Essay preview
Risk Pooling in Health Care Finance
Peter C. Smith and Sophie N. Witter
Centre for Health Economics
University of York
York YO10 5DD
United Kingdom
Report prepared for the World Bank Workshop
Resource Allocation and Purchasing in Health: Value for Money, Reaching the Poor World Bank, Washington DC, May 14-15 2001
Revised November 2001
Phone
Fax
E-mail
+ 44 1904 433779
+ 44 1904 433759
[email protected]
Acknowledgements
The authors would like to thank Jack Langenbrunner, Maureen Lewis, Alex Preker and Paul Shaw of the World Bank, Philip Davies of the World Health Organization, and participants at the workshop for comments.
Risk Pooling in Health Care Finance
Contents
RISK POOLING IN HEALTH CARE FINANCE...............................................................................................................1 CONTENTS ....................................................................................................................................................................................1 EXECUTIVE SUMMARY..........................................................................................................................................................1 RISK POOLING IN HEALTH CARE FINANCE...............................................................................................................3 1
INTRODUCTION...............................................................................................................................................................3
2
WHY RISK POOLING IN HEALTH CARE FINANCE? ......................................................................................3
3
APPROACHES TO RISK POOLING..........................................................................................................................7 3.1
3.2
3.3
3.4
4
NO RISK POOLING ............................................................................................................................................................. 7 UNITARY RISK POOL ......................................................................................................................................................... 9 FRAGMENTED RISK POOLS............................................................................................................................................. 10 INTEGRATED RISK POOLS............................................................................................................................................... 12
PRACTICAL ISSUES .....................................................................................................................................................14 4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
THE INSTITUTIONAL FRAMEWORK FOR RISK POOLING.............................................................................................. 14 M EMBERSHIP CRITERIA FOR RISK POOLS .................................................................................................................... 15 SIZE OF THE RISK POOLS ................................................................................................................................................ 15 SETTING CAPITATION PAYMENTS................................................................................................................................. 16 VARIATIONS IN THE BENEFIT PACKAGE ...................................................................................................................... 18 RETROSPECTIVE RISK SHARING.................................................................................................................................... 19 OVERLAPPING RISK POOLS............................................................................................................................................ 20 PAYMENT SYSTEMS AND RISK SHARING...................................................................................................................... 21
5
RISK POOLING IN LOW AND MIDDLE INCOME COUNTRIES................................................................22
6
RISK POOLING AND SYSTEM BEHAVIOUR.....................................................................................................28 6.1
6.2
6.3
BEHAVIOURAL RESPONSES IN SMALL RISK POOLS..................................................................................................... 28 EFFICIENCY OF SUPPLY.................................................................................................................................................. 29 QUALITY OF CARE .......................................................................................................................................................... 30
7
LOCAL CIRCUMSTANCES ........................................................................................................................................30
8
CONCLUDING COMMENTS......................................................................................................................................32
REFERENCES .............................................................................................................................................................................33 A.
TECHNICAL APPENDIX .............................................................................................................................................38 A.1 M ODELLING THE NUMBER OF CLAIMS............................................................................................................................ 38 A.2 M ODELLING THE SIZE OF CLAIMS.................................................................................................................................... 39 A.3 SOME ELABORATIONS....................................................................................................................................................... 40
Risk Pooling in Health Care Finance
Executive Summary
Pooling is the health system function whereby collected health revenues are transferred to purchasing organisations. Pooling ensures that the risk related to financing health interventions is borne by all the members of the pool and not by each contributor individually. Its main purpose is to share the financial risk associated with health interventions for which there is uncertain need.
The purpose of the report is to (i) identify modalities of pooling particularly relevant to developing countries; (ii) assess the impact of various ways of pooling on health system performance; (iii) provide guidelines to design functional pooling arrangements in a variety of health systems context; and (iv) present best practice.
Risk pooling is required because of the large uncertainty in the magnitude and timing of an individual’s health care expenditure needs (section 2). It implies three redistributive functions: from the rich to the poor, from the healthy to the sick, and from the productive to the unproductive stage of the life cycle.
