What is a Health System? Why Should We Care?
By
William C. Hsiao
K.T. Li Professor of Economics and Health Policy
Harvard School of Public Health
August, 2003
What Is A Health System? Why Should We Care?
Abstract
Health system reforms and health system comparisons have been popular topics
of discussion for the policy and research communities. Yet, there is no clear concept and
definition for a health system. As a result, comparisons are often made between apples
and oranges, resulting in confused discourses and misleading conclusions.
This paper argues that for policy and economic research purposes, it is most
useful to conceptualize a health system as a set of relationships in which the structural
components (means) and their interactions are associated and connected to the goals
the system desires to achieve (ends). The model identifies three common goals and
five means that nations use to achieve their goals. The differences in the structural
components may explain the variety of observed system outcomes.
Keywords: health systems comparisons, health system performance, reforms, policy
model
Acknowledgement
The paper benefited greatly from discussions with my colleagues Peter Berman,
Michael Reich, Marc Roberts, and Winnie Yip, and from the comments of the
participants of the annual World Bank’s Flagship Courses since 1997. Carrie Thiessen
improved the paper immensely by her insightful comments and her able research
assistance. Any error remains the sole responsibility of the author.
Introduction
Globally, the policy and research communities have heatedly debated health
system reforms. Health systems have been dissected, analyzed, evaluated and compared.
However, there is no common and consistent answer to the question what is a health
system? The term ‘health system’ has been defined differently for different purposes.
The ambiguous concepts and meanings of a health system have caused confusion in
public debate and misled policy deliberations. Policy makers have a specific interest in
the development of an adequate and consistent definition that will enable them to
understand what instruments (interventions) are likely to improve the performance of a
health system. At the same time, they want to learn from the “better” systems to reform
their own. Researchers want to investigate what structural components cause the varied
outcomes.
Health systems have been conceptualized and defined in various ways.
Traditionally, health systems were described in terms of capacity indicators and activities
(e.g. number of hospital beds, physicians and nurses, government programs.) [1], [2].
Roemer also argued that a health system should be described by five characteristics:
productive resources, organization of programs, economic support mechanisms,
management methods and service delivery. However, his conceptualization of health
system does not adequately explain why these categories of activity matter or what
difference it makes when the configuration of these characteristics varies. Hurst took a
different approach, describing health systems as a series of fund flows and payment
methods between population groups and institutions [3]. Both approaches are
informative, but neither explains why and how a particular system produces a set of
outcomes.
Another body of literature presents a health system as a set of functional
components. Londono and Frenk [4] argued a system consists of four functions:
financing, delivery, modulation and articulation. Applying this concept to health systems
1
financed through social insurance, they proposed a new organizational model to carry out
these functions. Anne Mills [5] also conceptualized health systems as loose framework of
actors and functions. The functions she identified are financing, regulation, resource
allocation, and service provision. While these approaches help classify and analyze a
health system by its internal functions, they do not make explicit what goals the functions
aim to achieve, how the functions effectuate them, how the functions interrelate, or how
variations in organizing the functions affect outcomes.
The World Health Organization’s World Health Report 2000 [6] defined health
systems by the boundary of activities they encompass. Unlike the approaches discussed
above, the majority of the Report focused on the performance (ultimate outcomes) of
health systems and performance measurement. The Report described health system
functions (stewardship, resource creation, service provision, and financing), emphasizing
the stewardship role of the government. However, the Report did not adequately address
the relationships between the key functions and health system performance. More
importantly for policymakers, it does not explain why a particular system yields a given
outcome, what features of that system contributed the most to producing the outcome, or
how one could restructure the system to achieve a preferable outcome.
To investigate this why and how, health economists have largely applied
economic theories of supply and demand to model and analyze actions in the various
markets that comprise the health system [7], [8]. A health system can be conceptualized
on at least two levels: macro and micro. The macro-level focus is on overall dimensions
of health sector, the total size, shape, and functioning of the “elephant,” that is the health
sector, while the micro-level explores behavior and dynamics of individual firms and
households [9], [10]. Ideally, the aggregated behavior of individual households and firms
predicted by microeconomic theory would explain macro-level phenomena. However, at
least a dozen markets compose the health sector, and the interactions among them are not
well understood or adequately studied. Consequently, microeconomic theory has offered
little insight into or explanation for macro-level outcomes such as overall health status
2
[11]. Moreover, microeconomic theory has not been able to offer adequate explanations
for major structural features that are common to most health systems and that influence
macro-outcomes.
This paper’s objective is to develop an analytical framework that models the
systemic aspects of a health system, i.e. the major components of a health system that are
related and can explain aggregate outcomes. Stated another way, it is a causal model
whose major components (i.e. explanatory variables) can largely account for observed
outcomes (i.e. dependent variables). Such a model can assist us in understanding the
major factors that may explain varied system outcomes, provide a framework to compare
health systems and test hypotheses, and offer instruments for policymakers to manage
their health systems’ performance.
