisations, how they function (and
more importantly, why they fail), is fundamental to implementing change in health systems.
Change is an essential part of any health service provision. As new technologies, diseases and different political
priorities emerge, so health care must evolve and transform to meet the needs of its population. Robust strategies
are needed to manage change carefully.
Health-care systems exist in a world with finite resources. Practitioners will be all too aware of resource constraints
and their effects on the capacity of organisations to provide health care. Understanding the basics of finance and
management accounting provides them with insights that are essential for establishing and sustaining health
services.

449

Public [Link] 449 3/6/08 [Link]

Public [Link] 450 3/6/08 [Link]
 5A I ndividuals, Teams and Groups

5A
Individuals, Teams and Groups

5A.1 Individuals, groups and team dynamics 451 5A.8 Effective communication 462
5A.2 Creativity and innovation 454 5A.9 Interactions between clinical and
5A.3 Interprofessional learning 455 managerial professionals 462
5A.4 Personal management skills 456 5A.10 Power and authority 464
5A.5 The effective manager 458 5A.11 Professional accountability 464
5A.6 Leadership and delegation 458 5A.12 Changing behaviour 467
5A.7 Negotiation and influencing 459

Understanding individuals, teams/groups and their development

When things are going well, the benefits of working with others are clear: stimulation of ideas, enthusiasm,
enjoyment, interest, increased productivity and a greater sense of reward. However, there are times when working
with other people leads to conflict and stifled creativity, and impedes the progress of a project.
Public health practitioners often find themselves in the position of working with people from different professional
groups, sometimes across organisational boundaries. A knowledge of management principles, team dynamics and
interactions between professional groups is vital to understanding how groups function. It can also offer ways to
analyse and address conflict, and enables us to modify our own attitudes and practices accordingly.

5A.1 INDIvIDUALS, GROUPS AND TEAM DYNAMICS

Motivation, creativity and innovation in individuals, and its relationship to group and team dynamics
People work differently when they are working alone or in groups. At its best, group work results in outcomes
of a quality that the individuals working alone could never have produced. However, it is important that the
composition, dynamics and leadership of the group be considered so as to motivate its members towards working
together as productively as possible.
Groups may be convened formally (in order to perform a specific task) or informally (by individuals on the basis of
a common interest). They are collections of people who:

451

Public [Link] 451 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

• Interact with one another

• Are psychologically aware of one another
• Perceive themselves to be in a group (Schein and Bennis 1965).

GROUP MOTIVATION
Group motivation depends primarily upon individual motivation. Within a group, individuals will adopt one or more
roles, as determined by a compromise between:
• How an individual wants to behave
• How other group members expect them to behave
• The group task.
Belbin (1996) described eight roles that should be represented by the members of an effective team, and can be
remembered with the mnemonic ICE FIRST: see Table 5A.1.1. A ninth role of specialist is sometimes added, i.e. a
person with a high level of skill in one given discipline. Note that team members may fulfil more than one role.

Table 5A.1.1 Belbin’s eight roles for an effective team

Implementer/company worker Make things happen
High degree of self-discipline
Deliver on time
Coordinator/chair Default chairperson
Step back to see the big picture
Evaluator/monitor Fair and even-handed observers and judges of what is going on
Can become almost machine like
May have difficulty inspiring themselves or others to be passionate about
their work
Finisher/completer Perfectionists
Strong sense of duty
Complete painstaking and unpleasant tasks if they believe that this will
improve quality
May frustrate their team mates
Innovator/plant Come up with unusual and innovative solutions to problems
Resource investigator Vigorously pursue contacts and opportunities
Excellent networkers
Tend to lose momentum towards the end of a project
Shaper Eager individuals that provoke their team into action
May be insensitive to the feelings and perceptions of others
Team worker Ensure that everyone in the working group is getting along
Good listener and diplomat
Talented at smoothing over conflicts

452

Public [Link] 452 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

Groups can be motivated by means of:

• Feedback to remind the team members of the importance of their group’s task, together with their individual
roles within it
• Leadership: a good leader develops team spirit and elicits a high level of commitment from team members.

INNOVATION
Amabile (1998) has argued that three components are needed for creativity in companies:
1. Motivation (especially intrinsic motivation, i.e. satisfying own needs)
2. Expertise (technical, procedural and intellectual knowledge)
3. Flexible thinking (how flexibly and imaginatively people approach problems).
Creativity itself can be thought of as a three-step sequence consisting of an input, a process and an output (i.e.
innovation): see Figure 5A.1.1.

Input: Process: Output:

ideas group innovation

interaction

Figure 5A.1.1 Sequence for creativity

TEAM DYNAMICS
Tuckman (1965) described a four-stage process through which newly formed teams progress and develop: see Box
5A.1.1. A fifth stage, adjourning, was added in 1975 and refers to the disbanding of the team after its task has
been completed.

Box 5A.1.1
Forming Tasks and rules are established
Resources are acquired
Reliance is placed on the leader
Storming Internal conflict
Members resist the task emotionally
Norming Conflict is settled
Cooperation develops
Views are exchanged
Norms (i.e. new standards) are developed
Performing Teamwork is achieved
Flexible roles are developed
Solutions are found and implemented

453

Public [Link] 453 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

TEAMWORK
Charles Handy (1978) maintains that the ideal team not only achieves formal goals (i.e. those of its organisation),
but also achieves informal goals [such as Pilowski and Spence (1997) satisfying its members’ psychosocial needs].
The ideal team is described as having the traits listed in Box 5A.1.2.

Box 5A.1.2
Organisation Common purpose
Clearly defined task
Clear team objective
Members Members have specific expertise
Members know their roles
Teamwork Members support each other
Members complement each other in skills and personalities
Members are committed to accomplishing the task
Leadership Leader who coordinates and takes responsibility

5A.2 CREATIvITY AND INNOvATION

Barriers to, and stimulation of, creativity and innovation (e.g. by brainstorming)
Creativity is one of the key strengths that public health professionals can bring to their working environment. By
focusing on the ‘bigger picture’, the public health viewpoint is well placed to challenge established practices in a
positive way.

STIMULATION OF CREATIVITY
The following techniques have all been used by successful organisations to stimulate creativity and innovation
among staff:
• Brainstorming
• Suggestion boxes
• Team away-days
• Mind-mapping
• Reward.
Brainstorming is a method for promoting creativity in which team members ‘think out loud’ in order to solve a
problem. The process is conducted in an atmosphere that is conducive to independent thought and discussion. It
is held to be a politically incorrect term by some, who prefer to use the term ‘thought shower’. The Plain English
Campaign considers that this suggestion ‘reaches the point of real ridicule’. National organisations representing
people with epilepsy or mental illness do not consider the term ‘brainstorm’ to be politically incorrect (see www.
[Link]/[Link]?i=100&s=1111).

454

Public [Link] 454 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

DIFFUSION OF INNOVATIONS
New ideas are likely to spread rapidly if they are:
• Compatible with existing systems
• Simple
• Amenable to being trialled
• Demonstrably more efficacious.
An example of an innovation that was rapidly propagated is that of crack cocaine: crack met all four of the
conditions above and its use spread rapidly among existing cocaine users.

BARRIERS TO CREATIVITY
The following factors hamper innovation:
• Uncertainty
• Over familiarity
• Fear of change
• Team conflict.

5A.3 INTERPROFESSIONAL LEARNING

Learning with individuals from different professional backgrounds
Several developments in health systems have led to an increased focus on interprofessional learning in recent years:
• Changing professional boundaries and requirements for clinical practice
• Changing health-care structures
• Emphasis within health care on multi-professional teams and clinical networks.
Interprofessional learning is more than simply sharing classes or training sessions with other professionals.
According to Humphris and Hean (2004), this requires sessions that enable students to learn ‘with, from and about
one another’. They also note that this requires more emphasis on small-group learning rather than large, didactic
teaching sessions.
Eng In England, interprofessional education has gained increasing momentum since early commitments in the NHS
Plan (DH 2000). The Bristol Royal Infirmary Inquiry (a high-profile enquiry into children’s heart surgery at Bristol
Royal Infirmary from 2001) made a number of recommendations to promote interprofessional learning from as early
as possible in clinical training to reduce ‘damaging intertribal rivalries’ in the NHS. The Department of Health’s
(2002) strategy to promote interprofessional learning, Working Together – Learning Together (see [Link])
includes a commitment to include interprofessional learning in:
• Education and training for all health professionals
• Undergraduate and pre-registration programmes, through to continuing professional education
• Practice placements as well as in the classroom.
A number of initiatives have been supported and funded, both nationally and locally, to support interprofessional
and common learning in health professional education.
See Box 5A.3.1.

455

Public [Link] 455 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

UK Box 5A.3.1
Example: Expansion of the public health workforce beyond medicine
Many disciplines are involved in delivering public health. In the past, a formal route for doctors to train in
public health existed but the expertise of other professional groups was not always recognised or regulated.
Public health in the UK has changed in recent years to enable individuals from a range of non-medical
backgrounds to enter formal specialist training programmes. Both medics and non-medics follow a common
recruitment process and training programme to become public health consultants.
Once they have completed training, doctors remain registered with the General Medical Council. A new
organisation, the UK Public Health Register, has been developed to maintain a register of non-medical public
health professionals who have completed specialist training or have been assessed through a portfolio of work as
competent to work as a public health consultant.
Reproduced from the UK Voluntary Register for Public Health Specialists: [Link]/[Link];
Faculty of Public Health: [Link].
There are advantages and disadvantages involved in moving away from education in single professional groups
towards interprofessional learning: see Box 5A.3.2.

Box 5A.3.2
Advantages Disadvantages
Improve communication between different professions Affect professional support by reducing traditional
professional networks and support systems
Reduce the formation of professional ‘silos’
Resource intensive to reconfigure training and
Promote clinical debate through opening issues to
education programmes to provide interprofessional
different professional perspectives
learning
Improve teamwork through enhanced appreciation of
Lack of evidence that interprofessional learning
the roles of other staff
produces the outcomes envisaged
Improve patient care by developing care centred on the
Long lead time for evaluation of pre-registration
patient rather than on professional structures
professional learning impact (10–12 years before
Enhance capacity by expanding clinical roles as required students are practising consultants)
by the situation and needs of the team

5A.4 PERSONAL MANAGEMENT SKILLS

Personal management skills (e.g. managing time, stress, difficult people, meetings)
Effective managers need skills in managing people and projects, but also in managing their own resources. The
skills needed to manage other people are the same skills required to manage ourselves: the ability to plan, delegate,
organise, direct and control.

TIME MANAGEMENT
Time management is a key component of self-management. Time can be managed by:
• Prioritisation into long-, middle- and short-term (today’s) plans
• Organising offices, workstations and desks
• Managing paperwork and emails systematically
• Delegation.

456

Public [Link] 456 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

Covey (1989) identifies a set of behaviours associated with efficient time management:
• Being proactive
• Beginning with the end
• Putting first things first
• Thinking win–win
• Seeking first to understand, then to be understood
• Synergising
• Sharpening the saw (taking time off).

STRESS MANAGEMENT
Several factors promote stress, including increased competition, de-regulation and rapid change. Methods of dealing
with stress include:
• Recognising the symptoms of stress in yourself and others
• Understanding the factors that cause stress
• Applying a range of strategies to avoid, reduce and manage stress.
Well-run meetings reduce stress, save time and improve the effectiveness of the organisation. They tend to have the
following features:
• Carefully planned with a realistic and succinct agenda
• Well chaired
• High participation before the meeting
• Participants are aware of the contribution that the meeting will make to the managerial process of the
organisation.
External facilitators can improve the effectiveness of meetings by:
• Taking the chair (an independent facilitator focuses participants’ thoughts on the matters at hand)
• Providing training for the chairperson and meeting participants
• Helping committees or boards to establish protocols for working effectively together
• Helping meetings decide the level of issues for which decisions can be made within the meeting, and therefore
what topics should be discussed elsewhere.

MANAGING CONFLICT
The three principal methods for avoiding conflict are:
1. Negotiation
2. Mediation
3. Group arbitration.
Difficult situations can often be avoided by the careful use of communication skills to improve the delivery of
bad or unpleasant news. By recognising personality types (especially your own personality characteristics), it is
possible to identify the people with whom you are likely to clash and to develop strategies for dealing with them.
When conflict does occur, it may be prevented from escalating further by means of the following conflict–resolution
principles:
• Mutual respect
• Identification of shared values
• Honesty
• Shared objectives
• Combating disinformation.

457

Public [Link] 457 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

5A.5 THE EFFECTIvE MANAGER

An effective manager can help improve the quality and productivity of an organisation by means of delegation,
feedback and listening.

DELEGATION
See Section 5A.6.

FEEDBACK
Feedback is crucial for the regulation of any system. It is the process whereby elements of the output of the system
are returned to its input so as to regulate further output. Likewise, the feedback that a manager provides to an
employee influences the performance of that employee.
The way in which feedback is delivered is pivotal, and managers need to practise giving feedback (an activity that
many will actively avoid). A good manager will provide feedback only on specific, observable behaviours and will
ensure that the feedback is provided non-confrontationally – perhaps as part of Iles’s (2005) criticism sandwich
(good news, bad news, more good news).

LISTENING
Iles (2005) argues that the success of organisations can be assessed by the speed with which bad news travels
upwards through the management hierarchy. One way to ensure that the senior management team are kept informed
of developments is to use the technique of management by wandering around (MBWA). Developed at the Hewlett-
Packard Corporation, managers employing MBWA set aside time in their diaries each week to walk through their
departments and engage in impromptu discussions. Although very simple (it has been described as the ‘technology
of obvious’), MBWA is an extremely effective concept – particularly at times when the organisation is facing financial
difficulties or reorganisation.
During their walk, managers should:
• Listen to what employees are saying
• Explain organisational policy face to face with employees
• Be prepared to offer on-the-spot assistance to employees.

5A.6 LEADERSHIP AND DELEGATION

There are many definitions of leadership. Most encompass the notion that a leader influences others to follow.
Different models of leadership are described in detail in Section 5C.1.

LEADERSHIP
As well as communicating shared visions to their employees, successful leaders are able to foster an environment
within their organisation that encourages:
• Appropriate risk taking
• Recognition and reward of success
• Empowerment that allows other leaders to emerge.
Leadership and management are separate concepts, with the former being a component of the latter. The principal
aim of a manager is to maximise the output of the organisation through:
• Organisation
• Planning

458

Public [Link] 458 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

• Staffing
• Leading
• Controlling.
In certain circumstances (e.g. within highly motivated groups) it is arguable whether leadership is required at all.

DELEGATION
Delegation is a key skill of an effective manager. It is the process of assigning authority and responsibility for a
specific activity to another person. While a manager may delegate tasks, the ultimate responsibility for these tasks
cannot be delegated.
As well as its most obvious benefit of sharing the burden of tasks, delegation can serve other purposes, such as:
• Reinforcing the role of leaders through promoting involvement
• Developing skills in team members.
However, many managers struggle to delegate successfully, fearing that a job will not be done properly by anyone
else. Iles (2005) describes three rules that the effective manager should ensure when delegating any task:
1. The manager must be confident that the employee understands the task
2. The employee and manager must both be confident that the employee has the skills and resources necessary
3. The manager must provide feedback to the employee.

DELEGATION OF GOALS
Drucker (1950) described the technique of management by objectives (MBO). Managers following this method
delegate goals rather than tasks. Employees are set a target to meet, but are free to choose the tactics and
strategies that they will use to accomplish it. MBO has the following advantages:
• Managers avoid becoming so engrossed in day-to-day events that they lose sight of the organisation’s
objectives
• All employees participate in the strategic planning process
• The organisation’s performance can be readily measured against defined objectives.

5A.7 NEGOTIATION AND INFLUENCING

Because of their relative lack of resource and position power (see Section 4C.9), public health practitioners must
often rely on their negotiation skills to achieve their aims through influencing other people. The processes of
negotiating and influencing both involve considering a situation from various points of view.

NEGOTIATION
Negotiation is the skill of resolving situations where two parties have conflicting desires. A negotiator investigates
the situation with the aim of finding a solution that is acceptable to both sides. The most effective negotiating
style will depend on the situation: it is influenced by the desire to meet your own needs and those of the other
party.

DIFFERENT NEGOTIATING STYLES

See Figure 5A.7.1.

459

Public [Link] 459 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

Desire to meet your own needs

HIGH
Win–lose Win–win
(‘Competing’) (‘Collaborating’)

Part win–part win

(‘Compromising’)

Lose–lose Lose–win
LOW

(‘Avoiding’) (‘Accommodating’)

LOW HIGH
Desire to meet the other party’s needs

Figure 5A.7.1 Negotiating styles. Reproduced from LSHTM Organisational Management notes

INFLUENCE
To influence is to mobilise resources that modify the behaviour of others. There are three components to influence:
• Conformity (individuals change their behaviour in order to adhere to social norms)
• Compliance (request for a change in behaviour)
• Obedience (in order to change behaviour).
According to Covey
(1989), every person Circle of
has a circle of concern Circle of concern
(i.e. a range of issues
in which the person has
emotional involvement)
and a smaller circle of Circle of Circle of
influence (i.e. a smaller influence influence
range of issues that the
person has the power to
alter): see Figure 5A.7.2.

concern

Figure 5A.7.2 Covey’s PROACTIVE FOCUS REACTIVE FOCUS

circles of concern and Positive energy enlarges Negative energy
influence. Reproduced the circle of influence reduces the circle of
from Covey (1989) influence

460

Public [Link] 460 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

People who are proactive concentrate their efforts on their circle of influence – thereby enlarging it and thus
becoming more powerful. In contrast, reactive people focus their attention on their circle of concern – to the
demise of their circle of influence.

PUBLIC HEALTH ADVOCACY

Advocacy is the field within public health that aims to develop and shape public opinion in a strategic way, using
whatever methods are the most effective in accomplishing this end. Even when the underlying goal being advocated
is not contentious (e.g. increasing road safety), the means used to achieve it may be highly controversial (e.g.
graphic television adverts that are designed to shock the audience).
Chapman (2004) identifies the following 10 steps that should be addressed when acting as a public health advocate:
1. Identify the public health objectives
2. Attempt to find a win–win outcome
3. Identify the key decision-makers and how they can be influenced
4. Identify the strengths and weaknesses of both sides of the argument
5. Set out the media objectives
6. Choose how to frame the key issue
7. Identify symbols and ‘word pictures’ to illustrate the argument
8. Compose ‘sound bites’
9. Personalise the topic by addressing the issue from the perspective of the ordinary citizen
10. Mobilise large numbers of sympathisers rapidly.
See Box 5A.7.1.

Box 5A.7.1
Example: Achieving cuts in global greenhouse gas emissions to halt climate change
While the science behind climate change is now widely accepted, the changes in policy required to reduce carbon
dioxide emissions have not followed. One of the reasons for this is that the changes necessary are seen as being
‘bad for business’. In October 2006, former head of the World Bank, Sir Nicholas Stern, produced a report on the
potential effects of climate change on the global economy. The report is a good example of advocacy because it:
1. Identifies decision-makers for the changes required to address climate change: the target audiences for
the report were the business and economic communities
2. Understands how to influence these groups: the report considered the effects of climate change on
business. The arguments are presented to the decision-makers by someone who is credible to that
community (Stern is a high-profile economist) rather than by an environmentalist or humanitarian
3. Re-frames the issue in these terms: the report highlights that it makes economic sense for the global
economy to tackle global emissions now rather than leaving them unchecked until 2050
As a BBC commentary notes:
‘The overall message of the report is not fundamentally new. In its 2001 report the Intergovernmental Panel on
Climate Change (IPCC) calculated costs in the same ballpark .… The acid test of Stern is whether his economist’s
language can bring about the fundamental shift in political and economic direction which other financial
analyses – as well as arguments posited on human rights, poverty alleviation, environmental services, health and
simple concern for the natural world – have failed to do.’
Reproduced from BBC (30 October 2006). Expert reaction to Stern review, available online at [Link]/1/hi/
business/[Link], BBC (30 October 2006); Climate costs: The next generation, available online at [Link].
[Link]/1/hi/sci/tech/[Link].

461

Public [Link] 461 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

5A.8 EFFECTIvE COMMUNICATION

Principles, theories and methods of effective communication (written and oral) in general, and in a management
context
See also Sections 2H.9 and 6C.
In a managerial context, the functions of communication within an organisation relate to:
• Production (direction, coordination and control of activities)
• Innovation (stimulation of change and development of new ideas)
• Maintenance (preservation of the values and relationships that bind the organisation).

COMMUNICATION METHODS
See Table 5A.8.1.

Table 5A.8.1 Communication methods

Formal Communications that are routed through a so-called ‘official channel’, e.g. a written memorandum
from a chief executive to the directors of an organisation
Informal Information is passed among colleagues in an unstructured way
Diagonal No obvious line of authority exists through which the information may be communicated
vertical This is the principal channel by which strategy, policy and tactics permeate from decision-makers
down through the organisation to the frontline – and by which news from the frontline is fed back
(see Section 5A.5)
Oral Forms of verbal communication include speaking to another person over the telephone, or face to
face in a discussion, debate, interview, presentation or meeting
Written A form of verbal communication where the message is documented in writing, e.g. email, letters,
fax
Non-verbal There is no spoken language, e.g. eye contact, body language, sign language
visual A form of non-verbal communication where information is displayed in various ways, e.g. tables,
advertisements. People express visual cues through their gestures and appearance (e.g. clothing,
hair style, make-up). These may affect how other people receive their verbal and non-verbal
communication
Internal There are various methods of internal communication that can be used in an organisation. These
include notices, bulletins, newsletters, tannoy, fax, letters, telephone, memos, email, instant
messaging, intranet pages, face to face, reports, memoranda, etc.
External Communication with outside bodies can be conducted using similar mechanisms to internal
communication. However, the routes employed are likely to be more formal and less spontaneous,
e.g. face-to-face communication is more likely to be in the setting of an arranged meeting than as
a chat over coffee

5A.9 INTERACTIONS BETWEEN CLINICAL AND MANAGERIAL PROFESSIONALS

Interactions among managers, doctors and others
Strong links between managers and clinicians are essential in the delivery of health services. However, there is often
perceived to be a conflict between the role of clinicians (particularly doctors) and that of managers. Public health

462

Public [Link] 462 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

practitioners cross both managerial and clinical spheres. While this places practitioners in a strong position to align
the two professional agendas, it may also place people in the middle of clinician/manager conflicts.
Tensions between doctors and managers are often described in relation to finance and workload but various
authors have characterised fundamental distinctions between the groups, right from the time that they enter their
professions: see Table 5A.9.1.

Table 5A.9.1 Distinctions between doctors and managers

Doctors Managers
Professional cultures Strong established professional identity Emerging identity
(see Section 4B.3)
Training and learning Specialised, prescribed body of knowledge No specific qualifications essential for role
Recognised qualifications essential to
practice
Career path Formal training route Different ways to enter profession
(although formal NHS training exists)
Low turnover (consultants often in posts
for many years) High turnover (22% of NHS chief
executives change jobs in 3 years)
Responsibilities Individual patient Systems
Clinic/practice Budgets

It is simplistic to describe differences just in terms of doctors and managers. Relationships between different
clinical staff and management vary. Moreover, managers working in health services have often had clinical
backgrounds and some clinicians adopt management roles in addition to clinical responsibilities. As Degeling
et al (2003) highlight, nurses, doctors working as clinicians and those working in management positions show
stereotypical differences in their views on the key elements of health service modernisation: see Table 5A.9.2.

Table 5A.9.2 Views on the key elements of health service modernisation

Medical Medical General Nurse Nurse
clinicians managers managers managers clinicians
Recognise connections between Oppose Support Equivocal Support Oppose
clinical decisions and resources
Transparent accountability Oppose Support Support Support Oppose
Systematisation Oppose Oppose Support Support Equivocal
Multidisciplinary teams Oppose Oppose Equivocal Support Support
Reproduced from Degeling et al (2003).
Interactions between clinicians and managers can be improved by:
• Aligning corporate and clinical goals
• Involving clinicians in management decisions, e.g. resource allocation
• Involving managers in clinical decisions, e.g. clinical pathway development
• Shared education and training sessions
• Further developing clinical leadership.
For an example from England, see Box 5A.9.1.

463

Public [Link] 463 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

Box 5A.9.1
Example: The Professional Executive Committee – clinical leadership in primary care trusts
When primary care trusts (PCTs) were set up in England, their management was divided between a trust board
and a professional executive committee (PEC). The PEC is chaired by a practising clinician, often but not always
a general practitioner, and the majority of its members are clinicians practising in primary or community care.
PECs were set up to ensure that clinicians were involved in making decisions about local service priorities. In
turn, they enable clinicians to take responsibility for making decisions with financial implications and to become
involved in some of the pressures faced by PCTs. While PECs provide a mechanism to involve clinicians in PCT
management, the committees have not been without their problems:
• Some primary care clinicians remained disengaged with PCTs
• The PEC role has been seen as ‘vague’ and sometimes overlapped with the trust board
• PECs can be resource intensive; as well as providing the administrative support for the meetings, PEC costs
also include the costs of locum cover for the primary care contractors involved in those meetings
The function of PECs has recently been reviewed following other changes to NHS structures.
Reproduced from [Link]/assetRoot/04/13/89/37/[Link].

5A.10 POWER AND AUTHORITY

Theoretical and practical aspects of power and authority, role and conflict
Power is the ability to make choices or influence outcomes. See Section 4C.9 for details.
Authority can be seen as the right to make decisions and give orders. According to Weber (1958), authority
manifests itself in three ways: see Box 5A.10.1.

Box 5A.10.1
Type of authority Description Example
Traditional Authority is derived from preserved customs The medical Royal Colleges rely on traditional
authority authority
Charismatic Authority comes from the personality and Nelson Mandela. However, Adolf Hitler was
authority leadership qualities of the individual which also a charismatic leader; charisma is not
inspire obedience and loyalty from others always a force for good
Rational–legal Authority is derived from powers that are The Chief Medical Officer of a country is
authority bureaucratically and legally attached to afforded this type of authority
certain positions

5A.11 PROFESSIONAL ACCOUNTABILITY

Professional accountability – clinical governance, performance and appraisal
Professional accountability involves making individuals and organisations responsible for the quality of the service
that they deliver. Systems of professional accountability have been strengthened in many countries in recent years.
UK When the NHS was first established, clinical professionals (particularly doctors) were held accountable for
their actions solely through professional bodies such as the General Medical Council. In recent years, clinicians have
partly ceded their high degree of professional autonomy and self-regulation in favour of:

464

Public [Link] 464 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

• A regulatory system that involves members of the public

• Explicit standards of accountability such as clinical governance.

ROLE OF PROFESSIONAL BODIES

UK Several organisations are involved in setting and maintaining clinical standards for doctors working in the
NHS. These include:
• Postgraduate Medical and Training Board (PMETB)
• National Clinical Assessment Service (NCAS)
• Medical Royal Colleges
• General Medical Council (GMC).
As well as the GMC, there are eight other major professional regulatory bodies for health care in the UK. These
include the General Dental Council (GDC), the Nursing and Midwifery Council (NMC) and the Royal Pharmaceutical
Society of Great Britain. Each of these bodies:
• Maintains registers of accredited professionals
• States how professional competence should be maintained
• Holds hearings when serious professional misconduct is alleged.
In the future they will require ongoing evidence of competence to practise.
In the UK, public health specialists may apply to join the UK Public Health Register ([Link].
uk). This is analogous to the public health specialist registers of the GMC and GDC, and its aim is to ensure high
standards of public health practice so as to protect the public.
NZ The Medical Council of New Zealand requires that all registered medical practitioners participate in approved
continuing professional development activities. For most practitioners this takes the form of an approved
re-certification programme specific to their field of practice. Other doctors (particularly those in training positions)
are required to have a formal collegial relationship with a nominated peer working in their field. Under the Health
Practitioners Competency Assurance Act, this system of re-certification is being extended to other groups of health
practitioners.
Accreditation of health-care agencies remains voluntary in New Zealand. A standardised system is provided by
Quality Health New Zealand.

CHANGES IN PROFESSIONAL ACCOUNTABILITY

Changes in professional accountability have arisen for a number of reasons:
• Some serious and high-profile service failures have prompted suggestions that current systems for accountability
were insufficient
• Recognition that the drive towards containing costs and reducing waiting times in health services in the early
1990s may have been at the expense of clinical quality
• Changes in the autonomy afforded to professionals in many fields
• Introduction of business practices into health, e.g. corporate governance.
Eng In England, recommendations to modernise professional regulation followed the Fifth Report of the Shipman
Inquiry. This was a large-scale investigation into the reasons why Dr Harold Shipman, a GP who worked near
Manchester, was able to murder an estimated 250 of his patients without detection over the course of several years.
The recommendations from this enquiry are legion, but in summary they are designed to:
• Align systems of organisation-wide and professional regulation to ensure patient safety
• Strengthen systems of re-validation and fitness to practise
• Support the safe expansion in roles of professionals in regular contact with patients.

465

Public [Link] 465 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

Following publication of this report, the Chief Medical Officer for England published a set of recommendations
entitled Good Doctors, Safer Patients, aimed at improving systems of medical regulation (see [Link]/
assetRoot/04/13/72/78/[Link]).

APPRAISAL
Appraisal is a non-threatening, confidential dialogue that occurs between a manager and an employee, aimed at
exploring and deciding:
• Progress in attaining objectives that were previously agreed
• Setting new objectives
• Identifying developmental needs.
Appraisal enables employees to achieve more from their work and to develop specific competencies. Its use is now
widespread within the public and private sectors.

CLINICAL GOVERNANCE
Eng In the NHS, the profile of professional standards was raised in 1998 with the introduction of clinical
governance. The government’s consultation document A First Class Service: Quality in the new NHS (1998), defined
clinical governance as:
‘A framework through which NHS organisations are accountable
for continually improving the quality of their services and
safeguarding high standards of care by creating an environment
in which excellence in clinical care will flourish.’
Patient–professional partnership
The Health Act 1999 enshrines clinical governance as a statutory
duty for all NHS organisations. Its introduction required a change
in the culture of accountability, away from considering clinical
standards the responsibility of single professional groups such as Seven pillars of clinical governance
doctors, towards regarding it as the responsibility of the entire
organisation, from the cleaners to the chief executive. 1. Clinical effectiveness
The Department of Health clinical governance support team 2. Risk management effectiveness
outlines seven ‘pillars’ of clinical governance, which are 3. Patient experience
underpinned by five ‘foundation stones’ of practice. Capping the 4. Communication effectiveness
framework is the partnership between patients and professionals.
5. Resource effectiveness
See Figure 5A.11.1.
6. Strategic effectiveness
Clinical governance is based on the principle that good systems 7. Learning effectiveness
of clinical care improve patient outcomes. Clinical governance
involves demonstrating that these systems and practices are
in place and are working well. For example, in terms of risk
management, clinical governance would require a demonstration
that an incident-reporting system (see Figure 5A.11.2) was Five foundation stones of practice
working effectively.
1. Systems awareness
2. Teamwork
3. Communication
Figure 5A.11.1 Seven pillars of clinical governance model. 4. Ownership
Adapted and reproduced with permission from NHS Clinical 5. Leadership
Governance Support Team (1999), see [Link]/
downloads/Seven_Pillars.pdf
466

Public [Link] 466 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

Adverse incident Record on Report incidents

or incident form centrally
near miss?

Any systemic risks Audit incident

identified? forms

Reduce risk through

• Training
Reduced no. of
• Information Re-audit
adverse incidents/ Figure 5A.11.2
• Guidance records
near misses? Demonstration of an
• Equipment redesign effective incident reporting
• Process redesign system of risk management
for clinical governance

5A.12 CHANGING BEHAvIOUR

Behaviour change in individuals and organisations
Binney and Williams (1995) describe ‘types’ of attitude that people may have towards change: see Box 5A.12.1.

Box 5A.12.1
Missionaries Pleased to embrace change – they adopt it, adapt rapidly and actively encourage others
to do so
Believers Understand the merits of the changes, believe in them – but are a little more cautious,
since they can see both benefits and the risks of the changes
People who pay lip Acknowledge that change is probably necessary but are typically not active in supporting
service or adopting it
Hiders and refugees Ignore or try to hide from the change – often through fear or lack of interest
Members of the Actively try to block the changes
underground
resistance
Honest opponents Declare their resistance – they openly challenge the need for change
Emigrants Simply leave, wanting nothing to do with the changes, preferring to seek their
employment elsewhere

ADOPTING CHANGE
Rogers (1995) developed the ‘diffusion’ model to explain how people generally move towards change: see Box 5.12.2
and Figure 5A.12.1.

467

Public [Link] 467 3/6/08 [Link]

 5A I nd i v idua l s, Teams and G roups

Box 5A.12.2
Innovators First to embrace change
Early adopters Part of the first sizeable ‘wave’ of people who take up change, innovation – many of them
becoming committed disciples of the change and innovation in question in due course
Early majority Typically they have ‘watched and waited’ before either seeing the benefits and/or getting
the confidence to take up the change themselves
Late majority Follow in due course – they are less change oriented, slower to respond, need more
convincing
Laggards Those who really do not show an interest, do not want to ‘get involved’

Innovators
Early Early Late
adopters majority majority Laggards
2.5% 13.5% 34% 34% 16%

Figure 5A.12.1 Adoption curve. Reproduced from Rogers (1995)

ALTERING BEHAVIOURS
Theories from the behavioural and social sciences provide a platform for understanding why people engage in
health-endangering behaviour or health-protective behaviour. The application of these theories can explain and
predict behaviour, and can guide the design and implementation of health promotional activity. A summary is given
in Table 5A.12.1 but see Section 2H.4 for more detail.

