Applied Health Economics and Health Policy (2023) 21:205–224

[Link]

SYSTEMATIC REVIEW

Analysing the Efficiency of Health Systems: A Systematic Review

of the Literature
Rahab Mbau1 · Anita Musiega1,2 · Lizah Nyawira1 · Benjamin Tsofa3 · Andrew Mulwa4 · Sassy Molyneux3,5 ·
Isabel Maina6 · Julie Jemutai1 · Charles Normand7,8 · Kara Hanson9 · Edwine Barasa1,5

Accepted: 4 December 2022 / Published online: 28 December 2022

© The Author(s) 2022

Abstract
Background Efficiency refers the use of resources in ways that optimise desired outcomes. Health system efficiency is a prior-
ity concern for policy makers globally as countries aim to achieve universal health coverage, and face the additional challenge
of an aging population. Efficiency analysis in the health sector has typically focused on the efficiency of healthcare facilities
(hospitals, primary healthcare facilities), with few studies focusing on system level (national or sub-national) efficiency. We
carried out a thematic review of literature that assessed the efficiency of health systems at the national and sub-national level.
Methods We conducted a systematic search of PubMed and Google scholar between 2000 and 2021 and a manual search
of relevant papers selected from their reference lists. A total of 131 papers were included. We analysed and synthesised evi-
dence from the selected papers using a thematic approach (selecting, sorting, coding and charting collected data according
to identified key issues and themes).
Findings There were more publications from high- and upper middle-income countries (53%) than from low-income and
lower middle-income countries. There were also more publications focusing on national level (60%) compared to sub-national
health systems’ efficiency. Only 6% of studies used either qualitative methods or mixed methods while 94% used quantitative
approaches. Data envelopment analysis, a non-parametric method, was the most common methodological approach used,
followed by stochastic frontier analysis, a parametric method. A range of regression methods were used to identify the deter-
minants of health system efficiency. While studies used a range of inputs, these generally considered the building blocks of
health systems, health risk factors, and social determinants of health. Outputs used in efficiency analysis could be classified
as either intermediate health service outputs (e.g., number of health facility visits), single health outcomes (e.g., infant mor-
tality rate) or composite indices of either intermediate outputs of health outcomes (e.g., Health Adjusted Life Expectancy).
Factors that were found to affect health system efficiency include demographic and socio-economic characteristics of the
population, macro-economic characteristics of the national and sub-national regions, population health and wellbeing, the
governance and political characteristics of these regions, and health system characteristics.
Conclusion This review highlights the limited evidence on health system efficiency, especially in low- and middle-income
countries. It also reveals the dearth of efficiency studies that use mixed methods approaches by incorporating qualitative
inquiry. The review offers insights on the drivers of the efficiency of national and sub-national health systems, and highlights
potential targets for reforms to improve health system efficiency.

* Edwine Barasa
ebarasa@[Link]
Extended author information available on the last page of the article

Vol.:(0123456789)
206 R. Mbau et al.

a system. Productivity, a concept related but different from

Key Points for Decision Makers efficiency, refers to the ratio of outputs to inputs in a pro-
duction system [10]. This paper focuses on the efficiency of
Most papers used quantitative methods that focused on health systems rather than productivity.
quantifying the level of efficiency and its determinants. It has been estimated that 20–40% of health system
Few papers used a qualitative approach and provided spending globally is wasted through inefficiency [11]. Such
more information about how certain factors might affect inefficiencies haemorrhage resources within the health sec-
health system efficiency, and also identified determinants tor and impede progress towards UHC [11]. While generat-
of efficiency that are not easily quantifiable. ing additional resources for health is crucial, improving the
System level efficiency is affected by factors both outside use of available resources in the health sector is argued by
the health system and within the health system; these some as one of the a promising strategies towards expanding
include demographic and socio-economic characteristics the fiscal space for health [11–13]. This is important both for
of the population, macro-economic characteristics of the LMICs that face constrained fiscal space as the support from
national and sub-national regions, population health and donors reduces, and HICs as an aging population increases
wellbeing, the governance and political characteristics of resource requirements in the health sector.
these regions, and health system characteristics. Efficiency measurement is therefore a key dimension
of health system performance assessment. It requires the
identification of the boundaries of the entity under scru-
tiny ranging from micro (provider-patient level) to meso
(organisational) to macro (national or global) levels of the
health system [4, 14]. The chosen level of analysis should
1 Introduction reflect an entity that will take accountability for the level
of performance identified by the analysis [4]. Efficiency is
Health system efficiency is a priority concern for policy increasingly assessed in healthcare, but most of these stud-
makers globally in the face of mounting health system ies analyse efficiency at the meso-level of the health system
expenditures [1–3]. Attainment of efficiency demonstrates involving healthcare organisations (such as hospitals and
good stewardship through good use of available resources health centres) [4, 10, 15], with fewer examining the national
and elimination of wastage [1, 4]. It also inspires the will- or subnational level [10].
ingness of governments and citizens to contribute resources Understanding efficiency within the health system and
towards Universal Health Coverage (UHC) [4]. Improved the associated inputs, outputs/outcomes and determinants of
efficiency is recognised as a desirable goal of the health sys- efficiency can influence policy formulation and managerial
tem [5] and an intermediate objective of health-financing decision-making [4]. This literature review aims to synthe-
policies that contributes towards the attainment of health sise existing empirical evidence on efficiency at the health
system goals [6]. It is also a major criterion for priority set- system level (national and sub-national rather than health
ting by decision makers [7]. Given the scarcity of health- facility level) to increase the understanding of the conceptu-
care resources, it is imperative that health systems, in both alisation and determinants of health system efficiency.
high income (HIC) and low- and middle-income coun-
tries (LMICs), orient their operations towards using their
resources efficiently to optimise the achievement of stated
health system goals and promote financial sustainability in
2 Methods
the long-term towards UHC [3, 4].
Efficiency refers to the extent to which system objectives
2.1 Literature Search Strategy
are met given the resources invested in the system [1]. Two
We followed the Preferred Reporting Items for Systematic
types of efficiency, technical and allocative efficiency, have
Review and Meta-Analysis (PRISMA) guidelines [16] in
been distinguished [8]. Technical efficiency (TE) is achieved
the reporting of this literature review. We searched two
when resources are allocated such that outputs are maxim-
electronic databases: PubMed and Google scholar for pub-
ised for a given level of inputs, or inputs are minimised for
lished and grey literature using specific subject headings
a given level of outputs [9]. Allocative efficiency (AE) is
and free text terms. We ran the searches separately for these
achieved when resources are allocated such that outputs are
databases. Search terms included ‘efficiency’, ‘technical
maximised for a given level of input cost, or input costs are
efficiency’, ‘data envelopment analysis’, ‘stochastic fron-
minimised for a given level of outputs [9]. Allocative and
tier analysis’, ‘health system’, ‘health sector’, ‘nation’, ‘sub
technical efficiency together make the 'overall' efficiency of
nation’, ‘country’, ‘region’ and ‘state’. These search terms
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 207

were used as key words in google scholar. We developed 2.3 Quality Appraisal

a Boolean algorithm to search PubMed was developed
(Fig. 1). The last literature search was done in December We used the critical appraisal skills programme (CASP)
2021. tool to assess the quality of the identified publications [17].
CASP tool uses a checklist approach to assess the adequacy,
2.2 Article Selection trustworthiness and relevance of the evidence reported in
the publications [18, 19]. The CASP tool helps to reflect
The electronic literature search identified 11,030 publica- the character of the studies included in the review [17]. The
tions. Of these, 10,899 were rejected based on our eligibil- results of the quality appraisal are indicated in Table 2. All
ity criteria (Table 1). We retained 131 articles that met our the publications were of sufficient quality to be included in
inclusion criteria. We only included publications that met the review.
the inclusion criteria. We included publications that reported
on empirical research on efficiency of health systems above 2.4 Data Extraction
the meso-level of the health system. We defined these lev-
els to include jurisdictions such as sub national, national, Two authors (RM and EB) performed data extraction of the
regional and international health systems. We included retrieved publications using a thematic analysis approach.
publications that were published in English due to time and Specifically, RM and EB first extracted data from a sub-
resource constraints that would otherwise be required for the set of selected papers (25/131). The data extracted by the
translation of non-English publications. We did not use any two authors was compared to establish concurrence on the
restrictions on publication year, publication status, country extraction approach. RB thereafter concluded data extrac-
income classification or study design. The publications that tion of the remaining papers. Thematic analysis is an ana-
met the inclusion criteria were imported into EndNote X8. lytic process that involves a systematic process of sifting,
The article selection process is summarised in a search flow sorting, coding and charting collected data according to key
diagram in Fig. 2. A total of 131 publications were retrieved issues and themes [20]. The first step in this analysis pro-
and reviewed. cess involved familiarisation with the publications through

Fig. 1  Key words

Table 1  Inclusion and exclusion criteria

Criteria Include Exclude

Does the article include the key search terms? Yes No

Do the authors explicitly state their aim, and provide evidence on efficiency at sub national, regional, Yes No
national or international level?
Publication language English Non-English
Empirical versus non-empirical Empirical Non-empirical
Analyses efficiency in the health sector versus analyses efficiency in other sectors (non-health sector) Health sector Non-health sector
Publication year 2000–2021
Study design (qualitative, quantitative or mixed methods) No limit
Study context (Country income classification) No limit
Publication status (published/grey) No limit
208 R. Mbau et al.

