Health Care and Insurance Schemes and Health Outcomes: A Global Comparative Analysis
Jessie Smith Nibley Doctors, Death Panels and Democracy Seminar Fall 2010 Professor Ruger January 18, 2011
Electronic copy available at: http://ssrn.com/abstract=1796342
Health Care and Insurance Schemes and Health Outcomes: A Global Comparative Analysis I. INTRODUCTION One of the largest and most urgent problems facing countries today—especially the United States—is health care. With an aging population and seemingly hard limits on resources such as doctors, organs, and medicine, each nation must attempt to distribute the available resources in a fair and efficient way without breaking the bank and, ideally, without selling out the health of its population. Nations’ attempts at creating workable health care systems result in regulatory and insurance schemes that vary widely across the globe, as countries form policies based on national values, institutions, and interests.1 Differing health outcomes among nations suggest that all health care delivery and insurance mechanisms are not equal. Some nations rely almost exclusively on the private sector for insurance and health care services, while others place health care and insurance squarely under government control, with wildly divergent results. These different approaches and vastly different health outcomes prompt the question: which is the best system? This paper attempts to compare the health care and insurance systems in twenty-four countries with the health outcomes achieved by those systems in order to answer the question: which health care delivery and insurance systems work best? Though it is difficult to determine exactly to what extent any given country’s health outcomes are influenced by its health policies, this paper assumes that health outcomes depend at least partly on the policy choices a nation makes for the provision of health care and insurance. And, although it may not result in a slam-
See Kieke G.H. Okma, et al., Six countries, six health reform models? Health care reform in Chile, Israel, Singapore, Switzerland, Taiwan and The Netherlands at 2.
Electronic copy available at: http://ssrn.com/abstract=1796342
dunk winner among the possible policies, this analysis will at least provide a crude comparison from which general observations may be extracted. Health policy experts have shied away from such a comparison in the past, acknowledging that “Data and methodological constraints preclude evaluation of the health outcome effects of specific policy interventions and/or health systems’ features on health outcomes.”2 This paper acknowledges the difficulty inherent in such an undertaking, but, rather than avoiding the problem altogether, aims simply to produce a rough analysis.
II. METHODOLOGY To compare different policies, it is necessary, first, to group countries with similar health care delivery and insurance mechanisms into broad categories, and, second, to compare the groups in terms of health outcomes.
A. Health System Categories The categories, taken from a publication of the Organisation for Economic Cooperation and Development (OECD), exist along a continuum from entirely private provision of insurance to entirely public, with variation in gate-keeping, budget constraints, and choice of providers.3 These categories are best represented in the following diagram, in substantially the same form in which it is found in the OECD analysis:4
George J. Schieber, et al., Health Care Systems in Twenty-Four Countries, Health Affairs 22, 34 (1991). Note that the twenty-four countries examined in Schieber’s article are not exactly the same as those studied here and were not chosen based on their inclusion in that article and that the subject of Schieber’s analysis is altogether different. 3 Organisation for Economic Co-operation and Development, Health Care Systems: Efficiency and Policy Settings, Executive Summary at 15 (2010). 4 Id. Note that the United States, which was left out of the OECD’s analysis due to lack of participation in its survey, has been added to the diagram, and all countries with developing or emerging economies have been omitted, as explained more fully in the following paragraphs.
Figure 1. Categorization of Countries.
Reliance on Market Mechanisms Mostly Public Provision and Public Insurance
private insurance for basic coverage
public insurance for basic coverage
no gate-keeping; ample choice of providers
private insurance beyond basic coverage; some gate-keeping
little private insurance beyond basic coverage; no gate-keeping
limited choice of providers; soft budget constraint
ample choice of providers; strict budget constraint
Germany Netherlands Switzerland United States
Australia Belgium Canada France
Austria Greece Japan Korea Luxembourg
Denmark Finland Portugal Spain
Ireland Italy New Zealand Norway United Kingdom
From left to right, these broad categories trend generally from more private delivery of health insurance and less government constraint to more public provision of health insurance and more government constraint. For ease of discussion, the groups will be referred to by number, 1 to 6, from left to right. The OECD explains the categories as follows: Group 1 countries rely almost exclusively on market mechanisms for regulation of insurance coverage and service provision.5 Group 2 and 3 countries have public basic insurance coverage and market mechanisms that regulate provision of services; they differ in their use of gate-keeping and the availability of private health insurance beyond the basic coverage.6 Group 4 countries provide users with choice among providers; the systems use gate-keeping, and the private supply of insurance is extremely
Organisation for Economic Co-operation and Development, Health Care Systems: Efficiency and Policy Settings, Executive Summary at 13–14 (2010). 6 Id. at 14.