468

Public [Link] 468 3/6/08 [Link]

 5A I ndividuals, Teams and Groups

Table 5A.12.1 Summary of theories for altering behaviour related to health

Theory Focus Key concepts
Stages of change Individual’s readiness to change or Pre-contemplation
model attempt to change towards healthy
Contemplation
behaviours
Decision/determination
Action
Maintenance
Health belief model Person’s perception of the threat of a Perceived susceptibility
health problem and the appraisal of
Individual level

Perceived severity
recommended behaviour(s) for preventing
or managing the problem Perceived benefits of action
Cues to action
Self-efficacy
Social learning Behaviour is explained via a three-way, Behaviour capability
theory dynamic reciprocal theory in which
Reciprocal determinism
personal factors, environmental influences
and behaviour continually interact Expectations
Self-efficacy
Observational learning
Reinforcement
Organisational Concerns processes and strategies for Problem definition (awareness stage)
change theory increasing the chances that healthy
Initiation of action (adoption stage)
policies and programmes will be adopted
and maintained in formal organisations Implementation of change
Organisational level

Institutionalisation of change
Diffusion of Addresses how new ideas, products and See above (Adopting change, p 467)
innovations theory social practices spread within a society or
Characteristics that determine rate of
from one society to another
change include:
• Relative advantage
• Compatibility
• Complexity
• Trialability
• Observability

469

Public [Link] 469 3/6/08 [Link]

Public [Link] 470 3/6/08 [Link]
 5B Understanding Organ isations, Their Function and Structure

5B
understanding Organisations, Their Function
and Structure

5B.1 Organisational environments 471 5B.4 Social networks and communities of

5B.2 Stakeholder interests 471 interest 474
5B.3 Interorganisational relationships 473 5B.5 Influences on organisations 475

In public health practice there are many opportunities to learn how different organisations work. This could be
through positive experiences such as effective liaison with stakeholders. Equally, it can come from clashes of
cultures and structures, such as those that occur following organisational mergers: these too can provide valuable
insights into organisations.
This chapter provides the tools for practitioners to develop systematic approaches to understanding organisations.
These will be of use when forming partnerships, managing organisational change and adapting to a changing
environment.

5B.1 ORGANISATIONAL ENvIRONMENTS

Understanding the internal and external organisational environments – evaluating internal resources and organisational
capabilities
See Section 5C.2.

5B.2 STAKEHOLDER INTERESTS

Identifying and managing internal and external stakeholder interests
A stakeholder is a person or group that has an interest (a ‘stake’) in the outcomes of a project or organisation. This
interest may stem from:
• Professional interest
• Personal reasons
• Democratic representation
• Commitment to achieving a particular outcome.

471

Public [Link] 471 3/6/08 [Link]

 5B Unde r s t a nd ing O rganisations, Their Function and Structure

Consulting stakeholders at the preliminary stages of

project planning tends to garner their support, and the HIGH
input from stakeholders invariably improves the quality
of projects. The support of stakeholders can be helpful
in securing additional resources, which in turn increases Keep Manage
the likelihood that the project will succeed. satisfied closely

STAKEHOLDER ANALYSIS

Power
The process of stakeholder analysis involves identifying
a comprehensive list of stakeholders (both internal Monitor
Keep
(minimum
and external), and plotting their degree of interest and informed
effort)
relative power: see Figure 5B.2.1.
The reaction of the stakeholders to the project may then
be anticipated according to their perceived motivations
from any ideological, strategic or financial interests in LOW Interest HIGH
the project. If the reaction of a particular stakeholder
is not likely to be positive, consideration can then be
given to what alterations might win their support. Where Figure 5B.2.1 Stakeholder analysis. Reproduced
winning such support is unrealistic, consideration can from: Thompson R, [Link]. Available online at:
be given to how their opposition may best be managed, [Link]/pages/article/newPPM_07.htm
together with a contemplation of the powers of leverage
that they may potentially exert.

LEVERS OPERATED BY STAKEHOLDERS

The powers held by stakeholders differ between internal and external actors.

INTERNAL STAKEHOLDERS
Internal stakeholders, such as managers and individual employees, have their own interests that they will tend
to pursue. For example, middle managers might seek promotion and employees may seek more favourable working
conditions. All internal stakeholders possess a degree of negative power that they may be driven to use in order to
impede the implementation of a particular strategy. These powers include threats of:
• Resignation
• Industrial action
• Refusal to relocate to another team or site.

ExTERNAL STAKEHOLDERS
Central government can exert influence on organisations through taxation, government spending, legal action,
regulation and threatened changes in the law.
Community and pressure groups can exert influence by:
• Publicising activities that they regard as unacceptable
• Campaigning for changes in the law
• Refusing to cooperate with the organisation
• Conducting illegal actions such as sabotage.

472

Public [Link] 472 3/6/08 [Link]

 5B Understanding Organ isations, Their Function and Structure

5B.3 INTERORGANISATIONAL RELATIONSHIPS

Structuring and managing interorganisational (network) relationships, including intersectoral work, collaborative
working practices and partnerships
There is growing academic and managerial interest in the field of interorganisational relationships (such as strategic
alliances, joint ventures and social networks). This is because of their potential to benefit organisations at both
the micro and the macro levels: see Box 5B.3.1.

Box 5B.3.1
Micro Individual employees benefit from discussing professional practices with people from related fields. This
additional insight may enable them to perform better
Macro Possibility for innovation and efficiency gains

TYPES OF INTERORGANISATIONAL RELATIONSHIPS

Barringer and Harrison (2000) classify the principal types of interorganisational relationships as shown in Box
5B.3.2.

Box 5B.3.2
Joint venture Two or more organisations pool a portion of their resources to form a separately owned
venture. Advantages include economies of scale and of scope, as well as opportunities to
innovate and to launch projects more rapidly than would otherwise be possible
Networks These are collections of organisations that organise joint projects by means of informal
arrangements rather than legally binding contracts. Such networks often operate in a hub-
and-spoke configuration where a central organisation coordinates the others, with each
organisation focusing on its particular specialty
Consortia Here, organisations with a common need come together to form a new entity that satisfies
that need on their behalf. For example, a group of neighbouring health authorities might
form a human resources consortium because it would be more costly for the individual
health authorities to operate their own
Alliances This is an arrangement, often informal and short term, between two or more organisations
that establishes an exchange relationship but no new entity is formed
Interlocking Here an executive director of one organisation sits on the board of another organisation.
directorates Such arrangements can help spread innovation and cooperation among organisations

Eng An example of interlocking directorates may be found in current government policy that encourages primary
care trusts and local authorities to make joint appointments of Directors of Public Health. These appointments
facilitate interorganisational work aimed at improving health and wellbeing, and are supported by the following
initiatives:
• Statutory requirements placed upon local authorities to promote wellbeing
• Shared national targets
• Shared Choosing Health agenda.
An example of the need for interagency collaboration is shown in Box 5B.3.3.

473

Public [Link] 473 3/6/08 [Link]

 5B Unde r s t a nd ing O rganisations, Their Function and Structure

Box 5B.3.3
Example: WHO Global Strategy on Diet, Physical Activity and Health (2004)
In an effort to reduce the global burden of non-communicable disease, the WHO released a strategy in May 2004
aimed at improving diet and promoting physical activity.
There was widespread recognition that the effort to promote physical activity at a population level required
interagency collaboration and partnership working. Health promotion would succeed only by working closely with
non-health sector agencies such as departments of transport, urban planning, education and sport.
Reproduced from World Health Organization (2006) Global Strategy on Diet, Physical Activity and Health, see
[Link]/dietphysicalactivity/en/.

5B.4 SOCIAL NETWORKS AND COMMUNITIES OF INTEREST

See also Section 4A.10.

SOCIAL NETWORKS
A social network is a group of people or organisations that are connected through social bonds. Relationships are
viewed in terms of:
• Nodes (individual actors such as family members, neighbours, friends and colleagues)
• Ties (ranging from casual acquaintances to close familial ties).
In its simplest form, a social network is a map of all of the relevant ties between the nodes. These concepts may be
displayed as a social network diagram, in which nodes are marked as the points and the ties as lines. The network
can also be used to determine the social capital of individual actors (see Section 4A.10).

DUNBAR’S NUMBER
The maximum size of social networks is consistently found to be around 150 people (Dunbar 1992). This is known
as Dunbar’s number and it represents the maximum number of individuals with whom any one person can maintain
stable relationships. The number was calculated in 1992 using a regression equation based on data from 38 primate
genera, and it is thought to be determined by characteristics of the neocortex. Dunbar’s number has subsequently
been researched in the contexts of anthropology, sociology, statistics and organisational management.

ONLINE SOCIAL NETWORKS

Social networking also refers to a category of internet applications to help connect friends, business partners
or other individuals together using a variety of tools. These applications, known as online social networks, are
growing rapidly.

COMMUNITIES OF INTEREST
Communities of interest are groups of people with a common concern who may not necessarily be linked in terms
of their location, profession, socioeconomic status or other characteristics. Members of a community are often
scattered across a country or across the globe, and they come from many walks of life. Communities of interest have
flourished with improved access to the internet. They offer members the opportunity to engage in discourse and
critical thinking about topics from the perspective of their common interest.

474

Public [Link] 474 3/6/08 [Link]

 5B Understanding Organ isations, Their Function and Structure

ExAMPLES OF COMMUNITIES OF INTEREST

• Older people experiencing isolation and poverty: older people’s partnership
• People with learning disabilities: learning disability partnership
• Disabled people
• Asylum seekers and refugees
• Ex-offenders
• Homeless people.

5B.5 INFLUENCES ON ORGANISATIONS

Assessing the impact of political, economic, sociocultural, environmental and other external influences
See Section 5C.2.

475

Public [Link] 475 3/6/08 [Link]

Public [Link] 476 3/6/08 [Link]
 5C Management and Ch ange

5C
Management and Change

5C.1 Management models and theories 477 5C.3 Performance managment 482
5C.2 Frameworks for managing change 479

Technological advancements, organisational restructuring and evolving populations mean that change is a constant
feature of health-care systems. Public health practitioners have an important role in managing change in health-care
systems to ensure that they continue to meet organisational goals and objectives. This chapter provides some key
models of leadership and frameworks for managing change. These models can act as useful tools to evaluate how
change has been managed in the past but can also be applied to addressing current situations.

5C.1 MANAGEMENT MODELS AND THEORIES

Understand the basic management models and theories associated with motivation and leadership and be able to apply
them to practical situations and problems
Motivation can be regarded as a measure of a person’s drive to initiate and persist in a given behaviour. Employees
of an organisation may be highly able to perform a task, but if they are unmotivated then they will not fully
dedicate their abilities to the job in hand.
In a managerial context, leadership is seen as the shaping of goals and development of ideas. It involves a
connection with employees at an emotional level.

MOTIVATION
All organisations aim to promote motivation. This is fundamental to the role of managers in the workplace, namely
to get things done through employees. There are several different theories of motivation in the workplace. Two
major management theorists – Maslow and Hertzberg – have produced seminal theories to describe and predict
what motivates individuals in their place of work.

477

Public [Link] 477 3/6/08 [Link]

 5C Ma na ge me nt and Change

MASLOW’S HIERARCHY THEORY

The hierarchy of human needs is among the most widely employed managerial theories. It states that:
• Only unsatisfied needs influence human behaviour
• Needs are ordered according to importance and complexity
• In the long run, people will be motivated by higher-level needs only once their lower-level needs have been
satisfied (this is called pre-potency)
• The higher the person is up the hierarchy, the more individuality, humanity and psychological health they express.
See Box 5C.1.1.

Box 5C.1.1
Hierarchy Description Work context
Self- Instinctive human need to make the Promotion, opportunities for creativity/
actualisation most of one’s unique abilities* innovation
Self-esteem Subjective appraisal of a person as Job title, reviews, appraisals
intrinsically positive or negative
Love and Affection and happiness in the Professional associations, social events,
belonging workplace supportive manager
Safety Absence of danger Company pension, substantive contract
Physiology Basic needs such as warmth and shelter Pay
*Definition of ‘self-actualisation’ is controversial.

At the higher levels of the hierarchy, respect and recognition become much more powerful motivators than financial
reward.
Advantages and disadvantages of the hierarchy theory are listed in Box 5C.1.2.

Box 5C.1.2
Advantages Disadvantages
Identifies individuals who fail to progress beyond the Overly individualistic
lower levels of the hierarchy
No allowance for altruism
Highlights how basic problems (e.g. workplace
temperature) can inhibit motivation
Makes intuitive sense

Although Maslow stresses that not everyone experiences needs in this order, the very concept of an order of needs
is disputed. For example, the need for warmth and shelter in a homeless person does not preclude them from
simultaneously having strong needs for love and belonging.

HERZBERG’S TWO-FACTOR THEORY

This theory (also known as the motivator–hygiene theory) contends that certain workplace factors lead to
job satisfaction, while others cause dissatisfaction. Factors are divided into motivators (which give positive
satisfaction) and hygiene factors (which do not give positive satisfaction but the absence of which causes
dissatisfaction): see Box 5C.1.3. Managers should aim to maximise both groups.

478

Public [Link] 478 3/6/08 [Link]

 5C Management and Ch ange

Box 5C.1.3
Motivators Hygiene
Varied work Good pay
Responsibility Good working conditions
Recognition Job security

LEADERSHIP
Leadership theory has been a topic of study throughout human history, and there are over 100 definitions of
leadership. Management and leadership are sometimes used interchangeably, but Mullins (2005, p 283)
distinguishes leadership from management in the following way:
‘Management may arguably be viewed more in terms of planning, organising, directing and controlling the activities
of subordinate staff. Leadership, however, is concerned more with attention to communicating with, motivating,
encouraging and involving people’.
In line with this definition, contemporary studies of leadership focus on change management and on empowering
others. See Box 5C.1.4.

Box 5C.1.4
Participative These models (by Likert and others) argue that participative styles of leadership lead to
theories increased job satisfaction and improved performance. An example is management by walking
around (MBWA) (see Section 5A.5)
Contingency These theories all argue that the most effective leadership style depends on the context. For
theories example, the managerial grid described by Blake can be used to determine whether a boss-
centred or a subordinate-centred approach will work better
Instrumental These contend that the leader’s behaviour patterns (e.g. participation, delegation) can
theories promote effective performance from others
Charismatic These include inspirational and transformational leadership styles. By enthusing others using
theories values and vision, the leader raises confidence in others. Charismatic leaders include those
who are not appointed to authority but assume leadership in other ways
vMC model In this model, leaders are seen as possessing the following qualities: vision, Management
skills and Commitment. The three qualities are required in different proportions depending
on the task at hand (e.g. vision is relatively unimportant in accountancy)

5C.2 FRAMEWORKS FOR MANAGING CHANGE

Critical evaluation of a range of principal frameworks for managing change
The main frameworks that are applied to change management are summarised in Table 5C.2.1.

Table 5C.2.1 Frameworks for managing change

Stage of change management Analysis framework
Current situation 7S analysis and PESTELI analysis
Reason for change SWOT analysis
Implementing change Organisational development and action research

479

Public [Link] 479 3/6/08 [Link]

 5C Ma na ge me nt and Change

Figure 5C.2.1 demonstrates how the PESTELI and the

McKinsey 7S
McKinsey 7S tools can be used to support the central
SWOT analysis, i.e. PESTELI and 7S can be used to Strategy Staff Shared values
ensure that each of the four components of SWOT Structure Style
(Strengths, Weaknesses, Opportunities and Threats) is Systems Skills
considered systematically and in sufficient depth.

McKINSEY 7S
This analysis identifies the strengths and weaknesses
of the organisation, by considering the links that exist S W
between the seven factors listed in Box 5C.2.1, each O T
of which begins with a letter S: strategy, structure,
systems, staff, style, shared values and skills. It is
useful for assessing internal factors that influence
performance.
PESTELI
Box 5C.2.1
Political
Strategy Action leading to allocation of
resources Economic

Structure Hierarchy and interconnections Sociological

Systems Procedures, processes including Technological

information flows Ecological
Staff Personnel Legislative
Style Style of key managers and how goals Industry
are achieved
Shared Guiding concepts shared by Figure 5C.2.1 Combining McKinsey 7S and PESTELI
values organisation’s members for SWOT analysis
Skills Capabilities of key personnel and
organisation as whole

There is conflicting evidence regarding the usefulness of the 7S model: its strengths and weaknesses are listed in
Box 5C.2.2.

Box 5C.2.2
Strengths Weaknesses
Survey of US companies in the 1970s using the 7S A repeat survey 5 years later showed that two-thirds
framework showed that the top 62 companies shared of the previously top companies were no longer at the
common characteristics peak
Expert opinion generally regards the 7S approach as Ignores discord and rebellion within an organisation,
being useful which also shape the organisational culture
Dual emphasis on ‘soft’ and ‘hard’ factors Tendency to focus on similarities between Ss and not
conflicts

480

Public [Link] 480 3/6/08 [Link]

 5C Management and Ch ange

In health care, the 7S schema is often used to assess how changing one ‘S’ can impact on another

SWOT ANALYSIS
A SWOT analysis considers both internal and external factors, listing the strengths, weaknesses, opportunities and
threats facing the organisation. These should be made with reference to the medium-term requirements of the
environment in which the organisation operates. Strengths and weaknesses of SWOT analysis are listed in Box
5C.2.3.

Box 5C.2.3
Strengths Weaknesses
Most widely used tool by UK businesses Review of its use found that it generated lists of factors that
were:
Considers both internal and external factors
• Too long
• Too general
• Often meaningless
Rare that the output of the analysis is actually used
Tendency to focus on the process and not the outcome

PESTELI ANALYSIS
This tool (see Box 5C.2.4) is used to identify potential opportunities and threats. The acronym was initially PEST
(political, economic, sociological and technological). Ecological, legislative and industry factors were added later to
give a more complete description of the environment in which the organisation operates. It is useful for identifying
factors that may help or impede progress but is often performed as a stand-alone activity.

Box 5C.2.4
Political factors Affecting performance and options
Economic influences Financial resources available and market factors
Sociological trends Demographic changes, attitudes and beliefs
Technological innovation Equipment, methods and approaches
Ecological factors How the organisation interacts with the wider environment
Legislative requirements Relevant laws that affect the organisation
Industry analysis Attractiveness of the industry

ORGANISATIONAL DEVELOPMENT
In this approach it is the alteration of employees’ on-the-job behaviour that is seen as key to implementing wider
organisational change. Interventions are used to prompt desired behaviours; these can be targeted at the individual,
group or organisational level. They include changes in:
• Technology
• Physical setting
• Goals and targets.

481

Public [Link] 481 3/6/08 [Link]

 5C Ma na ge me nt and Change

Strengths and weaknesses of the organisational development approach are listed in Box 5C.2.5.

Box 5C.2.5
Strengths Weaknesses
Several reviews and meta-analyses showing One review showed that a positive outcome was seen in 38% of
positive outcomes cases, a negative outcome in 10% of cases, but no change in
over 50% of cases.

5C.3 PERFORMANCE MANAGEMENT

An understanding of the issues underpinning the design and implementation of performance management against goals
and objectives
Performance management is a form of holistic people management, which should be:
• Integrated (linking people management, and individuals and teams)
• Strategic (addressing broad issues and setting long-term goals).
For performance management to be successful, organisations need to develop a culture in which employees and
teams take responsibility for their own contributions. The key steps involved are summarised in Box 5C.3.1.

Box 5C.3.1
Goal setting Managers define their expectations and set strategic departmental and
organisational goals
Agreement of a developmental plan Plans agreed between managers and employees or teams
Continuous monitoring Contemporaneous feedback and formal reviews

DESIGN
Like clinical audit, performance management is a cyclical process rather than an isolated event.

OBJECTIvES AND PERFORMANCE STANDARDS

Objectives or goals are expressed as:
• Targets to be met (e.g. achieving financial balance by the end of the financial year)
• Tasks to be completed by specified dates (e.g. by January 2004, 98% of patients are to spend under 4 hours in
the emergency department)
• Ongoing targets, called performance standards (e.g. maintaining the intraoperative wound rate below a certain
value).
Objectives should be SMART, i.e.
• Specific
• Measurable
• Achievable
• Realistic
• Timescale set for completion.

482

Public [Link] 482 3/6/08 [Link]

 5C Management and Ch ange

TEAMS
Team performance can be managed in connection with activity levels, outputs, patient satisfaction and financial
results. Teams should agree their objectives and receive feedback in the same way as if they were individuals. Team
members can contribute towards team evaluation through peer review.

360-DEGREE FEEDBACK
The concept of 360-degree feedback became popular in the 1990s. It involves collecting anonymous opinions and
data from a number of sources about a person. Sources include:
• Individual’s line manager
• Employees whom the individual line manages
• Peers (internal and external to the organisation)
• Self-assessment.
A 360-degree comment has the potential to provide a more rounded commentary that is less prone to bias than an
assessment conducted by one individual.

IMPLEMENTATION
Performance management is difficult to implement and requires engagement by all members of an organisation –
particularly line managers. The evidence suggests that performance management is well regarded by employees and
managers alike, especially insofar as it emphasises personal development. However, performance-rating schemes
often involve considerable amounts of red tape, which can be very time-consuming. Implementation of performance
management requires:
• Training (especially initially) of managers and employees
• Clarification of the definition of ‘performance’
• Understanding the organisation’s performance culture
• Stressing the personal benefits to individual employees of participation in the process
• Remembering that performance management is a tool for line managers whose success depends on their ability
to use it effectively.

483

Public [Link] 483 3/6/08 [Link]

Public [Link] 484 3/6/08 [Link]
 5D Understanding the Theory and Process of Strategy Development

5D
understanding the Theory and Process of
Strategy Development

5D.1 Developing health-care strategy 485 5D.6 Guideline development 496

5D.2 Theories of strategic planning 487 5D.7 Integrated care pathways 497
5D.3 Health service development and planning 489 5D.8 Consultation about health services 498
5D.4 Health service funding 489 5D.9 Historical development of personal health
5D.5 Risk management 494 services and of public health 498

There is a continued drive, both within health systems and externally, to improve, standardise and quality assure
health care. A number of subjects covered in this chapter seek to achieve this aim, from strategy implementation
to risk management and guideline development.
The organisation and funding of health systems are instrumental to their development. A range of diverse
approaches to health system funding and organisation across different countries is summarised and compared.

5D.1 DEvELOPING HEALTH-CARE STRATEGY

Strategy communication and strategy implementation in relation to health care
Health-care strategies are medium- or long-term action plans that are designed to improve health or to focus
on corporate priorities (such as containing health-care costs, reducing waiting lists or improving recruitment
and retention of staff). In order to have an impact, a strategy needs to be communicated effectively to local
stakeholders and decision-makers.

COMMUNICATION
In strategic communication, the following should be considered:
• Audience
• Strategic objectives
• Key message
• Media for conveying the message: often a range of communication channels will be employed, including
press releases, annual reports and direct communication to stakeholders. See also Section 5A.8 for effective
communication methods.

485

Public [Link] 485 3/6/08 [Link]

 5D Unde r s t a nd ing the Theory and Process of Strategy Development

Strategy communication is more than presenting a ‘finished product’. It is not enough for an individual or team to
write a strategy document and to send it to stakeholders and decision-makers. Communication mechanisms should
ensure that people responsible for delivery or affected by implementation of the strategy should be appropriately
involved in:
• Agreeing strategic objectives
• Producing a baseline assessment of the current state of health-care provision
• Committing resources
• Monitoring implementation and delivery.
Communication will depend on the stage of strategic development and the audience being targeted. It typically
includes the elements described in Table 5D.1.1.

Table 5D.1.1 Communication involved in strategy development

Stage of development Audience and agencies Communication methods and media
Strategy formation Commissioners Working groups
Partner organisations Surveys
Experts and specialists Focus groups
Practitioners Meetings
Service users
Dissemination All stakeholders Report publication (hard copy and on
websites)
Newsletters
Media releases
Public events
Monitoring implementation Commissioners Executive boards
Scrutiny and review boards Local strategic partnerships
Public Public reports

IMPLEMENTATION
Approaches to strategy implementation include all-out attack, and inside-venture and strategic alliances: see Box
5D.1.1.

Box 5D.1.1
All-out attack In this approach (Kono 1984) all current strategic plans are abandoned and replaced with
a new strategy. This typically occurs when a large organisation acquires, or is merged with,
another
Inside-venture This tactic (Zahra 1991) relies on the power of internal rewards and innovation. An
approach organisation encourages employees to be innovative and, where their ideas are feasible,
employees are rewarded with their own project team to develop the innovation
Strategic These include efforts such as joint ventures with other bodies: partnerships where two or
alliances more organisations pursue a set of agreed goals while remaining independent

486

Public [Link] 486 3/6/08 [Link]

 5D Understanding the Theory and Process of Strategy Development

Box 5D.1.2 describes the steps employed in implementing a national health strategy in England, the National
Service Framework for Coronary Heart Disease.

Eng Box 5D.1.2

Example: National Service Framework (NSF) for Coronary Heart Disease
The Department of Health published a 10-year national strategy for heart disease in 2000.
The stated goals were linked to the wider strategies, Our Healthier Nation and the NHS Plan, and concerned
reducing inequalities and improving outcomes for people with heart disease or at risk of heart disease through
12 standards. These were based on a baseline assessment of heart disease provision in England and reviews of
available evidence.
An expert external reference group developed the strategy involving clinicians, managers, epidemiologists and
patient representatives.
Dissemination and awareness-raising used a range of means including:
• At publication: press launch and publication of the report on the internet and in the health service circular
to all NHS organisations and local authorities
• After publication: reports on progress and guides for implementation
The strategy’s implementation was promoted through:
• Milestones with dates that were included in the strategy to indicate how NHS organisations were expected to
progress towards meeting the standards
• Additional funding that was announced to support delivery of the strategy
• The standards were linked to national performance indicators, e.g. stop-smoking targets were a key part of
the NHS star rating system for primary care trusts
• Parts of the NSF have been subject to regular review through Healthcare Commission/Audit Commission
assessments
Reproduced from the Department of Health, National Service Framework for Coronary Heart Disease (2000),
available online at: [Link]/PolicyAndGuidance/HealthAndSocialCareTopics/CoronaryHeartDisease/fs/en.

5D.2 THEORIES OF STRATEGIC PLANNING

Strategic planning is an activity through which an organisation confronts the major decisions that it faces. A
decision is not deemed strategic merely by being important. Rather, strategic decisions or issues must fulfil the
following criteria:
• Define the institution’s relationship to its environment
• Generally take the whole organisation as the unit of analysis
• Depend on inputs from a number of functional areas
• Provide direction for, and constraints on, administrative and operational activities throughout the institution.
Theories of strategic planning include the three-step models and the ranked order approach.

THREE-STEP MODELS
Two models of strategy development in common usage are STP and DST: see Boxes 5D.2.1 and 5D.2.2. As can be
seen, these models use similar steps in a different order.

487

Public [Link] 487 3/6/08 [Link]

 5D Unde r s t a nd ing the Theory and Process of Strategy Development

Box 5D.2.1
STP process
Situation Describe the current situation and the factors that caused it to be thus
Target Outline the ideal state of affairs and define specific goals and objectives that describe this state
Path Map a possible route to the goals and objectives of the ideal state

Box 5D.2.2
DST process
Draw Describe the ideal state of affairs
See Describe the current situation and the gap between the ideal state and the current state
Think Outline the specific actions needed to close the gap between the current state and the ideal state

RANKED ORDER APPROACH

This approach involves defining the strategy in terms of a priority list. The following components of the strategy are
first defined:
• Policies
• Plans
• Actions
• Goals
• Objectives
• Ideal state
• Strategies
• Tactics.
These items are then ranked into a hierarchy, such that:
• The item in a lower rank explains how the item immediately above it will be achieved
• The item in a higher rank addresses why the item immediately below it needs to be achieved.
In this way the top rank objective (TRO) does not address ‘why’, and therefore by definition this is the crux of the
strategy upon which all else depends.

ACTION PLANNING
Once a strategy has been developed, the next stage in its implementation is the writing of an action plan. This must
cover the following items:
• Name of the strategy
• Actions (tangible components of what will be done)
• Location
• Personnel (including the line-management structure of the programme)
• Timing (start and completion dates)
• Resources (staff, supplies, information, other resources)
• Audit (progress measurement and reporting)
• Benefits to the public of implementing this strategy
• Performance rewards (if any)
• Contingency plans.

488

Public [Link] 488 3/6/08 [Link]

 5D Understanding the Theory and Process of Strategy Development

5D.3 HEALTH SERvICE DEvELOPMENT AND PLANNING

The aims of health service improvement are to:
• Enhance the quality of patient care
• Improve strategic outcomes
• Contribute to improved public health.

HEALTH SERVICE DEVELOPMENT

Continual health care reform is necessary to ensure that services are effective, evidence based and appropriate for
local needs. Drivers of change are outlined in Table 5D.3.1.

Table 5D.3.1 Drivers of change in health services

Driver Example
Advances in technology New drugs, equipment, procedures, screening programmes and health care
settings
New information flows Production of cost per item: can be used to inform clinicians of the financial
impact of their decisions
National policy pronouncements In the UK NICE makes assessments of both effectiveness and value-for-money
Demand management Referral centres that act as a second gatekeeper to secondary care
Public perception Rationing debate, particularly with regard to novel high-cost drugs

HEALTH SERVICE PLANNING

The degree to which health-care provision is planned varies between countries. However, even in the USA (which
is generally regarded as an unregulated health market) there is an element of planning for large costly facilities, so
that wasteful over-provision is avoided.
Eng In England, regional strategic health authorities are responsible for detailed planning. Their areas of
responsibility are summarised in Table 5D.3.2.

Table 5D.3.2 Regional health authority responsibilities

Planning Developing plans for improving health services in their area
Regulating Assuring the quality and performance of local health services
Capacity building Developing health services to provide adequate services for the population’s needs
Priority setting Ensuring that national priorities are included in local health service

5D.4 HEALTH SERvICE FUNDING

Methods of organising and funding health services and their relative merits, focusing particularly on international
comparisons and their history
The funding and organisation of health care should be arranged so as to maximise efficiency and equity: see Box
5D.4.1.

489

Public [Link] 489 3/6/08 [Link]

 5D Unde r s t a nd ing the Theory and Process of Strategy Development

Box 5D.4.1
Efficiency Equity
Allocative efficiency (benefits exceed the costs) Financial equity (financial burden faced is proportional
to ability to pay)
Operational efficiency (scarce resources used to their
best advantage) Equity of opportunity (if not equality of access, equality
of utilisation)
Reducing health inequalities

ORGANISATION OF HEALTH SERVICES

A health organisation can be described in terms of its commissioner, its setting, the way in which it is accessed,
as well as its financial flows.

COMMISSIONER
Eng The purchaser of health care determines which services are made available to patients. In England, the
commissioner has traditionally been the local health authority or board, acting on behalf of the government. The
level at which commissioning occurs is now being diversified, ranging from GPs who can choose to hold budgets for
the provision of certain services, through groups of GPs, local health trusts, and up to supra-regional commissioners
(who commission super-specialised services on behalf of a group of health trusts).
Wal In Wales, the commissioners of hospital, community and primary care services are local health boards.
Internationally, commissioners include mutual societies (as in France) or managed care organisations (as in the
USA).

HEALTH-CARE SETTING
UK In the UK, primary health care is typically delivered in GP surgeries. Services based in the community are
provided by primary care trusts in England, or by combined hospital and community trusts in Wales. Hospitals
provide secondary and tertiary care in both inpatient and outpatient settings.
EU In many continental European countries, specialists work in town-centre offices rather than in outpatient
departments of hospitals. Polyclinics also exist where several specialists share premises and diagnostic services.
Novel health-care settings include telephone-based services, hospital at home, and walk-in centres in shopping
streets or transport hubs.

ACCESS TO HEALTH CARE

UK The UK, in common with some other countries, has a system of single-registration in which members of the
public are entitled to register with a single GP. Although individuals are free to change GP at any time, they may not
be registered with more than one GP concurrently. Advantages and disadvantages of single registration are listed in
Box 5D.4.2.

490

Public [Link] 490 3/6/08 [Link]

 5D Understanding the Theory and Process of Strategy Development

Box 5D.4.2
Advantages of single registration Disadvantages of single registration
Gate-keeping role avoids over-investigation and Less consumer choice for patients
reduces pressure on secondary care
Conflict of interest: GP is caring for the
Continuity of care individual patient and bearing in mind
implications for all other patients
Single person who receives all correspondence

GPs in the UK act as gatekeepers to elective secondary care. In contrast, members of the public in countries such
as France are free to access secondary care direct. This is less cost-effective for the system but is popular with the
public.

FUNDING OF HEALTH SERVICES

The five principal forms of funding health care are:
• Taxation
• Social insurance*
• Private insurance
• Out-of-pocket expenses
• Charities.
*A (more-or-less) compulsory insurance system in which employers and employees contribute to a fund. The
government pays contributions for people out of work.
Funding of health services (particularly through taxation) can be described as progressive or regressive. A tax is
said to be progressive where a higher percentage of income is paid as income rises; a regressive tax is one that
charges a lower percentage of income as income rises. Health-care systems that are funded through direct taxation
(e.g. social insurance) tend to be more progressive than those based on direct payments or private insurance.

SHORTCOMINGS OF HEALTH-CARE SYSTEMS

A number of drawbacks inherent to health-care systems are described below.

CONSUMER MORAL HAZARD

Where the consumer does not face the full cost of a health service, there is a tendency to over-use services in the
knowledge that the insurer (private or public) will foot the bill. This can lead to both over-consumption of health
care when ill and a lack of engagement with preventive health care.

PROvIDER MORAL HAZARD

This occurs where a service becomes ignorant of costs to consumers, or where the remuneration system offers
incentives to over-provide care (e.g. fee for service where the insurer always pays).

ADvERSE SELECTION
Where purchasers of health insurance are aware of their own personal risk, then those with low risk will tend not to
purchase insurance – which is designed (and priced) to cover people of average risk. The average risk level of those
remaining will rise, as will the premium to be paid, thereby exacerbating the problem. As a result, those at low risk
will be uninsured, and those at high risk will be priced out of the market.

491

Public [Link] 491 3/6/08 [Link]

 5D Unde r s t a nd ing the Theory and Process of Strategy Development

USER CHARGES
Introduction of cost sharing reduces utilization of health services by patients, but does so disproportionately among
lower-income groups. Furthermore, demand is reduced for both effective treatments and minor ailments. Where
supplier-induced demand exists, user charges may not reduce overall health expenditure. In lower-income countries
there may be no alternative to user charges for raising funds for health care.