Fig. 2  Search flow diagram

reading and re-reading. This formed the beginning of the the data by category, (2) identify patterns and linkages in
abstraction process. We used a data extraction form struc- our data, that is, findings that were similar and recurrent
tured (Additional File 2) in line with the review question and across selected papers, and (3) make comparisons across the
efficiency concepts. This form was used as a data registry papers, to identified both converging and diverging findings
and a guide for the identification of inputs, outputs and deter- and seek explanations for this.
minants of efficiency within the health system. Second, after
familiarisation, we applied codes, developed inductively and
deductively, to data that we interpreted as important and 3 Results
relevant. We then grouped similar codes into categories or
themes drawing upon: (a) a priori issues (those informed 3.1 Characteristics of Selected Publications
by the original research aims and researchers’ knowledge
in the subject area), (b) emergent categories identified by The list of the selected publications is provided in Online
the authors in the retrieved literature, and lastly, (3) ana- Supplementary Material, File 1. Empirical literature on
lytical themes arising from similarities across the identified health system efficiency has expanded noticeably over the
codes. We then charted the data into a framework matrix years with the highest number of the retrieved literature
using Microsoft Excel. This allowed us to: (1) summarise published in 2018 (Fig. 3). However, most of these studies
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 209

Table 2  Quality appraisal

Appraisal criteria Yes Somewhat No/not
applica-
ble

1. Was there a clear statement of the aims of the research? 131

2. Is the methodology used for the study appropriate for addressing the research goal? 131
3. Was the research design appropriate to address the aims of the research? 131
Has the researcher justified the research design?
4. Is the recruitment strategy appropriate for the study aims? 131
Researcher explained how the inputs/outputs/study participants were selected?
Discussion around recruitment, i.e., why some inputs/outputs/participants were not chosen?
5. Was the data collected in a way that addressed the research issue? 131
If the setting for data collection was justified?
If it is clear how data were collected?
If the researcher has made the methods explicit?
6. Have the authors addressed areas of potential bias during the formulation of research questions, data 100 25 6
collection and data analysis?
For qualitative studies, is there researcher reflexivity?
7. Have ethical issues been taken into consideration? 9 122
Informed consent and confidentiality
Approval from ethics committee?
8. Was the data analysis sufficiently rigorous? 131
In-depth description of the analysis process?
Clarity of the development of themes/categories
Are contradictory data considered?
9. Is there a clear statement of findings? 131
Explicit findings
Adequate discussion of evidence for and against the researcher arguments
Credibility of finds (triangulation, respondent validation, more than one analyst), findings are discussed in
relation to the original research question)
10. How valuable is the research? 131
Researcher discusses the contribution of the study to existing knowledge and understanding
If they identify new areas where research is possible?
If the researchers have discussed whether or how the findings can be transferred to other populations?

Fig. 3  Number of publications published by income level classification and year

210 R. Mbau et al.

(53%) presented findings of health system efficiency in upper population-based and individual-based care, respectively
middle-income and high-income countries, while 25% of the [43, 44]. Both subunits contributed towards the efficiency
studies focused exclusively on lower middle-income coun- of the overall health system [43]. In addition to inputs and
tries and low-income countries, and another 19% focused on outputs, the efficiency of the health system as a production
countries across all income groups (Table 3). “unit” was thought to be affected by contextual factors from
Sixty percent of all the retrieved publications examined within and outside of the health sector. These factors had
efficiency at the national/country level. These included stud- different labels, including exogenous factors, explanatory
ies that examined a single country health system or several factors, and determinants of efficiency.
country health systems such as OECD countries [21], World
Health Organization member states [22], Eastern European 3.3 Methods Used to Analyse Efficiency
countries [23], Asian countries [24, 25], Latin America
and Caribbean countries [26] and Sub Saharan Africa [27]. Of the selected papers, 123 (94%) used purely quantitative
Forty percent of the publications examined efficiency at sub approaches, five (4%) used purely qualitative approaches,
national levels such as: (1) provinces in China [28, 29] and and another three (2%) used mixed methods approaches.
South Africa [30]; regions in Saudi Arabia [31] and Switzer- Quantitative approaches were used to measure the level and
land [32]; municipalities in Brazil [33] and Finland [34]; and determinants of efficiency. Qualitative approaches were
districts in India [35], Zambia [36] and Mozambique [37]. used to examine study participants’ perceptions about the
objectives of the health system [41, 45] and existence and
3.2 Conceptualisation of Efficiency at Health nature of health system inefficiency and its determinants
System Level in the Retrieved Literature [41, 45–48]. Beyond identification, qualitative approaches
provided explanations of the relationship between identi-
Following existing production literature described by Far- fied determinants and health system efficiency [36, 48, 49].
rell [8], the majority of the authors of the retrieved litera- Seventy-two (57%) of the publications that used pure quan-
ture explicitly defined efficiency as the extent to which titative approaches or mixed methods used cross-sectional
desired health system goals were achieved given existing quantitative data to estimate the level of efficiency in the
resources [38–41]. The literature conceptualised a health health system. The remaining 54 (43%) of these papers used
system as a production system that transformed inputs into panel data with authors such as [27, 50, 51] indicating that
desired outputs [42]. In most of the studies, this produc- panel data offer more accurate estimations of efficiency
tion system was considered as a single unit. In two studies, because of the richness of the data and consideration of the
however, the health system was perceived to be composed effect of time [52] precludes the need to impose assumptions
of two subunits—a public health system (non-health facil- on the error terms likely to be correlated with time. Of the
ity-based health promotion and prevention services) and a papers that used panel data, 36 (67%) used the Malmquist
medical care system (i.e., healthcare facilities) that offered productivity index (MPI) approach to measure efficiency

Table 3  Descriptive statistics of Characteristic Descriptive statistics

the retrieved literature
Country of focus Pro-
portion
(%)

Publication by income level High-income countries only 30

Upper middle-income countries only 8
High-income and upper middle-income countries only 15
Lower middle-income countries only 21
Low-income countries only 2
Low-income and middle-income countries only 4
All income levels 19
Level of health system National 60
Sub-national (e.g., municipality, district, province, region 40
health authority)
Type of data Quantitative data 94
Qualitative data 4
Mixed (quantitative and qualitative) 2
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 211

changes over time, while 18 (33%) included time as a covari- into three broad categories: health system building blocks,
ate in a regression analysis. Five publications (4%) employed social determinants of health, and health risk factors. Among
qualitative approaches [36, 46, 53–55], while two studies the health system building blocks, finances were the most
(2%) used a mixed methods approach by combining both common input, with 68% of the studies using this variable
qualitative and quantitative methods [47, 49]. in the production function. This was followed by human
Efficiency measurement in the retrieved literature was resources for health (66%) and medical equipment (54%).
done using non-parametric (data envelopment analysis- In some of the studies, the number of beds was used as a
DEA and Free disposal hull technique) and parametric proxy for capital investment in health production [62, 64,
methods (stochastic frontier analysis-SFA). DEA was the 65] because direct measurement of capital in healthcare
most used technique for measuring efficiency. DEA is a was found to be problematic [62]. The number of health
non-parametric linear programming method that assess the facilities was only used in 22% of the studies. Education, a
relative efficiency of production units by obtaining the ratio social determinant of health, was used as an input in 15% of
of a weighted sum of the outputs of a productive unit to the studies. Health risk factor characteristics used as inputs
a weighted sum of its inputs [56]. The DEA technique is included tobacco and alcohol consumption (5%).
relevant in the health sector given the complex nature of The choice of inputs used in assessing efficiency was
health systems where multiple inputs are utilised to produce informed by various reasons. These included evidence of
multiple outputs. A key limitation of DEA is that its results use of the input variables in previous efficiency studies,
may be influenced by measurement error or statistical noise availability of the data, positive relationship with the out-
given that DEA is non-stochastic. DEA ascribes deviations puts, frequency of data reporting on the variable, direct
from the frontier entirely to inefficiency, even though these involvement of the input in the production of health, input
may be due to measurement errors. DEA was exclusively that would allow cross-country comparisons of efficiency or
used in 95 (76%) of the selected papers and used in combi- whether the input could be standardised across the system to
nation and compared with free disposal hull (FDH) or SFA allow comparison. It also included whether the input vari-
in two (2%) papers respectively. SFA was the second most able could be consistently measured across the units being
common approach, used exclusively in 23 (18%) papers and assessed, whether the influence of the variable on efficiency
in combination and compared to FDH in one (1%) paper. was within the control of the health system, and based
SFA is a parametric method that uses regression analysis to on economic theory and wider literature, and opinions of
estimate the production frontier, measuring the efficiency of experts and stakeholders in the system.
a unit using the residuals from the estimated equation [57].
Its key advantage over DEA is that SFA explicitly accounts 3.3.2 Outputs and Outcomes, Their Definition and Reasons
for measurement error. The DEA model decomposes the Why They Were Chosen
error term in a stochastic error component and an additional
error term that represents systematic inefficiency. SFA is Outputs used in the reviewed literature fall into three cat-
used because it accounts for random disturbances in the data egories: intermediate health service outputs, single health
[58]. Qualitative data were analysed using thematic analysis outcomes, or composite indices of either intermediate out-
[36]. puts or health outcomes. While several authors indicate that
Determinants of health systems efficiency were identified a general consensus in existing literature puts health status
in 72 (55%) of the selected papers. Methods used for the of the population [60] as the single most important output
quantitative identification of the determinants of efficiency of the system [66], its measurement has, however, remained
include: Bayesian linear regression [59, 60], Tobit regres- difficult [60]. As indicated by [23], the distinction between
sion [60, 61], truncated regression model [62], and multiple output and outcome is often blurred leading authors to use
regression analysis [63]. These methods were second stage the two terms interchangeably.
analysis in DEA or SFA. The list of outputs and outcomes identified in the liter-
ature is provided in Table 5. Seventy percent of the pub-
3.3.1 Inputs, Their Definition and Reasons Why They Were lications included more than one output variable in their
Chosen assessment of efficiency. Of the health outcome variable
used, mortality rates and life expectancy were the most
Inputs were defined as resources required to facilitate the common (51%). Mortality rates were considered a good
production function of the health system [21, 60]. These summary measure of overall population health [67] as well
resources were considered to be within the control of the as the closest measurable indicator of the stated health sys-
managers in the health system [33]. The list of inputs identi- tem objectives [41]. Common intermediate health outputs
fied in the literature is provided in Table 4. While different used included outpatient and inpatient workload measures
studies used different inputs, the inputs could be classified and maternal and child health services utilisation measures.
212 R. Mbau et al.