limited.7 Group 5 and 6 countries have heavily regulated public systems; they differ in gatekeeping arrangements and the strictness of budget constraints.8
B. Health Outcomes The relative strength or weakness of each country group will be assessed according to health outcomes based on data provided by the OECD and the World Health Organization (WHO). To avoid at least some of the pitfalls of analyzing health outcomes, which may be influenced by a host of factors including financial resources, countries whose economies were classified as “emerging and developing” by the International Monetary Fund as of 2008 (the latest year for which data regarding health outcomes are available) will not be considered. 9 This decision is intended to prevent outlying data from nations with weaker economies from skewing the aggregate results of the analysis. The countries will be compared by group average (mean) on five specific, measurable health outcomes. The measurable health outcomes were chosen based on ease of comparison and availability of reliable data. The OECD and the WHO compile data each year on a variety of topics, including health, and these reports were the sources of the raw numbers used in this analysis. For each country, the most recent available values were used—nearly all of the data is from 2008, and no data is from earlier than 2006. The health outcomes to be compared are (1) life expectancy at birth; (2) infant mortality; (3) under-5 mortality; (4) adult mortality; and (5) deaths from cancer. A sixth value will be analyzed as well: total expenditure on health as a percentage of each country’s gross domestic product. This number does not measure health
Id. Id. 9 See International Monetary Fund, World Economic Outlook Database April 2010—WEO Groups and Aggregates Information, available at http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/groups.htm#oem; International Monetary Fund, World Economic Outlook Database – Changes to the Database, available at http://www.imf.org/external/pubs/ft/weo/data/changes.htm.
outcomes but will allow for more useful discussion of the data and possible avenues for health system reform. These metrics (aside from expenditures) all measure life or death—an easily determined indication of a health care system’s success, but a flawed one for many reasons. Though, on a fundamental level, the goal of health care is to keep people alive, there are many health outcomes besides life and death that might give a more complete or even more accurate picture of a policy’s success or failure. These might include overall quality of life, inequities in health outcomes, or overall health of a population based on incidence of various diseases such as diabetes or heart disease or rates of obesity. Including these health outcomes would likely create a fuller and more useful picture of the strengths and weaknesses of the systems being compared, but reliable measurements (or, in many cases, any measurements at all) are not available or not quantifiable enough to be used in this analysis.
III. DATA Presented below are the values, provided by the OECD and the WHO, for the five selected health outcome metrics and expenditure, along with the mean value, calculated specifically for this analysis, for each group and for each metric.
Figure 2. Raw data.
Country Life Infant Expectancy10 Mortality11 Under-5 Mortality12 Adult Mortality13 Cancer Deaths14 Expense15
Life expectancy at birth in years. All values in this column, except for group means, are taken from Life expectancy at birth, total population, HEALTH: KEY TABLES FROM OECD, 2010, available at http://www.oecdilibrary.org/social-issues-migration-health/life-expectancy-at-birth-total-population_20758480-table8. 11 Deaths per 1,000 live births. All values in this column, except for group means, are taken from Infant Mortality, HEALTH: KEY T ABLES FROM OECD, 2010, available at http://www.oecd-ilibrary.org/social-issues-migrationhealth/infant-mortality_20758480-table9.