HISTORY
UK UNITED KINGDOM
Prior to the establishment of the NHS in 1948, patients generally had to pay for their own health care. Various
charitable hospitals used to operate (e.g. the Royal Free Hospital in North London) and some local councils ran
hospitals for their population: but provision was by no means universal. This all changed on 5 July 1948 when the
Health and Housing Minister, Aneurin Bevan, founded the NHS. It was based on a cooperative that the coalminers
ran in his hometown of Tredegar, South Wales, and followed the following principles:
• Services were provided free at the point of use
• Services were financed from central taxation
• Everyone was eligible for care (including foreign visitors and temporary residents).
The original structure of the NHS was tripartite: hospital service, primary care and community services. In the
1950s, rising costs led to out-of-pocket charges for prescriptions and dental treatment. To this day, these remain the
major exceptions to the NHS being free at the point of use (although, as of 2006, the Welsh Assembly was planning
to abandon prescription charges in Wales).
Other important NHS developments are summarised in Table 5D.4.1.

Table 5D.4.1 Key NHS developments

1950s More equitable distribution of hospitals
More medical staff
Outpatient departments established
1960s More equitable distribution of GPs
Primary care teams established
Shift away from large mental hospitals
1970s English NHS re-organised: regional health authorities established
Financial pressures mount
1980s General managers appointed
Internal market established
1990s Fundholding GPs purchase care for their own patients
Hospitals become semi-autonomous trusts

INTERNATIONAL COMPARISONS
A short comparison of health systems in eight countries is presented in Table 5D.4.2.

492

Public [Link] 492 3/6/08 [Link]

 5D Understanding the Theory and Process of Strategy Development

Table 5D.4.2 International comparisons

Country Organisation Funding provision
Australia General Universal free health care to all. Complex range of Public health insurance
providers and regulators system (Medicare) tax
Primary care Self-employed GPs levy, which reimburses
approximately 80% of
Secondary Public hospitals (70% acute beds) and private hospitals health-care fees
care (which Medicare part subsidises)
Canada General Public funding but private providers. Provincial National health insurance
government and Royal Colleges regulate providers plan (Medicare) that
Primary care GPs paid on fee-for-service basis covers health services

Secondary Mostly not-for-profit private hospitals

care
France General Health care provided by private and public hospitals, and National health insurance
by private practitioners system funded by tax
and compulsory social
insurance from employers
Primary care Open access to generalists and to specialists with no and employees
gate-keeping GPs
Most citizens have
supplementary mutual
Secondary Inpatient care provided by public and private hospitals insurance funds that cover
care (profit and not for profit) cost-sharing out-of-pocket
Outpatient care mainly provided by private specialists in expenses
their own offices
Germany General Over 90% covered by statutory health insurance and General taxation and
remainder by private insurance. Self-regulating health- social insurance fund
care system
Primary care Free access to office-based doctors (generalists and
specialists) with full range of diagnostics
Secondary Outpatient care provided by office-based doctors, who act
care as gatekeepers to elective inpatient care
England General Primary care provided by GPs. Hospitals mainly publicly General taxation
owned. Government is (currently) main purchaser and
provider of health care
Primary care GPs act as gatekeepers. Through Practice-based
Commissioning they can commission specialist care for
their patients. Walk-in clinics and NHS Direct (helpline
service) act as alternative forms of primary care
Secondary Access is through emergency ambulance, referral from a
care GP or self-referral to an A&E department. Hospitals are
semi-autonomous trusts
Table contd overleaf

493

Public [Link] 493 3/6/08 [Link]

 5D Unde r s t a nd ing the Theory and Process of Strategy Development

Table 5D.4.2 contd

Country Organisation Funding Provision
USA General Health-care provision is mostly by private providers. Voluntary private care
Government funds health care through four routes: funded by individuals and
their employers. Medicare
Medicare (people aged 65+)
covers elderly, disabled
Medicaid (very-low-income people) and end-stage renal
patients. Medicaid covers
Veterans
selected categories of the
State Children’s Health Insurance Program poor. Significant numbers
Primary care Except in Certain managed care plans, family doctors with no health insurance
have no gatekeeper role
Secondary Variety of private, non-profit and public hospitals
care
NZ General Costs of medical care largely funded by the state General taxation revenue
Primary care Primary health organisations (PHOs) employ GPs and Capitated funding with
other staff co-payments for working
age adults
Laboratory services and pharmaceuticals are largely
provided at no cost to consumers Dental care on a fee-for-
service basis
Secondary Public and private hospital systems. Private sector mainly Taxation
care provides elective surgical services
National public insurer
covers costs of injury
treatment
Patient charges in private
system
Hong Kong General A mixture of public and private General taxation
Public sector is highly subsidised, providing primary
through to tertiary care services
Private sector is run as a business model, with services
provided by generalists and specialists
Primary care 72% by private practitioners, 28% by public general
outpatient clinics
Secondary 82% by public (hospital authority) hospitals, 18% by
care private hospitals

5D.5 RISK MANAGEMENT

Providing health care is a risky undertaking. There are risks of different kinds to the:
• Patient
• Practitioner
• Provider
• Commissioner.

494

Public [Link] 494 3/6/08 [Link]

 5D Understanding the Theory and Process of Strategy Development

Risk management involves identifying, monitoring and minimising these risks through a range of means. In England,
systems of clinical governance (see Section 5A.11) provide the framework for organisational risk management.

RISKS TO PATIENTS
Patients trust health-care organisations to improve their health. However, patients are harmed in around 10% of
hospital admissions. Patient safety is a particular concern because:
• There are risks associated with all types of health care
• Patients can be more vulnerable to existing hazards (e.g. many people carry MRSA in their nasopharynx, but
immunocompromised patients are at greatest risk from infection).
Risks to patients cannot be eliminated but they can be minimised by ensuring that systems are reviewed and
questioned regularly, e.g. by critical event audits and by learning from complaints. In England the Chief Medical
Officer chaired a working group to devise recommendations to reduce adverse events in the NHS, which led to the
report, An organisation with a memory: see Box 5D.5.1.

UK Box 5D.5.1
An organisation with a memory (2000)
This provided the platform for patient safety systems in the UK. It recommended systems to enable the NHS to
learn lessons from previous incidents by:
• Focusing less on human error and more on systemic factors
• Learning from risk management in industry, particularly aviation
• Reporting incidents
• Analysing trends from reported incidents
• Ensuring that lessons learned from incidents are implemented

UK The National Patient Safety Agency was established in 2001 to promote systems of learning from incidents. The
agency collates information on incidents in the NHS and shares lessons learned through:
• Agency’s National Reporting and Learning System (NRLS)
• Patient Safety Observatory in the Agency, which synthesises information from incident reports sent by the NRLS,
clinical negligence claims, data from death registrations, hospital activity and national surveys.

RISKS TO PRACTITIONERS
Important elements of quality insurance include:
• Ensuring that clinicians are immunised against infectious diseases
• Working in a safe environment (e.g. one that follows COSHH regulations)
• Keeping up to date.

RISKS TO THE ORGANISATION

Poor quality is a threat to any organisation. In addition to reducing risks to patients and practitioners, organisations
can reduce their own risks by:
• Ensuring high-quality employment practice (including locum procedures and reviews of individual and team
performance)
• Providing a safe environment (including estates and privacy)
• Ensuring adherence to safety standards and established policies.

495

Public [Link] 495 3/6/08 [Link]

 5D Unde r s t a nd ing the Theory and Process of Strategy Development

Associated organisations (such as GP cooperatives, community pharmacists and residential care homes) should be
covered by clinical governance frameworks through agreeing to comply with the standards of the organisations with
which they are associated.
UK In the UK the clinical governance arrangements are complemented by and integrated with a strong risk
management framework (including maintenance of risk registers introduced after the Turnbull committee made
a report on corporate governance in the wake of the financial collapse of Bearings Bank). This allows clinical
governance risks to feature highly in the ways that NHS organisations manage risk.

NEGLIGENCE
Negligence claims are now a feature of all health-care services. They are expensive, lengthy and undesirable for all
concerned.
Eng In England, the NHS Litigation Authority is responsible for handling all claims made against the NHS in the
country. According to the NHS Litigation Authority in 2006–2007:
• There were just under 9000 claims of negligence against NHS bodies
• Clinical negligence claims cost £579.3 million including damages to patients and the legal costs to the NHS
• On average it took just under 1½ years to deal with a clinical claim.

5D.6 GUIDELINE DEvELOPMENT

The concept of evidence-based medicine stipulates that guidelines based on scientific evidence should take
precedence over individual judgement. In order to be acceptable to clinicians, however, clinical guidelines need to
be credible and their development must be seen to have been accountable. See Box 5D.6.1.

Box 5D.6.1
Attributes of good guidelines
• Validity
• Reliability
• Clinical applicability
• Clinical flexibility
• Clarity
• Multidisciplinary process
• Scheduled review
• Good documentation
Reproduced from Field and Lohr (1990).

Instruments such as those used by SIGN (Scottish Intercollegiate Guidelines Network) and AGREE (Appraisal of
Guidelines, Research and Evaluation in Europe) are based on these founding principles.

APPRAISAL CRITERIA
The Appraisal of Guidelines for Research and Evaluation in Europe instrument uses the criteria outlined in Box 5D.6.2.
These show the steps that should be taken when developing guidelines.

496

Public [Link] 496 3/6/08 [Link]

 5D Understanding the Theory and Process of Strategy Development

Box 5D.6.2
Scope and purpose Clear definitions of the guideline objective, clinical question and group of
patients to whom the guideline applies
Stakeholder involvement Range of professionals, together with patient involvement
Rigour of development Systematic appraisal of evidence, explicit consideration of benefits and risks,
evidence of external review prior to publication
Clarity and presentation Specific, unambiguous recommendations
Applicability Target users clearly defined; costs and other barriers discussed. Auditing criteria
outlined. Guideline piloted
Editorial independence Editors independent of funding body. Conflicts of interest recorded

5D.7 INTEGRATED CARE PATHWAYS

An integrated care pathway (ICP) is a multidisciplinary outline of anticipated care. It sets out explicit standards
relating to how a patient with a specific condition is expected to progress through a clinical encounter. Typical
timeframes for each step of the pathway are set out. It is important that ICPs should not be too rigid: there must
be room for some degree of clinical freedom to meet the particular needs of individual patients. Hospitals favour
ICPs because they have the potential to deliver health care at improved quality and lower cost. Advantages and
disadvantages of ICPs are listed in Box 5D.7.1.

Box 5D.7.1
Advantages Disadvantages
Facilitate clinical governance Danger of being too rigid
Reduce unwarranted variations in patient care Potential to disempower patients and carers
Tool for introducing clinical guidelines into everyday Fail to account for the unique biology of the patient,
usage with their special circumstances
Improve risk management
Incorporate organisational strategy into patient care
Avoid disputes over professional boundaries

The ICP document itself becomes the record of all care. Clinicians are required to write in the document rather than
in free-hand clinical notes. The ICP document sets out:
• Timeframes (e.g. how mobile the patient should be on day 3 post-surgery)
• Decision guidance (e.g. what should happen if a patient develops a postoperative infection)
• Observations (type, frequency and interpretation)
• Investigations (which tests should be performed on what day)
• Referral criteria (e.g. what should trigger a referral to a dietician)
• Outcome measures (e.g. patient’s bowels should have opened by day 4 post-surgery).

DEVELOPMENT OF INTEGRATED CARE PATHWAYS

Writing a good ICP is a laborious process that requires a committed facilitator to coordinate the developmental
steps. The process can be markedly accelerated if senior nursing and medical figures offer their public support.

497

Public [Link] 497 3/6/08 [Link]

 5D Unde r s t a nd ing the Theory and Process of Strategy Development

STEPS IN THE DEvELOPMENT OF AN ICP

The steps involved in writing a new ICP include drafting the following:
• Process map (the current patient ‘journey’ is documented, followed by the ideal patient journey)
• Sequence of steps (ideal patient journey is divided into a series of stages)
• Specific responsibilities (roles of the different members of the multidisciplinary team are divided)
• Problem areas (careful attention is paid to potential variations in the patient journey).
Health services often begin by designing ICPs for conditions that are relatively common and potentially
straightforward. Common examples include:
• Hip and knee replacement
• Stroke
• Coronary artery bypass graft (CABG)
• Heart failure
• Pneumonia
• Caesarean section.

EVALUATION OF INTEGRATED CARE PATHWAYS

The evaluation of ICPs is important because, although they have the potential to reduce average length of
stay and to encourage evidence-based practice, they do require additional resources for their development and
implementation. Assessing the impact of an ICP can be difficult since they involve many separate actions within
a complex package of care. In a review of the literature (Bandolier Forum: Care Pathways, [Link])
successful ICPs tended to:
• Examine the external evidence for individual technologies
• Combine this with local knowledge and experience and conditions
• Involve a number of different disciplines
• Measure the results of the actions
• Have information systems feeding back to the team on a timely basis
• Amend the pathway in the light of results.

5D.8 CONSULTATION ABOUT HEALTH SERvICES

Public and carer consultation and involvement in health service planning
The public can be involved in health services in a range of ways, from simply feeding back experiences of their own
care to taking part in the delivery of new models of health services. Individuals can be involved as:
• Consumers of health care or carers
• Leaders or members of community groups (e.g. minority ethnic or religious groups)
• Representatives of groups with specific health interests (e.g. breast cancer support groups).
Consultation and involvement in health care are described in more detail in Section 2I.4.

5D.9 HISTORICAL DEvELOPMENT OF PERSONAL HEALTH SERvICES AND OF PUBLIC HEALTH

HISTORY OF PERSONAL HEALTH SERVICES

See Section 5D.4.

498

Public [Link] 498 3/6/08 [Link]

 5D Understanding the Theory and Process of Strategy Development

HISTORY OF PUBLIC HEALTH

See Table 5D.9.1.

Table 5D.9.1 Major events in the history of public health

Approximate Event
date
450 bc Hippocrates studies medicine as a discipline in itself: records clinical experiences (e.g. obese
people more prone to disease; differentiates epidemic and endemic diseases) and hypothesises
about disease causation (e.g. the four humours: blood, phlegm, black bile and yellow bile)
1300 City of Venice introduces quarantine whereby incoming ships from ports infected with plague
were required to sit at anchor for 40 days before docking
1500 Fracastorius writes about contagion in the context of syphilis
1650 Sydenham carefully describes a series of diseases and hypothesises about miasma (i.e. ‘bad air’
as the cause of disease)
1650 Gaunt analyses Bills of Mortality to describe disease patterns
1670 Leeuwenhoek first visualises bacteria using a microscope
1800 Jenner deliberately vaccinates a boy with the pus from cowpox sores, thereby immunising the
boy against smallpox
1840 Chadwick notes differences in life-expectancy between the social classes. His campaign leads
to the Public Health Act 1848 which established a Central Board of Health with powers to
supervise street cleaning, refuse collection, water supply and sewerage disposal
1845 Snow identifies the cause of an outbreak of cholera in London as infected water coming from a
water pump in Soho. Snow arranges for the handle of the Broad Street pump to be removed, and
thereby terminates the outbreak
1850 Farr working as Registrar General (UK) uses statistical analysis as the basis for sanitary reforms
1875 Pasteur confirmed germ theory, produced artificial vaccines and described the process of
pasteurisation (heat treatment of food to reduce its load of microorganisms)
1875 Koch defines four postulates that must be met to signify disease causation by a microorganism
1948 Bevan establishes UK’s National Health Service
1974 Lalonde report sets out the health field concept: health is determined by the environment,
lifestyle, biology and health care
1977 Alma Ata WHO Assembly sets out the vision of ‘Health for All by 2000’
1980 Black Report (made infamous by the failed attempt of the incoming Conservative government to
suppress publication) outlines the correlation between social class and infant mortality rates,
life-expectancy and inequalities in the use of medical services

499

Public [Link] 499 3/6/08 [Link]

Public [Link] 500 3/6/08 [Link]
 5E Financing Health Care

5E
Finance, Management Accounting and Relevant
Theoretical Approaches

5E.1 Cost of health services 501 5E.3 Methods for audit of health-care spending 508
5E.2 Paying for services 502

Good health is often described as being priceless. Health care is an expensive commodity and, as such, adequate
funding is essential for the adoption of new technologies and the continuation of existing programmes. Public
health practitioners have an important role in promoting the use of effective, efficient technologies and successful
programmes to benefit the population. Practitioners therefore need a strong grasp of the methods of financial
allocation and service commissioning, and require a good insight into the audit of health-care spending.

5E.1 COST OF HEALTH SERvICES

Linkages between demographic information and health service information – its public health interpretation and
relationship to financial costs
Demographic information can be thought of as a measure of need, and health service information as a measure of
performance, and therefore they are proxies for certain outcomes.
Data linkage occurs when data about one set of features (e.g. demography) are linked to another (e.g. use of the
health service or use of the A&E department). Such linkage (of demographic and health service data) can be used to
assess the responsiveness of a health system to need.
Box 5E.1.1
DEMOGRAPHIC INFORMATION
Population size, movement and projections
This includes the items listed in Box 5E.1.1. Age, sex
Ethnicity
Deprivation Employment status
Home ownership
Income
Educational attainment

501

Public [Link] 501 3/6/08 [Link]

 5E F i na nc i ng Heal th Care

HEALTH SERVICE INFORMATION Box 5E.1.2

This includes the items listed in Box 5E.1.2. Tertiary care Specialised units
Secondary care Inpatient, A&E or outpatient activity
ADVANTAGES OF DATA LINKAGE
Length of stay
Linking demographic to health service information
has both public health and financial benefits. Re-admission rates
Acute bed occupancy
PUBLIC HEALTH
Primary care Referrals to secondary care
• Highlights particular health or health-
care issues and prompts further research or Consultation rates
investigation (e.g. identify inequalities in Prescribing information
access, use of services)
• Evaluates progress by local agencies in
improving health and reducing inequality (e.g.
evaluating mental health service provision
according to ethnicity, unemployment)
• Looks ahead to give early warning of public health problems (e.g. rising prevalence of childhood obesity)
• Uses health equity audits, needs assessment and health impact assessment.

FINANCIAL
• Identify long-term savings (average length of stay and re-admission rates – recognising areas of inefficiency
benchmark against neighbouring PCTs, regions, national rates)
• An indication obtained of the intensity of resource utilisation can be used to regulate the level of activity
contained within a service level agreement or contract.

WEAKNESSES OF DATA LINKAGE

While the benefits of data linkage vastly outweigh the disbenefits, some weaknesses of linking demographic and
health service information include:
• Problems with data quality, reliability, access, completeness and lack of data capture
• Poor data quality can widen inequality/inequity and worsen public health
• Poor estimation of costs due to flawed data can lead to over- or under-spending on services
• Data protection requirements can make linkage difficult to negotiate, particularly where several parties act as
guardians for particular data sources.

5E.2 PAYING FOR SERvICES

Budgetary preparation, financial allocation and service commissioning
Budget preparation is an important mechanism for ensuring organisational and financial management, and is crucial
for achieving strategic goals. Accounting may be financial or managerial: see Box 5E.2.1.

502

Public [Link] 502 3/6/08 [Link]

 5E Financing Health Care

Box 5E.2.1
Financial Financial accounting is a specialised field and relates to the use of accounting information
accounting for reporting to external bodies for auditing purposes
Managerial Managerial accounting is less technically complex and is both more amenable and more
accounting useful to the public health practitioner. By using a combination of historical data and
estimated data, managerial accounting can be used to guide:
• Day-to-day operations
• Future operations
• Organisational strategies

BUDGETARY PREPARATION
Budgets follow organisational priorities and can guide spending and decision-making. Budgets follow the fiscal year,
which varies between countries: see Box 5E.2.2

Box 5E.2.2
Country Fiscal year
UK*, Hong Kong, Canada 1 April–31 March
Australia, New Zealand 1 July–30 June
Ireland 1 January–31 December
USA 1 October–30 September
*In the UK, the fiscal year for personal tax affairs runs from 6 April to 5 April.

In order to ensure that decisions reflect both economic realities and remain sensitive to the strategic mission of the
organisation, budgetary preparation must involve both financial and managerial staff. The process begins several
months before the end of the financial year, and involves the steps shown in Box 5E.2.3.

Box 5E.2.3
Service data Collate service performance data (and compare against the objectives that were
set for the year)
Fiscal data Collate fiscal performance data (and compare against the budget that had been
set for the year)
Cost per unit Calculate the cost per unit of service by dividing the cost of the service by the
number of patients seen
Service objectives Determine service objectives for the forthcoming year based on the
organisation’s strategic plan
Costs projections Estimate the costs required to achieve these objectives
Revenue/expense comparison Compare revenue and expense projections
Budget setting Prepare monthly budget breakdowns that reflect anticipated cash flows (not
simply the full budget divided into 12 equal parts)

503

Public [Link] 503 3/6/08 [Link]

 5E F i na nc i ng Heal th Care

FINANCIAL BALANCE
Private sector organisations will, at different times, choose to:
• Incur a deficit (e.g. when investing prior profits into new developments)
• Realise a surplus (e.g. when establishing an operating reserve to guard against future cash flow shortfalls)
• Simply break even.
UK In contrast, government financial orders require NHS organisations to remain in continuous financial balance
on every day of every year. This rule is designed to guard against bankruptcy and at the same time ensure that
the organisation does not build up large surpluses rather than investing revenue in services for patients. It does,
however, restrict flexibility of public bodies such as NHS organisations compared with the private sector. By using
3-year budget cycles, private companies can borrow for the future but this option is not available to NHS trusts.
Trusts that are granted foundation status are, however, afforded additional financial flexibility.

COSTINGS
Budgeting must include the costs of staff, supplies and other resources (see also p 435). Managerial and financial
staff should consult each other to ensure that all resources required for a service are considered. Historical or
published costings can be used – with adjustments made for any impending cost changes.
In health care, staff costs typically amount to two-thirds or more of the expense budget, so this budget line item
must be considered particularly carefully. Recruitment of new staff is especially costly.
See Box 5E.2.4.

Box 5E.2.4
Type of cost Costs at a given level of activity Example
Direct Costs incurred exclusively for that output Disposable surgical equipment
Indirect Costs shared across several outputs Autoclave that sterilises equipment from several
operations
Overheads Costs shared across the entire organisation Cost of the organisation’s press officer

If expenses need to be reduced to maintain financial balance, it is helpful to determine what each programme would
cost at different service levels. A fixed percentage cut across all services is seldom the most effective way to reduce
overall expenses.
See Box 5E.2.5.

Box 5E.2.5
Type of cost Costs at changing levels of activity Example
Fixed Indispensable predetermined expense Salary of a hospital’s chief nurse
Semi-fixed Predetermined expenses that are fixed in the short term Total salaries bill
Semi-variable Cost that is related to output but not directly Cost of equipment maintenance
proportional
variable Cost that is directly proportional to an output Cost of materials

504

Public [Link] 504 3/6/08 [Link]

 5E Financing Health Care

FINANCIAL ALLOCATION
UK When the NHS was founded, it assumed responsibility for voluntary hospitals, distribution of which across the
country was patchy. Initially, funds had to be allocated in order to provide for services in the hospitals newly taken
over. Until the 1970s, these geographical inequalities in NHS funding persisted because hospitals broadly received
historical funding, i.e. the funding that they received in the previous year plus an allowance for growth.
In 1971 the Crossman formula was introduced in an attempt to rationalise the geographical allocation of resources.
The formula was based on three variables:
• Population (adjusted for age and sex)
• Hospital beds (adjusted according to medical specialty)
• Cases (inpatient, outpatient and day-case).
Although transparent, this formula did not explicitly allocate resources on the basis of need and made no
adjustment for deprivation.
Crossman was followed by the RAWP formula (Resource Allocation Working Party), which was based on the principle
of ‘equal opportunity of access to health care for people at equal risk’. RAWP was the first weighted capitation
formula to be used. Its intention was to distribute financial resources based on:
• Population size
• Need for health care based on standardised mortality ratios
• Unavoidable costs of providing health-care services.
The formula has subsequently been gradually adjusted (using small area census data, adjusted according to
the square root of SMR, etc). Sudden swings from historical funding to weighted capitation funding would be
destabilising. However, the NHS is committed to eradicating this historical legacy so current financial allocations are
made according to a combination of weighted capitation, recurrent baselines, difference from target and the pace-
of-change policy. For details see Section 4D.3.

CURRENT FINANCIAL ALLOCATION

UK Planned NHS spending in the UK for 2007–08 was £105.6bn – a rise from £65.4bn p.a. from 5 years earlier.
Eng The bulk of funds flows from the Department of Health as ‘Unified Revenue Allocations’ to PCTs, which then
fund primary and secondary care as shown in Box 5E.2.6.

MARKET FORCES FACTOR

Since a given amount of expenditure purchases different amounts of health care in different parts of the country, an
adjustment called the market forces factor (MFF) is applied to the funding formula. The purpose of the MFF is to
equalise the purchasing power of health authorities with regard to unavoidable variations in costs that are directly
related to location.

COMMISSIONING OF SERVICES
Eng Since the time when the NHS was first established, the process for defining the services to be provided has
evolved from one of lobbying by interested parties, through planning of district hospital services to be provided
across the board, onto the separation of purchasing from providing, and into the current model of commissioning.

505

Public [Link] 505 3/6/08 [Link]

 5E F i na nc i ng Heal th Care

Eng Box 5E.2.6

Allocation Fraction of total Basis of distribution Weighted by
(%)
Hospital and community 82.76 Population Age
health services
Needs
Market forces
Rurality
Prescribing 14.07 Population Age
Needs
Cash-limited general 2.50 Population Age
medical services
Needs
Market forces
HIv/AIDS 0.67 Population Age
Needs
Market forces

In the health context, the term ‘commissioning’ is used to describe the process by which a purchaser of health care
(e.g. a health authority or a GP) identifies a local need for a service and then procures a service that meets this
need. Commissioning is important because it:
• Identifies local health-care needs
• Exposes any deficiencies in current provision
• Determines the level of investment required to meet unmet needs
• Favours health-care services that are cost-efficient
• Ensures unambiguous agreements between purchasers and providers
• Guides workforce development.

PRINCIPLES FOR COMMISSIONING

The Department of Health identifies six principles that should guide commissioners of service. Each of the six is
associated with a tool from the public health armamentarium: see Box 5E.2.7.

Box 5E.2.7
Commissioning principles Public health tool
Population needs Health needs assessment
Local service gaps Health care evaluation
Equity Health equity audit and equality impact assessment
Evidence based Literature review and critical appraisal
Partnerships Change management analysis
value for money Economic evaluation

506

Public [Link] 506 3/6/08 [Link]

 5E Financing Health Care

FUNDHOLDING AND PRACTICE-BASED COMMISSIONING

UK In the early 1990s, organisations that previously both purchased and provided health services were split. This
involved the establishment of separate:
• Provider trusts (e.g. acute hospital trusts and mental health trusts); and
• Health authorities (Wales and England) or health boards (Scotland and Northern Ireland) with a role confined
to the purchasing of services. Later, in 2002–03 this responsibility moved to PCTs (England) and local health
boards (Wales), which also provided services such as district nursing and community physiotherapy.
In the mid-1990s, the UK government experimented with the concept of fundholding general practices in which
GPs were given a budget from which to purchase services for their patients including drugs and hospital care. In
1997 the incoming Labour government scrapped the fundholding initiative because of inequities: fundholding
practices had disproportionately large budgets compared with non-fundholding practices, and were found more
frequently in affluent areas. However, fundholding did achieve a relative reduction in prescribing costs and in
hospital referrals between fundholders and non-fundholders.
Eng In 2004 the UK government announced a new system of local commissioning. Under this system, called
practice-based commissioning, GP practices have the right to identify new providers of health care (including
in-house arrangements or other primary care providers) to offer their patients. Practice-based commissioning allows
practices to choose to commission a variety of services, and any cost savings relative to traditional (hospital) care
are made available for reinvestment by the practice.

SPECIALISED COMMISSIONING
Random fluctuations in health-care activity present a relatively larger risk to small commissioners (such as a GP
practice) than they do to large commissioners (such as a regional health authority). For example, in a GP list, a
single patient with haemophilia requiring surgery could consume as much in health-care resources in a year as all
other patients combined. In order to cope with inevitable over-spending and under-spending from one year to the
next, commissioners can create a risk pool. This is a form of insurance for commissioners, which over-spending
commissioners may access subject to explicit criteria.
UK The National Commissioning Group (NCG) covers risk-sharing arrangements for rare or expensive conditions.
Until 2007, this was known as the National Specialist Commissioning Advisory Group (NSCAG). It exists to assist the
groups listed in Box 5E.2.8.

Box 5E.2.8
Interested party Benefit
Patients Improving access to rare services
Health-care planners Restricting the number of specialist centres so as to maintain high levels of expertise
Commissioners Smoothing out risk
Providers Cash flow to support rare and expensive treatments
Specialists Focus point for discussion about service development

The 42 current services cover conditions such as:

• Craniofacial surgery service for congenital craniofacial disorders
• Epidermolysis bullosa service for children
• Extracorporeal membrane oxygenation (ECMO) service for neonates, infants and children
• Heart and lung transplantation service for adults and children

507

Public [Link] 507 3/6/08 [Link]

 5E F i na nc i ng Heal th Care

• Liver transplantation service for adults and children

• Secure forensic mental health service for young people.

5E.3 METHODS FOR AUDIT OF HEALTH-CARE SPENDING

The purpose of financial audit is to provide an independent and objective opinion on the performance of
organisation with regard to:
• Financial control
• Risk management
• Governance.

INTERNAL AUDIT
In order to maintain credibility, auditors must strive for integrity, objectivity and confidentiality. A health-care
organisation may be audited by:
• In-house audit teams
• In-house audit team with support from external contractors
• External ‘whole-internal-audit’ service.

EXTERNAL AUDIT
UK The National Audit Office (NAO) is an independent body that reports directly to Parliament. One of its roles
is to assess the efficiency and effectiveness with which public sector bodies use their resources. The NAO audits
the summary accounts of the NHS. In addition it appoints external auditors who audit the underlying health
organisations.

FRAUD DETECTION
Fraud detection is, in general, a line-management issue rather than a direct responsibility of auditors. In the UK the
NHS Counter Fraud Service is charged with tackling all fraud and corruption in the NHS.

508

Public [Link] 508 3/6/08 [Link]

 Section 6
SKILLS TESTED AT PART A

Sections 1–5 cover the knowledge basis of public health. But practitioners also need a broad range of skills to
function effectively. These skills are those of research (design and interpretation of studies), manipulation of
information (data processing, presentation and interpretation) and communication.
Skills cannot be learnt through reading, but require application and practice. However, the use of the tips, formulae
and principles in Section 6 will help practitioners as they develop these abilities.

509

Public [Link] 509 3/6/08 [Link]

Public [Link] 510 3/6/08 [Link]
 6A R esearch Design and Critical Appraisal

6A
Research Design and Critical Appraisal

6A.1 Skills in the design of research studies 511 6A.3 Drawing conclusions from research 513
6A.2 Critical appraisal 511

Being able to design and appraise research is a fundamental public health skill. Clearly, in offering study design
advice, there is a balance to be struck between a scientific ideal and the pragmatics of conducting the study. Your
effectiveness will depend on being able to sell the benefits of a good method coupled with the diplomacy of being
able to achieve this within the realities of running a health service.
There is a flow in conducting research from the original source idea to a research hypothesis (aims, outcomes
measurable?) and finally to the measurement of outcomes.
You need to consider the make-up of the research team (skills and personality types), as well as the ethical
dimensions and how the research will be funded.

6A.1 SKILLS IN THE DESIGN OF RESEARCH STUDIES

See Box 6A.1.1.
An imperative to consider when designing a study is the target audience for the results:
• To whom will they be addressed?
• How do you intend to communicate with this audience?
• How much will this cost?
• What do you expect to be the results of your communication?

6A.2 CRITICAL APPRAISAL

Ability critically to evaluate papers, including the validity of the use of statistical techniques and the inferences drawn
from them
See also Appendix A.

511

Public [Link] 511 3/6/08 [Link]

 6A R e s e a rc h Design and Critical A ppraisal

Box 6A.1.1
State the aim of study Estimation of parameters (risk)
Association between factor and outcome
Evaluation (intervention such as new treatment)
Hypothesis Evidence-based assumption to be tested by conducting the study
Choose an appropriate study design Depends on the aim:
Prevalence studies such as surveys, ecological or descriptive
studies: ‘How common is this?’
Case–control studies: ‘What caused this?’ (particularly rare disease)
Cohort studies: ‘What effect does this have?’ (How does the level of
the risk factor affect the outcome?)
Interventional/experimental studies such as an RCT: ‘What happens
if …?’ (apply the intervention and observe the outcome, i.e. test
the hypothesis)
Consider what resources are available
Ensure that the study is ethical
Choose study populations, sampling Use of comparison groups, i.e. controls (see Section 1A.7)
strategies and allocation methods
Sampling methods (see Section 1A.25)
Allocation (see Section 1A.26)
Measures to reduce errors (see Section Chance (sample size, power)
1A.9)
Bias (randomisation, blindness, standardisation, etc)
Confounding (restriction, matching, stratification)
Exposure measurement in the context of Attention to aspects of data collection and processing (including
observational research coding, data entry, cleaning, quality control, etc)

Critical appraisal is the systematic process of assessing and interpreting evidence. It involves judging a paper or
study with regard to three factors:
• validity (could the findings be explained by chance, bias or confounding?)
• Results (are the statistical methods sound?)
• Relevance (are the findings useful to your organisation?).
An underlying principle of appraisal is expressed as ‘just because it is published does not make it true’.
The process should be objectively balanced. It must be neither overly deferential nor harshly critical of the authors.

CRITICAL APPRAISAL SKILLS PROGRAMME

The Critical Appraisal Skills Programme (CASP) developed by Milton Keynes PCT can be viewed at [Link]/
casp. This contains the appraisal tools listed in Box 6A.2.1, which can be downloaded free, then used for critically
appraising papers.

512

Public [Link] 512 3/6/08 [Link]

 6A R esearch Design and Critical Appraisal

Box 6A.2.1
• Systematic reviews
• Randomised controlled trials
• Diagnostic test studies
• Cohort studies
• Case–control studies
• Economic evaluation studies
• Qualitative research studies

In addition, the CASP workbook and CD-ROM package (which can be purchased through the website) offer practical
experience of critically appraising papers similar to those found in the UK MFPH Part A examination.

