07

CASE-CONTROL
STUDY DESIGN

KSU COLLEGE OF MEDICINE

2019 - 2020

ACKNOWLEDGMENTS
TEAM LEADER
SARAH ALENEZY
TEAM MEMBERS
HAIFAA TALEB
RAWAN MISHAL
REVIEWER
YAZEED AL-DOSSARE
 Original, refrased. Our notes. Doctors notes. Golden notes.

CASE-CONTROL PAGE 02

TABLE OF CONTENTS

how to
Design of
conduct a
case-control
case-control
studies
study ?

Case
Control
Study
Design
confounding
Determine
in a
Study
case-control
Subjects:
study

LECTURE OBJECTIVES

By the end of this lecture, I am able to:

Describe the design of case-control studies.
QU

Identify steps for conducting case-control studies.

IZ

Identify issues in the design of case-control studies.

Describe the strengths and weakness of case-control studies.
 CASE-CONTROL PAGE 03

Design of case-control studies

All studies

Descriptive (PO) Aanalytical (PICO)

Generate hypotheses

Case report
Experimental Observational

Tests hypotheses
Case series Randomized
clinical trials
(RCTs) Group data Individual
Cross-sectional
(survey)
Cross-sect
Ecological ional
(analytical)
Qualitative

Cohort

Case-control
 CASE-CONTROL PAGE04

Design of case-control studies

A case-control study is a study that compares patients who have a disease or outcome
of interest (cases) with patients who do not have the disease or outcome (controls),
and looks back retrospectively to compare how frequently the exposure to a risk
factor is present in each group to determine the relationship between the risk factor
and the disease.

The study starts from the outcome

unlike cohort study that starts
When to conduct a case-control study from the exposure and goes
forward
● The outcome of interest is rare

● Multiple exposures may be associated with a single outcome This part is

important
● funding or time is limited

(1) To investigate cause-effect when experimental trials (e.g. RCT) are not ethical or
feasible, (lung cancer and smoking)

(2) To investigate cause-effect when cohort studies are expensive or non-feasible e.g.
(to investigate etiology of rare disease e.g. cancer)

How to conduct a case-control study ?

steps in conducting 1- define a source population
a case-control study

e.g.I take the medical 2- determine study subjects:”cases”

records and select all the pts (case-subjects:they have the disease or
with this disease“ regardless outcome of interest)
the gender and age”

3- determine study subjects: “controls”

The most
(control-subjects: they DO NOT have the
challenging
disease or outcome of interest)
step

4- decide on the ratio of cases to controls

Why ratio? Bc the cases are rare so
the sample size will be so small so
we increase it by the controls
 CASE-CONTROL PAGE05

How to conduct a case-control study ?

e.g. old patients with a rare 5- decide on matching cases
lung cancer dont match with
young ppl even if they
and controls
smoke too

6- estimate sample size

7- select cases and controls

8- measure exposure (risk factors)

9- analyze the data

past present

exposed cases
asses exposure history

(diseased)
determine “cases” and

unexposed
source
population
“controls”

exposed controls (no

disease)

unexposed
 CASE-CONTROL PAGE06

Determine Study Subjects: “Cases”= WHO IS THE CASE

[Link] of cases [Link] for cases

Cases admitted to or
1) establish a “standard case
hospital-

discharged from a hospital,

definition”: adopt a “standard
based

clinic or any health care facility.

diagnostic criteria”.

2) set inclusion and exclusion

criteria: area of residence, age,
● death certificates with gender, etc.
recorded cause of death.
population- based

● disease registries(e.g. 3)decide on the type of cases:

cancer registry).
● incident cases in a going - incidence cases(newly
cohort study. diagnosed cases)
● cases reported or
diagnosed during a - prevalent cases (people who
survey or surveillance may have had the disease for
system. some time).
● employment records.
 CASE-CONTROL PAGE07

Determine Study Subjects: “Controls” = WHO IS

THE CONTROL

hospital-based controls community-based controls

advantages disadvantages advantages disadvantages

1. Subjects are easily 1. Differential 1. reduction of 1. time and money

accessible. hospitalization selection bias. consuming. Some
patterns may people want money
2. Patients usually introduce selection for participation.
bias. the admission 2. generalization
have time to
participate. criteria for the cases of study results is
might be different from more valid. 2. may suffer low
3. Patients are often the admission criteria participation rate.
for the Controls
motivated to 3. may provide
cooperate with convenient 3. cases and
investigators. 2. difficult to blind control may
disease status from control of
extraneous exhibit differential
4. Controls and cases cases and controls
(confounding) recall of prior
may be drawn from
3. an underestimate variables. exposures.
similar social and
geographical of the study effect
environment. In the may result if control’s
small cities i expect disease is
that the determinants etiologically similar
are similar in a certain to cases’ disease.
hospital, however this
is not applicable in the
big cities
5. Differential recalls
of prior exposure is
likely to be
minimized.
 CASE-CONTROL PAGE08

Determine Study Subjects: “Controls” = WHO IS

THE CONTROL

Selection of Controls:
- The ideal “controls” are the healthy ones (very challenging!)