Table 4  Summary table of inputs identified in the retrieved literature

Category of inputs Examples of inputs and their units

Health systems building blocks

Monetary value of economic outputs Per capita Gross Domestic Product (GDP)
Annual GDP
Financial value of resources available in the health Total health expenditure per capita (public and private)
sector Total health expenditure as a percentage of GDP
Private health expenditure as a percentage of per capita GDP, and public health expendi-
tures as a percentage of per capita GDP
Per capita government expenditure on health
Human resources for health Number of physicians per capita; physicians’ density; number of salaried physicians
Number of nurses per capita
Number of pharmacists
Total number of health workers per 1000 population
Number of dentists employed in a clinic in a year
Number of specialists or resident medical specialists per 100,000
Number of paramedical staff per 1000 population; per capita paramedical staff
Physical medical infrastructure Total number of hospital beds; number of hospital beds per capita
Number of magnetic resonance imaging (MRI) units per one million population
Number of computerised tomography (CT) scanners
Medical products Prescription drugs
Number of pharmaceutical patents
Health facilities Number of primary healthcare centres
Per capita health facilities
Health facility density
Number of long-term care facilities; residential care facilities
Social determinants of health
Education Primary school education; enrolment in primary education
Share of population with secondary school education
Average years of schooling
Average years of schooling in population older than 15 years
Average years of schooling of population over 25 years old
Average school age (as a proxy for the level of education in each country)
School expectancy years-expected years of schooling
Health risk factors
Tobacco consumption Annual tobacco consumption per capita
Percent of population fifteen years and older who smoked daily; percentage of smokers in
adults (over 15 years of age)
Alcohol consumption Alcohol consumption per adult person (over 15 years of age), litres of pure alcohol per year

Several studies used composite indices as output/outcome The most common criterion that informed the choice of
measures. For example Tandon et al. [68] used a weighted outputs used in a study was evidence of common use of the
average of health system goals using disability adjusted life variable in previous studies [44, 60, 65, 71–73]. This was
expectancy (DALE), health inequality, responsiveness level, indicated in 40% of the retrieved literature. Other criteria
responsiveness distribution and fair financing. Tandon et al. applied to select outputs included: (1) use of the variable by
[69] created an outcome index by combining five indicators the ministry of health to monitor efficiency of the health sys-
on immunisation coverage, skilled birth attendance, iodised tem, for example, the hospital bed occupancy rate in Zambia
salt content, catastrophic expenditure and life expectancy [36]. (2) Relevance to millennium development goals related
while Achoki et al. [70] use a composite metric for maternal to reduction of maternal mortality and child mortality such
and child health services made up of diphtheria, pertussis, as institutional delivery rate [74] or under-5-year-old mortal-
tetanus vaccine-3 doses (DPT3) and measles immunisations, ity rates. (3) Relevance to the national government priorities
skilled birth attendance and malaria prevention. such as primary healthcare agenda in India [74] or increased
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 213

Table 5  Summary table of outputs identified in the retrieved literature

Category of outputs Examples of outputs and their units

Single measures of health outcomes

Mortality rates Infant mortality rate
Infant survival rate or inverse of infant mortality rate
Neonatal mortality rate per 1000 live births
Maternal mortality ratio; maternal mortality rate
Under 5 years of age mortality rates
Adult mortality rate
Adult survival rate
Mortality rate from treatable causes
Life expectancy Life expectancy in years at birth for males and females
Potential Years of Life Lost (PYLL)
Composite measures of health outcomes
Life expectancy Health Adjusted Life Expectancy (HALE) in years at birth
Disability Adjusted Life Expectancy (DALE)
Average healthy life years
Intermediate health outcomes
Outpatient workload Number of outpatient visits
Outpatient consults in a year
Number of emergency room visits per person
Inpatient workload Number of inpatient days per person
Bed utilisation/occupancy rate
Patient days
Number of inpatient discharges
Hospital discharge rate
Average length of stay
Number of inpatient admissions
Hospital workload Total patients; number of patient cases
Annual medical visits
Incidence of disease Tuberculosis (TB) incidence rate; incidence of TB per 100,000 people
People newly infected with HIV
Financial risk protection Out-of-pocket expenditure as a percentage of total health expenditure
Maternal and child health services utilisation Immunisation rates
Antenatal services utilisation rate
Number of institutional deliveries
Caesarean section rate
Utilisation of diagnostic services Number of radiology patients
Number of laboratory investigations
Utilisation of surgical services Inpatient surgical procedures per 1000 population
Cardiac bypass procedures per 100,000 population
Number of surgical operations
The total number of patients receiving surgery a year
Quality of care Average nursing time
Waiting time
Treatment success rate TB treatment success rate
The number of malnourished detected and cured in the regional hospitals
The number of malaria cases treated in the healthcare institutions
Number of live births per 1000 population
214 R. Mbau et al.

number of live births in Thailand [75] or access to quality score. For example, the most recent regional analysis of the
and effective healthcare in Canada [41]. (4) Availability of efficiency the country health systems reported a mean effi-
data [52, 76]. (5) Robustness of the indicators [77–79]. (6) ciency of 80% (range 31–100%) for 45 African countries
Objectivity of the variables [80]. Relevance of the variable [27], 92% (range 81–91%) for 46 Asian countries [83], 93%
to the context [81]. (7) Routine collection of the data and its (range 51–93%) for 27 Latin American countries, and 83%
ability to allow for cross-unit comparison [66, 82]. (range 54–94%) for 28 European countries [84]. The most
recent global analysis of the efficiency of 140 country health
3.3.3 Exogenous or Environmental Variables, Their systems reported a mean efficiency of 93% (range 71–100%),
Definition and Reasons Why They Were Chosen with the following regional means: African countries (86%),
Asian countries (95%) South American countries (95%), and
Exogenous variables refer to the factors that are not directly European countries (96%) [85]. An efficiency score of 100%
related to the resources in the sector in question but may denotes an efficient health system, while a score below 100%
have an effect on the relationship between the inputs and means that there is scope to improve efficiency by either
outputs of that sector [69]. These variables are recognised producing more output or reducing inputs to achieve a score
as the third variable for inclusion in efficiency measurement of 100%.
along with inputs and outputs [68]. Exogenous variables
were thought to capture heterogeneity and explain some of 3.5 Factors Affecting the Efficiency of Health
the differences or dispersion in the efficiency levels of units Systems
under analysis [39]. Fifty-six percent of the retrieved publi-
cations considered the influence of exogenous variables on 3.5.1 Demographic Characteristics of the Population
the efficiency of the units under consideration. However,
only one author [53] provided a conceptual framework that Several population/demographic characteristics were
shows the influence of environmental variables on a health found to determine health system technical efficiency.
system’s production function. The list of these variables is One of these was population density. Some studies found
provided in Table 6. that a high population density of a country or sub-national
Exogenous variables were chosen based on: (1) evidence unit (region/district etc) was associated with increased
of use in previous studies [15, 21, 59, 77]; this was the most technical efficiency. For instance, a study of the primary
common reason given by a third of all the authors who used healthcare system in Chile found that a high population of
exogenous variables in their analysis. (2) Completeness and primary healthcare catchment areas increased the techni-
consistency of reporting of the variable in question for the cal efficiency regional health systems [61]. Ahmed et al.
units under consideration [72]. And lastly, (3) evidence of assessed the technical efficiency of the health systems
their potential influence on efficiency [33, 62, 65]. Table 6 of 46 Asian countries and found that countries having
outlines the categories of exogenous variables used in the more than 200 people per square kilometre were more
analysis. It will be evident that some factors are used either technically efficient compared with the countries with ≤
as inputs or as exogenous variables in different studies. 100 population per square kilometre. Higher population
densities increased the technical efficiency of regional
3.4 Efficiency of Health Systems health systems by reducing distances to populations and
making it easier for health systems to organise and utilise
It is challenging to summarise and/or compare findings from their services infrastructure, and by reducing per capita
the literature on the efficiency of health systems because cost of healthcare [33]. However, some studies reported
of heterogeneity of methods. This includes differences in a negative association between population density and
approach (qualitative and quantitative), selection of inputs, health system technical efficiency. For instance, a study
outputs, exogenous variables, and models. For instance, a of Finnish municipalities found that large populations
sensitivity analysis of an efficiency analysis of 141 coun- reduced the technical efficiency of municipalities and
tries originally conducted by the World Health Organization speculated that this could be because other factors related
(WHO) found that country rankings and efficiency scores to population size such as quality differences, bureau-
were sensitive to the definition of efficiency and choice of cratic inefficiency, or unmeasured outputs [34]. A study
model specification [63]. Qualitative papers focused on in Kenya found that the technical efficiency of county
health system stakeholders’ views about the existence of health system was negatively associated with population
inefficiency and sources of inefficiency in health systems. density, and speculated that this was likely because higher
These are summarised in the next section. Quantitative population densities were not matched with healthcare
approaches reported the level of health system efficiency resources and hence compromising health outcomes [86].
as a proportion (with a range of 0–100) or an inefficiency Another factor that was explored was the rural/urban
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 215