The arguments in favour of risk pooling in health care reflect equity and efficiency considerations. The equity arguments reflect the view society does not feel that it is fair that individuals should assume all the risk associated with their health care expenditure needs. The efficiency arguments arise because pooling can lead to major improvements in population health, can increase productivity, and reduces uncertainty associated with health care expenditure. There are four classes of approach to risk pooling (section 3): • no risk pool, under which all expenditure liability lies with the individual; • unitary risk pool, under which all expenditure liability is transferred to a single national pool;
• fragmented risk pools, under which a series of independent risk pools (such as local governments or employer-based pools) are set up;
• integrated risk pools, under which fragmented risk pools are compensated for the variations in risk to which they are exposed.
The report argues that the unitary risk pool is an ideal, but that numerous practical difficulties in making it operational. Instead, it is likely that most health systems will use a system of fragmented pools, and that methods of integration are therefore required. The report offers some guidance on implementation (section 4). Numerous choices must in principle be made, including:
• the institutional basis for risk pools (geography, employment sector, employment status, and so on);
• the criteria for membership of a risk pool;
• the size of risk pools;
• whether or not the risk pools are competitive;
1
•
•
•
whether or not contributions are mandatory;
whether financial contributions are community rated or risk rated; the extent to which health care users retain some expenditure risk (in the form of user charges);
• the extent to which there are financial transfers between risk pools; • the extent to which the risk pools are protected from unpredicted variations in expenditure needs by some higher level pooling;
• the freedom given to risk pools to choose variations in packages of care, membership entitlement and financial contributions.
Eight practical issues that arise in seeking to design a system of pooling are discussed: the institutional framework for risk pooling arrangements, membership criteria for risk pools, the size of the risk pools, setting capitation payments, variations in the benefit package, retrospective risk sharing, overlapping risk pools, and payment systems and risk sharing. Experience in low and middle income countries is summarized (section 5). We report heavy reliance on out of pocket payments, great difficulty in collecting premiums, partial and fragmented risk pooling (where it exists), a preponderance of small risk pools, weak attempts at integration, a pattern of resource allocation which is highly dependent on history, a weak stewardship function and considerable tensions between local and central powers. The design of risk pools may have serious implications for system behaviour (section 6). There are particular dangers when risk pools are fragmented through devolution. We discuss possible impacts on adverse behavioural responses in small risk pools, on the efficiency of supply, and on quality.
The optimal design of risk pooling arrangements depends heavily on local circumstances (section 7). However, there exist a number of universal indicators of variations in behaviour between risk pools which can offer strong prima facie evidence of the extent to which risk pools have been successfully integrated.
We conclude (section 8) that in principle the desirability of some form of risk pool integration in health care should be uncontested for most situations. Rather, the debate should surround how it can be practically implemented in low and middle income countries, particularly when a reliance on community financing and user charges leads to fragmented risk pools. We highlight the many considerations that apply to any choice, and emphasize that system design will be contingent on local circumstances. We nevertheless conclude that there will be few circumstances in which some form of risk pool integration cannot be successfully introduced and strengthened.
2
Risk Pooling in Health Care Finance
1 Introduction
This report examines risk pooling in health care finance, with particular reference to developing economies. It was commissioned by the World Bank as part of a project on Resource Allocation and Purchasing in Health. Pooling is the health system function whereby collected health revenues are transferred to purchasing organisations. Pooling ensures that the risk related to financing health interventions is borne by all the members of the pool and not by each contributor individually. Its main purpose is to share the financial risk associated with health interventions for which there is uncertain need.
The purpose of the report is to (i) identify modalities of pooling particularly relevant to developing countries; (ii) assess the impact of various ways of pooling on health system performance; (iii) provide guidelines to design functional pooling arrangements in a variety of health systems context; and (iv) present best practice. The report covers issues related to purchasing only to the extent that they directly relate to the risk pooling function. The report is arranged as follows. The next section examines the rationale for risk pooling. It is followed by a discussion of the various types of risk pooling that exist. The following section examines the issues that arise when seeking to implement risk pooling in practice. Experience in low and middle income countries is then discussed. The report ends with an assessment of the effectiveness of alternative risk pooling arrangements, and a discussion of the implications for health care policy.