Germany Netherlands Switzerland United States Group 1 Mean Australia Belgium Canada France Group 2 Mean Austria Greece Japan Korea Luxembourg Group 3 Mean Iceland Sweden Group 4 Mean Denmark Finland Portugal Spain Group 5 Mean Ireland Italy New Zealand Norway UK
80.2 80.3 82.2 77.9 80.15 81.5 79.8 80.7 81.2 80.8 80.5 80 82.7 79.9 80.6 80.74 81.3 81.2 81.25 78.8 79.9 79.3 81.2 79.8 79.9 81.5 80.4 80.6 79.7
3.5 3.8 4 6.7 4.5 4.1 3.4 5.1 3.8 4.1 3.7 2.7 2.6 4.1 1.8 2.98 2.5 2.5 2.5 4 2.6 3.3 3.5 3.35 3.1 3.7 4.9 2.7 4.7
4 5 5 8 5.5 5 5 6 4 5 4 3 3 5 3 3.6 3 3 3 4 3 4 4 3.75 5 4 6 3 6
78 68 60 107 78.25 63 86 70 87 76.5 75 75 65 76 79 74 56 62 59 90 94 90 72 86.5 73 61 72 67 78
156.6 175.6 138.7 157.9 157.2 154.3 169 162.6 161.97 150.3 150 138.4 153.5 153.9 149.22 161.6 146.8 154.2 199.4 132.6 151.3 151.2 158.63 172.9 157.6 164.2 154.8 170.7
10.4 8.9 10.8 15.7 11.45 8.9 9.4 10.1 11 9.85 10.1 9.6 8 6.3 7.1 8.22 9.3 9.1 9.2 9.8 8.2 10 8.5 9.125 7.6 8.7 9 8.9 8.4
Deaths by age 5 per 1,000 live births. All values in this column, except for group means, are taken from World Health Organization, World Health Statistics 2010, at pp. 48–55. 13 Probability of dying between 15 and 60 years per 1,000 population. All values in this column, except for group means, are taken from World Health Organization, World Health Statistics 2010, at pp. 48–55. 14 Deaths from cancer per 100,000 population. All values in this column, except for group means, are taken from Deaths from Cancer, HEALTH: KEY T ABLES FROM OECD, 2010, available at http://www.oecd-ilibrary.org/socialissues-migration-health/deaths-from-cancer-total-population_20758480-table13. 15 Total expenditure on health as percentage of GDP. All values in this column, except for group means, are taken from World Health Organization, World Health Statistics 2010, at pp. 130–137.
Group 6 Mean
These data are most illustratively presented by individual health outcome metric. With this in mind, following are graphical representations of the mean values for each of the five health outcomes for side-by-side comparison of the country groups. Obviously, a higher value is better for life expectancy, while a lower value is better for all other metrics.
Figure 3. Life expectancy comparison.
81.5 81 80.5 80 79.5 79 1 2 3 4 5 6 80.15 79.8 80.8 80.74 80.42 81.25
Figure 4. Infant mortality comparison.
(deaths per 1,000 live births)
5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 4.5 4.1 3.35 2.98 2.5 3.82
Figure 5. Under-5 mortality comparison.
(deaths by age 5 per 1,000 live births)
6 5 4 3 2 1 0 1 2 3 4 5 6 3.6 3 3.75 5.5 5 4.8
Figure 6. Adult mortality comparison.
(probability of dying between 15 and 60 years per 1,000 population)
100 80 60 40 20 0 1 2 3 4 5 6 78.25 76.5 74 59 86.5 70.2
Figure 7. Cancer deaths comparison.
(per 100,000 population)
170 165 160 155 150 145 140 1 2 3 4 5 6 149.22 157.2 154.2 164.04 161.9666667 158.625
Figure 8. Expenditure on health comparison.
Expenditure on Health
(as percentage of GDP)
14 12 10 8 6 4 2 0 1 2 3 4 5 6 11.45 9.85 8.22 9.2 9.125 8.52
IV. ANALYSIS The comparison charts above must be analyzed with an understanding of the limits of this study and the problematic nature of the issue itself. The comparative analysis is hindered by the inescapably small sample size, the imperfect categorization of countries, and the impossibility of eliminating factors other than delivery and insurance constraints that may have important effects on health outcomes. Sample sizes for analyses such as this are limited to the number of developed countries for which health outcome data is available. Certainly, a larger sample of countries in each category would produce more complete and interesting results, but more countries simply do not exist. The categories created by the OECD are insightful, and this analysis was in part spurred by the possibility of divergent health outcomes resulting from the specific characteristics identified by the OECD; however, it is of course true that there are many other possible ways of categorizing even the countries that were included in this study, and the “science” of placing nations in discrete boxes is an imperfect one. Lastly, it seems clear that an