The paper represents the culmination of several years of research, initiated a
decade ago [12], [13], [14], [15] and builds upon the work of other researchers. The
paper is organized in four sections. The first section examines the fundamental principles
used for modeling health systems. Applying these principles, we clarify and answer the
question, “what is a health system?” Section II presents the final goals of a health
system. The next section discusses the five fundamental structural components of a
health system in some detail. The last section summarizes how this model can assist
policymakers, researchers and the public engaged in the search to structure better health
systems.
I. What is a Health System?
Health systems, like other socioeconomic systems, evolve in unique historic,
cultural and political contexts. Nonetheless, every system is structured by state actions or
non-actions to serve certain social purposes. The system exists and evolves to serve
societal needs. Simply put, a health system is a means to an end. Applying a long-
standing paradigm in industrial organization economics, we hypothesize that the
structural components of the system affect the behavior of individuals and firms in that
3
system, and that their behavior and interactions determine the observed outcomes. Under
this paradigm, a health system is a set of relationships in which the means (i.e. structural
components) are causally connected to the ends (i.e. goals.) In this context, then, we have
to analyze the goals and structural components of a health system.
What goals do nations want their health systems to achieve? A myriad of
programmatic goals has been discussed in the literature. Every evaluation study of health
programs specifies the goals by which the program will be assessed. However, goals are
heterogeneous, depending on the purpose of a program. Some programs seek to increase
average health status, some to maximize efficiency, some to prevent impoverishment,
some to improve quality of service. They are not all ultimate goals. Some are
intermediate outcomes or some pertain only to a selected disease or population.
At the systemic level, we must clarify what ultimate outcomes matter to a nation
and distinguish them from intermediate outcomes. While the latter are important and can
affect the final outcomes, they are only intermediary and partial results. We examined
multiple countries’ health-related legislation, policy papers, and reports to identify
the explicit and implicit goals of their health systems. The goals thus identified are:
improving health, financial risk protection, and public satisfaction.
As for the means, there are many possible structural variables that have some
power to explain observed outcomes. How can we sort out which ones are essential and
which ones are peripheral? We use three criteria. First, since our aim is to develop a
model that is useful for policy analysis, we will examine and select only those structural
variables that can be altered by policy. Because we are developing an ends-oriented
model, we focus on the elements that can be used as policy instruments to achieve
societal goals for the health sector. We exclude those variables that cannot be changed
except in the long term, such as culture. Finally, taking advantage of many nation’s
policy “experiments” to improve their health system’s outcomes, we identify plausible
explanatory variables based on empirical observations. The key means, which we call
4
control knobs, include financing, payment, macro-organization of health care delivery,
regulations and persuasion.
In sum, a health system, defined for policy purposes and economic research, is a
set of relationships in which the primary variables are causally associated and linked with
the outcomes. We limit the variables to those that can serve as policy levers. Using this
set of criteria, we propose a new health system definition.
“A health system is defined by those principal casual
components that can explain the system’s outcomes. These
components can be utilized as policy instruments to alter the
outcomes.”
We adopt the WHO’s description of the boundary of the health system as “all the
activities whose primary purpose is to promote, restore, or maintain health.” [6].
II. GOALS
Health Status
What socioeconomic ends are served by health systems? Despite the fact that
nations structure their health care systems very differently, most nations do share certain
basic beliefs: one, good health is of intrinsic value to people; and two, certain health
services are necessary to sustain life and to relieve intense suffering. Although some
researchers have argued that health maximization should be considered the sole goal of a
health system [16], there is now consensus that health systems have multiple purposes.
We have identified two additional health systems goals common to most countries.
Financial Risk Protection
The first of these is financial risk protection. National health insurance systems
explicitly places risk protection as a final goal. The earliest health insurance systems
5
such as the German krankenkassen began as a facet of national program to minimize the
risk of absolute impoverishment among the working class due to disease, disability, and
unemployment [17]. National Health Service systems implicitly offer risk protection by
providing free (or nearly free) expensive ambulatory and hospital services. The British
National Health Services can trace its roots to the Poor Law of 1911 [18]. The Beveridge
Report, which provided the framework for the NHS, recommended the development of a
social insurance scheme to provide a “minimum income needed for subsistence in all
normal cases” [19].
Many countries include “affordability” as a policy objective of the health system.
The affordability of a good is defined by the consumer’s ability to purchase it without
excessive financial burden. Health care is characterized by uncertainty of high medical
costs, hence affordability is determined by the extent of the insurance function of the
health system. Therefore, countries’ “affordability” objective is more properly defined as
the goal of attaining adequate financial risk protection for citizens.