2 Why Risk Pooling in Health Care Finance?
In contrast to many of life’s necessities, an individual’s need for health care is intrinsically uncertain. Whilst an individual’s expenditure on (say) food is regular and largely predictable, that same individual’s expenditure on health care is to a large extent unknowable, both in magnitude and timing. It is therefore intrinsically difficult for an individual to make financial provision for episodes of sickness, or even chronic health care needs. Furthermore, if (as is generally accepted) most individuals are risk averse, then they would value arrangements that protect them from this uncertainty in expenditure.
Of course, notwithstanding this large uncertainty, it is often the case that the health care expenditure needs of individuals can to some extent be predicted. Other things being equal, older people tend to have higher spending needs than younger people (except for the very young), and people with chronic health conditions tend to have higher spending needs than healthy individuals. To the extent that characteristics such as age and health status can be measured, it becomes possible to make improved predictions of health care spending for a particular individual. Developments such as genetic testing offer the prospect of refining further such predictions.
3
However, as Newhouse et al [1] and others have shown, even if it were possible to measure every conceivable factor contributing to an individual’s health care spending, it would still be possible to predict only about 20% of the variation in annual expenditure. And in practice, even if excellent information systems are available, that figure rarely rises above 10% [2]. The uncertainty in health care expenditure has two elements: its timing and its magnitude. Suppose first that all individuals were expected to incur the same health care expenditure over the course of their lifetimes, and that the only uncertain element is the timing of that expenditure. Then in principle financial provision for health care could be made in the form of identical individual health care funds, that are used as required over the course of the individual’s lifetime, and that are run down to zero by death. Such a fund could be endowed on all citizens at birth, or could be financed by regular contributions from the individual, say over a working lifetime. The role of the individual fund is to protect the citizen from lumpy expenditure needs of uncertain timing, and to offer unhindered access to necessary health care when the need arises. In effect, the annual risks associated with health care expenditure are pooled across an individual’s lifetime.
This sort of principle underlies the use of medical savings accounts in countries such as Singapore [3]. In a similar vein, there has been some examination in developed countries of the extent to which an individual’s death is the major indicator of the timing of health care expenditure, on the grounds that the year before death is known to be a period of particularly intense use of health care, but the findings suggest that this is only one of many important predictors of expenditure [4].
In practice, the timing of expenditure is only one element of the uncertain need for health care confronting individuals, and there are also substantial variations in the lifetime expenditure on health care associated with different individuals. We might characterize the spectrum of lifetime expenditure as moving from “healthy” individuals (low lifetime expenditure) to “unhealthy” individuals. In practice, individuals with chronic medical conditions might be expected to be towards the unhealthy end of the spectrum.
Society could in principle take the view that the pursuit of health and consumption of health care are matters strictly for the individual to arrange, and offer the individual no intervention from the broader community to compensate for variations in health care expenditure needs. Such an extreme individualistic position has rarely been adopted in practice. Instead, to a greater or lesser extent, all systems of health care implicitly pool the risks associated with individual health care needs. 1 The World Health Organization defines risk pooling as “the practice of bringing several risks together for insurance purposes in order to balance the consequences of the realization of each individual risk” [5].
1.
It is important to note that in the English language the concept of “risk” is ambiguous, and has at least two quite distinct connotations relevant to this paper: risk defined as the relative propensity to incur health care expenditure (in the sense of a "high risk patient"); and risk defined as the unpredictable variability associated with a particular expected level of expenditure. The creation of risk pools refers to both of these definitions. It amalgamates individuals with different expected health care needs (definition 1), but also reduces the per capita variability in total health care needs (definition 2). Both of these notions of risk are of fundamental importance in the design, management and performance of the health care system. If there is any doubt about the intended meaning, we use the expression “variability” when employing the second sense of the word.