Table 6  Summary table of exogenous variables identified in the retrieved literature

Category of variables Examples of variables and their units

Population/demographic characteristics Population size and density Population density, people per square kilometre
Population growth Population growth rate
Rural-urban population distribution Proportion of urban population as a percentage of
total population
Proportion of rural population as a percentage of
total population
Age structure The proportion of the population under age 6 years
The proportion of enrolled inhabitants over age
65 years
Proportion of population aged 0–14 years
Socio-economic characteristics of the population Employment status Unemployment rate
Economically active population
Long-term unemployment
Income distribution Gini coefficient
Poverty index
Income level Per capita income
Educational attainment The level of primary school enrolment in the
country
Average years of schooling in the adult population
Literacy levels in rural and urban areas
Literacy rate in percentage
Proportion of out-of-school children
Access to basic sanitation amenities Population covered by individual household
latrines
Percentage of the population with access to clean
water
Percentage of the population with access to sanita-
tion facilities
Health facilities with water
Health system characteristics Health expenditure Total health expenditures per capita
Public health expenditure per capita
Total health expenditure as a share of GDP
Public healthcare expenditure as a percentage of
total health care expenditure
Out-of-pocket healthcare expenditure
Access to health providers Proportion of rural, urban, and other public health
facilities each municipality runs respectively
Distance to the closest reference hospital
Share of public sector in the provision of service
Degree of private provision: breakdown of doctors
and hospital or private status
Utilisation of health services Annual referrals rate to specialists
Annual home visits rate
Distribution of health service provision Proportion of the medical services in primary
medical facilities (%)
Regulation on healthcare users Patient choice among providers
Gate keeping
216 R. Mbau et al.

Table 6  (continued)
Category of variables Examples of variables and their units

Regulation on insurers Regulation of prices paid by third-party payers

Degree of delegation to insurers
Regulation on providers Purchaser-provider split
Regulation of prices billed by providers
Political and governance environment Political stability
Degree of decentralisation
Measure of democratisation and freedom of politi-
cal unit
Provider payment methods and incentives The reimbursement system; physician remunera-
tion methods
Health and wellbeing Lifestyle risk factors Tobacco consumption
Alcohol consumption
Overweight population
Physically inactive population
Happiness Happiness index
Disease prevalence HIV prevalence rate
Three or more chronic conditions
Macro-economic characteristics Country/ sub national level income level Average income
Gross domestic product
Gross national income per capita

distribution of the population. There is a general finding In addition to average income levels in a country, the
that regions with low urbanisation rates are likely to be distribution of incomes was also found to determine health
less technically efficient [34, 35, 87]. This was because, system technical efficiency. High income inequality and
among others, lower urbanisation was associated with poverty was associated with reduced technical efficiency.
lower unemployment rates and lower income levels that Bekarogu and Heffley found that increased poverty and
affect healthcare utilisation [88]. Population age structure income inequality affected the technical efficiency of
was also explored; high proportions of the very young health system by reducing the overall level of health sys-
(children) or the very old reduced the technical efficiency tem outcomes. A related socio-economic characteristic
of health systems because these vulnerable populations was employment status, where high unemployment rates
increased the cost of healthcare [33, 61]. were associated with reduced health system technical effi-
ciency [34].
3.5.2 Socio‑Economic Characteristics of the Population/ Several studies found that access to basic sanitation and
Social Determinants of Health clean water increased the technical efficiency of health sys-
tems. This was because improved sanitation improved health
Several socio-economic characteristics of the population outcomes, which was linked to improved technical efficiency
were examined. Some studies reported that improved socio- of the system. For example, Grigoli and Kapsoli [90] found
economic status of the population is positively associated that the percentage of the population with access to sanita-
with health system technical efficiency. For instance, several tion services was associated with an increase in technical
studies found that increased per capita income of a country efficiency, while Hassan et al. [91] found that the rate of
or regions population was associated with increased tech- access to drinking water decreased the incidence of water-
nical efficiency of the health system [89]. However, some related diseases such as cholera, fever and malaria, and was
studies reported a negative association between population associated with increased technical efficiency.
income per capita and health system technical efficiency. Increased literacy was associated with increased technical
This was thought to be because health systems whose catch- efficiency of health systems [26, 38, 87, 92]. For example,
ment populations had higher income per capita were char- Ahmed et al. found that Asian countries with higher literacy
acterised by higher levels of overprovision and higher costs levels have higher health system technical efficiency. This
of care. was thought to be because educated people more easily
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 217

transform health information and knowledge into health health insurance firms, was negatively associated with health
outcomes [87, 89]. system efficiency [53, 77]. The level of health expenditure
also had an impact on health system efficiency. Total health
3.5.3 Macro‑Economic Characteristics expenditure as a share of GDP was positively associated
with the technical efficiency of health systems [38, 42, 65,
Findings on the effect of the size of a country’s economy on 100]. The role of availability of funds was also highlighted
health system technical efficiency were mixed. Some studies in Kenya [47, 48]. This was thought to be because greater
found that higher country per capita gross domestic product healthcare spending was essential in improving health out-
(GDP) was associated with a more technically efficient deliv- comes [65]. However, some studies found that higher lev-
ery of healthcare [93–96]. This was thought to be because els of total health expenditure can be negatively associated
increased country wealth could translate to increased invest- with efficiency when the health system is characterised by
ments in the health sector as well as other sectors that impact unnecessary care and/or higher costs of care [14, 15, 40, 48].
on social determinants of health, with improved health and The source of funding for the health sector was also shown
quality of life having a positive impact on overall health to affect technical efficiency. The share of public spending
outcomes. For instance, some studies found that countries on healthcare was positively associated with health system
with good road infrastructure and good access to electricity technical efficiency [14]. Further, Increased population cov-
were associated with increased technical efficiency of health erage with a prepayment health-financing mechanism (such
systems [87]. However, other studies found that higher GDP as health insurance) was associated with increased technical
per capita was associated with lower technical efficiency efficiency of health systems [101]. An assessment in China
of health systems. This was thought to be because of the found that provinces with a high proportion of out-of-pocket
increased cost of healthcare because of overprovision [15, payments had lower technical efficiency [102]. However,
61, 97] and higher relative prices of healthcare in richer some studies on the efficiency of OECD (Organisation for
countries [98]. Economic Co-operation and Development) countries [98,
101] have found that out of pocket payments in the form of
3.5.4 Health and Wellbeing of the Population co-payments were positively associated with health system
efficiency in contexts that have adequate population cov-
Several aspects of the health and wellbeing of the popula- erage with prepayment mechanisms. This was because co-
tion affected the technical efficiency of the health system. payments disincentivized unnecessary use of care. Public
Generally, higher prevalence of chronic disease was associ- finance management arrangements also influenced health
ated with reduced health system technical efficiency. For system efficiency. Enhanced capacity to execute budgets,
instance, Novignon and Lawanson found that HIV/AIDS flow of funds directly to providers, timeliness of funds dis-
negatively affects technical efficiency of health systems in bursements to local authorities and health facilities, the flex-
Africa, with a similar finding reported in Kenya [86]. Allin ibility of budgets, and the autonomy of local authorities and
et al. found that an increase in the proportion of people with health facilities over resources enhanced efficiency [45, 47,
chronic conditions by 10% would decrease the technical effi- 48].
ciency score by between 10 and 18% in regional health sys- With regard to the purchasing function of the health sys-
tems in Canada. Further, health systems that serve popula- tem, how healthcare providers were paid also affected health
tions with high levels of health risk factors such as smoking, system efficiency. For instance, prospective payments such
alcohol consumption and obesity were likely to be less tech- as capitation, rather than fee-for-service payments, were
nically efficient [15, 21, 59, 99]. For example, Bekaroglu and founds to be positively associated with health system effi-
Heffley found that a high consumption of alcohol increases ciency in some studies because they disincentivised unnec-
inefficiency by causing premature ill health and death. A essary care and provided purchasers with better control
high prevalence of chronic disease and health risk factors over costs [103]. In the Democratic republic of Congo, the
reduced health system outcomes and increased healthcare introduction of a zero-margin policy for drug sales in the
costs with negative impacts on health system efficiency. public sector reduced the incentive of healthcare providers
to prescribe unnecessary medicines [53]. The design and
3.5.5 Health System Characteristics implementation of benefit packages also affects health sys-
tem efficiency. Chile, Mexico and Uruguay improved the
Several characteristics of health system functions were efficiency of their health systems by prioritising health ser-
found to determine the efficiency of health systems. First, vices that are cost-effective in their benefit packages [53].
how health systems are financed affected health system The efficiency of health systems was also found to be
efficiency in several ways. The fragmentation of financ- affected by how users interacted with the health service
ing arrangements, and specifically the presence of multiple providers. Health systems where patients exercised choice
218 R. Mbau et al.