CRITICAL APPRAISAL OF STATISTICAL TECHNIQUES

The number of statistical tests available is infinite, so you may well not be familiar with the exact analysis used in
the paper that you are appraising. This need not be intimidating: simply interpret the statistical test in terms of a p
value and confidence interval.
Other issues to address when critically appraising statistical techniques include:
• Do the figures in the tables add up?
• Was a power calculation used?
• Were multiple comparisons made without using the Bonferroni correction?
• Are confidence intervals given?
• Is a p value quoted for null results?
• Is the absolute risk reduction (as well as the relative risk reduction) quoted?
• Can you suggest an alternative statistical test that might have been used?

6A.3 DRAWING CONCLUSIONS FROM RESEARCH

Ability to draw appropriate conclusions from quantitative and qualitative research
In a public health examination, ask yourself why has this paper been put into today’s examination? Are there any
traps? Any conclusions or inferences that should not be drawn? Any biases or fundamental flaws? Think of all the
possible explanations other than the orthodox.
Conclusions drawn from research should clearly state whether the findings support or refute the hypothesis posed.
Their public health relevance can then be ascertained by considering whether the findings:
• Justify or prove the effectiveness of a programme
• Serve to refine an existing theory
• Can be used to develop a new theory.
The BMJ ([Link]/advice/[Link]) recommends that conclusions should be structured as follows:
• Statement of principal findings
• Strengths and weaknesses of the study
• Strengths and weaknesses in relation to other studies, discussing important differences in results
• Meaning of the study: possible explanations and implications for clinicians and policy-makers
• Unanswered questions and future research.

513

Public [Link] 513 3/6/08 [Link]

Public [Link] 514 3/6/08 [Link]
 6B Drawing Conclusions from Data

6B
Drawing Conclusions from Data

6B.1 Drawing conclusions from data 515

6B.1 DRAWING CONCLUSIONS FROM DATA

Ability to sort and manipulate data and to draw appropriate conclusions from quantitative and qualitative data
The ability to handle data and draw appropriate conclusions is a fundamental skill for all public health practitioners.
In some instances, this may involve presenting data that practitioners have collected and analysed themselves. At
other times, it requires the conclusions and data presented by others to be critically reviewed.
Basic summary statistics and graphical techniques can be used to highlight trends and make comparisons within
quantitative data. For example, frequencies may be illustrated using a bar or pie chart (for categorical data) or using
a histogram (for continuous data).

TRANSFORMING RAW DATA INTO MEANINGFUL INFORMATION

See also Section 1B.8.
Well-presented tables, figures and graphs enable the reader to identify patterns and contrasts in the data that would
otherwise not be immediately apparent.

TABLES
When displaying data in tables, aim to follow the following rules:
• Sort in a meaningful order (e.g. largest to smallest) rather than random or alphabetical order
• Label the table correctly: rows and columns (with units), together with the title for the table itself
• Two significant figures usually provide sufficient information for the reader
• Rates are often more useful than numbers for comparing data (but the denominators must be comparable).

GRAPHS
When generating graphs, follow the principles shown in Box 6B.1.1.

515

Public [Link] 515 3/6/08 [Link]

 6B D ra w i ng C oncl usions from D ata

Box 6B.1.1
Clarity Label the graph (title, axes, units) completely but succinctly
Use scales appropriately – fill as much of the graph’s space with data
Maintain convention by starting the axes of the graph from zero
Ensure that the colour or pattern of lines or bars clearly differentiates each category
Simplicity Consider Edward Tufte’s ink-to-data ratio: use the least ornate format of data to convey the
message (i.e. table rather than bar chart, bar chart rather than pie chart)
Use few gridlines
Use three-dimensional designs only where a two-dimensional design would not be appropriate
Be selective with content: only include relevant data
Transparency Provide a source (and date) for the data wherever possible
If data have been transformed (e.g. divided into categories or calculated as percentages), then
also provide the raw data

WHICH GRAPH TO USE?

See Table 6B.1.1 and Section 1B.8.

Table 6B.1.1 Summary of display graphs

Type of display Strengths Weaknesses
Table Display data details that Difficult to visualise
would be lost in charts relationships and trends
or text
Categorical Pie chart Displays percentages Difficult to gauge size of pie
data within a whole slices by eye
Easily understood, Difficult to compare
popular in lay materials, proportions across two pie
e.g. newspapers charts
Bar graph Summarises large Scaling effects (i.e. when
6 amounts of data in a one variable is much greater
5 visual form than others, this reduces
Count or percentage

4 the scale of the graph and

Trends and relationships
3 therefore makes it difficult
easy to visualise
2 to visualise small changes in
1 Relatively simple, so other variables)
0 accessible to a range of
A B C
Category
audiences

516

Public [Link] 516 3/6/08 [Link]

 6B Drawing Conclusions from Data

Type of display Strengths Weaknesses

Univariate Box and whisker Median, range of data Relatively complex – not
numerical points easily identifiable readily accessible for wide
data range of readers
Useful to compare two or
more sets of data Exact values not retained

Histogram Indicate mean, median Does not display exact

and mode numbers, only category
bands
Illustrate shape of
frequency distribution
(e.g. symmetrical or
skewed)

Bivariate Line graph Clear display of Temptation to extrapolate

numerical relationships between beyond data points
data several dependent
Difficult to distinguish
variables and an
between lines (particularly if
independent variable
printed on black and white
Displays trends (e.g. or poorly photocopied)
changes over time)
Scaling effects – can bias
results (e.g. on a 0–100
scale, a change of two
points looks small; whereas
on a scale of 30–40, a
change of two points looks
large)
Scatterplot Shows minimum, Hard to visualise results/
maximum and outliers trends and relationships for
large data-sets
Illustrates relationship
between variables

Table contd overleaf

517

Public [Link] 517 3/6/08 [Link]

 6B D ra w i ng C oncl usions from D ata

Table 6B.1.1 contd

Type of display Strengths Weaknesses
Other Flow chart Illustrates a series of Potential to over-simplify
steps in a procedure, processes or procedures
Eligible
patients decision or other
(n = …) ‘stepwise’ process
Essential in RCTs to show
Not randominsed the proportion of original
(n = …)
units retained by the end
of the study

Randomisation

Treatment Control
(n = …) (n = …)

Follow-up Follow-up
(n = …) (n = …)

Withdrawn Withdrawn
(n = …) (n = …)

Completed Completed
(n = …) (n = …)

Maps: Shows relationships Potential to distort (e.g.

• Qualitative: mind maps/organisational between areas large rural areas with
maps small populations appear
Allows rates to be
• Coded mapping (e.g. MapInfo) bigger on a map than small
compared between and
urban areas with a high
among regions and
population)
countries

COMMON PITFALLS IN GRAPHICAL DISPLAY

There are several ways in which a poor graphical display can cloud or distort the message that it intended to convey.
These include:
• ‘Chart junk’ – unnecessary graphics or text on charts
• Overuse of three-dimensional or complex graphical designs that obfuscate the main message and can bias the
presentation (e.g. pie slices at the back of a three-dimentional pie chart look smaller than those at the front)
• Incorrect or insufficient labelling
• Scaling problems: too small to see trends; no zero; distorted scales – the ‘gee whiz’ graph (see below)
• Spurious comparisons (unequal denominators; numbers given rather than rates)
• Too many data points
• Too few data points.

ExAMPLE 1: THE EFFECTS OF SCALING ON DATA PRESENTATION (‘GEE-WHIZ’ GRAPHS)

Depending on how the data in Box 6B.1.2 are presented, the effectiveness of the New Year diet plan can appear very
different: see Figure 6B.1.1.

518

Public [Link] 518 3/6/08 [Link]

 6B Drawing Conclusions from Data

Box 6B.1.2
Individual’s weight by week following New Year diet plan
1 Jan 8 Jan 15 Jan 22 Jan 29 Jan
Weight (kg) 73 72.5 72 70 69.5

74
70
73
60
72
50
Weight (kg)

Weight (kg)
71
40

70 30

69 20

68 10

67 0
1 Jan 8 Jan 15 Jan 22 Jan 29 Jan 5 Feb 12 Feb 19 Feb 26 Feb 1 Jan 8 Jan 15 Jan 22 Jan 29 Jan

Week Week

‘Gee whiz’ – amplified change: Normal scaling – reduced change:

Most of the y axis scale is missing (67–74 kg) All of the scale is shown (y axis 0–75 kg)
2 months shown; x axis: compressed, relative to vertical x axis only 1 month

Figure 6B.1.1 Illustration of how the gee-whiz graph amplifies change

ExAMPLE 2: CHOOSING APPROPRIATE GRAPHICAL DESIGN FORMATS

See Box 6B.1.3 and Table 6B.1.2.

Box 6B.1.3
Area Ethnicity (%)

White South Asian African/Caribbean

A 81 6.1 3.2
B 63 21 6.0
C 42 45 11

519

Public [Link] 519 3/6/08 [Link]

 6B D ra w i ng C oncl usions from D ata

Table 6B.1.2 Choosing a graphical design format

Ethnicity breakdown across three
Ethnicity areas
breakdown across three areas
90 90
80 80 White South Asian African/Caribbean
White South Asian African/Caribbean
70 70
60 60
50 50 C C
40 40
30 30

Area

Area
20 20 B B
10 10C C
0 B0 B
A
White

White
A A
South Asian

South Asian
African/
Caribbean

African/
Caribbean
0 20 40 0 60 20 80 40 100 60 80 100
Percentage by area Percentage by area

Weaknesses Strengths
Design Three-dimensional, complex design Two-dimensional, simple design
Takes up more space, confuses message Enables easy comparison of relative population
constituents
Labelling No title Title describing what the graph shows
No label or units on y axis, obtrusive Axes labelled with units
labelling of x axis
Legend at the bottom of graph, taking up least room
possible
Scaling Size disparity between categories: difficult Scale appropriate for category sizes – enables
to see relative differences in each ethnic comparison between three areas
category, cannot read off the values

DRAWING CONCLUSIONS FROM DATA DISPLAYS

See also Appendix A.
When presented with data, it is tempting to focus immediately on the detail. Instead, it is more effective to
describe observations in a logical order, such as:
1. What type of display (chart/plot) is it?
2. What information is it attempting to show? Title, x and y axes and scale
3. What shapes do you observe? (e.g. bell?, skewed?, linear increase?, sharp increase or spike?, J, S shapes)
4. What similarities/differences do you notice?
• Range of data – highest/lowest
• Trends
5. How do you interpret what you see? Are there alternative explanations?
6. What caveats would you note?
• Quality of the presentation (simplicity, clarity, transparency)
• Quality of data handling, manipulation
• Comprehensiveness of the data (other information that you would need to strengthen your interpretation).
See Boxes 6B.1.4 and 6B.1.5.

520

Public [Link] 520 3/6/08 [Link]

 6B Drawing Conclusions from Data

Box 6B.1.4
Example: interpreting a line graph

GP registered patients (15 –75 years) with a recorded BMI of 30 or over

Dr A Dr B Dr C Dr D

25
Percentage patients*

20

15

10

0
2004 2005 2006
Year

*Registered patients with recorded BMI 30� as a proportion of total GP registered population

Box 6B.1.5
Suggested commentary
This is a line graph, showing the proportion of patients registered with four different general practices with a
recorded body mass index (BMI) of over 30 (clinically obese) in a 3-year period (2004–2006).
In all practices the proportion of patients considered clinically obese rose from 2004 to 2006. In 2004, the
proportion ranged from about 6% in Dr D’s practice to 14% in Dr A’s practice. In 2006 the proportion ranged
from just under 9% in Dr C’s practice to 21% in Dr A’s practice.
The increase was approximately linear for Dr C’s practice, with an increase of roughly 1% for each year. The
increase was steeper in Dr D’s practice in 2004–2005 than in 2005–2006. In contrast, the increase for Dr A’s and
Dr B’s practices was greater in 2005–2006 than in 2004–2005.
The graph indicates the following:
• There appears to be a clear trend of growing obesity prevalence in all four practices
• The proportion of obese patients varies by practice and this variation has increased since 2004, possibly
indicating increasing health inequalities within the area
However, the data should be interpreted with caution because:
• The graph indicates the proportion of patients registered with the GP with a recorded BMI of 30 and above;
it may simply show that recording of BMI has increased. It would have been more useful to use a different
denominator for the percentage calculations, e.g. people with a BMI 30+ as a percentage of all patients
registered with the practice where BMI was recorded
• The data have been divided into just two categories – BMI of 30 and above and BMI below 30. It would be
helpful to have information on the raw values to find the range and dispersal of BMI by practice over time
• Only 3 years of data are presented here; data from previous years would be helpful to forecast with greater
confidence how the prevalence of obesity will develop over time

521

Public [Link] 521 3/6/08 [Link]

 6B D ra w i ng C oncl usions from D ata

DATA ANALYSIS PACKAGES

For both quantitative and qualitative data (particularly large datasets), statistical packages are now available to
accelerate the process of data manipulation and display: see Table 6B.1.3. However, their effective use still relies on
fundamental understanding of what the data mean.

Table 6B.1.3 Statistical packages for quantitative and qualitative data

Qualitative data Packages such as NUD*IST, MaxQDA and NVivo allow textual data to be searched and sorted.
Segments of interest can then be noted and marked with code words. The packages allow
analyses to be performed upon these code words which can be saved, exported or analysed
further
Quantitative Most data need to be processed before they can be analysed. Data columns (variables) need to
data be checked for accuracy, and may be re-arranged, re-coded or re-ordered. Some tables may need
to be combined with others, especially if the data are from a relational database. Programs such
as Microsoft Access™ can do much of this, although programming skill is required
A most versatile programme, in terms of ease of use and the fact that it is to be found on most
PCs, is Microsoft Excel™. This is very useful for re-shaping and re-presenting data. Its graphs
and tables are easy to produce, a particular bonus for novice users
In epidemiology, EpiInfoTM can create questionnaires, store the data collected, manipulate,
analyse and display the results
In general public health, the following statistical packages are often used: SASTM, SPSSTM and
StataTM. The relative merits of these packages depend on the type of data, the analysis that
needs to be performed and the user’s technical abilities
Geographical Geographical Information Systems (GIS) such as MapInfoTM offer a range of ways of
data manipulating and displaying information with a geographical component

522

Public [Link] 522 3/6/08 [Link]

 6C Written Presentation Skills

6C
Written Presentation Skills

6C.1 Written presentation skills 523 6C.3 Presenting to different audiences 526
6C.2 Preparation of papers for publication 524

6C.1 WRITTEN PRESENTATION SKILLS

Communicating clearly and effectively is a vital public health skill. It can make the difference between your efforts
leading to significant changes and your efforts leading to a report gathering dust on a shelf.
Throughout the MFPH Part A examination (and its equivalents), candidates are tested on their written presentation
skills. It is a temptation in the examination setting to skimp on the preparation and presentation of answers, but
to do so is a false economy. Always ensure that you have addressed each of the following three principles before
committing pen to paper: preparation, organisation and customisation.

PREPARATION
Before you start writing, consider the issues listed in Box 6C.1.1.

Box 6C.1.1
Brief What are you writing for?
What are the constraints, e.g. time, word count?
Audience Who is your intended audience? Different language, structures and content will be appropriate for
different audiences. Think about your audience’s:
• Familiarity with the subject matter
• Education/understanding (is technical language appropriate? Will it alienate your audience or
will it provide you with credibility?)
• Culture
• Point of view

ORGANISATION
Ways to organise your writing include those shown in Box 6C.1.2.

523

Public [Link] 523 3/6/08 [Link]

 6C W r i t t e n P resentation Skil l s

Box 6C1.2
Structure Use a clear structure for your written presentations (in the rest of this chapter some standard
formats and structures for particular audiences and purposes are given). In an examination, it is
absolutely vital to sketch this out before you start writing an answer
Subheadings Use subheadings to help your reader navigate through your writing, and to help you keep to
the topic at hand
Lists Use lists and bullets interspersed throughout your prose

CUSTOMISATION
Ensure that the language that you use is appropriate for your audience. Aim for the characteristics shown in Box
6C.1.3.

Box 6C.1.3
Clarity and Use the simplest language appropriate for your audience: choose short words rather than long;
simplicity make sentences as short as possible
Precision and Technical language and abbreviations (which should be defined at the first usage) may be
brevity appropriate. Only use them if they provide a more precise way of expressing yourself in fewer
words, and you anticipate that your audience will be familiar with them
Neutral and Avoid using words or phrases that could be construed as pejorative or insulting by some of your
respectful audience. Common pitfalls include the term ‘innocent victims’ (implying that there are some
language guilty ones); defining groups by disease or characteristics (e.g. schizophrenics or insomniacs, as
opposed to people with schizophrenia or sleeping problems), and sexist terms (e.g. workmen)

CONCLUSIONS
If you have time, check and summarise what you have written. Note, however, that in an examination it is
debatable whether an extensive summary should be included. This is because conclusions do not contain any new
material and therefore do not attract new marks.
1. Refer back to original instructions periodically: does what you have written meet the original brief?
2. Although the executive summary will normally be at the start of your document, write the summary last,
after reviewing what you have written. This will ensure that it closely reflects what follows in the rest of the
document.
3. Read what you have written. In an ideal world, it is useful to leave written work overnight before checking
and submitting it. Clearly, in an examination this will not be possible.
4. If you have written electronically then use a spelling and grammar checker. This is not a substitute for
reading the document yourself but it will pick up many errors and may suggest simpler sentence structures.
5. Remove any embedded comments and tracked changes before submitting a document electronically.

6C.2 PREPARATION OF PAPERS FOR PUBLICATION

Most journals will have their own requirements on what they will accept for submission. If you are re-submitting
an article to a different journal because of a rejection, then the article should be thoroughly re-drafted and
re-formatted for the new journal. Goulding (2003) summarises the main points to consider for different types of
scientific publication.
A paper should be structured along the lines shown in Box 6C.2.1.

524

Public [Link] 524 3/6/08 [Link]

 6C Written Presentation Skills

Box 6C.2.1
Title
Keyword list
Used by the indexing and abstracting services, in addition to those already present in the title
Abstract
(Most journals specify a length, typically not exceeding 250 words)
Principal objectives and scope of the investigation
Summarise the results
Principal conclusions
Introduction
Context, background, literature review
Why was there a need to conduct the study? Introduce pertinent literature
Methods
Major study design elements:
• Sample
• Measurements
• Statistical models and testing
Include ethics approval
Results
Combine the use of text, tables and figures to condense data and highlight trends – refer to publisher’s
guidelines for preparing tables and figures
Discussion (see also Section 6A.3)
Generalisations that can be drawn
How findings compare with the findings of others or expectations based on previous work
Any theoretical/practical implications of your work
Consider the limitations of work
References
Check the publication’s style
Reference list should contain all references cited in the text
Include with each reference details of the author(s), year of publication, title of article, name of journal or book,
and place of publication of books, volume and page numbers.
Be consistent in the use of journal abbreviations
Authorship
Order: modern journals have strict rules regarding authorship and contributorship
Acknowledgements
Grant-awarding body
Clerical support, etc.
Conflicts of interest

525

Public [Link] 525 3/6/08 [Link]

 6C W r i t t e n P resentation Skil l s

SUBMISSION PROCESS
When dealing with reviews:
• Organise the final version of the paper and all ancillary data carefully before submission
• Incorporate any helpful comments and re-submit.
Choose the journal carefully, ensuring:
• Relevance to the subject
• Potential audience
• Impact factor.

6C.3 PRESENTING TO DIFFERENT AUDIENCES

Preparation of material for different audiences, including expert and non-expert audiences, the media and information
handling and use of media in advising the public about health services, disease prevention and health promotion

SLIDE PRESENTATIONS
PowerPointTM is the most common way of delivering a professional presentation. For slides to be effective, it is
important that the presenter produces effective slides, and prepares for and performs the presentation in a format
appropriate to the audience.

PRODUCING EFFECTIvE SLIDES

1. Minimise content per slide: 3–5 bullets, 5–6 words long (or use a simple graph, figure or picture)
2. Ensure that the presentation is readable:
• Font: size 20+; sans serif; high contrast with the background
• Template consistent on all slides
• Inconspicuous background colour and design
3. Use slide animation sparingly if at all.

PREPARATION
1. Practise the presentation (e.g. audio or video recording)
2. Time its length
3. Arrive early on the day and check that your presentation runs on the projector and computer. Better still,
bring your own tried and tested equipment.

PRESENTATION
1. Talk to the audience not the screen
2. Use slide text as key points; do not just read off the screen or from a script
3. Use a loud, low-pitched, slow voice
4. Combine with other formats (e.g. interact with the audience; include video; provide handouts).

PRESS RELEASES
General principles for press releases are described in Table 6C.3.1.

526

Public [Link] 526 3/6/08 [Link]

 6C Written Presentation Skills

Table 6C.3.1 Principles for press releases

Information pyramid Put the most important information at the beginning of the release, with less important
material further down. (Imagine that the release could be ‘cut’ at any point from the
bottom of the page upwards)
Keep it short • Short sentences (~20 words)
• Short paragraphs (~2 or 3 sentences each)
• Short release (~2 pages in total, with notes to editors)
Active voice Use the active, not passive, voice (i.e. say ‘Researchers found that …’ not ‘It was found
that …’)
Non-technical Wherever possible (e.g. use ‘link’ not ‘epidemiological association’; use ‘breathing’ not
language ‘pulmonary’)
Messages Use the release to include established public health messages, e.g. effects of smoking

The format and content of a release will vary to some extent depending on the organisation producing it and on the
story itself. A standard layout is shown in Box 6C.3.1.

INTERVIEW BRIEFINGS
Be clear about what information and what impression you want to leave with the audience. You should plan your
own responses in advance and prepare for related issues – or topical matters – that may also be asked about. Always
consider the perspective of the interviewer and that of the audience.
Organisations are likely to have their own format for briefings, but the following kinds of information will typically
be generated in consultation with the communications officer.

ARRANGEMENTS
• Interview for: (station/programme/presenter/newspaper)
• Time
• Date
• Telephone/studio
• Live/pre-record.

KEY POINTS
Identify three key points and ensure that they are:
• In plain English, suitable for a general audience
• Short and memorable.
Note that it is useful here to have ‘bridging phrases’ handy. These acknowledge the question that was asked but also
ensure that the key points are covered, e.g. ‘and this leads me on to …’ and ‘… but the real issue is …’.

BACKGROUND
• Target audience of radio station/newspaper, e.g. ABC1 women, teenagers/young people, professional men …
• Agenda and point of view of the media (e.g. anti/pro public health issue?)
• Any other people due to be interviewed? Their viewpoint(s)?
• Topic(s) of the interview: some media will send through a list of questions/topics for the interview in advance if
requested.

527

Public [Link] 527 3/6/08 [Link]

 6C W r i t t e n P resentation Skil l s

Box 6C.3.1.
Organisation
Date
Press release
TITLE, e.g. NEW RESEARCH LINKS SMOKING AND COT DEATH
EMBARGOED UNTIL … HOURS, DATE (or FOR IMMEDIATE RELEASE)
Subtitle (e.g. ‘Reducing parents’ smoking may cut baby deaths’)
Paragraph 1
Cover: Who? What? Where? When?
For example, researchers (who) at X University (where) have linked cot deaths with smoking (what) in a paper
published today (when)
Paragraph 2
More details, answering ‘How?’
For example, over 1000 families answered various questions about their lifestyle, health and living conditions
Paragraph 3
More detail/Why?
For example, researchers believe the effects of smoking could be ... or It is too early to understand why smoking
has these effects
Paragraph 4
Quote from someone in authority or connected with the study
For example, X Director of Public Health, Jane Smith, said, ‘This could help reduce cot death …’ OR
John Smith, who led the study, said, ‘This tells us something new about cot death’.
-ENDS-
Notes to editors:
• Contact details for more information
• Background information: e.g. x babies die from cot death every year in the UK; smoking is the biggest
preventable cause of death

CORRESPONDENCE
LETTERS
See Box 6C.3.2.

EMAILS
Emails combine the immediacy of face to face communication with the permanence of traditional written
correspondence and the audience of broadcast media. Email used well can be invaluable. Email used badly can,
at best, be ignored and, at worst, alienate. The style and manner of email correspondence (see Table 6C.3.2) can
ensure that emails are effective and well received.

528

Public [Link] 528 3/6/08 [Link]

 6C Written Presentation Skills

Box 6C.3.2
Headed paper containing:
Organisation
Reply address
Date
Recipient address
Dear … [title/name as they have written to you]
Re: [subject of the letter]
Paragraph 1 Thank you for your letter of ….
Paragraph 2 Acknowledge concern/query …
for example, ‘I acknowledge that X is a particular issue …’
Paragraph 3 Background and evidence base for your point of view
for example, ‘X services are provided currently …’
Finish This is the current situation …
We shall, of course, keep the situation under review.
Further contact Please contact me if I can be of further help.
Yours sincerely*
Name, Qualifications
Position
*Use ‘Yours faithfully’ for letters addressed to Dear Madam or Dear Sir

STYLE
• Be concise: people receive hundreds or thousands of emails per month and can be impatient with unnecessarily
long and uninterrupted text.
• Be sensitive: the tone of communication implicitly conveyed through speech and even handwriting is lost in
emails – it can be easy to offend through overly terse text, or misplaced humour.
Retain the formality of traditional written correspondence when emailing people professionally:
• Address people whose first name you have not been given as Ms, Mr, Dr, Professor
• Follow rules of grammar, punctuation and sentence structure
• Use conventional spellings, few abbreviations, no emoticons or text language.

RESPONDING TO EMAILS
Reply within 24 h wherever possible. Never send an email too hastily:
• If responding to an offensive message, wait until you are calm enough to respond politely before sending a
response
• Check the contents for errors
• Observe good email etiquette (see [Link]).

529

Public [Link] 529 3/6/08 [Link]

 6C W r i t t e n P resentation Skil l s

Table 6C.3.2 Rules for email

Use always Use sparingly
Email signature: set as a template including the following: Attachments: at best unnecessary
attachments are just ignored, at worst
• Disclaimer and confidentiality message (viruses, if received in
they could alienate recipients by clogging
error)
email boxes or transmitting viruses
• Your full name and position
• Organisation name, phone number and postal address
Subject header: this can determine whether your email is even Urgent priority (marks emails as ‘!’):
opened. Good subject headers are concise, but indicate what the the recipient is unlikely to respond any
message will contain. Ideally they should also give some insight into quicker to these emails if there is no
what the recipient needs to do. obvious reason for its use
Bcc for mass mailings: ensure that recipients do not need to scroll Reply to all: ensure that emails are sent
down a list of names before the message (write your own email only to those who need to see them
address in the To: field and all other names in the Bcc field)
Spell checker Formatting and graphics: html formats,
e.g. font styles, bullets, tables, may
not be retained when sent to different
systems

STRATEGY DOCUMENT
1. Current position (identify issues) – where are we now?
2. Future priorities and objectives (including targets, evidence base) – where do we want to be?
3. Strategy (include time table, implementation) – how are we going to get there?
4. Monitoring and evaluation – how will we know we are there?

REPORT TO MANAGEMENT
GENERAL PRINCIPLES
• Keep to around four sides A4
• Why is this paper written, and why now?
• Main points only: you need to make it clear what exactly it is that you are asking management to decide
• Intersperse text with bullet points
• Use subheadings to break up the text.
A standard layout is shown in Box 6C.3.3.

530

Public [Link] 530 3/6/08 [Link]

 6C Written Presentation Skills

Box 6C.3.3
Example: Standard report template
Report to: name of management board
Report from: name, position, department
Title: covering the main subject of the paper in non-technical language
Date:
Purpose: what the board should do: i.e. for information, for approval of recommendations, for discussion
Executive summary (and recommendations if appropriate)
5–7 sentences summarising the report
Enough information needs to be included for someone to read the executive summary only
Background
• Set out the context
• What is known from policy and/or research
• Issue under question
Heading(s) specific to subject
Consider the audience:
• Lay members (use non-medical language, outline relevant medical principles)
• Responsibility of the board (e.g. if commissioning, consider cost and contracting issues)
Recommendations/options
Make clear:
• Who is responsible for implementing what
• Timescales involved
• Resource implications of options

531

Public [Link] 531 3/6/08 [Link]

Public [Link] 532 3/6/08 [Link]
 6D Formulae R equired to Pass Part A

6D
Formulae Required to Pass Part A

6D.1 Sensitivity and specificity 533 6D.8 Chi-squared for a 2 ¥ 2 table 537
6D.2 Positive and negative predictive power 534 6D.9 McNemar’s test 538
6D.3 Numbers needed to treat 536 6D.10 Standardisation – direct and indirect 539
6D.4 Relative risk 536 6D.11 Weighted averages 539
6D.5 Odds ratio 536 6D.12 Confidence intervals and standard errors
6D.6 Attributable risk fraction 536 of the mean 540
6D.7 Applications of standard error 537

This chapter lists all of the formulae that candidates must learn and be prepared to apply in the UK MFPH Part
A examination. Formulae found elsewhere in the book that are not listed here are included only to facilitate
understanding.

6D.1 SENSITIvITY AND SPECIFICITY

See Box 6D.1.1.

533

Public [Link] 533 3/6/08 [Link]

 6D F o r mu l a e Req uired to Pass Part A

Box 6D.1.1

Sensitivity = proportion of those who truly have the disease and are picked up by the test
= the doubly positive cell / sum of both truly positive cells
a
=
a+c

Truth
POSITIvE NEGATIvE
POSITIvE a b a+b
Test result
NEGATIvE c d c+d
a+c b+d

Specificity = proportion of those who truly do not have the disease and are left alone by the test
= the doubly negative cell / sum of both truly negative cells
d
=
b+d

Truth
POSITIvE NEGATIvE
POSITIvE a b a+b
Test result
NEGATIvE c d c+d
a+c b+d

6D.2 POSITIvE AND NEGATIvE PREDICTIvE POWER

See Box 6D.2.1. See also Section 2C.2.

534

Public [Link] 534 3/6/08 [Link]

 6D Formulae R equired to Pass Part A

Box 6D.2.1

Positive predictive power = proportion of those testing positive who truly have the disease
= the doubly positive cell / sum of both test positive cells
a
=
a+b

Truth
POSITIvE NEGATIvE
POSITIvE a b a+b
Test result
NEGATIvE c d c+d
a+c b+d

Negative predictive power = proportion of those testing negative who truly do not have the disease
= the doubly negative cell / sum of both test negative cells
d
=
c+d

Truth
POSITIvE NEGATIvE
POSITIvE a b a+b
Test result
NEGATIvE c d c+d
a+c b+d

535

Public [Link] 535 3/6/08 [Link]

 6D F o r mu l a e Req uired to Pass Part A

6D.3 NUMBERS NEEDED TO TREAT

See Box 6D.3.1. See also Section 1A.22.

Box 6D.3.1

1
Number needed to treat (NNT) =
Absolute risk reduction

6D.4 RELATIvE RISK

Relative risk is calculated as the risk ratio, the rate ratio, the odds ratio (for case–control studies) and the
standardised mortality ratio (SMR) (in an occupational setting). In cohort studies the time period must be stated
(because the risk of dying will always be 100% in the long run for both groups). See Box 6D.4.1. See also Section
1A.10.

Box 6D.4.1

Risk ratio = (Risk of disease in exposed) ∏ (Risk of disease in non-exposed)

Rate ratio = (Incidence rate in exposed) ∏ (Incidence rate in non-exposed)
Odds ratio = (Odds of exposure in cases) ∏ (Odds of exposure in controls)
SMR = (Number of cases observed) ∏ (Number of cases expected) ¥ 100%

6D.5 ODDS RATIO

The odds ratio is a measure of relative risk used in reporting case–control studies. See Box 6D.5.1. See also Section
1A.10.

Box 6D.5.1

Odds ratio = (Odds of exposure in cases) ∏ (Odds of exposure in controls)

6D.6 ATTRIBUTABLE RISK FRACTION

See Box 6D.6.1. See also Section 1A.10.

Box 6D.6.1

Population attributable risk

Population attributable risk fraction =
Rate of disease in population
Attributable risk
Attributable risk fraction =
Risk in exposed group

536

Public [Link] 536 3/6/08 [Link]

 6D Formulae R equired to Pass Part A

6D.7 APPLICATIONS OF STANDARD ERROR

Standard error and confidence interval of a proportion and of a difference in proportions
See Boxes 6D.7.1 and 6D.7.2. See also Sections 1B.3 and 1B.7.

Box 6D.7.1

Proportion

Standard error (proportion) =

95% confidence interval = sample value ± (1.96 ¥ standard error)

Box 6D.7.2

Difference in proportions

Standard error (difference in proportions) =

95% confidence interval = sample value ± (1.96 x standard error)

6D.8 CHI-SQUARED FOR A 2 ¥ 2 TABLE

Box 6D.8.1

See Box 6D.8.1. Chi-squared is used only for actual numbers: not proportions, percentages, etc.
1. For each observed number calculate the expected number
2. Subtract the expected number from the observed number (O – E)
3. Square the result and divide this by the expected number (O – E)2 ∏ E
4. X2 = total of these results for all cells, i.e. sum of (3)
5. Look up X2 (using degrees of freedom = [rows – 1] ¥ [columns – 1]) to find the p value.
See also Section 1B.12.

537

Public [Link] 537 3/6/08 [Link]

 6D F o r mu l a e Req uired to Pass Part A

6D.9 McNEMAR’S TEST

Box 6D.9.1

(A – B)2
c2 =
A+B

Box 6D.9.2
Example of McNemar’s test: calculating significance of difference between matched smokers trying to give
up receiving nicotine replacement and receiving placebo
Pairs of smokers were matched according to age, sex and ethnicity. The first in the pair received nicotine
replacement patches and the second in the pair received a placebo patch. All participants were assessed at
6 weeks after beginning the patches.