- it is crucial to select control group/s from people who we are certain not to have
got the specified disease/condition.
The most important thing is that the
- Aim of selecting controls: controls don't have the disease of interest

Is to compare the exposure rate among both cases and controls (e.g. % smoking
among cases and controls).
Then to confirm/refute if that the risk factor has occurred more frequently in the
cases than in the controls using the measurement of association..

Determining
controls

you have to be careful

that they don't have the
risk factors of this
disease
 CASE-CONTROL PAGE09

Decide on the ratio of cases to controls

There will be a
question from this part
● The ratio of cases to control should be at least and ideally 1:1.

● However, in many situations we may not be able recruit a large number of cases
and it may be easier to recruit more controls for the study.
● It has been suggested that we can increase the number of controls to increase
statistical power (if we have limited number of cases) of the study.

● Increase in the ratio lead to increase in “study precision”: 1:2, 1:3, 1:4 The highest
thing is 4 controls for 1 case
● Further increase in the ratio is associated with little increase in study precision
relative to the cost involved (i.e will not add much to the study power but will add
to the cost!)

Decide on matching cases and controls

A major concern in conducting a case-control study is that cases

and controls may differ in characteristics or exposures other than
the one that has been targeted for study.
● An approach to deal with this challenge: Matching!
● Matching: The process of selecting the controls so that they
are similar to the cases in certain characteristics
(confounders), such as age, race, gender, socioeconomic
status, and occupation.
● Matching reduces the possible confounding effect
● Matching on several characteristics is not advisable as
i. Creates difficulties in finding controls
ii. Requires more complex statistical analysis
iii. May result in overmatching
 CASE-CONTROL PAGE 10

Analysis in case-control studies

- The odds ratio (OR) is used in case-control studies to estimate the strength of the
association between exposure and outcome.

- Note that it is not possible to estimate the incidence of disease from a case control
study unless the study is population based and all cases in a defined population are
obtained.

- The odds ratio is a measure of the odds of disease in the exposed compared to the
odds of disease in the unexposed (controls) and is calculated as: OR = ad/bc.

- OR interpretations: OR>1, OR=1, OR<1

vaping and pulmonary illness

Case-control study of vaping and pulmonary illness among 100 cases and 400 controls.
Exposure: vaping Outcome: pulmonary illness

cases controls Total

vaping 60 100 160

no vaping 40 300 340

Total 100 400 500

 CASE-CONTROL PAGE 11

Vaping and pulmonary illness

cases controls Total

vaping 60 a 100 b 160

no vaping 40 c 300 d 340

Total 100 400 500

OR = Odds of exposure among cases (a/c)

Odds of exposure among controls(b/d)

= ad / bc
= (60X300) / (100X40) = 4.5 what does 4.5 means??

- Those who vape are 4.5 times more likely to develop pulmonary illness than
non-vaping

Formulation of a clearly defined hypothesis,

case, and sources

- Clearly defined hypothesis: a case-control study should begin with the formulation
of a clearly defined hypothesis.

- Case definition: It is essential that the case definition is clearly defined at the outset
of the investigation to ensure that all cases included in the study are based on the
same diagnostic criteria.

- Source of cases: The source of cases needs to be clearly defined

 CASE-CONTROL PAGE 12

Measuring exposure status

● In case-control studies, the measurement of exposure is established after the

development of disease.

● As a result is prone to both recall and observer bias.

● Various methods can be used to ascertain exposure status, including not just by
asking questions :
● Standardized questionnaires
● Biological samples
● Interviews with the subject
● Interviews with spouse or other family members
● Medical records
● Employment records
● Pharmacy records
● The procedures used for the collection of exposure data should be the same for
cases and controls.

Bias in case control study

1. Selection bias: Selection bias occurs when the persons in one group are
different on some factor (other than disease)

2. Ascertainment bias: may arise because:

● Cases may recall exposure better

● Investigators may search for exposure more thoroughly in cases
● Different data collection instrument may be used for the controls

3. confounding.
 CASE-CONTROL PAGE 13

confounding in a case-control study

The two groups differ in some characteristic which is associated with both the
outcome and exposure being studied

● A confounding variable is one that can influence both the exposure and the
outcome
● E.g. in relation between vaping and pulmonary illness, cigarette smoking is a
likely confounder
● Males who use vapors are more likely to smoke, and smoking is strongly
associated with pulmonary illness. Age could be another confounder

E.g., is vapors use associated with pulmonary illness :

Cases: all males admitted to hospital with pulmonary illness aged 20-49 in region X

● Controls: a random sample of resident males in region X; age: 20-49 who

have not had pulmonary illness
● Exposure: vapors use during 3 months prior to interview
● Data collection: personal interview of cases and controls – cases in hospital,
controls telephone interview
● Potential bias: cases are younger than controls. Age is related to both
exposure (vapors use) and outcome (pulmonary illness)

strengths and weakness

strength weakness

Cost effective relative to other analytical studies such

as cohort studies. Particularly prone to bias; especially
selection, recall and observer bias
● Case-control studies are retrospective, and
• Case-control studies are limited to
cases are identified at the beginning of the
examining one outcome.
study; therefore there is no long follow up
period (as compared to cohort studies). • Unable to estimate incidence rates of
● Efficient for the study of diseases with long disease (unless study is population bas
latency periods.
● Efficient for the study of rare diseases. • Poor choice for the study of rare
● Good for examining multiple exposures. exposures.