of health providers were associated with higher technical participation in healthcare decision making [104]. Effective
efficiency [59]. Gate keeping by primary-care providers, performance monitoring and accountability was found to
where a patient is required to have a referral from a gen- improve health system efficiency in Canada [46]. Leadership
eral practitioner for non-emergency access to a specialist, and management practices and capacity were also thought
enhanced health system efficiency by aligning the level of to be a determinant of health system efficiency [45, 46, 53].
specialisation and cost of healthcare with healthcare needs, An assessment of the technical efficiency of 27 Latin Ameri-
and reducing healthcare costs [103]. However, some stud- can and Caribbean countries found a positive association
ies found that gate keeping could reduce health system effi- between governance quality and system technical efficiency
ciency in settings where primary-care physicians had lim- [26]. Governance quality in the study was defined as a mul-
ited ability to coordinate the follow-up of patient care, or tidimensional index that included measures of government
in settings where the health system’s capacity to provide effectiveness, voice and accountability, rule of law, regu-
secondary care was limited [77]. Inadequate health system latory quality, political stability and absence of violence/
capacity to provide specialised care resulted in long waiting terrorism, and control of corruption. Further, assessments
times, and increased the utilisation of emergency depart- of the technical efficiency of WHO member country health
ments and hospitalisations and hence resulting in ineffi- systems found that an increase in democratisation and free-
ciency [77]. The effectiveness of gate keeping in enhancing dom was associated with increased health system technical
health system efficiency was also dependent on whether it efficiency [39, 94]. Corruption has also been found to be
was accompanied by interventions to improve the availabil- associated with reduced technical efficiency [69, 93].
ity and quality of secondary-care services [77]. Further, an The availability and distribution of health system hard-
interaction between price regulation and gate keeping has ware such infrastructure, equipment and health commodi-
been reported. It has been observed that when healthcare ties were associated with increased technical efficiency of
prices are regulated, gate keeping may reduce efficiency by health systems [83]. Inadequate availability of input led to
incentivising excessive specialisation of healthcare profes- an inefficient mix of inputs with negative impacts on health
sionals to access higher fees [60]. It also incentivises general system efficiency. For example, an assessment of the tech-
practitioners to make unnecessary referrals of patients to nical efficiency of Asian country health systems found that
specialised care so as to minimise their (general practitioner) the density of hospital beds had a positive association with
input costs [60]. technical efficiency [83]. An assessment in Canada found
On health governance, strong regulation of health sys- that increased inequitable distribution of health workers was
tem functions, and specifically price regulation, medicine associated with increased technical efficiency of national and
use, and health workforce regulation were associated with sub-national health systems [15]. In Ethiopia, an increase
increased technical efficiency [53, 59, 60]. In China and in the number of primary healthcare facilities that was not
El Salvador, the introduction medicines regulations that matched with an increase in the number of health workers
strengthened price regulation, generic prescribing, and resulted in inefficiency [53]. The level and distribution of
the enforcement of national essential drugs lists improved health workers affected health system efficiency. National
health system efficiency [53]. Improved coordination in the and sub-national health systems that had inadequate num-
health sector, including the coordination of donor initiatives, bers of health workers were less efficient [97, 105]. Pro-
was also associated with improved health system efficiency curement inefficiencies were also identified. This included
[45, 53]. The Democratic Republic of Congo and Zambia fragmented procurement of health commodities, high pro-
realised improvement in health system efficiency by align- curement prices, and supply chain challenges leading to
ing and coordinating donor support with health sector pri- delays in deliveries and stock-outs, and procurement cor-
orities, and coordinating health sector planning, budgeting ruption [45, 47, 53]. Verhoeven et al. [98] found that high
and resource allocation to reduce waste and duplication [36, spending on pharmaceuticals was associated with lower
53]. Beyond health sector coordination, multisectoral coor- health system efficiency. This was thought to be because
dination and partnerships to tackle social determinants of high pharmaceutical expenditure crowded out other health-
health were thought to improve efficiency [46]. Some studies care inputs and hence reduced the efficient use of health
reported that decentralisation of health functions was asso- resources.
ciated with higher technical efficiency of national and sub- Overprovision of health services (long lengths of stay,
national health systems [59]. An assessment of the technical high referrals, high doctor consultations, high admission
efficiency of healthcare systems of selected middle-income rates, inappropriate drug (such as antibiotic) use), an aspect
countries found that technical efficiency was enhanced by of quality of care, was associated with reduced technical effi-
decentralisation, which enhanced the delivery of care in ciency [15, 61, 97]. For example, Chai et al. found that nega-
rural areas, and improved the responsiveness of health sys- tive association of admission rates with technical efficiency
tems to community needs through improved community implied that a resource-intensive hospitalisation service use
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 219

was harmful to health system technical efficiency. Ramirez- missed non-quantifiable, software aspects of health systems.
Valdivia et al. [61] found that increasing annual referrals On the other hand, qualitative approaches provided more
to specialists increases the intechnical efficiency score. In information about how certain factors might affect health
China, the inappropriate use of drugs reduced health system system efficiency. They also identified determinants of effi-
efficiency [53]. ciency that are not easily quantifiable—software factors
The organisation of care to prioritise lower level basic such as the role of leadership and management practices,
care primary healthcare is associated with increased health and health sector coordination. This highlights the need
system efficiency. Ding et al. [28] found that an increase for mixed methods approaches that incorporate the use of
in the proportion of medical services in primary facilities qualitative methods to undertake in-depth assessments of the
would increase the technical efficiency of provincial medi- interplay of factors that determine health system efficiency.
cal centres. The share of essential/basic services in benefit While studies used a range of inputs, these were generally
packages was positively associated with health system effi- the building blocks of health systems, health risk factors, and
ciency [53, 98]. This was because essential/basic services social determinants of health. The common justification for
were more cost-effective compared to advanced/expensive the use of specific inputs and outputs in efficiency analysis
care. Further, health systems with a high share of basic was the fact that it had been used by a similar analysis in
care health workers (rather than specialists) were likely to other settings and data availability. This finding is similar to
be more efficient [15, 59]. Policy reforms with a focus on that of a review efficiency analysis of primary-care systems
expanding primary and community-based care, and engag- [107]. A limitation to this input/output selection approach
ing the community were shown to improve the technical is the likelihood that the selected inputs/outputs may not be
efficiency of healthcare systems in OECD countries [43]. relevant and suitable to characterising specific health sys-
Further reforms geared on enhancing access to healthcare for tems. Exercises to engage health system decision makers and
the disadvantaged and vulnerable, and reducing inequality in implementors in specific contexts to understand their health
access to healthcare services was associated with increased systems and identify context-appropriate inputs and outputs
technical efficiency [43]. for efficiency analysis should be explored. Such efforts will
require and inform the strengthening of data systems for
health system performance monitoring and evaluations in
4 Discussion specific contexts.
The findings on factors that determine the efficiency of
Previous reviews of health system efficiency have either health systems highlight several issues. First, that the effi-
focused on the methods used in the analysis [106–108] or on ciency of national and sub-national health system is partly
synthesising the evidence on health facility level efficiency determined by factors not easily influenced by health system
[107, 109]. This study presents a systematic review of empir- policy makers. Broader contextual factors such as the demo-
ical studies that have examined the efficiency of national or graphic, individual and household socio-economic, macro-
sub-national health systems. Our findings show that there economic, and governance and political system characteris-
were more publications from high and upper middle-income tics of the national and sub-national unit are outside of the
countries than from low-income and lower middle-income control of health system policy makers. This emphasises
countries. One of the factors that may have contributed to the fact that health systems are part of and are affected by
this is the availability of rich cross-country data on high the larger society that forms their contexts and underlies the
income countries such as the ones held by OECD and EU, need for countries to strengthen multisectoral coordination
and health system observatory databases. It underlies the and approaches to health [110, 111].
need for LMICs to set up mechanisms to collate and curate Second, the efficiency of national and sub-national health
health system data that could be used for monitoring health systems is also affected by the general health and wellbe-
system performance. ing of populations. This underlies the need for health policy
All the identified studies assessed technical efficiency and makers to prioritise investments in preventive and promo-
its determinants and used quantitative approaches except for tive interventions that reduce the risk of disease, and the
five studies that used qualitative approaches and three that management of chronic conditions to reduce their burden on
used mixed methods and considered efficiency more broadly health systems. Health system reforms aimed at promoting
(beyond technical). While quantitative approaches quantified efficiency should not only focus on optimising health sys-
the level and determinants of health system efficiency, they tem functions to provide care for the sick, but also prioritise
did not provide insights on the mechanisms by which deter- interventions to prevent and promote health and wellbeing
minants interact with the technical efficiency of health sys- of the population. Investing in preventive and promotive
tems. Quantitative approaches were also limited to assessing health interventions has been shown to be cost effective
quantifiable factors (typically health system hardware), and [112] and prioritising preventive and promotive health has
220 R. Mbau et al.