First in the pair

(nicotine replacement)
Still Not Total
smoking smoking
Second in the Still smoking 86 25 111
pair (placebo)
Not smoking 56 17 73

Total 142 42 184

1. Ignore the concordant cells

First in the pair

Still Not Total
smoking smoking
Second in Still smoking 86 25 111
the pair
Not smoking 56 17 73

Total 142 42 184

(A – B)2
2. Assign A to be 25, and B to be 56 for the equation c2 =
A+B
(25 – 56) 2
3. c2 = = 11.9
25 + 56

Since 11.9 is greater than 3.84 (1.962), it can be concluded that there is a significant difference at the p = 0.05
level between the matched pairs, i.e. that there is a difference between placebo and nicotine replacement

538

Public [Link] 538 3/6/08 [Link]

 6D Formulae R equired to Pass Part A

See Box 6D.9.1. McNemar’s test is used for matched analyses. McNemar’s test is used only for actual numbers: not
proportions, percentages, etc.
1. Ignore the concordant cells
2. Treat one of the discordant cells as A, and the other discordant pair as B (it makes no difference which way
round these are assigned)
3. Calculate c2 using the formula in Box 6D.9.1
4. Assess significance at the p = 0.05 level using the 3.84 cut-off.
An example is shown in Box 6D.9.2.

6D.10 STANDARDISATION – DIRECT AND INDIRECT

Standardisation is necessary to make fair comparisons between populations of differing demographic structures,
where simply using crude mortality or morbidity rates would be misleading.

DIRECT STANDARDISATION
See Box 6D.10.1.

Box 6D.10.1

1. Begin with a reference population*

2. Break down the size of the reference population into individual age bands
3. Take the age-specific mortality rates for the comparator population, and multiply them by the size
weighting of the reference population
4. Sum the values in (3) to obtain the age-standardised mortality rate.
*One of the populations being compared, their average, or an outside population.

INDIRECT STANDARDISATION
See Box 6D.10.2.

Box 6D.10.2

1. Start with the stratum-specific death rates of a standard population (e.g. European Standard Population)
2. Use these to calculate expected number of deaths in the study population, according to its age and sex
structure
3. Add up the expected number of deaths for each age band
Observed deaths
4. SMR = ¥ 100%.
Expected deaths

6D.11 WEIGHTED AvERAGES

See Box 6D.11.1 and, for an example, Box 6D.11.2.

539

Public [Link] 539 3/6/08 [Link]

 6D F o r mu l a e Req uired to Pass Part A

Box 6D.11.1

where x–1 = mean of sample 1, n1 = number in sample 1

and x– = mean of sample 2, n = number in sample 2
2 2

Box 6D.11.2
Example: weighted averages
Two A&E departments are striving to meet a government target of seeing and treating their patients within 4 h
of presentation. In the month of October, Department A sees 80% of its patients within 4 h, and Department B
sees 90% of its patients within 4 h. During that month, Department A saw 3690 patients and Department B saw
2697 patients.

Arithmetic mean = 85% of patients seen within 4 h

Weighted mean =

= (80 x 3690) + (90 x 2697)

(3690 + 2697)
= 84.2%

6D.12 CONFIDENCE INTERvALS AND STANDARD ERRORS OF THE MEAN

See Box 6D.12.1 and Section 1B.3.

Box 6D.12.1

Standard error (mean) =

95% confidence interval = sample value ± (1.96 ¥ standard error)

where s = standard deviation for the sample, n = number in the sample

540

Public [Link] 540 3/6/08 [Link]

 Appendix A R evision Tips

Appendix A
Revision Tips

The revision tips here are generally focused on the UK MFPH examination, although the principles of exam technique
will apply to any exam.
The UK MFPH Part A examination is notorious for its low pass rate: barely one candidate in five passes at some
sittings. The elaborate marking algorithm ensures that candidates who do not perform consistently well in all
questions will be heavily penalised. Accordingly, we advise you to ensure that:
• Your knowledge of the syllabus is broad rather than deep
• You divide your time in the examination across the questions (and sub-sections of questions) proportionately
according to the marks available for each sub-question.

EXAMINATION STRUCTURE
UK Passing the exam (known formally as the ‘Part A Examination for Membership of the Faculty of Public Health of
the Royal Colleges of Physicians of the United Kingdom’) entitles you to apply for:
• Diplomate membership of the Faculty of Public Health (DFPH)
• Entry to the Part B examination – success in which leads to full membership of the Faculty (MFPH).
Although the Part A examination actually consists of four papers (one each morning and afternoon over two
consecutive days), the Faculty’s marking scheme refers simply to papers I and II, one on each day.

DAY 1
Session Paper Duration Description Details Marks Timing
am Paper Ia 2½ hours Short-answer Research methods (epidemiology, 60 6 questions in
questions statistics, qualitative research, 150 min
(SAQs) health information sources), health
= 25 min per
promotion and health protection
question
pm Paper Ib 1½ hours SAQs Medical sociology, social policy, 40 4 questions in
health economics and organisational 90 min
management of health care
= 22½ min per
question

541

Public [Link] 541 3/6/08 [Link]

 A p p e nd i x A R evision Tips

DAY 2
Session Paper Duration Description Details Marks Timing
am Paper 2½ hours Critical Candidates are provided with a 50 50 marks in 150 min
IIa appraisal and journal paper to read (usually
= allow
discussion from the BMJ)
approximately 1 hour
to read and digest
the paper, then
Questions include composing
roughly 20 min per
a structured abstract of the
10 marks
paper, critically appraising the
paper and then addressing more
general questions on the topic
pm Paper 1½ hours Data handling Candidates are provided with 50 50 marks in 90 min
IIb and strategy evidence in the form of raw
= roughly 20 min per
writing data, tables of data, maps or
10 marks
graphs

Questions involve
interpretation of these data
(which may include statistical
manipulation) and then the
writing of a policy document
based on the above

Each of the two papers carries 100 marks, so that the examination is marked out of 200 in total. There is a complex
marking algorithm which contains six hurdles. A candidate who stumbles on any of these hurdles will fail the
examination.

542

Public [Link] 542 3/6/08 [Link]

 Appendix A R evision Tips

Across exam
Has candidate scored YES NO
100/200 or more in total?

Paper l
Has candidate scored 50/100? YES NO

Passed 7 questions?
YES NO

Is more than one FAIL

question marked 0, 1 or 2? NO YES

Paper ll
Has candidate scored 50/100? YES NO

Is a question marked 20/50 or less? NO YES

PASS

Reproduced from [Link]/exams/downloads/PtI_marking_algorithm.pdf.

In order to pass the examination overall, candidates must pass both paper I and paper II. However, if candidates
pass only one of the papers (score 50/100 or more) but also obtain an overall score of at least 100/200, then that
paper can be banked. Once a paper has been banked, candidates need only attempt the remaining paper when they
re-sit the examination on a subsequent occasion.
In practice the intricacies of the algorithm are irrelevant, and the bottom line is that you must avoid doing badly in
any question.

REVISION
The amount of time required for revision obviously depends on your prior knowledge and experience, and on your
personal revision style. That said, most successful candidates seem to devote at least 4–6 months to revision. A
good way to gauge how much work you will need to do is to familiarise yourself with the syllabus and with one of
the past papers and examiners’ comments. We would advise you not to look at the most recent paper until a week
before the exam, when you should use it as a mock under self-imposed examination conditions.
On page 544 is a suggested revision schedule that worked for us – and which you might want to use or adapt to
your personal revision style.

543

Public [Link] 543 3/6/08 [Link]

 A p p e nd i x A R evision Tips

Revision Subject Details

phase matter
Initial Syllabus • Read the syllabus and scan through the contents of this book
Sample past • Read a sample past paper (not the most recent), and the associated
paper examiners’ comments to gauge the required standard
Early Epidemiology • Ensure that you are au fait with all of the epidemiological components of the
syllabus. Epidemiology is the backbone of the examination and it is tested
not only in the epidemiology questions in paper Ia, but also throughout
paper II
• Study Chapter 1 of this book and read the recommended epidemiology text
below
Mid Cover the • Study the rest of this book
whole syllabus • Consult the further reading as recommended below
• Work your way through the CASP workbook and CD-ROM (crucial for paper IIa)
Past papers Again setting aside the most recent past paper, work your way through the
previous five or six past papers as follows:
• For each question, write down the introductory sentence that you would use
and the essay structure that you would employ
• Compare what you have written with the examiners’ key points and comments
Late Context You can add weight to your answers by adding material and examples from the
following:
• Selected editorials from the last few months’ editions of the BMJ (see Dr
Edmund Jessop’s weekly reading list [[Link]])
• Items on the websites of the following organisations: King’s Fund (www.
[Link]), Food Standards Agency ([Link]), National
Statistics ([Link]), National Institute for Health and
Clinical Excellence ([Link]), Healthcare Commission (www.
[Link]), Health Protection Agency ([Link]),
Department of Health ([Link] ) and the WHO ([Link] )
Week before Mock paper Exactly 1 week before the exam you should attempt the most recent past paper in
‘real time’ and under exam conditions. Write your answers longhand and use only
the materials that will be available to you in the examination itself ([Link].
uk/exams/part_1/exam_preparation.asp)
Finalise Consolidate and review your revision notes
Short-term Learn the following:
memory • Question timings/marks-per-minute rates for each paper
• Statistical formulae
• Key definitions (infant mortality, etc.)
• Essay structures

We know from personal experience that this book contains more than enough information to pass the MFPH Part A.
The books below were useful for us for addressing uncertainties.

544

Public [Link] 544 3/6/08 [Link]

 Appendix A R evision Tips

Subject area Further reading

Epidemiology Hennekens C, Buring J (1987) Epidemiology in Medicine. Philadelphia: Lippincott
Williams & Wilkins
Statistics Swinscow TDV, Campbell MJ (2002) Statistics at Square One. London: BMJ Books
Campbell MJ (2001) Statistics at Square Two. London: BMJ Books
Health information [Link]; [Link]
Health promotion Nutbeam D, Harris E (2004) Theory in a Nutshell: A Guide to Health Promotion Theory.
Maidenhead: McGraw-Hill Education
Environmental public Chapter within Donaldson LJ, Donaldson RJ (2003) Essential Public Health. Petroc
health Press
Communicable disease Hawker J et al (2005) Communicable Disease Handbook. Oxford: Blackwell Science
Medical sociology Scambler G (2003) Sociology as Applied to Medicine. Philadephia: WB Saunders
Qualitative research Green J, Browne J (2005) Principles of Social Research. Milton Keynes: Open
University Press
Health economics Wonderling D et al (2005) Introduction to Health Economics. Milton Keynes: Open
University Press
Health care management Iles V (2005) Really Managing Health Care. Oxford: Oxford University Press
Critical appraisal CASP: Evidence-based Health Care Workbook and CD-ROM. Update Software Ltd (www.
[Link])
Public health practice Pencheon D et al (2006) Oxford Handbook of Public Health Practice (2nd ed). Oxford:
Oxford University Press
Strategy writing Model answer Paper 2a, June 2005
Clear communication The Economist Style Guide (2003) London: Economist Books

EXAMINATION TECHNIQUE
Unfortunately the examination papers rarely start exactly at the published time: a delay of 5–10 min is usual.
This delay can make the timing of questions rather tricky, so you will need to spend the first minute or two of the
exam calculating the exact times at which you should start each question. Write down these times alongside each
question on the question paper. Note that the short answer questions (SAQs) are of different durations in paper Ia
(25 min each) and paper Ib (22½ min each).
You need to remain acutely aware of the time throughout the examination. Therefore, during the MFPH Part A
examination you should regard your watch as the equivalent of the rear-view mirror in a driving test: force yourself
to keep looking at it frequently. This is the only way to ensure that you allow a proportionate amount of time to
each question and sub-question.

PAPER I
As the questions in all parts of paper I are compulsory, there is little to be gained from reading through the whole
question booklet at the start. Instead we would suggest treating each SAQ as a mini-exam, and then starting each
new question afresh at the calculated time.
The first 5 minutes of each SAQ should be spent on planning. You are advised to use this time to:
• Re-read the question and underline the keywords

545

Public [Link] 545 3/6/08 [Link]

 A p p e nd i x A R evision Tips

• Brainstorm the question, i.e. write down the key elements that you think the examiners are expecting you to
cover, together with any ‘gems’ that you can throw in from your own work experience or from your revision
(particularly the contextual material that you may have read in the later stages of revision) to impress the
examiners
• Choose a structure (either one of the structures in Appendix B or a bespoke structure for that question)
• Craft the first sentence of your answer very carefully: first impressions count
• There is generally no need for a conclusion.
Note that where a question asks you to answer in relation to a ‘country of your choice’, you are expected to state
that country explicitly at the beginning of your answer.

PAPER II
In contrast to paper I, you should read through all of the questions as soon as you are allowed to turn over the
paper. This is because the questions build on each other and therefore give an indication of the examiners’ line of
thought.
For paper IIa we would advise you to set aside the first hour to read and digest the study that you are being
required to appraise critically. We suggest that you follow Dr Edmund Jessop’s advice (see [Link].
uk/partI_main.doc) of reading the following parts of the paper first:
• Title
• Last paragraph of the introduction
• First paragraph of the discussion
• Last paragraph of the discussion.
This will provide you with the gist of the paper (and therefore the bulk of the abstract that you will be required to
write) and you should try to make this completely clear in your mind before you read any further.
For paper IIb you should begin by studying and describing in a systematic way any data presented to you (see
6B1). Medically qualified candidates will be familiar with the standard way to report a chest radiograph, namely:
• Type of image
• Name and date of birth of the patient
• Date of examination
• Striking features
• Systematic approach to bones, soft tissues, zones of the lungs, etc.
For example, the report may read, ‘This is a postero-anterior chest radiograph of Mr David Jones (DOB 24/7/46) taken
on 14 June 2006. The most striking abnormality is a left-sided pneumothorax. The bones appear normal …’ .
You should adopt a similar methodical approach to whatever data source you are asked to describe in paper IIb –
paying particular attention to any axes, units or denominators shown.
Later in the paper – when it comes to writing the policy document – you should build on your answers to earlier
questions. You should refer to what you wrote earlier in the exam but you must not regurgitate the same material.

TECHNIQUE FOR SHORT ANSWER QUESTIONS

For each SAQ you must convince the examiners that you understand:
1. What exactly the question is asking
2. Why this question is being asked today (i.e. why it is important to contemporary public health)
3. How to set about tackling the question in a logical fashion.
We would encourage you to approach each SAQ by covering these three points. Always begin with a carefully
constructed sentence that teases out the issues in the question and explains why they are important to the world of

546

Public [Link] 546 3/6/08 [Link]

 Appendix A R evision Tips

public health today. Next, set out the structure that you are going to adopt to answer the question, i.e. the answer
framework.
You should then proceed to answer the question, writing out and underlining the individual headings contained in
your box as you go along.
Aim for a clear writing style that follows the advice of the Economist’s Style Guide, i.e. you should:
• Use short, snappy sentences
• Avoid trite turns of phrase
• Steer clear of unnecessary jargon.
The examiners encourage you to employ the ‘rationed use of bullet points, tables and diagrams’ to illustrate your
answers. You should aim to include relevant examples, and to name any eponymous theories or structures that you
use.
Example (January 2006, Paper Ia, Question 1): Describe how you would undertake a formal survey to
determine the prevalence of angina in a local area of a developed country (population 100 000), e.g. the
United Kingdom.

The public health importance of angina is that:

(a) it is an indicator of unmet need in relation to the treatment of ischaemic
heart disease (currently the commonest cause of death in developed
countries)
(b) in most cases, it is treatable either medically or through revascularisation
(surgically or percutaneously)
(c) differences in its prevalence can therefore be used to compare inequalities in
access to health care within and across populations.
1. Definitions (setting; angina; formal survey)
2. Steps:
• Obtaining a sampling frame
• Obtaining a sample
• Assessing presence of angina
• Calculation of prevalence
3. Strengths and weaknesses of method chosen

Definitions
I shall conduct this survey in the relatively deprived inner-London borough of
Islington, UK. Although there are many types of angina (e.g. Ludwig’s angina,
Prinzmental’s angina), for the purpose of this study, angina will be defined as
stable angina of cardiac origin, specifically …

547

Public [Link] 547 3/6/08 [Link]

Public [Link] 548 3/6/08 [Link]
 Appendix B Answer Frameworks

Appendix B
Answer Frameworks

A recurrent gripe mentioned in the examiners’ comments is that candidates’ answers lack structure. To avoid this,
candidates must ensure that – without exception – every word that they write during the examination fits into a
structure.
Answers can be organised using one of the frameworks below, or a framework can be custom-made for a particular
question during the examination. Either way, it should be made explicit to the examiners what framework is being
used (e.g. by drawing a box at the start of each answer and writing the framework for that answer inside the box).
The contents of the box should be repeated as headings within the answer, and these should be underlined.

GENERIC FRAMEWORKS
The following two structures can be used either as the stand-alone framework for an entire question (especially
if none of the more specific frameworks fits) or alternatively as the subheadings for a component of another
framework.

Jessop framework Areas of public health

Definition Health intelligence

Example Health services
Advantages Health promotion
Disadvantages Health protection

549

Public [Link] 549 3/6/08 [Link]

 A p p e nd i x B A nsw er Frameworks

PAPER I FRAMEWORKS
NEEDS ASSESSMENT

Needs for a population group Needs for a disease

Physical Public health Basic (food, water, shelter) Epidemiological Definition

Lifestyle (smoking, alcohol) Numbers
Screening Current set-up
Immunisations Alternative set-ups

Primary care Medical Comparative Other places

Dental Gold standard
Pharmacy
Corporate Central government
Secondary care Physical Health authority
Mental

Mental

Social

550

Public [Link] 550 3/6/08 [Link]

 Appendix B Answer Frameworks

HEALTH CARE EvALUATION

The names of Maxwell or Donabedian should be quoted if their frameworks are used.

Donabedian Maxwell

Structure Staff numbers A Access

Staff qualifications Acceptability
Bed capacity Appropriateness

Process Admissions E Equity

Efficiency/economy
Procedures
Effectiveness

Outcomes Survival
R Relevance

Quality of life

Ongoing evaluation (‘HADEPO’)

Health Public health indices
Access Equality/equity
Delivery Evidence-based practice
Efficiency Costs
Patient experience Questionnaires
Outcome Audit cycle

551

Public [Link] 551 3/6/08 [Link]

 A p p e nd i x B A nsw er Frameworks

HEALTH INFORMATION

Information sources

Mortality

Morbidity Morbidity information

Surveys Hospital Laboratory

Inpatient
Research Outpatient
A&E attendances
Non-NHS Fire/Police etc
Primary care Medical
Dental
Pharmacy
NHS Direct
Health information
systems
Registers Cancer
Capture Congenital abnormalities
Coding Transplants
Output Prostheses

552

Public [Link] 552 3/6/08 [Link]

 Appendix B Answer Frameworks

COMMUNICABLE DISEASE

Epidemiology of a communicable disease Control of a communicable disease

‘AgORMICS’ ‘PIDQuICS’

Agent Virus/bacteria/protozoa Prevention

Illness caused
Method of diagnosis Isolation

Occurrence In named country Disinfection

Seasonal pattern/sporadic/imported
Reservoir Quarantine

Mode of transmission Parenteral/Faeco-oral/other Immunisation

Incubation (Omit if unsure) Contacts

Communicability e.g. communicable while still Specific measures

excreting in stool

Acknowledgement: Edmund Jessop’s notes

Susceptibility e.g. infection confers resistance

Acknowledgement: Edmund Jessop’s notes

Outbreak investigation
CCDCs HATE IT
Count cases Test hypotheses
Control outbreak Epidemic confirmation
Diagnosis verified
Communication Identify cases
Surveillance enhanced Tabulate data
Hypothesis formulation
Additional microbiology samples sent Acknowledgement: NCL revision
course notes

553

Public [Link] 553 3/6/08 [Link]

 A p p e nd i x B A nsw er Frameworks

MISCELLANEOUS

Needs Harms Lalonde health Epidemiology

fields

Felt Source Time

Lifestyle
Expressed Path
Place
Health service
Normative Receptor
Person Age
Environment
Sex
Class
Genes
Ethnicity
Occupation

Intervention

PAPER IIA FRAMEWORKS

ABSTRACT

Abstract (taken from BMJ)

Objective
Design
Setting
Participants
Intervention
Outcome measures
Results
Conclusions

554

Public [Link] 554 3/6/08 [Link]

 Appendix B Answer Frameworks

CRITICAL APPRAISAL

RCT Meta-analysis

Screening Focused issue (population/ Screening Focused (population/

intervention/outcome) intervention/outcome)
Randomisation Randomised
Follow-up complete
----------------------------------------------------------- -----------------------------------------------------------
Fairness Triple blinding
Patient characteristics equal Rigour Relevant studies included
both groups
Databases
Identical management except
Reference lists
intervention
Personal contacts
Unpublished work
Results Treatment effect
Non-English work
Confidence intervals and p
value quoted Assessed quality
Reasonable to merge
(similar results)
Locally applicable Similar population
All potential outcomes
considered (e.g. QoL) Results Bottom line (odds ratio/
NNT)
Cost:benefit
Confidence intervals

Locally applicable Similar population and

setting
All outcomes
Cost:benefit

Acknowledgement: Critical Appraisal Skills Programme (CASP) [Link].

555

Public [Link] 555 3/6/08 [Link]

 A p p e nd i x B A nsw er Frameworks

PAPER IIB FRAMEWORKS

Health impact assessment Health needs assessment (HNA)

Screening (Define the problem) Objective

Steering group (Define the options) Setting

Negotiation (Options appraisal) Methods (HNA for a group or for

a disease – as in the
frameworks for paper I)
Implementation

Results
Monitoring

Conclusions
Evaluation

Dissemination of evaluation

Service specification Strategy

Pathway Diagnosis
Patient empowerment Current Where are we now?

Medication
Interventions Future Where are we going?

Clinical guidelines Best practice Strategy How do we get there?

Type and dose of drug

}
Skills Monitoring

Minimum activity Are we there?

Evaluation

Clinical governance
Skill mix
Accreditation

556

Public [Link] 556 3/6/08 [Link]

 Appendix B Answer Frameworks

Briefing Letter

Preamble Title Business letter layout

Author
Thank you

Audience
Acknowledge concerns
Date
Background to the issues raised
Purpose Strategy
Evidence
Methods

Statement of current situation

Results

Conclusions Situation will be kept under review

Executive summary Please contact again if

Introduction and background

Key issues

Options

Recommendations

Timetable

557

Public [Link] 557 3/6/08 [Link]

Public [Link] 558 3/6/08 [Link]
 Appendix C Top Fives

Appendix C
Top Fives

In the examination, you will need to quote standard criteria, classic studies and key names in order to demonstrate
your knowledge of the specialty of public health. We suggest that you use the lists below as a starting point for
creating your own lists which you then memorise prior to the examination.

EPIDEMIOLOGY
Concept More information in
Section
John Snow (1854) Epidemiological method: cholera and Broad 5D
Street pump
Richard Doll and Austin Bradford Doctors’ cohort: Smoking causes lung cancer
Hill (1954)
Archie Cochrane (1972) Evidence-based medicine 1A
David Barker (1987) Risk of coronary heart disease in adults is 2A
linked to in utero development
Geoffrey Rose (1992) Population-based prevention, and the 2H
prevention paradox

UK UK PUBLIC HEALTH POLICY

Concept More information in
Section
Chadwick (1842) Sanitation and health 5D
Tudor Hart (1971) Inverse care law 1C
Black (1980) Health inequalities are growing in the UK and 2I
are linked to social class
Acheson (1998) Reducing health inequalities under the new 2I
Labour government
Wanless (2004) Economic case for investing in public health 2I

559

Public [Link] 559 3/6/08 [Link]

 A p p e nd i x C Top Fives

HEALTH PROMOTION: FRAMEWORKS

Concept More information in
Section
Tannahill (1985) Health promotion = overlapping spheres of 2H
protection, prevention and education
Lalonde (1974) Health field concept 2H
Dahlgren and Whitehead (1991) Policy rainbow 2H
Evans and Stoddart (1990) Health field model 2H
Diderichsen and Hallqvist (1998) Social determinants 2H

HEALTH PROMOTION: MODELS

Concept More information in
Section
Hochbaum et al (1958) Health belief model 2H
Bandura (1970s) Social learning 2H
Prochaska and DiClemente (1984) Stages of change 2H
Beattie (1991) Dimensions of health promotion: authoritative 2H
– negotiated; individual – collective
Ewles and Simnet (1995) Health field concept 2H

SOCIOLOGY
Concept More information in
Section
Talcott Parsons (1951) The sick role 4A
Edwin Lemert (1967) Primary and secondary deviance 4A
Irving Goffman (1963) Stigma; institutionalisation 4A
Ivan Illich (1975) Iatrogenesis 4A
Emile Durkheim (1897) Social integration and suicide
John Rawls (1971) Social justice 4C

MANAGEMENT
Concept More information in
Section
Donabedian (1966) Health service quality: structure–process– 1C
outcome
Maxwell (1984) Health service quality: access, equity, 1C
efficiency, effectiveness, economy,
appropriateness, acceptability
Belbin (1996) Team roles 5A
Maslow (1943) Hierarchy of needs 5A, 5B
Handy (1987) Organisational types 5A

560

Public [Link] 560 3/6/08 [Link]

 Appendix C Top Fives

KEY STUDIES
STUDY DESIGN ExAMPLES
Study design Example Participants Major finding(s) More information
RCT Women’s health 64 500 women over 15 Contrary to observational Wassertheil-Smoller
initiative (2003) years study evidence, HRT does S et al 2003. JAMA
not protect against CHD 289:2673–2684
Cohort Whitehall (1967 Whitehall I: 18 000 male Risk of death from CHD [Link]/
onwards) civil servants linked to social status/ whitehallII
employment grade
Whitehall II: 10 000 male
and female civil servants
Case–control UK Childhood Cases: records of 30 000 Proximity to powerlines at Draper G et al 2005.
Cancer Study children with cancer birth linked to childhood BMJ 330:1290
(1999) leukaemia
Controls: children matched
for age, sex, area of birth
Cross- Health Survey for 6000 adults + 3000 Prevalence of risk factors [Link]/
sectional England (annual) children (in 2004 – random and health behaviours, pubs/healthsurvey
selection and boosted e.g. fruit and vegetable 2004ethnicfull/
sample in high minority consumption hse2004vol1/file
ethnic group areas)
Case reports Pneumonia Five homosexual men aged Abnormal epidemiology [Link]/
jirovec 29–36 of PCP, early sign of MMWR/preview/
pneumonia (PJP) emergence of AIDS mmwrhtml/june_5.
in Los Angeles htm
(1981)

EFFECTS OF DIET ON HEALTH

Study design Major finding(s) More information
North Karelia Intervention (before/after): Altering lifestyle (including Section 2E
(1972–1982) residents of North Karelia saturated fat intake) reduced
CHD mortality
Framingham Cohort: 5000 residents of Factors affecting risk of CHD [Link]/about/
(1948 …) Framingham, MA include cholesterol framingham/[Link]

7 Countries (1958– Ecological cohort: 11 000 Link between unsaturated fat Section 2E
1970) men aged 40–59 in 7 and lower risk of death from
countries in Europe and the CHD
USA
Intersalt (1988) Cross-section: 10 000 men Dietary salt levels linked to Section 2E
and women blood pressure
UK Women’s cohort Cohort: 35 000 women in UK Exploring links between diet [Link]/medicine/
(1993 …) and cancer ceb/NutEp/ukwcs/
[Link]

561

Public [Link] 561 3/6/08 [Link]

 A p p e nd i x C Top Fives

OTHER REVISION LISTS

UK SCREENING: UK NATIONAL SCREENING COMMITTEE SUMMARY
The condition: • An important health problem
• Epidemiology and natural history understood
• Cost-effective primary prevention interventions implemented first
The test • Simple, safe, precise and validated
• Distribution of test values in the target population should be known and a cut-off
level defined and agreed
• Acceptable to the population
• Policy on diagnostics for individuals with a positive test and choices available
• Criteria to select mutations to be covered by screening
The treatment • Effective treatment exists for patients identified through early detection
• Early treatment has better outcomes than late treatment
• Agreed criteria for which individuals to be offered treatment and what should be
offered
• Providers prepared to manage patients before programme starts
The programme: • Reduces mortality or morbidity
• Where screening aimed solely at providing ‘informed choice’, test accurately measures
risk, provides valuable information readily understood by the individual being screened
• Complete programme clinically, socially and ethically acceptable to health
professionals and the public
• Benefit outweighs harm (physical and psychological)
• Value for money
• Managing and monitoring arrangements in place
• Adequate resources available – staffing and facilities for testing, diagnosis, treatment
and programme management
• Other options for managing the condition considered
• Evidence-based information for potential participants explaining screening
consequences
• Anticipate public pressure for widening eligibility criteria. Decisions about screening
parameters should be justifiable to the public.
More information • Section 2C
• [Link]/pdfs/[Link]
• Wilson and Jungner (1968)

562

Public [Link] 562 3/6/08 [Link]

 Appendix C Top Fives

CAUSATION AND ASSOCIATION

Bradford Hill ‘viewpoints’ for studying causation (Bradford Hill 1965) (see Section 1A)

Strength High relative risk or odds

Consistency Similar results from several studies in different populations
Specificity Single cause produces a single effect
Temporality Cause must precede effect
Biological gradient Dose–response curve
Plausibility Biologically acceptable or relevant reason for the cause to produce effect
Coherence Does not conflict with current knowledge
Experimental Introduction or removal of putative cause leads to change in effect
Analogy Consistent with previous experience in similar situations

563

Public [Link] 563 3/6/08 [Link]

Public [Link] 564 3/6/08 [Link]
 Further R eading

Further Reading

CHAPTER 1A

BOOKS
Donaldson LJ, Donaldson RJ (2003) Essential Public Health. Oxford: Petroc Press. Contains readable summaries on a
wide range of subjects in the Part A syllabus.
Hennekens CH, Buring JE (1987) Epidemiology in Medicine. Philadelphia: Lippincott, Williams & Wilkins. Comprehensive
description of epidemiology concepts, study designs and analysis.
Kirkwood BR, Sterne J (2003) Essential Medical Statistics (2nd ed). Oxford: Blackwell Science. In-depth description of
statistics and their use in epidemiological studies including key elements required for Part A.
Petrie A, Sabin C (2005) Medical Statistics at a Glance (2nd ed). Oxford: Blackwell Publishing. Short summary covering
all key elements of statistics required for Part A with examples from epidemiological studies.
Swinscow TDV, Campbell MJ (2002) Statistics at Square One (10th ed). London: BMJ Books. Guide to statistics,
containing several exercises to practice calculations and embed understanding of statistical concepts. 9th edition
available online at: [Link]/statsbk.

ARTICLES
Northridge ME (1995) Public health methods – attributable risk as a link between causality and public health action.
American Journal of Public Health. 85:1202–4. Provides examples of calculations of attributable risk percent and
population attributable risk percent, and illustrates how it can be used to inform public health priorities.

WEBSITES
Centre for Evidence-Based Medicine − [Link]. Promotes evidence-based health-care and provides support and
resources to anyone who wants to make use of them.
Cochrane Collaboration − [Link]. Produces and disseminates systematic reviews of health-care interventions
and promotes the search for evidence in the form of clinical trials and other studies of interventions.
GreyNet – [Link]. Grey literature network service which facilitates dialogue, research and communication
between people and organisations in the field of grey literature.

565

Public [Link] 565 3/6/08 [Link]

 F u r t he r R e a d i ng

National Institute for Health and Clinical Excellence − [Link]. Organisation responsible for providing national
guidance on the promotion of good health and the prevention and treatment of ill health (merged with the Health
Development Agency in 2005).
National Research Ethics Service (NRES) − [Link]. NRES is part of the National Patient Safety Agency.
It coordinates certain parts of the research ethics approval process and provides research ethics approval process, and
guidance, training and support to local committees and applicants.
Office of National Statistics − [Link]. Compiles and publishes statistics on Britain’s population and
society at national and local level.
Wikipedia – [Link]/wiki/Main_Page. Reference resource that anyone can contribute to, containing within
its 1000 000 articles, information on many of the subjects in the Part A syllabus.

CHAPTER 1B

BOOKS
Campbell MJ (2001) Statistics at Square Two: Understanding modern statistical applications in medicine. London: BMJ
Books.
Kirkwood BR, Sterne J (2003) Essential Medical Statistics (2nd ed). Oxford: Blackwell Science. In-depth description of
statistics and their use in epidemiological studies including key elements required for Part A.
Petrie A, Sabin C (2005) Medical Statistics at a Glance (2nd ed). Oxford: Blackwell Publishing. Short summary covering
all key elements of statistics required for Part A with examples from epidemiological studies.
Pereira-Maxwell F (1998) A-Z of Medical Statistics: A companion for critical appraisal. London: Hodder Arnold.
Rowntree D (1991) Statistics without Tears: An introduction for non-mathematicians. London: Penguin Books Ltd.
Simple introduction to statistics, which assumes no previous knowledge of the subject.

ARTICLES
Altman DG, Bland JM (1996) Education and debate. Statistics notes: Comparing several groups using analysis of
variance. BMJ 312:1472−3.
Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ
315:629–34.
Harrison WN, Mohammed MA, Wall MK, Marshall TP (2004) Analysis of inadequate cervical smears using Shewhart
control charts. BMC Public Health 4:25.
Spiegelhalter DJ (2005) Handling over-dispersion of performance indicators. Quality and Safety in Health Care
14:347−51. Discusses funnel plots in the context of different statistical methods for dealing with variations in
available data.
Sterne JAC, Egger M, Davey Smith G (2001) Systematic reviews in health care: Investigating and dealing with
publication and other biases in meta-analysis. BMJ 323:101−5. Describes statistical and graphical methods for
detecting and correcting for bias.
Tekkis PP, McCulloch P, Steger AC et al (2003) Mortality control charts for comparing performance of surgical units:
validation study using hospital mortality data. BMJ 326:786–8.

566

Public [Link] 566 3/6/08 [Link]

 Further R eading

WEBSITES
Bandolier − [Link]/bandolier. Monthly journal and resources on evidence-based health care. The Learning
Zone section is particularly useful for Part A.
Children’s Mercy Hospitals and Clinics − [Link]/stats/definitions/[Link]. Explanation of
conditional probability.
Public Health electronic Library – [Link]. The health knowledge section contains succinct
summaries on key subjects tested at Part A.