Instead of risk protection, the WHR identified “fairness of financial contribution”
as a health system goal [6]. Fairness of financial contribution measures the share of
households’ non-food expenditure spent on health. It does not assess whether services
are affordable to the poor or how well all citizens are protected against financial
catastrophe. On the other hand, financial risk protection is precisely what concerns most
countries. As a result, recent international research focuses on financial risk protection as
a basic health system goal [20].
Public Satisfaction
It’s self evident that public satisfaction is a goal for political leaders and
policymakers of democratic societies. Even leaders of authoritarian states have to satisfy
the public in the long run. Economists often call this goal individual utility improvement
[21]. Governments are increasingly cognizant of the fact that the stability of the health
system is not assured without adequate public satisfaction. Blendon et al [22] concluded
that public dissatisfaction with health system performance contributes to political
6
pressure for health system reform: dissatisfaction with the status quo is highly correlated
with public opinion that the health care system requires fundamental change or complete
overhaul.
Governments have also relied on opinion polls to guide its policy decisions.
When the British National Health Services was debated in 1942, Beveridge stated:
“This desire is shown both by the established popularity of compulsory
insurance, and by the phenomenal growth of voluntary insurance against
sickness, against death and for endowment, and most recently for hospital
treatment. It is shown in another way by the strength of popular objection
to any kind of means test.” (emphasis added, [19])
More recently, UK explicitly undertook reforms to make services patient-centered and
implemented plans to monitor patient attitudes with surveys and focus groups. (NHS,
2000). At the same time, private insurance providers have adopted patient satisfaction
(and patient experience) as performance measures [23].
The WHR rejected satisfaction as a health system objective, arguing that
satisfaction confounds expectations with accurate assessment of the present
circumstances [24]. However, this component of expectation in public satisfaction is
precisely what helps defines the goals toward which reform is oriented.
The role of equity
Equity is widely defined as a health system objective, often expressed in terms of
“universal equal access to health care.” We consider it a principle be applied to the
achievement of the three goals of health status, financial risk protection, and consumer
satisfaction. In sum, there are two dimensions to each of the three goals: level and
distribution. We can illustrate these declared goals in Fig. 1, noting that these objectives
go beyond the usual concerns of economic analyses that tend to focus exclusively on
efficiency and remain silent on equity [25].
7
Fig. 1 around here
The goals are not entirely independent of each other. Greater achievement in one
goal may further another; likewise, poor performance in respect to a goal may limit
ability to attain another. For example, Blendon et al. conclude that inadequate financial
protection is one of the primary causes for public dissatisfaction with the American
health care system [22].
All nations’ common objective is to achieve multiple goals with a given resource
constraint. Every nation must make difficult trade-offs when it wants to achieve multiple
objectives with limited resources. A nation wrestles with two types of trade-offs: inter-
sectoral and intra-sectoral. First, a nation has to make trade-offs between health-system
goals (e.g. improving the health status of the population) and other economic, political
and social goals (e.g. providing education for all children). Consequently, the level and
distribution of health status, financial risk protection, and consumer satisfaction depend,
in part, on a nation’s economic resources. In common parlance, it depends on what is
affordable.
The second type of trade-off takes place when a nation tries to achieve different
goals within a health system. For example, on the margin a nation has to make trade-offs
between health status and public satisfaction (e.g. no waiting lines.) But rarely do nations
make these inherent trade-offs explicit. Historical processes and fundamental social
values create implicit boundaries to trading off different objectives, limiting the range of
available reform options. Health care systems in European nations, for example, are
deeply rooted in egalitarian traditions. Policy proposals violating this basic foundation of
solidarity have little overall appeal regardless of how much they would enhance
efficiency [26]. On the other hand, the health care system of the USA is rooted in
libertarian traditions. Compulsory health insurance to cover all Americans remains
elusive after more than sixty years of public debate [27].
We often confuse intermediate outcomes with the ultimate goals we care about.
Targeting health polices and programs to improve access, quality and/or efficiency are
8
important, but they are of derivative importance to the ultimate goals of a health system.
We are interested in pursuing higher technical quality of health services because it has a
positive effect on health status. Improving service quality of health services is desirable
insofar as it affects patient satisfaction and health outcomes. Maximizing allocative
efficiency enables improvements in health status and risk protection under budgetary
constraints. Ultimately, a nation’s success in attaining these intermediate outcomes
should be assessed in terms the extent to which they contribute to the final outputs. Fig.
2 illustrates the relationship between means, some intermediate outcomes and final goals
of a health system.
Fig. 2 around here
III Control Knobs (Means)
Many nations have tried different policy “experiments” to improve their health
systems’ performance. These ‘natural exper