been identified as an efficiency enhancing intervention by of equipment, investigations, and procedures [113]. Further,
the WHO 2010 world health report [113]. the introduction of essential medicines list to guide medicine
Lastly, health system arrangements offer several policy procurement was found to be efficiency enhancing in China
levers for improving the efficiency of national and sub- [114].
national health systems. These factors are key because they Policy options will, however, need to appropriate for con-
are under the direct control of health system managers and text and take a whole system view given that policy reforms
policy makers and can be leveraged to enhance health sys- are efficiency enhancing only if they are implemented in pol-
tem efficiency. On health governance, strengthening health icy environments that they are aligned with or coherent. For
sector leadership and management, enhancing health sector instance, while increasing the level of financing to the health
coordination as well as multisectoral coordination, decen- sector may enhance efficiency, this needs to be accompa-
tralisation of health functions, and introducing/strengthening nied by reforms to contain overprovision/unnecessary care
the regulation of health workers and healthcare costs/pricing or price increases. While gate keeping by primary healthcare
should be considered. These findings resonate with those providers may enhance health system efficiency in contexts
of other analyses. For instance, a review of 10 case studies with good availability and quality of secondary care ser-
of health system efficiency found that strengthening health vices, and controls that check against unnecessary referrals.
sector regulation, pharmaceutical pricing, and enhancing Further, cost-sharing appears to be efficiency enhancing in
donor coordination enhanced the efficiency of country health settings in context with adequate prepayment health financ-
systems [114]. With regard to health financing, interven- ing mechanisms. Policy interventions will therefore need to
tions include scaling-up prepayment health financing mech- be contextualised for specific settings.
anisms and reducing the level of out-of-pocket payments. A potential limitation of this review is the likelihood of
Purchasing and public finance management (PFM) reforms missing out key literature since we conducted searches on
that include defining evidence-based, cost-effective benefit only two databases. Further, this review presents a qualita-
packages, reforming payment systems to transition to pro- tive evaluation of determinants of efficiency rather thana
spective payment mechanisms, strengthening budget execu- quantification of their impacts. Further research could there-
tion and flexibility of budgets, the direct payment of health- fore include meta-analysis of the impact of interventions to
care providers (to improve provider autonomy, and flow of address health system efficiency.
funds), health facility autonomy, and ensuring patient choice
of health providers may promote the efficiency of health
systems. Yip and Hafez [114] found that the introduction 5 Conclusion
of evidence based benefit packages enhanced health system
efficiency in Chile and Uruguay while scaling population This review highlights the asymmetry of evidence on health
coverage with prepayment financing systems was efficiency system efficiency between HICs and LMICs, with most stud-
enhancing in Mexico. On health system hardware, invest- ies focusing on HICs. It underscores the need to carry out
ing in adequate levels of health workers, health commodi- studies to understand the levels and determinants of system
ties, and health infrastructure, as well as interventions to level health system efficiency in LMICs. The review also
ensure equity (including geographical) in their distribution reveals the dearth of efficiency studies that use mixed meth-
and access are key. Having an optimal mix of health sys- ods approaches by incorporating qualitative inquiry. While
tem inputs has been identified by other studies as efficiency the standard quantitative approaches determine the level of
enhancing [113]. On service delivery, policy options include efficiency and the factors that influence efficiency, they fall
re-orienting health systems to prioritise primary health care short in illuminating how and why certain factors influence
and strengthening community health systems and strength- health systems efficiency in certain contexts. There is there-
ening gate keeping and referral systems. Investments are fore a need for mixed methods approaches to deepen the
needed to support care at all levels, for example, increas- understanding of efficiency and its determinants in different
ing PHC financing, coverage of basic services, prioritising settings. Lastly, the review offers insights on the drivers of
health workers providing primary healthcare. In Ethiopia, the efficiency of national and sub-national health systems,
investment in health extension workers that provided com- highlights potential targets for reforms to improve health
munity health services to catchment populations improved system efficiency.
the efficiency of the health system by improving access to
Supplementary Information The online version contains supplemen-
primary healthcare [53]. Further, interventions to strengthen tary material available at [Link] oi.o​ rg/1​ 0.1​ 007/s​ 40258-0​ 22-0​ 0785-2.
quality of care and curb overprovision of unnecessary care
should be considered. The WHO report on the ten leading Acknowledgements This paper is published with the permission of
causes of efficiency identified examples that include inap- the director KEMRI.
propriate hospital admissions, medical errors, and overuse
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 221

Declarations 11. World Health Organization. Health systems financing: the path
to universal coverage. 2010.
Ethics approval and consent to participate Not applicable—this is a 12. Tandon A, Cashin C. Assessing public expenditure on health
review of literature. from a fiscal space perspective. Washington, DC: World Bank;
2010. p. 1–84.
Consent for publication Not applicable—this is a review of literature. 13. Powell-Jackson, T., K. Hanson, and D. McIntyre, Fiscal space
for health: a review of the literature. 2012: London, UK.
Availability of data and materials All the data used for this study, as 14. Karpa W, Leśniowska J. Efficiency of Health Care Systems:
well as the coding framework used, are publicly available as Online Stochastic Frontier Analysis Including Innovation Component.
Supplementary files. Prace Naukowe Akademii im. Jana Długosza w Częstochowie.
Pragmata Tes Oikonomias. 2014; (8): 159–67
15. Allin S, Grignon M, Wang L. The determinants of efficiency
Conflict of interests All authors declare no competing interests. in the Canadian health care system. Health Econ Policy Law.
2016;11(1):39–65.
Funding This work was funded by a MRC/FCDO/ESRC/Wellcome 16. PRISMA. PRISMA-TRANSPARENT REPORTING of SYS-
Trust Health Systems Research Initiative (HSRI) Grant no MR/ TEMATIC REVIEWS and META-ANALYSES. 2015. [Link]
R01373X/1. www.​prisma-​state​ment.​org/. Accessed 31 Dec 2017.
17. Mijovic H, McKnight J, English M. What does the literature tell
Authors’ contributions EB, JJ, KH, AM, LN, IM, SM, BS, and CN us about health workers’ experiences of task-shifting projects
conceived of the study. EB and RM collected data. EB and RM ana- in sub-Saharan Africa? A systematic, qualitative review. J Clin
lysed the data and wrote the first version of the manuscript. All authors Nurs. 2016;25(15–16):2083–100.
contributed to further analysis and writing of subsequent versions of 18. CASP UK. Critical Appraisal Skills Programme (CASP) Check
the manuscript. lists. 2017. http:// ​ w ww. ​ c asp- ​ u k. ​ n et/ ​ c asp- ​ t ools- ​ check ​ l ists.
Accessed 31 July 2017.
Open Access This article is licensed under a Creative Commons Attri- 19. Hannes K. Chapter 4: Critical appraisal of qualitative research
bution-NonCommercial 4.0 International License, which permits any in Noyes J, Booth A, Hannes K, Harden A, Harris J, Lewin S,
non-commercial use, sharing, adaptation, distribution and reproduction Lockwood C (editors), Supplementary Guidance for Inclusion
in any medium or format, as long as you give appropriate credit to the of Qualitative Research in Cochrane Systematic Reviews of
original author(s) and the source, provide a link to the Creative Com- Interventions. 2011. [Link]
mons licence, and indicate if changes were made. The images or other handb​ook-​guida​nce. Accessed 1 Sept 2015.
third party material in this article are included in the article's Creative 20. Ritchie J, Spencer L. Qualitative data analysis for applied
Commons licence, unless indicated otherwise in a credit line to the policy research. In: Alan B, Robert GB, editors. Analyzing
material. If material is not included in the article's Creative Commons qualitative data. The qualitative researcher’s companion. Oxon:
licence and your intended use is not permitted by statutory regula- Routledge; 1994. p. 173–94.
tion or exceeds the permitted use, you will need to obtain permission 21. Bekaroglu C, Heffley D. A multi-stage efficiency analy-
directly from the copyright holder. To view a copy of this licence, visit sis of OECD healthcare systems. J Manag Econ Res.
[Link] 2018;16(2):264–85.
22. Evans DB, et al. The comparative efficiency of national health
systems in producing health: an analysis of 191 countries.
World Health Organization. 2000;29(29):1–36.
23. Aristovnik A. Measuring relative efficiency in health and
References education sector: the case of East European countries. Actual
Probl Econ. 2012;10(136):305–14.
1. World Health Organization. The World Health Report 2000. 24. Kim Y, Park MJ, Atukeren E. Healthcare and Welfare Policy
Health systems: improving performance. 2000. Efficiency in 34 Developing Countries in Asia. Int J Environ
2. Jacobs R, Smith PC, Street A. Measuring efficiency in health Res Public Health. 2020;17:4617.
care: analytic techniques and health policy. Cambridge: Cam- 25. Singh S, et al. Application of DEA-Based Malmquist Produc-
bridge University Press; 2006. tivity Index on healthcare system efficiency of ASEAN Coun-
3. Hollingsworth B. Cost, production, efficiency, or effectiveness: tries. Int J Health Plan Manag. 2020;36:1236–50.
where should we focus? Lancet Glob Health. 2013;1(5):e249–50. 26. Moreno-Serraa R et al. An exploratory application of data
4. Cylus J, Papanicolas I, Smith PC. Health system efficiency. How envelopment analysis to the efficiency of health service cover-
to make measurement matter for policy and management. 2016. age and access. Results for Development Institute. 2012.
5. World Health Organization. Everybody's business. Strengthening 27. Novignon J, Lawanson A. Efficiency of health systems in sub-
health systems to improve health outcomes. WHO's framework Sahara Africa: a comparative analysis of time varying stochas-
for action. 2007. tic frontier models, MPRA paper 56897. Berlin: University
6. Kutzin J. Health financing policy: a guide for decision-makers. Library of Munich; 2014.
Health financing policy paper. Copenhagen, WHO Regional 28. Ding J, et al. Equity and efficiency of medical service systems
Office for Europe, 24, 2008. at the provincial level of China’s mainland: a comparative
7. Palmer S, Torgerson DJ. Definitions of efficiency. BMJ. study from 2009 to 2014. BMC Public Health. 2018;18(1):214.
1999;318(7191):1136. 29. Du T. Performance measurement of healthcare service and
8. Farrell MJ. The measurement of productive efficiency. 1957. association discussion between quality and efficiency: evi-
9. Coelli TJ, et al. An introduction to efficiency and productivity dence from 31 provinces of mainland china. Sustainability.
analysis. 2nd ed. New York: Springer Science & Business Media; 2017;10(1):74.
2005. 30. Ngobeni V, Breitenbach MC, Aye GC. Technical efficiency of
10. Hollingsworth B, Peacock S. Efficiency measurement in health provincial public healthcare in South Africa. Cost Eff Resour
and health care. Oxon: Routledge; 2008. Alloc. 2020;18(3):1–19.
222 R. Mbau et al.