CHAPTER 1C

BOOKS
Fulop N, Allen P, Clarke A, Black N (eds) (2001) Studying the Organisation and Delivery of Health Services: Research
methods. London: Routledge. Health services research from the perspective of several different research disciplines,
including action research, organisational psychology and operational research.
McPake B, Kumaranayake L, Normand C (2002) Health Economics: An international perspective. London: Routledge. An
introduction to health economics in three sections − covering supply and demand concepts and markets, economic
evaluation and a comparison of different health-care systems.
Pencheon D, Guest C, Melzer D, Muir Gray JA (eds) (2006) The Oxford Handbook of Public Health Practice (2nd ed)
(Oxford Handbooks Series). Oxford: Oxford University Press. Covers many essential public health topics including needs
assessment, advocacy and management issues with step-by-step descriptions and practical examples.
Wright J (ed) (1998) Health Needs Assessment in Practice. London: BMJ Books.

ARTICLES
Bradshaw J (1972) A taxonomy of social need. New Society March:640−3.
Carr-Hill RA (1992) The measurement of patient satisfaction. Journal of Public Health Medicine 14:236−49.
Lock K (2000) Health impact assessment. BMJ 320:1395−8.
McColl A, Roderick P, Gabbay J, Ferris G (1998) What do health authorities think of population based outcome
indicators? Quality in Health Care 7:90−7.
National Institute for Clinical Excellence (2002) Principles for Best Practice in Clinical Audit. Oxford: Radcliffe Medical
Press. Also available online at: [Link]/[Link]?o=29058. A comprehensive guide to practising clinical
audit in the health service.

WEBSITES
Department of Health − [Link]. Government department in England responsible for health-care policy and the
strategic direction of health care.
Healthcare Commission − [Link]. Responsible for reviewing the performance and quality
of NHS and private health care.
National Institute for Health and Clinical Excellence − [Link]. Responsible for providing national guidance
on the promotion of good health and the prevention and treatment of ill health.

567

Public [Link] 567 3/6/08 [Link]

 F u r t he r R e a d i ng

CHAPTER 1D

BOOKS
Green J, Thorogood N (2004) Qualitative Methods for Health Research. London: Sage Publications Ltd. A clearly
written, practical guide that contains case studies illustrating how to carry out qualitative research and some of the
practical and theoretical issues that need to be addressed.

CHAPTER 2A

ARTICLES
Ben-Shlomo Y, Kuh D (2002) A life course approach to chronic disease epidemiology: conceptual models, empirical
challenges and interdisciplinary perspectives. International Journal of Epidemiology 31:285−93. This editorial
accompanies a series of articles with a ‘life course’ theme. It outlines some of the major key conceptual issues around
life course epidemiology.

CHAPTER 2B

WEBSITES
Association of Public Health Observatories − [Link]/apho. For statistics and reports relating to health in
regional areas.
Clinical and Health Outcomes Knowledge Base − [Link]. For the compendium of 500 indicators enabling
health comparisons at regional and organisational levels.
Department of Health − [Link]. For national strategies, e.g. National Service Frameworks, statistics or guidance
relevant to England.
National Institute for Health and Clinical Excellence − [Link]. For nationally recommended cost-effective
interventions (technology assessments) and guidelines on the management of specific conditions.
World Health Organization − [Link]. For international guidelines and statistics. In particular, the global
burden of disease estimates are useful for key diseases nationally and internationally, available at: [Link]/
healthinfo/statistics/[Link].
Major charity websites, e.g.:
• Mental health: MIND − [Link]; Rethink − [Link]; Alzheimer’s Society − [Link].
uk
• Cardiovascular disease: British Heart Foundation − [Link]; Stroke Association − [Link]
• Long-term conditions: Diabetes UK − [Link]; Asthma UK − [Link]; Multiple Sclerosis
Society − [Link].

CHAPTER 2C

ARTICLES
Deeks JJ, Altman DG (2004) Diagnostic tests 4: likelihood ratios. BMJ 329:168–9.
Department of Health (2002) Screening/Case Finding: National Service Frameworks: A practical aid to implementation in
primary care. London: Department of Health. Available online at: [Link]/assetRoot/04/05/08/68/04050868.
pdf.

568

Public [Link] 568 3/6/08 [Link]

 Further R eading

Gaeta T (2005) Screening and Diagnostic Tests. eMedicine, WebMD. Available online at: [Link]/emerg/
[Link].
Holtzman NA, Shapiro D (1998) Genetic testing and public policy. BMJ 316:852−6.
Karimi A, Kadivar MR, Fararoee M, Alborzi A (2000) Active case-finding of communicable diseases in the south of the
Islamic Republic of Iran. Eastern Mediterranean Health Journal 6:487−90.
Marks D, Wonderling D, Thorogood M et al (2000) Screening for hypercholesterolaemia versus case finding for familial
hypercholesterolaemia: a systematic review and cost-effectiveness analysis. Health Technology Assessment 4:1−123.
Marteau TM, Dormandy E, Michie S (2001) A measure of informed choice. Health Expectations 4:99−108.

WEBSITES
Jarrett J (2004) Health economics of screening − [Link]/resources/ppt/jj_guys_talk.pps. Cambridge
Genetics Knowledge Park and University of East Anglia Presentation at Guy’s Hospital, London.
National Library for Health (NLH) − [Link]. NLH organises and provides access to the best available
evidence on screening.
Public Health electronic Library – [Link]/publichealth.
UK National Screening Committee (NSC) − [Link]. The NSC advises Ministers, the devolved National Assemblies
and the Scottish Parliament on all aspects of screening policy. The website has information on the evidence base for
recommended programmes and those in discussion. In particular, UK National Screening Committee’s Policy Positions
− [Link]/pdfs/policy_position_chart_july06%5B1%[Link].
Genome programmes of the US Department of Energy Office of Science - [Link]. In particular, Human
Genome Project Information: Ethical, Legal, and Social Issues − [Link]/sci/techresources/Human_Genome/
elsi/[Link].

CHAPTER 2D

BOOKS
Zimmern R (2001) Genetics in disease prevention. In: Pencheon D, Guest C, Melzer D, Muir Gray JA (eds) Oxford
Handbook of Public Health Practice. Oxford: Oxford University Press, pp 544−9. Concise account of public health
genetics, linking concepts to issues that public health practitioners may encounter in their work.

ARTICLES
Kaslow RA, Moser SA (2000) Role of microbiology in epidemiology: Before and beyond 2000. Epidemiologic Reviews
22:131−5. Commentary on the ways in which molecular biological techniques have been used in epidemiology.
Kirk M (2005) The role of genetic factors in maintaining health. Nursing Standard 20:50−4. Introduction to genetics
in public health with an emphasis on use in clinical practice.

WEBSITES
Centers for Disease Control and Prevention, Office for Genomics and Disease Prevention − [Link]/genomics.
US-based website providing information about human genetic developments and potential uses in improving health
and preventing disease at the population level.
Foundation for Genomics and Population Health − [Link]/. Contains primers on basic genetics,
including genetic epidemiology, principles, techniques and applications to health and disease suitable for health-
care professionals.

569

Public [Link] 569 3/6/08 [Link]

 F u r t he r R e a d i ng

CHAPTER 2E

ARTICLES
Department of Health (2004) Choosing Health: Making healthy choices easier. London: Department of Health.
Available online at: [Link]. Public Health White Paper for England, including a range of dietary
interventions to reduce the prevalence of obesity.

WEBSITES
British Nutrition Foundation − [Link]. Charity providing evidence-based nutritional knowledge and
advice.
Food Standards Agency − [Link]. Agency for the English government and the devolved
administrations in the UK providing advice, information, monitoring and enforcement about food safety. Its website
contains nutritional advice, food safety information and results of nutritional surveys.
WHO Global Strategy on Diet, Physical Activity and Health − [Link]/dietphysicalactivity/en. These webpages
provide information on risk factors, WHO strategy and effective interventions related to diet.
WHO Child Growth Standards − [Link]/childgrowth/standards/en/. These webpages provide information
on recommended child height, weight, body mass index, for children up to age 5 and the background to their
development.

CHAPTER 2F

BOOKS
Donaldson L, Donaldson RJ (2003) Essential Public Health (2nd ed). Oxford: Petroc Press.
Yassi A, Kjellstrom T, de Kok T, Guidotti T (2001) Basic Environmental Health. Oxford: Oxford University Press.

WEBSITES
BBC Weather Centre: Climate Change − [Link]/climate/evidence. Basic guide to climate change, covering the
evidence base, impact and policies to reduce global warming.
Communities and Local Government − [Link]. Government department in England created in May
2006 that replaces many of the functions of the revised Office of the Deputy Prime Minister.
Department for Environment, Food and Rural Affairs − [Link]. Government department developing policy
around the interests of farmers and the countryside, the environment and the rural economy.
Department for Transport − [Link]. Government department developing policy around transport, which
includes environmentally friendly and sustainable transport options.
Drinking Water Inspectorate − [Link]. Agency responsible for assessing and enforcing the quality of
drinking water in England and Wales.
Environment Agency − [Link]. Government agency in England and Wales responsible for
inspecting and regulating businesses, providing information and advice on environmental issues and taking action,
after an environmental incident such as a flood or pollution.
Food Standards Agency − [Link]. Government department that provides public information on food safety,
circulates food alerts and food legislation, and undertakes local authority audits and other enforcement activities to
assure food safety and quality.

570

Public [Link] 570 3/6/08 [Link]

 Further R eading

Friends of the Earth − [Link]. UK-based charity campaigning for initiatives to address climate change.
Health Protection Agency − [Link]. In particular, for checklists for chemical incident management, see:
[Link]/chemicals/checklists/htm.
Sustainable Development − [Link]. Website for the government’s sustainable
development unit, based in the Department for Environment, Food and Rural Affairs (defra), provides information
and updates on sustainable development issues.
Sustainable Development Commission − [Link]. The government's independent advisory body for
England on sustainable development.
UK Air Quality Archive − [Link]/archive/[Link]. Provides general information on the causes and
nature of air pollution and current levels of local air pollution condition across Britain.
UN Economic Commission for Europe − [Link]/env/[Link]. The UN Economic Commission for Europe’s
environment website covers the formation of international policies, countrywide inspections and standards, and
international treaties on the environment.

CHAPTER 2G

BOOKS
Chin J (ed) (2000) Control of Communicable Diseases Manual (17th ed). American Public Health Association. A
relatively concise handbook for the management of communicable disease and the definitive text in the USA.
Department of Health (1996) Immunisation Against Infectious Disease 1996 − ‘The Green Book’. London: Department
of Health. Revised versions are available online at: [Link]. Contains the most recent national advice on
immunisation with descriptions of the epidemiology and clinical features of immunisable diseases. Several areas,
e.g. MMR, have been updated since 1996.
Donaldson L, Donaldson RJ (2003) Essential Public Health (2nd ed). Oxford: Petroc Press.
Hawker J, Begg N, Blair I et al (2005) Communicable Disease Control Handbook (2nd ed). Oxford: Blackwell
Publishing. UK-based publication describing the clinical features of communicable diseases and providing guidance
for management.

WEBSITES
Department of Health − [Link]. The Department of Health sets health policy for England. Policies on issues
relevant to infection control on the website include:
• The Health Bill (2006) – includes Hygiene Code of Practice for the Prevention and Control of Healthcare
Associated Infections
• Saving Lives: A delivery programme to reduce healthcare associated infections including MRSA (2005)
• Matrons’ Charter: An action plan to cleaner hospitals (2004)
• Getting Ahead of the Curve: A strategy for combating infectious diseases (including other aspects of health
protection) (2002).
Health Protection Agency − [Link]. The statutory body in England with responsibility for protecting
the public from communicable diseases and chemical, poisonous or radioactive hazards, and preparing for new
and emerging threats to health, such as bioterrorism or new disease strains. The HPA’s website contains succinct
descriptions of many communicable diseases. Its online journal, CDR Weekly, provides updates on the incidence of
major communicable diseases across England.

571

Public [Link] 571 3/6/08 [Link]

 F u r t he r R e a d i ng

National Resource for Infection Control − [Link]. Maintained by City University and funded by the
Department of Health, the site provides resources and links to up-to-date, UK-based information on infection control
for health care staff.
Public Health electronic Library − [Link]. The health knowledge communicable disease section
includes summaries on outbreaks, surveillance and immunisation.
World Health Organization − [Link]. The WHO website has several factsheets on infectious diseases, provides
descriptions of WHO activities, and downloadable reports – see Section 2H.

BOOKS
Naidoo J, Wills J (2000) Health Promotion: Foundations for practice. London: Baillière Tindall. Comprehensive
textbook covering health promotion, from theory to practice with examples from practice.
Nutbeam D, Harris E (1999) Theory in a Nutshell: A guide to health promotion theory. Maidenhead: McGraw-Hill
Education. Very short, digestible summaries of major health promotion theories.
Downie RS, Tannahill C, Tannahill A (1996) Health Promotion: Models and values (2nd ed). Oxford: Oxford University
Press. Textbook describes the Tannahill health promotion framework, and covers the policy and practice of health
promotion from various different clinical and social perspectives.
O’Sullivan GA, Yonkler JA, Morgan W, Merritt AP (2003) A Field Guide to Designing A Health Communication Strategy.
Baltimore: Johns Hopkins Bloomberg School of Public Health/Center for Communication Programs. Available at:
[Link]/iudtoolkit/marketing_comm/[Link]. A series of steps and tools for effective health
communication, including health behaviour theories to audience segmentation and communication strategy.

WEBSITES
World Health Organization – [Link]. A subsidiary of the United Nations, the most important source of
guidance and information regarding international health and health promotion initiatives.
Action on Smoking and Health (ASH) – [Link]. A campaigning organisation working to eliminate the
harm from tobacco; its website contains information about the effects of tobacco and useful summaries of current
tobacco policy both in the UK and globally.

CHAPTER 2I

ARTICLES
Cancer Research UK SunSmart Campaign Strategy. Available online at: [Link]/healthyliving/
sunsmart/forprofessionals/campaignstrategy. A description of the origins, research, range of activities and
evaluation of a national social marketing campaign in England to reduce harm from the sun.
Wanless D (2004) Securing Good Health for the Population. London: HMSO. Available online at: [Link].
uk/consultations_and_legislation/wanless/consult_wanless04_final.cfm. Derek Wanless’s second report into health,
commissioned by the Treasury, provides a summary of key public health policies in England in the twentieth and
early twenty-first century, and summarises the evidence base around smoking, falls, obesity/physical activity and
salt.

572

Public [Link] 572 3/6/08 [Link]

 Further R eading

CHAPTER 3A

BOOKS
Pencheon D, Guest C, Melzer D, Muir Gray JA (eds) (2006) The Oxford Handbook of Public Health Practice (2nd ed)
(Oxford Handbooks Series). Oxford: Oxford University Press. Includes chapters on information collection and use.
In particular, see Chapters 1.1 Information, 1.2 Acute health trends: surveillance, 1.3 Longer term health trends:
registers and 5.2 Evaluating health care using routine data.

ARTICLES
Butler RN (1997) Population aging and health. BMJ 315:1082−4. This article covers trends in ageing and the impact
on the population, and appears in a themed issue of the BMJ devoted to ageing.
McMichael AJ (2002) Population, environment, disease, and survival: Past patterns, uncertain futures. Lancet
359:1145.

WEBSITES
Communities and Local Government (CLG) – [Link]. Formerly known as the Department of the
Deputy Prime Minister, CLG provides key ‘non-health’ information and policies relevant to public health, including
local government, housing/homelessness and transport.
Government Actuary’s Department – [Link]. Provides demographic information, life-tables and population
projections for the UK with descriptions of the methodologies used.
Office for National Statistics – [Link]. The primary source for information on Britain’s population,
lifestyles, economy and society at national and local levels. Statistical tables and commentaries on the statistics are
available.

CHAPTER 3B
The fast pace of change in England’s NHS means that few resources stay current for long and this is particularly true
for information sources. Most of the further reading suggested therefore comes from websites, rather than textbooks.
However, the degree of organisational change may mean that some of these sites become obsolete in the near
future.

BOOKS
Donaldson LJ, Donaldson RJ (2003) Essential Public Health (2nd ed). Oxford: Petroc Press. Key public health text
with useful and relatively current section on health and related information.

WEBSITES
NHS Connecting for Health − [Link]. The organisation responsible for the NHS IT
modernisation programme. The website contains information on many aspects of information use, communication
and storage within the NHS.
The Information Centre for Health and Social Care – [Link]. It commissions and produces a range of health-
care information relevant to public health, including:
• Primary care: Quality and Outcomes Framework (QOF) data relating to general practices’ performance on key
clinical and organisational domains

573

Public [Link] 573 3/6/08 [Link]

 F u r t he r R e a d i ng

• Surveys, e.g. Health Survey for England data

• Service statistics, e.g. maternity, community care.
Medicines and Healthcare products Regulatory Agency − [Link]. The MHRA is the government agency for
England responsible for ensuring that medicines and medical devices work and are acceptably safe.
National Prescribing Centre − [Link]. Formed by the Department of Health to promote high-quality, cost-
effective prescribing to relevant professionals and senior managers working in the NHS. Contains information on
medicines management, non-medical prescribing, education and development.
Office for National Statistics – [Link]. The primary source for information on Britain's population,
lifestyles, economy and society at national and local levels.
Prescription Pricing Authority − [Link]/[Link]. This is part of the NHS Business Services Authority and
processes all NHS prescriptions dispensed in England. The website contains information on prescribing volumes,
trends and costs in the NHS.
United Kingdom Association of Cancer Registries − [Link]. Organisations involved in collecting and coding
cancer information in the UK and Ireland.
World Health Organization, International Classification of Diseases (10th ed) − [Link]/classifications/icd/en.
This website provides a list of ICD-10 codes and subcodes.

CHAPTER 3C

BOOKS
Berg M (ed) (2004). Health Information Management: Integrating information and communication technology in
health care work. London: Routledge. Recent and international description of health information management
theories and implementation.
Deutsch T, Carson E, Ludwig E (1994) Dealing with Medical Knowledge: Computers in clinical decision making. New
York: Plenum Press.
Donaldson L, Donaldson RJ (2003) Essential Public Health (2nd ed). Oxford: Petroc Press. Provides a summary of
sources of health and health-care information available in England and Wales.

ARTICLES
Musgrave S (2004) Public health information systems – tools to widen their accessibility and impact. Healthcare
Computing 237−44. Available online at: [Link]/hc2004/P30_Musgrave.pdf.
Marshall T, Rouse A (2002) Resource implications and health benefits of primary prevention strategies for
cardiovascular disease in people aged 30 to 74: mathematical modelling study. BMJ 325:197−9. An important
example of a cost-effectiveness exercise (within the National Service Framework for Coronary Heart Disease).
Stevens A, Gillam S (1998) Needs assessment: from theory to practice. BMJ 316:1448−52.

WEBSITES
Association of Public Health Observatories − [Link]/apho. The umbrella body of the regional public health
observatories (PHOs) in England. The Association includes profiles on health themes such as determinants of health
and links to the regional PHOs, which collate and analyse local data to provide local health and health service
profiles for their areas.

574

Public [Link] 574 3/6/08 [Link]

 Further R eading

Centre for Health Economics, Health Policy Team, University of York − [Link]/inst/che/research/[Link]
Information on applied and methodological economics research relevant to health care.
Department of Health – [Link]. The website contains some information on the English NHS performance.
National Institute of Economic and Social Research − [Link]. In particular, O’Mahony M (2006) Outputs,
Inputs and Productivity in the NHS.
NHS Health Informatics Community − [Link]. Website relating to NHS informatics,
information and developments.
NHS Connecting for Health − [Link]. News and information on the National Programme for
IT, which aims to introduce modern computer systems throughout the NHS.
Public Health electronic Library − [Link].
Virtual classroom of Health Informatics − [Link]/virtualclassroom/[Link].

CHAPTER 4A

BOOKS
Halpern D (2005) Social Capital. Cambridge: Polity Press. Textbook that covers virtually all of the Part A syllabus
relating to medical sociology.
Scambler G (2003) Sociology as Applied to Medicine (5th ed). Philadelphia: WB Saunders.

ARTICLES
Wilkinson R, Marmot M (2003) Social Determinants of Health: The solid facts (2nd ed). Copenhagen: WHO. Available
at: [Link]/document/[Link]
Brimlow D, Cook JS, Seaton R (2003) Stigma and HIV/AIDS: A review of the literature. US Department of Health
and Human Services Health Resources and Services Administration. Available online at: [Link]/publications/
stigma/[Link].
Gray AJ (2002) Stigma in psychiatry. Journal of the Royal Society of Medicine 95:72−6.
Macintyre M, Telban B, Mitchell WE et al (2004) Book review forum on Gilbert Lewis’s A Failure of Treatment. Journal
of Ritual Studies 18:121−51. Available online at: [Link]/~strather/Gilbert%20Lewis%20Review%[Link].

WEBSITES
Equality and Human Rights Commission − [Link]/pages/[Link]. Contains educational
materials that illustrate the social model of disability and information relating to the three merged organisations,
the Disability Rights Commission, Commission for Racial Equality and the Equal Opportunities Commission.
Healthcare Commission − [Link]. Commission for Health Improvement. In particular,
see Unpacking the Patients’ Perspective: Variations in NHS patient experience in England (2004), available at: www.
[Link]/_db/_documents/[Link].
National Grid for Learning Cymru, Glossary of Sociological Terms – [Link]/vtc/ngfl/sociology/
detailed_glossary.htm. Authored by David Bown, assisted by Janis Griffiths.
Public Health electronic Library − [Link]. Contains short summaries on medical sociology applied
to Part 1.

575

Public [Link] 575 3/6/08 [Link]

 F u r t he r R e a d i ng

CHAPTER 4B

BOOKS
Gabe J, Bury M, Elston MA (2004) Key Concepts in Medical Sociology. London: Sage Publications. Lay language
overview of medical sociology.
Scambler G (2004) Sociology as Applied to Medicine (5th ed). London: WB Saunders. Standard medical sociology text,
covering the Part A syllabus.

CHAPTER 4C

BOOKS
Donaldson C, Gerard K, Milton C et al (2004) Economics of Healthcare Financing: The visible hand. Basingstoke:
Palgrave Macmillan. Readable explanation of modern health economic policy concepts and dilemmas.
Hogwood B, Gunn LA (1985). Policy Analysis for the Real World. Oxford: Oxford University Press, 52−3.
McPake B, Kumaranayake L, Normand C (2002) Health Economics: An international perspective. Routledge. An
introduction to health economics in three sections − covering supply and demand concepts and markets, economic
evaluation, and a comparison of different health-care systems.
Pencheon D, Guest C, Melzer D, Muir Gray JA (eds) (2006) Oxford Handbook of Public Health Practice (Oxford
Handbooks Series). Second Edition Oxford: Oxford University Press. Covers many essential public health topics,
including policy formation and priority setting descriptions, and gives practical examples.

ARTICLES
Department of Health (2003) Code of Practice on Openness in the NHS. Leeds: Department of Health. Available at:
[Link]/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_4050490.
Dollar D (2001) Is globalization good for your health? Bulletin of the World Health Organization 79:827−33.
Leon DA, Walt G, Gilson L (2001) International perspectives on health inequalities and policy. BMJ 322:591–4.
Maynard A (2001) Economics-based medicine: an evolving paradigm. Journal of the Royal College of Physicians of
Edinburgh 31(Suppl 9):16–7.
NICE (2006) Rapid review of the economic evidence of physical activity interventions. Available at: [Link]/
[Link]?o=528518.
Stevens A, Gabbay J (1991) Needs assessment needs assessment. Health Trends 23:20−3.

WEBSITE
Wikipedia, Procedural Justice − [Link]/wiki/Procedural_justice.

CHAPTER 4D

BOOKS
Farmer R, Miller D, Lawrenson R (2004) Lecture Notes on Epidemiology and Public Health Medicine (5th ed). Oxford:
Blackwell Publishing.
Jefferson T, Demicheli V, Mugford M (2000) Elementary Economic Evaluation in Health Care (2nd ed). London: BMJ
Books.

576

Public [Link] 576 3/6/08 [Link]

 Further R eading

ARTICLES
Cohen D (1994) Marginal analysis in practice: an alternative to needs assessment for contracting health care. BMJ
309:781−4.
National Institute for Health and Clinical Excellence (2006) Methods for Development of NICE Public Health Guidance.
Available at: [Link]/[Link]?o=299970.
Office of Health Economics (1999) The Economics of Health Care Scarcity (3rd ed). Available at: [Link]/
[Link].
Palmer S, Byford S, Raftery J (1999) Economics notes: Types of economic evaluation. BMJ 318:1349.
Palmer S, Raftery J (1999) Economics notes: Opportunity cost. BMJ 318:1551−2.
Palmer S, Torgerson DJ (1999) Economic notes: Definitions of efficiency BMJ 318:1136.
Sassi F, Le Grand J, Archard L (2001) Equity versus efficiency: a dilemma for the NHS. BMJ 323:762−3.
WHO (2001) About Health Systems Performance. Available at: [Link]/health-systems-performance/about.
htm#Why%20assess%20health%20system%20performance.

CHAPTER 5A

BOOKS
Iles V, Cranfield S (2005) Really Managing Healthcare (2nd ed) Maidenhead: Open University Press/McGraw-Hill.
Pencheon D, Guest C, Melzer D, Muir Gray JA (eds) (2006) Oxford Handbook of Public Health Practice (2nd ed)
(Oxford Handbooks Series). Oxford: Oxford University Press. Covers many essential public health topics including
management.

ARTICLES
Edwards N (2003) Doctors and managers: poor relationships may be damaging patients – what can be done? Quality
and Safety in Health Care 12:21–4.
Finch J (2000) Inter-professional education and team-working: a view from the education providers. BMJ
321:1138−40.
Garelick A, Fagin F (2005) The doctor−manager relationship. Advances in Psychiatric Treatment 11:241−50.
Barr H, Goosey D (2002) Interprofessional Education: Selected case studies. London: Department of Health. Available
at: [Link]/assetRoot/04/03/45/41/[Link].

WEBSITES
Public Health electronic Library health knowledge resources − [Link]. The website contains
summaries of a range of subjects relevant to Part A, including organisational management relating to health.
Brown B (2000) Organisational Communication Theories − [Link]/cit_courseware/research/[Link].
This is a summary of the information in Littlejohn S (1992) Theories of Human Communication (5th ed). California:
Wadsworth Publishing.
Evidence-based medicine − [Link]. In particular, Storey N (2001) What is Clinical
Governance? Volume 1, number 12. Available at: [Link]/ebmfiles/[Link].

577

Public [Link] 577 3/6/08 [Link]

 F u r t he r R e a d i ng

The Shipman Inquiry: Independent public enquiry into the issues arising from the case of Harold Fredrick Shipman -
[Link].
Managing stress − [Link]/positive/managing_stress.html. Website provided by Cambs
Mental Health Info.

CHAPTER 5B

ARTICLES
Miller TR (1996) Structuring networks for maximum performance under managed care. Healthcare Financial
Management 50:37−9.
Pluye P, Potvin L, Pelletier J (2004) Community coalitions and health promotion: Is it that important to develop an
inter-organisational network? Promotion and Education 11:17−24.

WEBSITES
Public Health electronic Library health knowledge resources − [Link]. The website contains
summaries on a range of subjects relevant to Part A, including organisational management relating to health.
Classics in the history of psychology. Intergroup Conflict and Cooperation: The Robbers Cave Experiment. Sherif M,
Harvey O, White BJ et al (1954/1961) − [Link]/Sherif. An internet resource developed by Christopher
D Green, York University, Toronto, Ontario.

CHAPTER 5C

BOOKS
Iles V, Sutherland K (2001) Organisational Change: A review for health care managers, professionals and researchers.
London: NCCDSO, London School of Hygiene and Tropical Medicine. The guide provides readable, concise summaries
of change management models and the experience of their application to health care.
Mullins LJ (2004) Management and Organisational Behaviour (7th ed). Harlow: Financial Times/Prentice-Hall. A
comprehensive management textbook covering the Part A syllabus and much more in considerable depth. Examples
come from a range of organisations, not just health care.
Silbiger S (1999) The 10-Day MBA: A step-by-step guide to mastering the skills taught in top business schools. London:
Piatkus Books.

CHAPTER 5D

BOOKS
Donaldson L, Donaldson RJ (2003) Essential Public Health (2nd ed). Oxford: Petroc Press.
Iles V, Sutherland K (2001) Organisational Change: A review for health care managers, professionals and researchers.
London: NCCSDO. A review of management literature as it relates to health care in the NHS.
McDougall A, Duckett P, Manku M, Robertson J (2003) International Health Comparisons: A compendium of published
information on healthcare systems, the provision of healthcare and health achievement in 10 countries. London:
National Audit Office. Available online at: [Link]/publications/Int_Health_Comp.pdf. A description of
health systems in ten industrialised nations.

578

Public [Link] 578 3/6/08 [Link]

 Further R eading

WEBSITES
Appraisal of Guidelines Research and Evaluation (AGREE) − [Link]/intro.
Bandolier − [Link]/bandolier. Monthly journal and resources on evidence-based health-care.
Scottish Intercollegiate Guidelines Network (SIGN) − [Link].

CHAPTER 6A

BOOKS
Petrie A, Sabin C (2005) Medical Statistics at a Glance (2nd ed). Oxford: Blackwell Publishing. Short summary
covering all key elements of statistics required for Part A with examples from epidemiological studies.
Silman AJ (2002) Epidemiological Studies: A practical guide (2nd ed). Cambridge: Cambridge University Press.
Contains a useful section on the design and application of epidemiological studies, and a diagram that illustrates
differences between observational studies.

CHAPTER 6B

ARTICLES
Stanford X (2002) Organizational Mapping: Knowing the pitfalls. Alberta: Stanford Solutions Inc. No. 59. Available at:
[Link]/updates/u59_f1.htm.

WEBSITES
Betty C Jung’s website − [Link]/[Link]. Charting and Graphing Data Presentation.
Edward Tufte’s website − [Link]/tufte/newet. Examples of clear presentations of data and of
misleading or confusing presentations of the same data.
Indiana University Southeast Basic Business Statistics, Chapter 2: Presenting Data in Tables and Charts (8th ed) −
[Link]/SRAUSC01/[Link]#286,29.
New York State, Department of Health: Tips for presenting data − [Link]/statistics/chac/[Link].
Presenting Data: Tabular and graphic display of social indicators − [Link]/gmklass/pos138/datadisplay/
sections/[Link]. Website created by Gary Klass, Illinois State University (2002). Good do’s and don’ts.
See also Chapters 6A.3 and 1B.8.

CHAPTER 6C

BOOKS
The Economist (2005) Style Guide: The bestselling guide to English usage. London: Profile Books.
Jessop E (2001) Writing to effect change. In: Pencheon D, Guest C, Melzer D, Muir Gray JA (eds) Oxford Handbook of
Public Health Practice. Oxford: Oxford University Press, 428−35.

WEBSITES
How to write medical information in plain English − [Link]/[Link]. Useful tips for clearer
writing and a list of medical terms explained in non-technical language.

579

Public [Link] 579 3/6/08 [Link]

Public [Link] 580 3/6/08 [Link]
 R eferences

References

4S Study (1994) Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the
Scandinavian Simvastatin Survival Study (4S). The Lancet 344:1383−9.
Acheson D (1998) Independent Inquiry into Inequalities in Health Report. London: TSO. Available online at: www.
[Link]/document/doh/ih/[Link].
Administration for Children and Families (2006) Head Start General Information website – [Link]/
programs/hsb/.
Amabile TM (1998) How to kill creativity. Harvard Business Review 76:76−87.
Anderson G, Sotir Hussey P (2001) Influencing government policy. In: Pencheon D, Guest C, Melzer D, Muir Gray JA
(eds) Oxford Handbook of Public Health. Oxford University Press, Chapter 3.3.
Anon (2006) Preventing coronary heart disease: Does Rose’s population prevention axiom still apply in the 21st
century? (editorial) BMJ 332:617−8.
Armstrong BG (1998) Effect of measurement error on epidemiological studies of environmental and occupational
exposures. Occupational and Environmental Medicine 55:651–6.
Armstrong D (2002) Clinical autonomy, individual and collective: the problem of changing doctors’ behaviour. Social
Science and Medicine 55:1771–7.
Avent ND, Chitty LS (2006) Non-invasive diagnosis of foetal sex: utilisation of free foetal DNA in maternal plasma
and ultrasound. Prenatal Diagnosis 26:598−603.
Baker R, Robertson N, Farooqi A (1995) Audit in general practice: factors influencing participation. BMJ 311:31−4.
Barker D (2003) The midwife, the coincidence, and the hypothesis. BMJ 327:1428−30. David Barker’s account of the
circumstances that led to the recording of a unique set of birth records at the beginning of the twentieth century,
which became known as the Hertfordshire Cohort. It was access to this dataset that led to the development of the
Barker hypothesis.
Barringer BR, Harrison JS (2000) Walking a tightrope: Creating value through interorganizational relationships.
Journal of Management 26:367−403.
Bandura A (1977). Social Learning Theory. New York: General Learning Press.
Barsky AJ, Borus JF (1995) Somatization and medicalization in the era of managed care. JAMA 274: 1931−4.