4
In the extreme, all health care resources available for people of all needs could be pooled, and access to health care determined solely by clinical need. If patients are charged no out-of-pocket fee for such access, this arrangement implies an implicit redistribution of resources from the healthy to the less healthy. This is the principle to which certain unitary systems of health care aspire, as for example in the UK National Health Service. The intention is that equal opportunity of access should be offered on the basis of clinical need only, regardless of any other individual characteristics, such as wealth or area of residence. In practice, all systems of health care exhibit a mix of these two extreme principles of no pooling and complete pooling. Any health care risk pool must be financed, but pooling implies that there is no reason why the magnitude of individual contributions to the finance should be related to health status or health care utilization. Rather, society must choose the extent to which individual financial contributions depend on financial means, health care utilization, or other factors. Whatever system is chosen, a crucial constraint is of course that the revenues received must be sufficient to provide the desired system of health care.
The World Health Organization [5] illustrates two of the redistributive issues implicit in risk pooling (from healthy to sick, and from rich to poor) by means of two stylized scenarios: • members might make equal financial contributions, but the pool effectively enables a transfer to be made from the relatively healthy to the sick (the risk pooling function). A community financing programme charging members a flat rate might be expected to function in this way.
• members might make equal use of health care, but by seeking differential financial contributions the pool effectively enables a transfer to be made from the rich to the poor (the income redistribution function). This is the aim of health financing systems that base contributions on income levels (for example, many social insurance programmes). To these we would add a third:
• members might make equal financial contributions and make equal use of health care across their lifetimes, but the pool enables a transfer to be made depending on the stage of the individual’s life cycle. This is the life cycle redistributive function of the risk pool. These concepts are illustrated in the RAP concept note as in Figure 1 [6].
5
e
e
e
c
r
u
o
Cross-subsidy from
rich to poor
o
Cross-subsidy from
low-risk to high-risk
Cross subsidy from
productive to non-productive
part of the life cycle
s
s
s
o
u
u
r
r
c
c
Figure 1: Pooling of revenues implies transfers (a) from healthy to sick (b) from rich to poor and (c) across life cycle
e
$
Low
risk
R
R
R
e
$
$
e
$
$
$
High
risk
Poor
Health risk
Rich
Income
Produ
ctive
Nonproduc
tive
Age
There are two broad categories of argument in favour of risk pooling in health care, reflecting equity and efficiency considerations. The equity arguments reflect the view society does not feel that it is fair that individuals should assume all the risk associated with their health care expenditure needs. Instead, all or some of that risk should be spread across a given risk pool. This implies an equity objective of offering equal access to health care for members of the risk pool in equal need, regardless of their personal circumstances. In developing countries there are two reasons why the equity argument is particularly acute. First, the pattern of burden of disease (still predominantly communicable diseases) is closely related to poverty: the poor (those least able to pay) are the ones most in need to treatment. Secondly, low absolute levels of income mean that even modest financial contributions can lead to inability to seek treatment, or adverse consequences from seeking treatment (such as reduced expenditure on other essential items, or indebtedness).
As well as offending most people’s notion of fairness, an absence of risk pooling is likely to be inefficient. Most obviously, risk pooling transfers health care resources to the poor, who are at the margin likely to be able to benefit more from health care than the rich. Pooling therefore can lead to major improvements in population health. Such health gain is likely to be desirable in its own right. Moreover, with no pooling, poorer citizens who could benefit from health care (and may thereby become more economically productive) might languish untreated and become a burden on society. Pooling can reduce or eliminate a large degree of uncertainty associated with health care expenditure, thereby leading to widespread improvements in individual utility. Treating and preventing communicable diseases is likely to be particularly efficient, both because of the high cost-efficiency of many interventions and the high returns to society. There
6
are numerous positive externalities from increasing access to prevention and treatment, both in terms of limiting the spread of infectious diseases like tuberculosis, and in wider economic terms (such as improved returns on education and higher workforce productivity).
3 Approaches to Risk Pooling
The nature of the risk pooling arrangements is a matter of policy choice, which will be heavily influenced by a nation’s circumstances and its policy priorities. In western European countries risk pools are frequently entire regions or nations, reflecting the equity objective of securing universal coverage, often referred to as the solidarity principle. In the US risk pools are more heterogeneous, being based on factors such as age (Medicare), poverty (Medicaid), or employment. In spite of the acute equity and efficiency considerations noted...