31. Mousa W, Aldehayyat JS. Regional efficiency of healthcare ser- 54. Nyawira L, et al. Examining health sector stakeholder percep-
vices in Saudi Arabia. Middle East Dev J. 2018;10(1):152–74. tions on the efficency of county ehalth systems in Kenya. PLoS
32. Schleiniger R. Regional quantity, productivity and efficiency Global Public Health. 2021;1(12):e0000077.
measures of the Swiss health care system. Swiss J Econ Stat. 55. Zeng W, et al. Understanding the performance of county health
2008;144(3):459–76. service delivery in Kenya: a mixed-method analysis. Health
33. Varela PS, de Andrade Martins G. Efficiency of primary health Policy Plan. 2021;00:1–11.
care spending by municipalities in the metropolitan region of 56. Charnes A, Cooper W, Rhodes E. Measuring of efficiency of
Sao Paulo: a comparative analysis pf DEA models. Rev Bus. decision making units. Eur J Oper Res. 1978;2:429–39.
2011;32(1):17–35. 57. Aigner D, Lovell C, Schmidt P. Formulation and estimation
34. Loikkanen HA, Susiluoto I. Cost efficiency of Finnish munici- of stochastic frontier production function models. J Econom.
palities in basic service provision 1994–2002. 2005. 1977;6:21–37.
35. Purohit BC. Efficiency in health care sector in Bihar (India): 58. Makiela K. State level efficiency measures for healthcare sys-
An exploratory analysis using DEA. Afr J Health Econ. tems. 2010.
2015;4(2):1–13. 59. Şenel T, Cengiz MA. A bayesian approach for evaluation of
36. Chansa C, et al. Exploring SWAp’s contribution to the efficient determinants of health system efficiency using stochastic frontier
allocation and use of resources in the health sector in Zambia. analysis and beta regression. Comput Math Methods Med. 2016.
Health Policy Plan. 2008;23(4):244–51. 60. de Cos PH, Moral-Benito E. Determinants of health-system
37. Anselmi L, Lagarde M, Hanson K. The efficiency of the local efficiency: evidence from OECD countries. Int J Health Care
health systems: investigating the roles of health administra- Finance Econ. 2014;14(1):69–93.
tions and health care providers. Health Econ Policy Law. 61. Ramírez-Valdivia MT, Maturana S, Salvo-Garrido S. A multiple
2018;13(1):10–32. stage approach for performance improvement of primary health-
38. Evans DB, et al. Comparative efficiency of national care practice. J Med Syst. 2011;35(5):1015–28.
health systems: cross national econometric analysis. BMJ. 62. De Nicola A, Gitto S, Mancuso P. A two-stage DEA model to
2001;323(7308):307–10. evaluate the efficiency of the Italian health system. 2011.
39. Kumbhakar SC. Efficiency and productivity of world health 63. Gravelle H, et al. Comparing the efficiency of national health sys-
systems: where does your country stand? Appl Econ. tems: a sensitivity analysis of the WHO approach. Appl Health
2010;42(13):1641–59. Econ Health Policy. 2003;2(3):141–7.
40. Herrera S, Pang G. Efficiency of public spending in developing 64. Purohit BC. Healthcare sector efficiency in Gujarat (India): an
countries: an efficiency frontier approach. Washington, DC: The exploratory study using data envelopment analysis. Healthc Low
World Bank; 2005. Resour Settings. 2016;4(1). [Link]
41. Allin S, et al. How can health system efficiency be improved in 5525
Canada? Healthc Policy. 2015;11(1):33. 65. See KF, Yen SH. Does happiness matter to health system effi-
42. Popescu C, Asandului L, Fatulescu P. A data envelopment analy- ciency? A performance analysis. Heal Econ Rev. 2018;8(1):33.
sis for evaluating Romania’s health system. Procedia Soc Behav 66. Sankar D, Kathuria V. Health system performance in rural India:
Sci. 2014;109:1185–9. efficiency estimates across states. Econ Political Wkly. 2004:
43. Ozcan YA, Khushalani J. Assessing efficiency of public health 39(13):1427–33
and medical care provision in OECD countries after a decade of 67. Sinimole K. Evaluation of the efficiency of national health sys-
reform. CEJOR. 2017;25(2):325–43. tems of the members of World Health Organization. Leadersh
44. Kujawska J. Efficiency of healthcare systems in european coun- Health Serv. 2012;25(2):139–50.
tries-the DEA network approach. Metody Ilościowe w Badaniach 68. Tandon A, et al. Measuring overall health system performance
Ekonomicznych. 2018;19:60–70. for 191 countries. Geneva: World Health Organization; 2000.
45. Nyawira L, et al. Examining health sector stakeholder percep- 69. Tandon A. Measuring efficiency of macro systems: an application
tions on the efficiency of county health systems in Kenya. PLoS to Millennium Development Goal attainment. 2005.
Glob Public Health. 2021;1(12):e0000077. 70. Achoki T, et al. Technical and scale efficiency in the delivery of
46. Allin S, et al. Improving health system efficiency: perspectives child health services in Zambia: results from data envelopment
of decision makers. Healthc Q. 2017;20(1):10–3. analysis. BMJ Open. 2017;7(1): e012321.
47. Moses MW, et al. Performance assessment of the county health- 71. Stefko R, Gavurova B, Kocisova K. Healthcare efficiency assess-
care systems in Kenya: a mixedmethods analysis. BMJ Glob ment using DEA analysis in the Slovak Republic. Health Econ
Health. 2021;6:e004707. Rev. 2018;8(1):6.
48. Zeng W, et al. Understanding the performance of county health 72. Dash U, Vaishnavi SD, Muraleedharan VR. Technical efficiency
service delivery in Kenya: a mixed-method analysis. Health in the use of healthcare resources: a case study of Tamil Nadu.
Policy Plan. 2022;37(2):189–99. Indian Econ Rev. 2008;43(1):69–82
49. Nannyonjo J, Okot N. Decentralization, local government capac- 73. Anton SG. Technical efficiency in the use of health care
ity and efficiency of health service delivery in Uganda. J Afr Dev. resources: a cross-country analysis. Ann Alexandru Ioan Cuza
2013;15(1):125–58. Univ Econ. 2013;60(1):1–12.
50. Hooda Y. Efficiency in expenditure on healthcare system at state 74. Kinfu Y, Sawhney M. Inefficiency, heterogeneity and spillover
level in India. Int J Adv Res Manag Soc Sci. 2018;7:34–41. effects in maternal care in India: a spatial stochastic frontier
51. Purohit BC. Efficiency of social sector expenditure in India: a analysis. BMC Health Serv Res. 2015;15(1):118.
case of health and education in selected Indian states. Healthc 75. Suraratdecha C, Okunade AA. Measuring operational efficiency
Low Resource Settings. 2014;2(1):26–34 in a health care system: a case study from Thailand. Health Pol-
52. Felder S, Tauchmann H. Federal state differentials in the effi- icy. 2006;77(1):2–23.
ciency of health production in Germany: an artifact of spatial 76. Canagarajah S, Ye X. Public health and education spending in
dependence? Eur J Health Econ. 2013;14(1):21–39. Ghana in 1992–98: issues of equity and efficiently, Vol. 2579.
53. Yip W, Hafez R, Organization WH. Improving health system effi- 2001; World Bank Publications.
ciency: reforms for improving the efficiency of health systems:
lessons from 10 country cases. World Health Organization. 2015.
Analysing the Efficiency of Health Systems: A Systematic Review of the Literature 223