581

Public [Link] 581 3/6/08 [Link]

 R e f e re nc e s

Beattie (1991) Knowledge and control in health promotion: a test case for social policy and social theory. In: Gabe
J, Calnan M, Bury M (eds) The Sociology of the Health Service. London: Routledge.
Beauchamp TL, Childress JF (2001) Principles of Biomedical Ethics (5th ed). Oxford: Oxford University Press.
Belbin RM (1996) Team roles at work. London: Butterworth-Heinemann.
Belsky J, Melhuish E, Barnes J et al (2006) Effects of SureStart local programmes on children and families: early
findings from a quasi-experimental, cross-sectional study. BMJ 332:1476.
Benzeval M, Judge K, Whitehead M (1995) Tackling Inequalities in Health: An agenda for action. London: King’s Fund.
Benzeval M (1995) Tackling inequalities in health. London: King’s Fund.
Binney G, Williams C (1995) Leaning into the Future: Changing the way people change organizations. London:
Nicholas Brealey Publishing Ltd.
Black D, Morris JN, Smith C et al (1980) Inequalities in Health: Report of a research working group (Black Report).
London: Department of Health and Social Security. Available online at: [Link]/history/[Link].
Blackburn H (1999) On the Trail of Heart Attacks in Seven Countries − [Link]/research/7countries/index.
shtm.
Blake R, Mouton J (1964) The managerial grid: key orientations for achieving production through people. Houston:
Gulf.
Bland M (2000) An Introduction to Medical Statistics (3rd ed). Oxford: Oxford University Press, 128.
Bonnefoy XR, Annesi-Maesano I, Moreno LM et al (2004). Review of evidence on housing and health. Fourth
Ministerial Conference on Environment and Health, Budapest, 23–25 June 2004 (EUR/04/5046267/BD/1). Available
online at: [Link]/document/HOH/[Link].
Bradshaw JR (1972) The taxonomy of social need. In: McLachlan G (ed) Problems and Progress in Medical Care.
Oxford: Oxford University Press, 69−82.
Brager G, Specht H (1973) quoted in World Health Organization (2002) Community Participation in Local Health and
Sustainable Development: Approaches and techniques. Available online at: [Link]/document/[Link].
Breslow NE, Day NE (1987) Statistical Methods in Cancer Research, Volume II: The Design and Analysis of Cohort
Studies. Lyon: International Agency for Research on Cancer.
Bunton R (2002) Health promotion as social policy. In: Bunton R, Macdonald G (eds) Health Promotion: Disciplines
and diversity (2nd ed). London: Routledge.
Buse K, Mays N, Walt G (2005) Making Health Policy. Maidenhead: Open University Press.
Carstairs’ Deprivation Index for Scottish Postcode Sectors, 1991 and 2001 – [Link]/EUDL/carstairs.
html.
Carstairs V, Morris R (1989) Deprivation, mortality and resource allocation. Community Medicine 11:364–72.
Cavanagh S, Chadwick K (2005) Health Needs Assessment: A practical guide. London: National Institute for Clinical
Excellence. Available online at: [Link]/[Link]?o=513203.
Centre for Evidence-Based Medicine (2007) Levels of Evidence – [Link].
Centers for Disease Control and Prevention (2003) Radiation Measurement Factsheet − [Link]/radiation/pdf/
[Link].
Chadwick E (1842) The Sanitary Conditions of the Labouring Population. Self published.

582

Public [Link] 582 3/6/08 [Link]

 R eferences

Chapman S (2004) Advocacy for public health: a primer. Journal of Epidemiology and Community Health 58:361−5.
Cochrane AL (1972) Effectiveness and Efficiency: Random reflections on health services. London: Nuffield Provincial
Hospitals Trust. Reprinted in 1999 for Nuffield Trust by the Royal Society of Medicine Press, London.
Cohen D (1994) Marginal analysis in practice: an alternative to needs assessment for contracting health care. BMJ
309:781−4.
Comptroller and Auditor General (2004) The Drug Treatment and Testing Order: early lessons. London: National Audit
Office. Available online at: [Link]/publications/nao_reports/03-04/[Link].
Cornwell J (1984) Hard-earned Lives: Accounts of Health and Illness from East London. London: Routledge.
Council of Europe (2006) Social Cohesion Strategy. Available at: [Link]/T/E/social_cohesion.
Covey SR (1989) The Seven Habits of Highly Effective People. London: Simon & Schuster.
Dahlgren G, Whitehead M (1991) Policies and Strategies to Promote Social Equity in Health. Stockholm: Institute for
Futures Studies.
Davis RL, Khoury MJ (2006) A public health approach to genomics and gene-based diagnostic tests.
Pharmacogenomics 7:331−7.
Degeling P, Maxwell S, Kennedy J, Coyle B (2003) Medicine, management and modernisation: a ‘danse macabre’. BMJ
326:649−52.
Department of Health (1998) A First Class Service: Quality in the new NHS. London: Department of Health. Available
online at: [Link].
Department of Heath (1999a) Patient and Public Involvement in the New NHS. London: Department of Health.
Available online at: [Link].
Department of Health (1999b) Saving Lives: Our healthier nation. London: The Stationery Office. Available online at:
[Link]/document/cm43/4386/[Link].
Department of Health (2000a) Coronary Heart Disease National Service Framework. London: Department of Health.
Available online at: [Link].
Department of Health (2000b) The NHS Plan: a plan for investment, a plan for reform. London: The Stationery Office.
Available online at: [Link].
Department of Health (2002) Getting ahead of the curve: a strategy for combating infectious diseases (including other
aspects of health protection). London: Department of Health. Available online at: [Link].
Department of Health (2003) Recognising the potential. In: ‘Our Inheritance, Our Future: Realising the potential of
genetics in the NHS’. London: The Stationery Office. Available online at: [Link].
Department of Health (2004a) Choosing Health: Making healthy choices easier. London: The Stationery Office.
Available online at: [Link].
Department of Health (2004b) A matron’s charter: An action plan for cleaner hospitals. London: Department of
Health. Available online at: [Link].
Department of Health (2004c) National Standards, Local Action: Health and social care standards and planning
framework 2005/06–2007/08. London: Department of Health. Available online at: [Link].
Department of Health (2005a) Resource Allocation: Weighted Capitation Formula (5th ed). Leeds: Department of
Health. Available online at: [Link].

583

Public [Link] 583 3/6/08 [Link]

 R e f e re nc e s

Department of Health (2005b) Saving Lives: a delivery programme to reduce healthcare associated infection including
MRSA: Skills for implementation. London: Department of Health. Available online at: [Link].
Department of Health (2006) NHS Allocations. London: Department of Health. Available online at: [Link].
Department of Health, South Africa (2004) Ethics in Health Research: Principles, structures and processes. Research
ethics guidelines – 2004. Pretoria: Department of Health. Available online at: [Link]/docs/factsheets/
guidelines/ethnics/[Link].
Department of Health Expert Group (Chairman, CMO) (2000) An Organisation with a Memory. Report of an expert
group on learning from adverse events in the NHS chaired by the Chief Medical Officer. London: The Stationery Office.
Department of Health, National Service Framework for Coronary Heart Disease (2000). The National Service Framework
for Coronary Heart Disease. London: Department of Health. Available online at: [Link].
Diderichsen F, Hallqvist J (1998) Social inequalities in health: some methodological considerations for the study of
social position and social context. In: Arve-Parès B (ed) Inequality in Health – a Swedish perspective. Stockholm:
Swedish Council for Social Research, 25−39.
Doll R, Hill AB (1954) The mortality of doctors in relation to their smoking habits: a preliminary report. BMJ
228:1451–5.
Donabedian A (1966) Evaluating the quality of medical care. Millbank Memorial Fund Quarterly 44:166−206.
Donaldson C, Gerard K, Mitton C et al (2004) Economics of Health Care Financing: The visible hand. Basingstoke:
Palgrave MacMillan.
Dorak MT (2007) Genetic Epidemiology – [Link]/epi/[Link].
Dosani S, Cross P (2004) How to pass MFPH. BMJ Career Focus 329:5.
Drucker P (1950) The New Society: The Anatomy of Industrial Order. New York: Harper.
Drucker P (1954) The Practice of Management. New York: HarperCollins.
Dunbar RIM (1992) Neocortex size as a constraint on group size in primates. Journal of Human Evolution 22:469−93.
Durkheim E (1897). Le suicide. Paris: Alcan.
Egger M, Davey Smith G, Schneider M, Minder C (1997). Bias in meta-analysis detected by a simple, graphical test.
BMJ 315:629–34.
Elbein SC (2002) Perspective: The search for genes for type 2 diabetes in the post-genome era. Endocrinology
143:2012−8.
Emberson J, Whincup P, Morris R et al (2004) Evaluating the impact of population and high-risk strategies for the
primary prevention of cardiovascular disease. European Heart Journal 25:484−91.
Etziou A (1973) Mixed scanning: A third approach to decision-making. In: Faludi A (ed) Reader in Planning Theory.
Oxford: Perganum Press, 218–20.
Evans R, Stoddart G (1990) Producing health, consuming resources. Social Science and Medicine 31:1347−63.
Ewles L, Simnet I (1995) Promoting Health: A practical guide (3rd rev ed). London: Ballière Tindall.
Ewles L, Simnet I (2003) Promoting Health: A practical guide (5th ed). London: Ballière Tindall.
Eyers JE (1998) How to do (or not to do) … Searching bibliographic databases effectively. Health Policy and
Planning 13:339−42.
Fagan TJ (1975). Nomogram from Bayes’s theorem (letter). New England Journal of Medicine 293:257.

584

Public [Link] 584 3/6/08 [Link]

 R eferences

Fiander M, Burns T (2000) A Delphi approach to describing service models of community mental health practice.
Psychiatry Services 51:656−8.
Field MJ, Lohr KN (eds) (1990) Institute of Medicine Committee to Advise the Public Health Service on Clinical
Practice Guidelines. Clinical Practice Guidelines: Directions for a new program. Washington DC: National Academy Press.
Fight Crime: Invest in Kids (2006) Head Start reduces crime and improves achievement. New York: Fight Crime: Invest
in Kids. Available online at: [Link]/reports/[Link].
Fleiss JL (1981) Statistical Methods for Rates and Proportions (3rd ed). New York: John Wiley & Sons, 77.
Foundation Trust Network (2005) New Voices, New Accountabilities: A guide to wider governance in foundation trusts.
London: Foundation Trust Network. Available online at: [Link]/ftn.
Framingham Heart Study (1948) A Project of the National Heart, Lung and Blood Institute and Boston University –
[Link]/about/[Link].
French JRP, Raven BH (1960) The bases of social power. In: Cartwright D, Zander A (eds) Group Dynamics. New York:
Harper and Row.
Friedson E (1959) Specialties without roots: the utilization of new services. Human Organization, 18:112–16.
Fuller E (ed) (2005) Smoking, Drinking and Drug Use Among Young People in England in 2004. Leeds: Health and
Social Care Information Centre. Available online at: [Link]/pubs.
Garces GT, Thomas D, Currie J (2002) Longer-term effects of HeadStart. American Economic Review 92:999–1012.
The Global Fund to fight AIDS, Tuberculosis and Malaria website –[Link]/en/.
Goffman E (1963) Stigma: Notes on the management of spoiled identity. London: Prentice-Hall.
Goulding N (2003) How to Prepare and Write for a Scientific Journal. Available online at: [Link]/
epm/students/ssm7/[Link].
Hadorn DC (1991) Setting health care priorities in Oregon. Cost-effectiveness meets the rule of rescue. JAMA
265:2218−25.
Hale J (2000) What contribution can health economics make to health promotion. Health Promotion International
15:341−8.
Hall R (2007) Psychology World: Between Subjects One-Way ANOVA example − [Link]/~psyworld/
[Link]; [Link].
Handy CB (1987) Understanding Organizations (Penguin Business) (3rd ed). London: Penguin.
Hanneman RL (1996) Intersalt: hypertension rise with age revisited. BMJ 312:1283−4.
Hart JT (1971) The inverse care law. The Lancet 1:405–12. Available online at: [Link]/history/
[Link].
He FJ, MacGregor GA (2004) Effect of longer-term modest salt reduction on blood pressure. Cochrane Database of
Systematic Reviews Issue 3: CD004937.
Health Protection Agency Centre for Infections (2005) Hepatitis C in England: The First Health Protection Agency
Annual Report 2005. London: Health Protection Agency. Available online at: [Link].
Health Protection Agency (2006) Tuberculosis Update, March 2006. The National TB Strain Typing Database – An
Update. Information for Action. Available online at: [Link].
Henkel GL (undated) North Karelia Project Shows the World How to Reduce Heart Disease – [Link]/
nwfwebsite/puska_heart.html. New World Finn website.

585

Public [Link] 585 3/6/08 [Link]

 R e f e re nc e s

Herzberg F (1968) ‘One more time: how do you motivate employees?’ Harvard Business Review 46:53–62.
Hethcote HW (2000) The mathematics of infectious diseases. SIAM Review 42:599−653.
Hill AB (1965) The environment and disease: association or causation? Proceedings of the Royal Society of Medicine
58:295−300.
HM Treasury (2007) Public service agreements. London: HSMO. Available online at: [Link]/pbr_csr/
psa/pbr_csr07_psaopportunity.cfm.
Hochbaum GM (1958) Public participation in medical screening programs: A socio-psychological study. PHS publication
No. 572. Washington, DC: US Government Printing Office.
Hoffman A, Lowe D, Rudd A et al (2004) National Sentinel Stroke Audit 2004. London: Royal College of Physicians.
Humphris D, Hean S (2004) Educating the future workforce: building the evidence about interprofessional learning.
Journal of Health Services Research and Policy 9:24–27.
Hunt K, Emslie C (2001) Commentary: The prevention paradox in lay epidemiology – Rose revisited. International
Journal of Epidemiology 30:442−6.
Hunt SC, Gwinn M, Adams TD (2003) Family history assessment: Strategies for prevention of cardiovascular disease.
American Journal of Preventive Medicine 24:136−42.
Iles V (2005) Really Managing Health Care. Oxford: Oxford University Press.
Illich I (1975) Medical Nemesis: The expropriation of health. London: Marian Boyars.
Intersalt Cooperative Research Group (1988) Intersalt: An international study of electrolyte excretion and blood
pressure. Results for 24 hour urinary sodium and potassium excretion. BMJ 297:319−28.
Irish Council for Bioethics. Website (2004) – [Link].
Jamison J, O’Reilly D, Carr-Hill, R, Stevenson M (2004) Equity of inpatient services by the two communities in
Northern Ireland. Journal of Health Services Research and Policy 9:34−8.
Jarman B (1983) Identification of underprivileged areas. BMJ 286:1705–8.
Jefferson T, Demicheli V, Mugford M (2000) Elementary Economic Evaluation in Health Care (2nd ed). London: BMJ
Books.
Johnston G, Crombie IK, Davies HTO et al (2000) Reviewing audit: barriers and facilitating factors for effective
clinical audit. Quality in Health Care 9:23−6.
Kahan B (2005/2006) Interactive Domain Model Best Practices: The pre-determinants of the social determinants of
health − [Link]/[Link]?content=archjottings#brkOctober05.
Kee F, McDonald P, Gaffney B (1997) Prioritising the cardiac surgery waiting list. The angina patient’s perspective.
Heart 77:330−2.
Keys A (1980) Seven Countries: A multivariance analysis of death and coronary heart disease. London: Harvard
University Press.
Keys AB, Keys M (1975) How to Eat Well and Stay Well: The Mediterranean Way. London: Doubleday.
King’s Fund (2006) Patients at Risk of Re-hospitalisation (PARR) Case Finding Tool − [Link]/current_
projects/predictive_risk/patients_at_risk.html.
Kirkwood B, Sterne J (2003) Essential Medical Statistics (2nd ed). Oxford: Blackwell Science.
Knott L (2006) Patient UK: Records and computers − [Link]/showdoc/40000769.

586

Public [Link] 586 3/6/08 [Link]

 R eferences

Kohlmann W, Gruber SB (2004, updated 2006) Hereditary Non-Polyposis Colon Cancer − [Link].
Koller T (2006) WHO Interview with Prof Goran Dahlgren − [Link]/socialdeterminants/
socmarketing/20060221_6.
Kono T (1984) Strategy and Structure of Japanese Enterprises. London: Macmillan.
Kotler P, Zaltman G (1971) Social marketing: an approach to planned social change. Journal of Marketing 35:3–12.
Künzli N, Kaiser R, Medina S et al (2000) Public-health impact of outdoor and traffic-related air pollution. A
European assessment. The Lancet 356:795−801.
Lakhani A, Coles J, Eayres D et al (2005) Creative use of existing clinical and health outcomes data to assess NHS
performance in England: Part 1 – performance indicators closely linked to clinical care. BMJ 330:1426−31
Lakhani A, Olearnik H, Eayres D (eds) (2007) Compendium of Clinical and Health Indicators. London: The Information
Centre for Health and Social Care/National Centre for Health Outcomes Development. Available online at: www.
[Link].
Lalonde M (1974) A New Perspective on the Health of Canadians: A working document − [Link]/ph-sp/
phdd/pdf/[Link].
Larson (1977) The Rise of Professionalism: A sociological analysis. Berkeley, CA: University of California Press.
Last JM (1963) The iceberg: ‘completing the clinical picture’ in general practice. The Lancet 2:28–31.
Lemert EM (1967) Human Deviance, Social Problems and Social Control. London: Prentice Hall.
Lewis G (2000) A Failure of Treatment. Oxford: Oxford University Press.
Lewis R, Hunt P, Carson D (2006) Social Enterprise and Community Based Care: Is there a future for mutually-owned
organizations in community and primary care? London: King’s Fund.
Lincoln P, Nutbeam D (2005) WHO and international initiatives. In: Davies M, Macdowell W (eds) Health Promotion
Theory. London: McGraw-Hill.
Lindblom C (1959) The science of muddling through. Public Administration Review 19:79–88.
Lo YM, Corbetta N, Chamberlain PF et al (1997) Presence of fetal DNA in maternal plasma and serum. The Lancet
369:440–2.
London Health Commission (2002) Health in London: 2002 Review of the London Health Strategy high-level indicators.
London: London Health Commission. Available online at: [Link]/pdf/[Link].
Manuel DG, Lim J, Tanuseputro P et al (2006) Revisting Rose: Strategies for reducing coronary heart disease. BMJ
332:659−62.
Marshall T (2000) Exploring a fiscal food policy: the case of diet and ischaemic heart disease. BMJ 320:301−4.
Marshall T, Kennedy E, Offutt S (2000) Exploring a fiscal food policy: the case of diet and ischaemic heart disease.
Commentary: Alternative nutrition and outcomes using a fiscal food policy. BMJ 320:304−5.
Marshall T, Rouse A (2002) Resource implications and health benefits of primary prevention strategies for
cardiovascular disease in people aged 30 to 74: mathematical modelling study. BMJ 325:197.
Maslow AH (1943) A Theory of Human Motivation. Psychological Review 50:370–96.
Maxwell RJ (1984) Quality assessment in health. BMJ 288:1470–2.
Maynard A (2001) Economic-based medicine: an evolving paradigm. Journal of the Royal College of Physicians of
Edinburgh 31 (Suppl 9):16−7.

587

Public [Link] 587 3/6/08 [Link]

 R e f e re nc e s

McCarthy MI (2004) Progress in defining the molecular basis of type 2 diabetes mellitus through susceptibility-gene
identification. Human Molecular Genetics 13:R33−41.
McConnachie A, Hunt K, Emslie C et al (2001) ‘Unwarranted survivals’ and ‘anomalous deaths’ from coronary heart
disease: prospective survey of general population. BMJ 323:1487−91.
McKeown RE, Weed DL (2002) Ethics in epidemiology and public health II. Applied terms. Journal of Epidemiology
and Community Health 56:739−41.
Mechanic D (1968) Medical Sociology. New York: Free Press.
Montori, VM, Smieja M, Guyatt GH (2000) Publication bias: A brief review for clinicians. Mayo Clinic Proceedings
75:1284−8.
Morgan O, Baker A (2006) Measuring deprivation in England and Wales using 2001 Carstairs scores. Health Statistics
Quarterly 31:28–33. Available online at: [Link]/articles/hsq/HSQ31deprivation_using_carstairs.pdf.
Mullins LJ (2005) The nature of leadership. In: Mullins LJ (2005) Management and Organisational Behaviour (7th
ed). Harlow: Financial Times/Prentice Hall.
Myer L, Morroni C, Link BG (2004). Impact of measurement error in the study of sexually transmitted infections.
Sexually Transmitted Infections; 80:318−21.
Naidoo J, Wills J (1998) Practising Health Promotion: Dilemmas and challenges. London: Baillière Tindall.
Naidoo J, Wills J (2000) Health Promotion: Foundations for practice (2nd edn). London: Baillière Tindall.
National Advisory Committee on Health and Disability (2003) Screening to Improve Health in New Zealand: Criteria to
assess screening programmes. Wellington: National Advisory Committee on Health and Disability.
National Center for Chronic Disease Prevention and Health Promotion (2006) Steps of an Outbreak Investigation −
[Link]/excite/classroom/outbreak/[Link].
National Human Genome Research Institute, National Institutes of Health (2007) Learning about Cystic Fibrosis −
[Link]/10001213.
National Institute for Clinical Excellence (2004) Guide to the Methods of Technology Appraisal. London: NICE.
Available online at: [Link]/niceMedia/pdf/TAP_Methods.pdf.
National Institute for Health and Clinical Excellence (2006) Methods for Development of NICE Public Health Guidance.
London: NICE. Available online at: [Link]/nicemedia/pdf/[Link].
National PCT database (2004) Measures of Need − [Link].
Newsom JT (2006) Distinguishing Between Random and Fixed: Variables, effects, and coefficients − [Link]/
IOA/newsom/mlrclass/ho_randfixd.doc.
NHS Clinical Governance Support Team (1999) Seven Pillars of Clinical Governance Model − [Link]/
downloads/Seven_Pillars.pdf.
NHS Litigation Authority (2007) About the NHS Litigation Authority − [Link]/Channels.
Northern Ireland Statistics and Research Agency (2005) Northern Ireland Multiple Deprivation Measure. Belfast: The
Stationery Office. Available online at: [Link]/aboutus/[Link].
Nutbeam D, Harris E (2004) Theory in a Nutshell: A guide to health promotion theory (2nd ed). Maidenhead: McGraw-
Hill Education.
Nutrional Advisory Group (1995) Recommendations for a Food and Nutrition Policy for Ireland. Dublin: Stationery
Office.

588

Public [Link] 588 3/6/08 [Link]

 R eferences

Office for National Statistics (2007) The National Statistics Socio-economic Classification. London: Office for National
Statistics. Available online at: [Link]/methods_quality/ns_sec/[Link].
O’Reilly D, Stevenson M (2004) Our data on the uptake of disability benefits a useful addition to census data in
describing population health care needs? Journal of Health Service Research Policy 9:165−70.
Parsons T (1951) The Social System. New York: Free Press.
Petrie A, Sabin C (2005) Medical Statistics at a Glance (2nd ed). Oxford: Blackwell Publishing.
PHLS Advisory Committee on Gastrointestinal Infections (2004) Preventing person-to-person spread following
gastrointestinal infections: guidelines for public health physicians and environmental health officers. Communicable
Diseases and Public Health 7:362−84.
Pilowski I, Spence ND (1977) Ethnicity and illness behaviour. Psychological Medicine 7:447−52.
Prochaska JO, DiClemente CC (1984) Self-change processes, self-efficacy and decisional balance across five stages of
smoking cessation. In: Advances in cancer control – 1983. New York: Alan R. Liss, Inc.
Prochaska JO, Velicer WF (1997) The transtheoretical model of health behavior change. American Journal of Health
Promotion 12:38−48.
Puska P, Salonen JT, Nissinen A et al (1983) Change in risk factors for coronary heart disease during 10 years of a
community intervention programme (North Karelia project). BMJ 287:1840−4.
Rawls J (1971) A Theory of Justice. Cambridge: Harvard University Press.
Richardson SW, Wilson MC, Guyatt GH et al (1999) Users’ guides to the medical literature: XV. How to use an article
about disease probability for differential diagnosis. JAMA 281:1214−9.
Robinson V, Segrott J (2002) Home Office Research Study 243: Understanding the decision-making of asylum seekers.
London: Home Office. Available online at: [Link]/rds/pdfs2/[Link].
Rogers E (1995) Diffusion of Innovation (4th ed). New York: Free Press.
Rootman I, Goodstadt M, Hyndman B et al (2001) Evaluation in health promotion: Principles and perspectives.
Copenhagen: WHO Regional Publications, European Series, No 92.
Rose G (1981) Strategy of prevention: lessons from cardiovascular disease. BMJ 282:1847−51.
Rose G (1992) The Strategy of Preventive Medicine. Oxford: Oxford University Press.
Rootman I, Goodstadt M, Hyndman B et al (2001) Evaluation in Health Promotion: Principles and perspectives.
European Series, No. 92. Copenhagen: WHO Regional Publications.
Rosenstock IM, Strecher VJ, Becker MH (1998) Social learning theory and the Health Belief Model. Health Education
Quarterly 15:175−83.
Salisbury DM (2005) Development of immunization policy and its implementation in the United Kingdom. Health
Affairs 24:744−54.
Sandman PM (1987) Risk communication: Facing public outrage. EPA Journal Nov:21−22. Available online at: www.
[Link]/articles/[Link].
Sassi F, Le Grand J, Archard L (2001) Equity versus efficiency: a dilemma for the NHS. BMJ 323:762−3.
Scambler G (ed) (1997) Sociology as Applied to Medicine (4th ed). London: Saunders.
Schein EH, Bennis WG (1965) Personal and Organizational Change through Group Methods: The laboratory approach.
New York: John Wiley & Sons.

589

Public [Link] 589 3/6/08 [Link]

 R e f e re nc e s

Scientific Committee of the Food Safety Authority of Ireland, Nutrition Sub-committee (2007) Salt and Health:
Review of the Scientific Evidence and Recommendations for Public Policy in Ireland. Dublin: Food Safety Authority of
Ireland. Available online at: [Link]/publications/reports/salt_report.pdf.
Scottish Executive (2004) Scottish Index of Multiple Deprivation: Overview. Edinburgh: Scottish Executive. Available
online at: [Link]/Topics/Statistics/SIMD/Overview.
Sihota S, Lennard L (2004) Health Literacy: Being able to make the most of health. London: National Consumer
Council. Available online at: [Link]/nccpdf/poldocs/NCC064_health_literacy.pdf.
Simon H (1947) Administrative Behaviour. A study of decision-making processes in administrative organisations (4th
ed). New York: The Free Press.
Smithies J, Adams L (1990) Community participation in health promotion. London: Health Education Authority.
Snow J (1855) On the mode of communication of cholera. London: John Churchill; 1–38. Available online at:
[Link]/EPI/snow/[Link].
Social Disadvantage Research Centre (SDRC), Department of Social Policy and Social Work at the University of
Oxford (2007) The English Indices of Deprivation 2007: Summary. London: Department for Communities and Local
Government. Available online at: [Link]/publications/communities/indicesdeprivation07.
Sproston K, Mindel J (eds) (2006) Health Survey for England 2004. Leeds: The Information Centre. Available online
at: [Link]/pubs.
Sylva K, Melhuish E, Sammons P et al (2004) The Effective Provision of Pre-School Education (EPPE) Project. EPPE
findings from the Early Primary Years. London: DFES. Available online at: [Link]/research/keyresearch/
eppe.
Tannahill A (1985) What is health promotion? Health Education Journal 44:167−8.
Tuckman BW (1965) Developmental sequence in small groups. Psychological Bulletin 63:384–99.
Tufte ER (2001) The Visual Display of Quantitative Information (2nd edn). Connecticut: Graphics Press LLC.
United Nations (2007) UN Millennium Development Goals. United Nations. Available online at: [Link]/
millenniumgoals.
Walker P (1999) Protecting and Using Patient Information: A manual for Caldicott Guardians. London: Department of
Health. Available online at: [Link]/assetRoot/04/06/81/36/[Link].
Walt G (2001) Global cooperation in international health. In: Merson M, Black RE, Mills AJ (eds) International Public
Health: Diseases, programs, systems, and policies. New York: Aspen Publishers.
Wanless D (2002) Securing our future health: Taking a long-term view. London: HM Treasury. Available online at: www.
[Link]/Consultations_and_Legislation/wanless/consult_wanless_final.cfm.
Wanless D (2004) Securing Good Health for the Whole Population. London: HMSO.
Wanless D (2006) Securing good care for older people: Taking a long-term view. London: King’s Fund.
Waterhouse J, Muir C, Correa P et al (eds) (1976) Cancer Incidence in Five Continents, Vol 3.3. Lyon: IARC, 456.
Watkiss P, Brand C, Hurley F et al (2000) On the Move: Informing transport health impact assessment in London.
London: Health Commission. Available online at: [Link]/pdf/[Link].
Weber M (1958) The Three Types of Legitimate Rule (translation). Quoted in: Tripartite classification of authority.
Wikipedia website at: [Link]/wiki/Tripartite_classification_of_authority. Berkeley: University of California.
Whincup PH, Owen CG, Sattar N, Cook DG (2005) School dinners and markers of cardiovascular health and type 2
diabetes in 13−16 year olds: cross sectional study. BMJ 331:1060−1.

590

Public [Link] 590 3/6/08 [Link]

 R eferences

Whitely L (2007) z, t and F tests − [Link]/spm/doc/mfd/[Link].

Wikipedia, the free encyclopedia. (2007) Molecular biology − [Link]/wiki/Molecular_biology
Willis CE, Kee F, Beverland P, Watson JD (2000) Urban rural differences in total hip replacements. The next stage.
Journal of Public Health Medicine 22:435−8.
Wilson JM, Jungner G (1968) Principles and Practice of Screening for Diseases. Geneva: WHO.
WHO (1978) Declaration of Alma-Ata, policy statement. Geneva: WHO. Available online at: [Link]/
AboutWHO/Policy/20010827_1.
WHO (1980) International Classification of Functioning, Disability and Health (ICF). Geneva: WHO. Available online at:
[Link]/classifications/icf/en.
WHO (1986) Ottawa Charter for Health Promotion. First International Conference on Health Promotion, Ottawa, 21
November 1986 – WHO/HPR/HEP/95.1. Geneva: WHO. Available online at: [Link]/hpr/NPH/docs/ottawa_
charter_hp.pdf.
WHO (2000) Health Systems Performance Report. Geneva: WHO. Available online at: [Link]/health-systems-
performance.
WHO (2003a) North Karelia Project: From Demonstration Project to National Activity. Geneva: WHO. Available online
at: [Link]/hpr/[Link].
WHO (2003b) International Migration, Health and Human Rights. Health & Human Rights Publication Series, Issue
No. 4. Geneva: WHO.
WHO (2004a) Global strategy: overall goal. Geneva: WHO. Available online at: [Link]/dietphysicalactivity/
goals/en/.
WHO (2004b) Global Strategy on Diet, Physical Activity and Health. Geneva: WHO. Available online at: [Link]/
dietphysicalactivity/en/.
WHO (2005a) The Bangkok Charter for Health Promotion in a Globalized World from: 6th Global Conference on Health
Promotion, Bangkok, Thailand, August 2005. Geneva: WHO. Available online at: [Link]/healthpromotion/
conferences/6gchp/bangkok_charter/en/.
WHO (2005b) Framework convention on tobacco control. Geneva: WHO. Available online at: [Link]/tobacco/
framework/en/.
WHO (2005c) Statement of the WHO on International Trade and Health. Geneva: WHO.
WHO (2006) Healthy Cities and urban governance. Geneva: WHO. Available online at: [Link]/healthy-
cities.
WHO (2007a) Ethical, Legal and Social Implications (ELSI) of Human Genomics. Geneva: WHO. Available online at:
[Link]/genomics/elsi/en/.
WHO (2007b) Water-Related Diseases. Geneva: WHO. Available online at: [Link]/water_sanitation_health/
diseases/en.
WHO (2008) International Classification of Diseases (ICD). Geneva: WHO. Available online at: [Link]/
classifications/icd/en/.
Wilkinson M (2006) Carbon Capture and Storage − [Link].
Wilkinson R, Marmot M (2003) Social Determinants of Health: The solid facts (2nd ed). Copenhagen: WHO. Available
online at: [Link]/document/[Link]. A short, readable summary providing evidence of the health
implications of social, cultural and economic circumstances.

591

Public [Link] 591 3/6/08 [Link]

 R e f e re nc e s

World Health Annual of Statistics (1991) Available online at: [Link]/whois/en.

Wright M (2005) Homosexuality and HIV/AIDS prevention: the challenge of transferring lessons learned from Western
Europe to Central and Eastern European Countries. Health Promotion International 20:91−8.
Zahra SA (1991) Predictors and financial outcomes of corporate entrepreneurship: An exploratory study. Journal of
Business Venturing 6:259−85.
Zborowski M (1952) Cultural components in response to pain. Journal of Society Issues 8:16−30.
Zoonekynd V (2007) Probability Distributions − [Link]/UNIX/48_R/[Link].