77. Hadad S, Hadad Y, Simon-Tuval T. Determinants of healthcare methodology analysis. 2011, Department of Economics, Univer-
system’s efficiency in OECD countries. Eur J Health Econ. sity of Bari
2013;14(2):253–65. 97. Chai P, Wan Q, Kinfu Y. Efficiency and productivity of health
78. Ibrahim MD, Daneshvar S. Efficiency analysis of healthcare systems in prevention and control of non-communicable diseases
system in Lebanon using modified data envelopment analysis. J in China, 2008–2015. Eur J Health Econ. 2021;22:267–79.
Healthc Eng. 2018;2018:2060138. [Link] oi.o​ rg/1​ 0.1​ 155/2​ 018/​ 98. Verhoeven M, Gunnarsson V, Lugaresi S. The health sector in the
20601​38. Slovak Republic: efficiency and reform. IMF Working Papers,
79. Porcelli F. Electoral accountability and local government effi- 2007. p. 1–25.
ciency: quasi-experimental evidence from the Italian health care 99. Jafarov E, Gunnarsson V. Efficiency of government social spend-
sector reforms. Econ Gov. 2014;15(3):221–51. ing in Croatia. Financial Theory Pract. 2008;32(3):289–320.
80. Lee S, Kim C. Estimation of association between healthcare 100. Gong G, et al. Has the efficiency of China’s healthcare system
system efficiency and policy factors for public health. Appl Sci. improved after healthcare reform? A network data envelopment
2018;8(12):2674. analysis and Tobit regression approach. Int J Environ Res Public
81. González E, Cárcaba A, Ventura J. Value efficiency analysis Health. 2019;16:4847.
of health systems: does public financing play a role? J Public 101. Wranik D. Healthcare policy tools as determinants of health-
Health. 2010;18(4):337–50. system efficiency: evidence from the OECD. Health Econ Policy
82. Spinks J, Hollingsworth B. Cross-country comparisons of techni- Law. 2012;7(2):197–226.
cal efficiency of health production: a demonstration of pitfalls. 102. Chai P, et al. Technical and scale efficiency of provincial health
Appl Econ. 2009;41(4):417–27. systems in China: a bootsrappin data envelopment analysis. BMJ
83. Ahmed S, et al. Measuring the efficiency of health systems in Open. 2019;9:e027539.
Asia: a data envelopment analysis. BMJ Open. 2019;9:e022155. 103. Bhat VN. Institutional arrangements and efficiency of health care
84. Moreno-Serra R, Anaya-Montes M, Smith PC. Potential determi- delivery systems. Eur J Health Econ. 2005;6(3):215–22.
nants of health system efficiency: evidence from Latin America 104. Nassar H, et al. Technical efficiency of health-care systems in
and the Caribbean. PLoS One. 2019;14(5): e0216620. selected middle-income countries: an empirical investigation.
85. Zarulli V, et al. Health care system efficiency and life expectancy: Review of economics and political science. 2020;5(4):267–87.
a 140-country study. PLoS One. 2021;16(7): e0253450. 105. Sousa A, et al. Measuring the efficiency of human resources for
86. Barasa E, et al. Level and determinants of county health system health for attaining health outcomes across subnational units in
technical efficiency in Kenya: two stage data envelopment analy- Brazil. World health report, 2006.
sis. Cost Eff Resour Alloc. 2021;19(78):1–11. 106. Varabyova Y, Müller J-M. The efficiency of health care produc-
87. Purohit BC. Efficiency in healthcare sector in Assam: a sub-state tion in OECD countries: a systematic review and meta-analysis
level analysis. Online J Health Allied Sci. 2016;14(4):1. of cross-country comparisons. Health Policy. 2016;120:252–63.
88. Prachitha J, Shanmugam K. Efficiency of raising health outcomes 107. Pelone F, et al. Primary care efficiency measurement using
in the Indian States. Chennai: Madras School of Economics; data envelopment analysis: a systematic review. J Med Syst.
2012. 2015;39:156.
89. Purohit BC. Efficiency of health care sector at sub-state 108. Hussey PS, et al. A systematic review of health care efficiency
level in India: a case of Punjab. Online J Health Allied Sci. measures. Health Services Res. 2009;44(3):784–805.
2009;8(3):212–25. 109. Asbu EZ, Masri MD, Naboulsi MA. Determinants of hospital
90. Grigoli F, Kapsoli J. Waste not, want not: the efficiency of health efficiency: a literature review. Int J Healthc. 2020;6(2).44–53.
expenditure in emerging and developing economies. Rev Dev 110. Bennett S, Glandon D, Rasanathan K. Governing multisectoral
Econ. 2018;22(1):384–403. action for health in low-income and middle-income countries:
91. Hassan CM, Djekonbe D, DRD T. The efficiency of the health unpacking the problem and rising to the challenge. BMJ Glob
system in Chad. 2018. Health. 2018;3(Suppl 4):e000880.
92. Hollingsworth B, Wildman J. The efficiency of health produc- 111. Tangcharoensathien V, et al. Multisectoral actions for health:
tion: re-estimating the WHO panel data using parametric and challenges and opportunities in complex policy environments.
non-parametric approaches to provide additional information. Int J Health Policy Manag. 2017;6(7):359–63.
Health Econ. 2003;12(6):493–504. 112. Merkur S, Sassi F, McDaid D. Promoting health, preventing
93. Lagravinese R, Paradiso M, Mastromarco C. Social heterogene- disease: is there an economic case? WHO Regional Office for
ity and efficiency in Italian health care system: a Simar–Wilson Europe: Denmark; 2013.
methodology analysis. Società Italiana degli Economisti Working 113. WHO. More health for the money, in World Health Report—
Paper September, 2011. 15: p. 2011. Health Systems Financing; the path to universal coverage. 2010,
94. Greene W. Distinguishing between heterogeneity and ineffi- World Health Organization: Geneva.
ciency: stochastic frontier analysis of the World Health Organi- 114. Yip W, Hafez R. Reforms for improving the efficiency of health
zation’s panel data on national health care systems. Health Econ. systems: lessons from 10 country cases; 2015. World Health
2004;13(10):959–80. Organization.
95. Ravangard R, et al. Factors affecting the technical efficiency of
health systems: a case study of Economic Cooperation Organi-
zation (ECO) countries (2004–10). Int J Health Policy Manag.
2014;3(2):63.
96. Lagravinese R, Paradiso M, Mastromarco C. Social heterogene-
ity and efficiency in Italian healthcare system: a Simar–Wilson
224 R. Mbau et al.

Authors and Affiliations

Rahab Mbau1 · Anita Musiega1,2 · Lizah Nyawira1 · Benjamin Tsofa3 · Andrew Mulwa4 · Sassy Molyneux3,5 ·
Isabel Maina6 · Julie Jemutai1 · Charles Normand7,8 · Kara Hanson9 · Edwine Barasa1,5

1 6
Health Economics Research Unit, KEMRI-Wellcome Trust Health Financing Department, Ministry of Health, Nairobi,
Research Programme, Nairobi, Kenya Kenya
2 7
Institute of Healthcare Management, Strathmore University, Centre for Health Policy and Management, Trinity College,
Nairobi, Kenya The University of Dublin, Dublin, Ireland
3 8
Health Systems and Research Ethics Department, Cicely Saunders Institute, Kings College London, London,
KEMRI-Wellcome Trust Research Programme, Kilifi, Kenya UK
4 9
County Department of Health, Makueni County Government, Faculty of Public Health and Policy, London School
Nairobi, Kenya of Hygiene and Tropical Medicine, London, UK
5
Centre for Tropical Medicine and Global Health, Nuffield
department of Medicine, University of Oxford, Oxford, UK