592

Public [Link] 592 3/6/08 [Link]

 Index

INDEX

360-degree feedback, management health impact assessment 556 Bandura (1977), social learning theory
483 health information 552 of health promotion 291
health needs assessment 556 bar charts 82
abdominal aortic aneurysm, screening, letters 557 Barnett formula 432
UK 166 needs assessment 550 Barthel ‘Index of Activities of Daily
academic paradigms 148 service specification 556 Living’ (ADL) 382
acceptability of services, assessing strategy 556 basic reproduction number (R0) 54
131–3 antenatal screening programme 165, Bayes’ theorem 103–4
Acheson Report (1998), healthcare 180 Beattie (1991), health promotion model
inequalities 312 appraisal, professional accountability 289
action planning 488 466 Beauchamp and Childress’s principles,
action research 139–44 Appraisal of Guidelines for Research and medical ethics 175
Activities of Daily Living (ADL) 382 Evaluation in Europe 496–7 behaviour change theories 463–4
adjustment see standardisation Arbuthnott formula 432 Belbin’s roles for an effective team 452
adult risk factor, life-course paradigm association Beveridge Report (1942), ‘giant evils’ in
148 and causation 26–8, 563 society, UK 417
adult screening programmes 166, 180–1 genetic studies 68 bias 26–7, 28–31
adverse selection, health insurance asthma 157–8 follow-up 30
491 asylum seeker 415 instrument 30
air pollution 215 attack rate 24 intervention studies 30
alcohol 209, 274, 285, 297 attributable risk 21–2, 23 meta-analysis 58
alternative therapies 390–2 attributable risk fraction formula 536 misclassification 29
Alzheimer’s disease and dementia 152–3 attributable risk per cent 21–2 observer 30
analysis, economic audit population data 334–5
cost–benefit 440–3 cycle 128–9 publication 62
cost–consequence 447 definition 128 screening 179
cost-effectiveness 438–9 health equity 506 selection 30
cost-minimisation 438 success criteria 129 volunteer 30
cost–utility 439 Australia binary data 10–11
analysis of variance (ANOVA) 93–4 cancer screening 167 binomial distribution 76
multivariate (MANOVA) 94, 95–6 censuses 327 birthweight, low 310
answer frameworks, MFPH Part A community care 392 Black Report (1980), health inequalities,
examination data linkage in Western Australia 358 UK 281, 413
briefing 549 health care provision and funding Black Triangle Scheme, prescribing, UK
communicable disease 553 493 356
critical appraisal 555 population pyramid 2005 331 blood pressure and salt intake 207
generic 549 autosomal inheritance, dominant and Bonferroni’s correction 88
health care evaluation 551 recessive 187 Bordetella pertussis 257

593

Public [Link] 593 3/6/08 [Link]

 I ndex

boxplot (box-and-whiskers plot) 83–4 Buncefield Oil Depot fire 233 national surveillance schemes and
Bradshaw’s categories, need 400 categories 230 infection control 238–9, 267–9
Brager and Specht (1973), health ladder Hazsub events, New Zealand 234 nosocomial (health care acquired)
316–17 national and local agencies, UK 232 infections 261–2
brainstorming 454 public information 231–2 sexually transmitted infections (STIs)
breast cancer risk minimisation 231 263–4
epidemiology 156–7 children’s screening programmes, UK vaccine-preventable diseases 256–60
genetics 186, 188 166 viral hepatitides 254–5
screening programmes 166–7 chi-squared (c2) test (2 x 2 table) 92–3, communication, effective methods 462
breastfeeding 311 537 see also leadership; negotiation and
briefing, answer framework, MFPH Part A Chlamydia trachomatis 263 influencing
examination 549 screening, UK 166 community care 392
Buncefield Oil Depot fire 233 cholera 248 community development 300, 319
Choosing Health (DH White Paper 2004) complementary therapy 390–2
Caldicott Guardianship, data 65 274–5, 281 confidence intervals 71–4
Campylobacter 247 chromosome structure 184 formula 540
Canada, health care provision and chronic obstructive pulmonary disease confidentiality 64
funding 493 158–9 confidential enquiry processes 129–30
cancer climate change 220, 461 confounding 27, 31–4
bowel screening 166, 167 clinical audit see audit Connecting for Health 357
breast screening 166, 167 clinical autonomy 395–6 Connecting for Health, NHS IT
cervical screening 166, 167 clinical governance 466–7 programme, England 371
prostate screening 166 clinical iceberg 389 consumer rights and choice 406–8
western forms of 156–7 clinical professionals, interaction with continuous data 10
carbon-offsetting 230 managerial professionals 462–4 Control of Substances Hazardous
cardiovascular diseases 154–6 Clostridium tetani 257 to Health (COSHH) (2002)
care pathway design 110 clustered data 41–2 Regulations, UK 225
Carstairs score, deprivation 123–4 Cochrane Collaboration 62–3 coronary heart disease (CHD) 154–5
case finding, screening 171–2 cohort component method, population Secondary Prevention Programme
case–control studies 37 estimation 334–5 361
control choice 47 cohort studies 38 correlation 99
nested 44–5 colorectal cancer correspondence
population-based genetic 68 control measures 194 emails 528, 529–30
categorical data, graphing 516 screening 166–7 letters 529
causality 27–8 combined model, predictive risk report to management 530–1
Bradford Hill criteria 28 algorithm 358 strategy documents 530
causation 26–8, 563 combining studies, meta-analyses Corynebacterium diphtheriae 256
censoring 8–9, 51 56–9 cost–benefit analysis (CBA) 440–3
censuses Commission for Social Care Inspection cost–consequence analysis 447
alternatives to traditional 328 (CSCI) 434 cost-effectiveness analysis 438–9
Australia 327 commissioning of services cost-minimisation analysis 438
conduct of censuses 325–6 allocation 505–6 cost–utility analysis 439
hard-to-count groups 327–8 fundholding and practice-based Cox’s regression 101
Hong Kong 327 commissioning 507 creativity and innovation 454–5
New Zealand 327 principles 505–6 critical appraisal
routine vs. ad hoc data 328–9 specialised 507 answer framework, MFPH Part A
small areas 327 Committee on Safety of Medicines (CSM) examination 555
UK 2001 census 326 356–7 Critical Appraisal Skills Programme
see also demography; population communicable disease (CASP) 512
central limit theorem 77 answer frameworks, MFPH Part A critical period model, life-course
central tendency (mean/median/mode) examination 553 paradigm 147
11 emerging infections 265–7 cross-sectional studies 36
cervical cancer gastrointestinal infection 247–53 cryptosporidiosis 249
epidemiology 156–7 immunisation programmes 241–4 cumulative incidence see risk
screening programmes 166–7, 431 international surveillance 240–1, CYP gene, genotyping test 192
chance, association and causation 26 271–2 cystic fibrosis 189
chemical incident management methods of control 241 cytochrome P450 CYP gene, genotyping
assessment 231 mortality 235 test 192

594

Public [Link] 594 3/6/08 [Link]

 Index

Dahlgren and Whitehead (1991), health diagnostic tests vs. screening 170–1 electronic data storage 65–6
‘rainbow’ 276–7 Diderichsen and Hallqvist (1998), social emails, correspondence 528, 529–30
data determinants frameworks 279–80 emerging infections
analysis, computer packages 522 diet see nutrition influences 265
central tendency 11 dietary reference values 204 pandemic influenza virus in humans
drawing conclusions from 513, diphtheria 256 (H5N1) 266–7
515–22 disability and handicap UK and European concerns 266
methods of collection, qualitative definitions 382 England, population pyramid 2005
140 measuring disability 382 330
qualitative 9–10, 139–44 measuring handicap 382 environment
quantitative 10–11 Disability Rights Commission 383 chemical incident management
recording 49 disability-adjusted life years (DALYs) 230–4
spread 11–12 444 climate change 220
summarising 9–11 discrete data 10 control legislation 223–4
see also routine data from health disease determinants of disease 211–18
statistics burden, measures 17 environmental determinants of disease
data linkage communicable 235–72 211–18
demographic and health service diagnosis and screening 163–82 hazard and risk 218–19
information 502 global burden of disease 149–50 housing conditions 221–2
pharmacovigilance 357–8 ‘important’ diseases, criteria 149–50 transport 228–30
Data Protection Acts 64–5 mortality, by WHO region 150 environmental impact assessment (EIA)
data sources public health knowledge 150–1 135–6
needs assessment 109, 111–13 see also specific diseases epidemic theory 52–6
UK equivalence 112 dispersion, measures of (statistics) 80–1 curve 54–5
death certificates and registration, UK distributive justice 401, 406 threshold 54
344 DNA structure and biochemistry 184–5 epidemiological measurement
decision analysis, health economics gel electrophoresis 197–8 common errors 19–21, 26–7
444–5 polymerase chain reaction (PCR) 196 statistical analysis 52
defined daily dose (DDD) 355 restriction enzymes 195 variation, sources of 18–19
delegation of goals 456 sequencing 197 epidemiological paradigms 147–8
Delphi technique, surveys 130–1 Donabedian’s framework 116–17, 551 epidemiology 1–68, 554
demand fluctuations 110 Down’s syndrome, antenatal screening, definition 3, 6
dementia 152–3 UK 165 denominator 7
demographic information, health costs drug misuse and health outcomes 209, descriptive 6–7
501–2 285 exception reporting 55–6
demography Drug Testing and Treatment Orders further reading 559
age distribution 330–2, 333 (DTTOs) 379 geographical patterns 55
basic concepts 329 draw see think (DST) process, strategic numerator 7
effects of change on health care planning 488 preventive strategies 133–4
341 Dunbar’s Number, social networks specific diseases 149–61
ethnicity 333–4 474–5 study design 35–40
gender differences 333 equity
population pyramids 330–2 early years 309–12 concepts 405, 431
see also censuses; population ecological fallacy 35 and efficiency 405–6, 431
Department of Health, England ecological studies 34–5 equity and equality, health care 126–8
infection control policies 268 economics see health economics errors, type I and type II 87–8
White Paper Choosing Health (2004) effect modification 32 Escherichia coli (vero cytotoxin
274–5, 281 Effective Provision of Pre-school producing) 251
depression 151–2 Education (EPPE) Project 311 estimation 78
deprivation, health effects 318–19 effects of demographic change on health ethics
deprivation measures 121–5 care 340–1 committees 66–7
composite indicators 125 efficiency screening programmes 175
descriptive studies 34–5 concepts 405, 430 ethnic groups, male-to-female ratio
determinants of health 275–80, 313–14 and equity 405–6, 431 333
deviance, primary and secondary 379 incentives to improve 435 European Standard Population 14
diabetes 160–1 electronic bibliographical databases evaluation of health services 361–2
as a polygenic disorder 190 59–60 Evans and Stoddart (1990), health field
screening, UK 166 limitations 59, 60 model 278–9

595

Public [Link] 595 3/6/08 [Link]

 I ndex

Evelina Children’s Hospital, London funding health services see health guidelines
317 economics appraisal criteria 496–7
evidence-based medicine (EBM) 60–1 funnel plots 62, 101–2 development of 496
Ewles and Simnett (2003), health further reading
promotion model 289–90 chapters 565–80 Haase and Pratschke index, deprivation,
exam technique see Membership of the revision tips 545 Ireland 124
Faculty of Public Health (MFPH) Part top fives 559–61 Haemophilus influenzae type b (Hib) 258
A examination hazard and risk, environmental 218–19
Expert Patient Programme 317–18, gastrointestinal infection 247–53 hazardous substances
378–9 risk of spread 245 disposal 213
external audit, spending 508 gel electrophoresis 197–8 (Hazsub) events, New Zealand 234
externality, economics 406 gender differences, demography 333 Head Start programme, USA 311
gene therapy 192–3 health and safety at work
Fagan’s nomogram, post-test probability gene–environment interactions 191–2 COSHH Regulations 225
174 General Medical Council 464 occupational health regulations 227
familial adenomatous polyposis (FAP) genetics 183–98 physical effects 225
194 epidemiological studies 67–8 psychological factors 225–6
familial polyposis coli 194 gene therapy 192–3 safety at work audits 227
family care 392 gene–environment interactions 191–2 UK agencies 226
family studies, genetic 67 genetic disease in relatives 193–4 health belief model (HBM) (1958),
fatty foods, taxation 298 genotypes and phenotypes 189 health promotion 290
fertility indices 5–6 health promotion and genetic disease health care
fetal anomaly scan 165 191–2 alternative sources 389–93
fetal sex determination, early gestation human 183–5 answer framework, MFPH Part A
186 inherited disease types 186 examination 551
finance, health services see health Mendelian inheritance 186–7 clinical autonomy 395–6
economics molecular biology 194–8 health care acquired infections (HCAI)
flow charts, design of 518 non-Mendelian inheritance 187–8 261–2
focus groups 139–44 polygenic disorders 189–90 history of 499
follow-up bias 30 screening, ethical, social and legal hospitals as social institutions 393–4
Food Standards Agency 214 implications 181–2 illness behaviour 396–7
dietary advice 206 variable gene penetrance 188 needs, assessment and evaluation
Forest plot 58–9 variation 185 107–37
formulae genotype 189 professions 394–5
attributable risk fraction 536 test, CYP gene 192 provision and funding, international
chi-squared (c2) (2 x 2 table) 537 Germany, health care provision and comparisons 493–4
confidence intervals and standard funding 493 psychology of decision-making 397
errors of the mean 540 global warming 220 shortcomings 491–2
McNemar’s test 538–9 goals, delegation of 456 health economics
number needed to treat (NNT) 42–3, gonorrhoea 263 analysis 437–8
536 GP contract (2003), NHS 395–6 appraisal 435–6
odds ratio 24, 25, 536 graphs (display of data) assessing performance 431–2
predictive power, positive and common pitfalls 518–20 audit of spending 508
negative 535 design of 516 benefits 436–7
relative risk 23–4, 536 drawing conclusions from 520–1 commissioning of services 505–8
sensitivity and specificity 534 which type to use 516–17 cost of services 501–2, 504
standard error, applications of 537 greenhouse gas emissions, cutting 461 decision analysis 444–5
standardisation, direct and indirect grey literature 60 economics screening 175–6, 178
13–16, 539 groups and team dynamics evaluation and priority setting 446–7
weighted averages 540 Belbin’s roles 452 financial balance 504
weighted capitation 432 changing behaviour 467–9 financial resource allocation 432–3,
Framework Convention on Tobacco creativity and innovation 454–5 505
Control (2005) 307 interaction between professionals funding health services 489–90, 491
France, health care provision and 462–4 health care systems and incentives
funding 493 motivation 452–3 433–5
fraud detection, spending 508 team dynamics 453 market forces factor (MFF) 505
Freedom of Information Act (2000) teamwork 454 paying for services 502–3
408–9 understanding individuals 451–2 principles 421–31

596

Public [Link] 596 3/6/08 [Link]

 Index

health education 284, 285 quality assessment and assurance immunisation programmes 241–4
communication in 293–6 125–6 immunity 237
health field model 278–9 shortcomings 491–2 immunoassays 270–1
health impact assessment (HIA) 136, structure, process and outcomes impact assessment 319–20
228 116–18 see also health impact assessment
answer framework, MFPH Part A utilisation and performance 110–13 incidence 8, 12–13
examination 556 health statistics 4–5 incineration of waste 212
transport in London 229 data handling 4 Index of Multiple Deprivation (IMD)
health inequalities and policy 413–14 routine data 4 122
health informatics 371 vital statistics 4 infectious diseases see communicable
answer frameworks, MFPH Part A health status 119 disease
examination 552 measurements of 352–4 influenza
health ladder 316–17 Health Survey for England 353 enhanced surveillance for pandemic
health needs assessment, answer health technology assessment 113–14 239
framework, MFPH Part A NICE Health Technology Appraisals influenza A/H5N1 239, 266
examination 556 114 information
health policy health-adjusted life-expectancy (HALE) applications 359–71
dilemmas 280–3 17 governance, NHS 66
further reading 559 Healthcare Commission (HC) 434 as measures of provision and usage
health promotion Healthy Cities (1988), WHO 307 362–4
appropriate settings 284–7 healthy worker effect 30 public access to 408–9
community development 300–1 hearing screening 165–6 use in planning and evaluation
determinants of health 275–80 heart disease 154–5 359–62
further reading 560 hepatitis A (HAV) 254 uses of mathematical modelling
health education 284, 285, 293–6 hepatitis B (HBV) 254 techniques 364–6
international initiatives 306–8 screening, NZ 167 Information for Action, global
legislation 296–8 serology 265 programme for vaccines and
models 287–92 hepatitis C (HCV) 255 immunisation, WHO 240
partnerships 301–5 hepatitis D (HDV) 255 information systems 362–3
policy dilemmas 280–4 hepatitis E (HEV) 255 information technology 370–1
programmes 298–300 hereditary non-polyposis colorectal informed choice, screening 176–7
responsibility for health 273–5 cancer (HNPCC) 194 informed consent 64
risk behaviour 292–3 Herzberg’s Two-Factor Theory, motivation inherited disease types 186
Health Protection Agency (HPA), UK 478–9 instrument bias 30
55–6, 238, 268 histogram 84 integrated care pathway (ICP) 434,
Centre for Infections (CfI) 238, 268 HIV/AIDS 264, 283, 308 497–8
NOIDS weekly report 238 spread along trade routes 417 intention-to-treat analysis 41
tuberculosis strain identification 198 Hong Kong internal audit, spending 508
WARNER notification 56, 238 censuses 327 International Classification of Disease
Health Protection Surveillance Centre health care provision and funding (ICD) 350–2
(HPSC), Ireland 239 494 international health regulations 240–1,
health ‘rainbow’ 276–7 population pyramid 2005 332 271–2
health service(s) hospital episode statistics (HES) 346 interprofessional learning 455–6
appropriate and acceptable services hospitals as social institutions 393–4 Intersalt study 207
131–3 housing conditions 221–2 interval censoring 8
comparative studies 115–16 human behaviour, methods of study interval data 10
consultation about 498 375–6 intervention studies 39–40
delivery and organisation 114 human genetics 183–5 allocation methods 48
development and planning 489 human immunodeficiency virus (HIV) bias 30
equity and equality 126–8 264 economics 447
funding 489–90, 491 hypothesis testing 79, 86–7 investment in health improvement
history of 499 418–19
indices used by 366 iatrogenesis 383 Ireland
modernisation, views of different health care acquired infections (HCAI) community care 392
professions 463 261–2 population pyramid 2005 331
need for 107–9 ideological dilemmas, health policy
organisation of 490–1 280–3 Jarman score, deprivation 123
planning and evaluation 359–62 illness behaviour 396–7 job satisfaction 478–9

597

Public [Link] 597 3/6/08 [Link]

 I ndex

kappa (k) statistic 94, 96–7 male-to-female ratio 333 meta-analysis 57–9
Keys’ Seven Countries study, cholesterol management bias 58
207–8 360-degree feedback 483 meticillin-resistant Staphylococcus aureus
frameworks for managing change (MRSA) 262
Lalonde Health Field Concept (1974) 479–82 microbiological techniques
276, 554 further reading 560 nucleic acid reference systems 270
land, environmental regulation 215 models and theories 477–9 traditional methods 269–70
landfill, waste disposal 212–13 performance management 482–3 migration
Large Analysis and Review of European management by objectives (MBO) 459 classification of migrants 415–16
housing and health Status (LARES), management by wandering around health effects 414–15, 416
WHO 222 (MBWA), Hewlett-Packard Millennium Development Goals, UN,
leadership Corporation 458 water shortages 223
changing behaviour 467–9 management skills, personal misclassification bias 29
delegation 459 conflict management 457 mitochondrial inheritance 188
feedback 458 stress management 457 modelling see mathematical modelling
interaction with other professionals time management 456–7 techniques
462–4 manager, effective see leadership molecular biology 194–8
management by objectives (MBO) managerial professionals, interaction moral hazard 491
459 with clinical professionals 462–4 morbidity data 345–9
management by wandering around Mann–Whitney U-test 94 mortality
(MBWA) 458 margins, economic 428–9 data 343–5
motivation theory 479 market failure 427 premature 16
power and authority 464 market forces factor (MFF) 505 tables (see life-tables)
professional accountability 464–7 Markov modelling 441 mortality and morbidity data 343–9
work environment 458–9 Maslow’s Hierarchy Theory, motivation mortality indices 5–6
see also communication; negotiation 477–8 calculation 9
and influencing mass communication 295–6 motivation
lead-time bias 179 maternal nutrition 310 Herzberg’s Two-Factor Theory 478–9
left-censoring 8 mathematical modelling techniques, uses leadership theory 479
legal aspects and legislation in planning 364–6 Maslow’s Hierarchy Theory 477–8
environmental control 223–4 Maxwell's dimensions of quality 126, motivator–hygiene theory 478–9
health promotion 296–8 551 MRSA 262
screening 176 McKinsey 7S organisation analysis multiple regression 98
length-time bias 179 479–80 linear 100
leptospirosis 266 McNemar’s test formula 538–9 logistic 99
letters measles 260 multivariate analysis of variance
answer framework, MFPH Part A measurements of health status 119–21, (MANOVA) 94, 95–6
examination 557 352–4 mumps 259
correspondence 529 mediating factor 31, 32 mutations 185
life-course paradigm Medicines and Healthcare Products Mycobacterium tuberculosis (MTB) 259
adult risk factor 148 Regulatory Authority (MHRA), UK
programming 147 356–7 National Air Quality Information Service,
lifestyle and health 199–210 Membership of the Faculty of Public UK 215
life-tables 336–7 Health (MFPH) Part A examination National Disaster Organisation, Australia
application of 337 answer frameworks 549–57 231
cohort 337 drawing conclusions from data National Disease Audits 116
period 336 515–22 National Health Service (NHS)
likelihood ratios, screening 172–5 examination structure 541–2 An organisation with a memory
linkage disequilibrium 68 examination technique 545–7 (2000), risk management 495
linkage studies 67 formulae required 533–40 history 492
local strategic partnerships (LSPs) marking algorithm 543 international comparisons 493–4
316–17 research design and critical appraisal National Institute for Health and Clinical
location, measures of (statistics) 79–80 511–13 Excellence (NICE)
logistic regression 99 revision tips 543–7 economic evaluation of physical
log-rank test 100–1 written presentation skills 523–31 activity interventions 418–19,
London Health Commission reports Mendelian inheritance 186–7 446
278 meningitis 258, 259 experimental study ranking 115
long-term conditions 390–2 meningococcal disease 258 health technology appraisals 114

598

Public [Link] 598 3/6/08 [Link]

 Index

National Sentinel Stroke Audit 116 early childhood 201 pertussis (whooping cough) 257
National Statistics SocioEconomic evidence for health-related goals 207 PESTELI organisation analysis 479–81
Classification (NSSEC) 121–2 factors affecting dietary choice 203 pharmacogenomics 192
need further reading 561 pharmacovigilance 356–8
Bradshaw’s categories 400, 401 long-term effects 200 data linkage 357–8
definition 399 national implementation programmes see also prescribing
distinctions between need and 204–6 phenotype 189
demand 427–8 pregnancy 201 phenylketonuria (PKU), prevention 191
and social justice 399–400 recommendations 204 pie charts 82
needs assessment regional dietary variations 209 placebo 30
answer framework, MFPH Part A short-term effects 200 planning and evaluation 359–62
examination 550 nutritional intervention 201–3 steps 360
comparative 108 breastfeeding 311 pneumococcal conjugate vaccine (PCV)
corporate 108 complex approaches 208–10 244
data sources 109, 111–13 low birthweight 310 point (period) prevalence 174
evidence-based 108–9 maternal nutrition 310 Poisson distribution 76–7
participatory 110 police road traffic accident reports
steps 109 obesity 201 (STATS19), UK 344
negligence 496 observational techniques, validity 50–1 policy
negotiation and influencing observer bias 30 balancing equity and efficiency
influence 460–1 occupation and health 405–8
negotiating styles 459–60 occupational health regulations 227 formation 410–11
public health advocacy 461 unemployment 227–8 and health inequalities 413–14
see also communication; leadership odds ratio 24, 25, 536 implementation problems 409–10
Neisseria gonorrhoeae 263 Official Information Act (1982), NZ 409 international influences 417
Neisseria meningitidis 258 opportunity cost, economics 429 investment in health improvement
nested case–control studies 44–5 ordinal data 10 418–19
neuropsychiatric conditions 151–4 organisations political ideology and policy formation
New Zealand development 481–2 411–12
censuses 327 frameworks for managing change polluter pays 224
Hazsub events 234 479–82 polygenic disorders 189–90
health care provision and funding performance management 482–3 polymerase chain reaction (PCR) 196
494 stakeholder analysis 472 population 340
Official Information Act (1982) 409 stakeholder interests 471–2 biases and artefacts in data 334–5
sources of morbidity information Ottawa Charter (1986) 281 cohort component method, resident
349 outbreak investigations 244–5, 246 populations 334–5
New Zealand index of socioeconomic outcome measures 52, 53 effects of change on health care 341
deprivation for individuals (NZiDep) output index 366 effects of fertility, mortality and
122, 124 migration 338
nominal data 10, 11 pandemic influenza estimation methods 334–5
non-probability sampling 46–7 enhanced surveillance for 239 historic changes 339–40
normal distribution 77 virus in humans (H5N1) 266–7 Immigration 339
North Karelia project, Finland 202 parallel and serial testing, screening life-tables 336–7
Northern Ireland Health and Social 169 projections 335–6, 337
Wellbeing Survey 353–4 parametric and non-parametric tests special groups 335
Northern Ireland Multiple Deprivation 88–96 stochastic predictions 335–6
Measure (NIMDM) 124 paramyxovirus 259, 260 time bomb 338–9
nosocomial (health care acquired) paratyphoid fever 252 see also censuses; demography
infections 261–2 Parkinson’s disease 154 population at risk 7
notifiable diseases 238, 268–9 participatory needs assessment 110 population attributable risk 22
nucleic acid probes and amplification partnerships for health promotion population pyramids 330–2
systems 270 301–5 port health regulations 272
null hypothesis (Ho) 26, 52 penetrance 188 positive predictive value 174
numbers needed to treat (NNT) 42–3, performance poverty 281–2
536 assessing, economics 431–2 power, interests and ideology 411–12
nutritional health management 482–3 power, statistics see statistics
assessment in populations 199–200 management indicators 360–1, 367–8 power and authority, leadership 464
deficiency syndromes 201–3 pertussis vaccine 244 precautionary principle, pollution 224

599

Public [Link] 599 3/6/08 [Link]

 I ndex

predictive power (positive and negative) protein–energy malnutrition (PEM) 201 replacement fertility, population
formulae 535 provider moral hazard 491 estimation 335
predictive risk modelling 365 Provincial Indices of Multiple Deprivation report writing
prescribing (PIMD), South Africa 124 appropriate language 524
data 354–5 psychology of decision-making 397 organisation 523–4
information, primary care 348–9 public access to information 408–9 preparation and presentation 523
units of measurement 355 public health, history of 499 preparation for publication 524–6
see also pharmacovigilance public involvement 315–8 presentations to different audiences
Prescribing Analysis and CosT (PACT) publication bias 62 526
data 354 publication of papers 524–6 press releases 527–8
electronic system (EPACT) 354–5 pulmonary disease 158–9 report to management 530–1
prescribing units (PUs) 355 strategy documents 530
presentations to different audiences qualitative data research evidence
526 appropriate use 141 ethics of 63–5
press releases 527–8 common errors 143–4 hierarchy of 61–2
prevalence 12–13 ethical issues 141–2 research studies
prevention paradox 134, 283–4 principal methods of collection critical appraisal 511–13
preventive strategies 139–40 design 511
epidemiological basis 133–4 summarising 9–10 drawing conclusions 513, 515–22
levels 134 validity, reliability and generalisation statistical techniques 513
population and high-risk approaches from 141–3 target audience 511
134–5 Quality Management and Analysis System resources
price elasticity, economics 426–7 (QMAS), primary care, UK 348 allocation 432–3
primary care quality of life (QOL), measurement priorities and rationing 402
planning and evaluation 359–62 119–20 respiratory diseases 157–9
practice profiles, quality improvement quality-adjusted life years (QALY) responsibility for health 273–5
364 440–3 restriction enzymes 195
prescribing information 348–9 quantitative data 10–11 revision tips
star rating system 368 examination structure 541–2
priorities and rationing 402–4 radiation examination technique 545–7
probabilistic population forecasting measuring 217 marking algorithm 543
336 monitoring and control 218 schedule 543–4
probability 69–75 non-ionising 217 suggested further reading 545
addition rule (OR) 70 Radiation Protection Division, UK 218 rheumatic fever 266
conditional 74–5 randomisation, intervention studies right-censoring 8
distributions 70–1 48–9 risk 21–3
multiplication rule (AND) 70 randomised controlled trials (RCTs) attributable 21–2, 23
post-test 174–5 58 management 494–6
pre-test (prevalence) 174 ranked order approach, strategic negligence 496
sampling 45–6 planning 488 relative 23–4, 536
theory, elementary 69–70 rate, epidemiology 8 Sandman’s concept 218–19
procedural justice 402 rate ratio 23 risk behaviour 292–3
Prochaska and DiClemente (1984), stages ratio, epidemiology 7 risk difference 23
of change health promotion model ratio data 10, 11 risk ratio 23
292 rationing resources RNA 185
professional accountability criteria 404 role of
appraisal 466 health care priorities in Oregon 404 doctor–patient 378
clinical governance 466–7 levels 403 medicine in society 378
professional bodies 464–5 prioritisation 402 the sick 377
recent changes 465–6 stakeholder involvement 403 Rose, Geoffrey, The Strategy of Preventive
Professional Executive Committees recall bias 30 Medicine 134
(PECs) 464 receiver operating characteristic (ROC) routine data from health statistics 4,
professionalisation 394–5 curves, screening 170 360
prognosis studies 51–2 refugees 415 improving reliability 370
proportion, epidemiology 7–8 regression strengths and weaknesses 369
proportional hazards regression see line equations 99 uses 369
Cox’s regression techniques 99–100 vs. ad hoc data in censuses 328–9
prostate cancer screening, UK 166 relative risk formulae 23–4, 536 rubella 260

600

Public [Link] 600 3/6/08 [Link]

 Index

safety at work audits 227 situation target path (STP) process, star rating system 368
salmonellae 251 strategic planning 488 statistical inferences 78–9
salt intake and blood pressure 207 small area analyses 35 statistics 69–106
sample size, statistics 97–8 smoking bayesian 103–4
sampling methods, population studies in enclosed public places 417 choice of test 104–6
45–7 Framework Convention on Tobacco distribution curves 76–8
Sandman’s concept, risk 218–19 Control (2005) 307 errors, type I and type II 87–8
scarcity health outcomes 209 funnel plots 101–2
economics 422 during pregnancy 310 graphical methods 82–5
resources 402 reducing among NHS staff 300 heterogeneity 101, 102
scatter plot 85 social capital 387 hypothesis testing 79, 86–7
schizophrenia 153–4 social determinants frameworks 279–80 location and dispersion, measures of
Scottish Health Survey 353 social justice 79–81
Scottish Index of Multiple Deprivation distributive justice 401, 406 making inferences 78–9
(SIMD) 123 and need 399–400 parametric and non-parametric tests
screening procedural justice 402 88–96
aims and principles 163–4 utilitarianism 400–1 power 26, 39, 83, 97–8
biases 179 social learning theory, health promotion probability 69–75
case finding 171–2 291 regression and correlation 98–101
developing policies 179–81 social marketing sample size 97–8
ethical, economic, legal and social behaviour change 314–15 screening tests 168–70
aspects 175–6, 181 and disease prevention 210 stem-and-leaf display 83
genetic 181–2 social networks 474–5 stigma 379–81
informed choice 176–7 sociology challenging, virtuous circle 381
likelihood ratios 172–5 disability and handicap 382–3 stochastic predictions, population
National Screening Committee, UK further reading 560 335–6
summary 562 human behaviour 375–6 strategy development
national screening programmes illness as a social role 376–9 answer frameworks, MFPH Part A
164–8, 180–1 medicine in society 383–4 examination 556
newborn 165, 167, 180 primary and secondary deviance communication 485–6
planning, operation and evaluation 379 guideline development 496–7
177–8 social capital and social epidemiology health service development and
statistical aspects 168–70 387 planning 489
vs. diagnostic tests 170–1 social factors in disease aetiology implementation 486–7
screening programmes 166 386–7 integrated care pathway (ICP) 497–8
selection bias 30, 179 social patterns of illness 384–6 risk management 494–6
self-care 390–1 stigma 379–81 strategic planning, theories of
self-help groups 392–3 South Africa, community care 392 487–8
sensitivity analysis 444 specificity formula 534 strategy documents 530
sensitivity formula 534 spread (range/variance/standard stress management 457
sequencing DNA 195–8 deviation) 11–12 stroke 156–7
service specification, answer framework, stages of change health promotion study design
MFPH Part A examination 556 model 292 epidemiology 35–40
service utilisation, primary and stakeholder further reading 561
secondary care indicators 363 analysis 472 scientific 115
Seven Countries study, cholesterol interests 471–2 service utilisation and performance
207–8 leverage by 472473 113
severe combined immunodeficiency standard error sun exposure and health outcomes 209
(SCID) 193 applications of 537 supply and demand, economics
sexual behaviour and health outcomes formula 540 curves 424–5
209 standardisation 13–16 principles 422–4
sexually transmitted infections (STIs) direct 13–14, 15 Sure Start Local Programme (SSLP) 312
263–4 formulae 539 surveillance principles 237
shigellae 250 indirect 14, 16 surveys
sickle cell disease 159 standardised mortality ratio (SMR) 14, Delphi technique 130–1
Sierra Leone, population pyramid 2005 24 documentation 49
332 Staphylococcus aureus, meticillin- survival analysis 51–2, 101–2
significant clusters 56 resistant 262 sustainability 220–1

601

Public [Link] 601 3/6/08 [Link]

 I ndex

Sustainable Development Commission USA Welsh Townsend score 123, 432

221 Head Start programme 311 whooping cough 257
SWOT organisation analysis 479–81 health care priorities in Oregon 404 Wilcoxon’s test 94
syphilis 264 health care provision and funding Wilson and Jungner’s criteria, screening
systematic review 56–7, 62 494 programme assessment 178
user and carer involvement 409 work see health and safety at work
tables (displaying data), design of 515 user charges 492 world
Tannahill (1985), health promotion utilitarianism, social justice 400–1 effects of demographic change on
model 287–8 health care 340–1
target setting 319–21 vaccination see immunisation population 340
targeted or universal health promotion programmes World Health Organization (WHO)
282–3 vaccine-preventable diseases 256–60 core obligations for member states
team working see groups and team validity 272
dynamics observational techniques 50–1 Framework Convention on Tobacco
test, statistical, choice of 104–6 questionnaire 50 Control (2005) 307
tetanus 257 value-laden statistic 16 genetic screening recommendations
three-step models, strategic planning variation 182
487–8 genetic 185 global burden of disease 149–50
threshold setting, screening 169 sources of 18–19 Global Strategy on Diet, Physical
time at risk 8–9 varicella vaccine 244 Activity and Health (2004)
incidence rate 8 vascular risk screening, UK 166 474
time series designs 43–4 veil of ignorance, social justice 401 Health Systems: Improving
time trend analysis 43 Vibrio cholerae 248 Performance (2000) 367
total institutions 393–4 viral hepatitides 254–5 Healthy Cities (1988) 307
Townsend score, deprivation 123 virtuous circle, challenging stigma 381 impairment, disability and handicap
traffic light indicators, practice profiles vision, school-aged screening 166 classification 119
364 volunteer bias 30 Information for Action, global
transport programme for vaccines and
reducing car usage 229 waiting times 110–11 immunisation 240
reducing environmental impact 230 waste disposal International Classification of Disease
sustainable 228–9 hazardous 213 (ICD) 350–2
trans-theoretical health promotion model incineration 212 Large Analysis and Review of
292 landfill 212–13 European housing and health
Treponema pallidum 264 management 212 Status (LARES) 222
t-test 89–91 water Ottawa Charter for Health Promotion
tuberculosis 259 inadequate supplies and sanitation (1986) 136
strain identification 198 222–3 social determinants of health 387
twin studies 67 monitoring and control 213 written presentation skills 523–31
type I and type II errors 87–8 pollution 213
typhoid 253 tests 214 years of life lost (YLL) 16–17
weighted averages formula 540 Yellow Card Scheme 356
uncertainty, statistics 70–1 weighted capitation formula 432
unemployment 227–8 Welsh Health Survey (WHS) 353 z-test 89, 91–2

602

Public [Link] 602 3/6/08 [Link]