array of factors may play into the health effects achieved by a country’s insurance and delivery systems: financial resources, genetic predisposition to certain maladies, aggregate lifestyle of a population, and so on. While the scope of this analysis does not include accounting for each of these other factors, lifestyle does not have as great an impact on infant and under-5 mortality as on other measures of health system success, and the study at least leveled the financial playing field among countries somewhat, omitting data from those countries with weaker economies. It should be noted, as well, that more money does not necessarily mean better health outcomes, as evidenced by the relatively high amount spent by the United States16 for relatively poor health outcomes.17 Even with these limits in mind, the side-by-side comparisons reveal at least three striking trends. First, there is a rough bell- or upside-down-bell-curve in most of the charts—countries in the outer-most groups generally did poorly, and countries in the center groups generally performed well. In terms of the country groups’ characteristics, extremes in reliance on public or private delivery and insurance appear to be bad for health outcomes, and moderation or a mix of public and private regulation appears to be good for health outcomes. Second, Groups 3 and 4, representing countries with no gate-keeping, performed consistently well across all categories, often the best of all the groups. This trend suggests that gate-keeping may lead to relatively poor health outcomes and may be a better explanation of the curve than the extreme/moderate attributes of the groups. Finally, the high expenditure averages in Groups 1 and 2 suggest that
See Figure 2, supra. See also Kaiser Family Foundation, Trends in Health Care Costs and Spending, March 2009 at 1; I. Joumard et al., “Health Status Determinants: Lifestyle, Environment, Health Care Resources and Efficiency,” OECD Economics Department Working Papers, No. 627 (OECD Publishing 2008), at 71 (noting that “costs efficiency is low in the United States”). 17 See Karen Davis et al., Mirror, Mirror on the Wall: How the Performance of the U.S. Health Care System Compares Internationally, 2010 Update at 3 (“Overall, the U.S. ranks last or next-to-last on all five dimensions of a high performance health system”).
systems based mainly on market regulation with private insurance are more expensive than mainly public systems. Each of these trends is discussed in greater detail below.
A. Moderation The most obvious characteristic of the group comparisons is the consistently better outcomes for those countries in the middle two groups—Groups 3 and 4. In no case did Group 1, 5, or 6—the most extreme approaches—achieve the best health outcome. One possible explanation for this disparity is that too much reliance on market mechanisms or too much government constraint leads to poorer health than some kind of middle ground. For Groups 3 and 4, the market influence appears to be on the provider side rather than the insurance side, as both groups rely on a public system of insurance at least for basic coverage. It is important to understand why Group 4 outperforms the others on the “mostly public” side of the spectrum and why Group 3 outperforms the others on the “mostly private” side. The difference between Group 4 and Groups 5 and 6 is in the amount of government control beyond simply providing insurance: Group 4 allows users ample choice of providers and has no obligatory or financially encouraged gate-keeping; Groups 5 and 6 go further, restricting access by gate-keeping as well as budget constraints and, in the case of Group 5, limiting the user’s choice of providers. On the other side of the spectrum, Group 3 differs from Groups 1 and 2 in that countries in Groups 1 and 2 allow consumers to obtain private insurance, either supplementary to a basic plan, as with Group 2, or to cover all expenditures, as with Group 1. Gate-keeping appears again as another key difference between Group 3 and Groups 1 and 2, as described more fully below.
The high performance of Groups 3 and 4, which split regulatory responsibility between public and private actors, may mean that exclusive control by either the public or the private sector has a negative impact on the health outcomes of a system. On the other hand, the fact that Groups 3 and 4 are also the only countries with no gate-keeping cannot be overlooked. A thorough discussion of the gate-keeping factor may be found in the next subsection. If a country does choose an end of the spectrum, the data indicate that it is better to err on the side of more public control rather than more reliance on market mechanisms. Group 1, with the most private provision and insurance, fared worse than Group 5 or 6 (or both), with the least private insurance, in every health outcome besides cancer deaths. It should be noted, however, that, as can be seen more clearly from the raw data in Figure 1, the health outcome values for the United States worsened the Group 1 averages considerably. With the United States data removed from the averages, the comparisons tend to look more like this: Figure 9. Infant mortality comparison with the United States omitted.
4.5 4 3.766666667 3.5 3 2.5 2 1.5 1 0.5 0 1 2 3 4 5 6 3.35 2.98 2.5 4.1 3.82
The resulting shape is not the obvious curve it was, but the fact remains that Groups 3 and 4 still outperform the other groups, and Groups 5 and/or 6, on the public extreme, still tend to outperform Group 1, on the private extreme.
B. Gate-Keeping Gate-keeping is a widely used cost-control measure.18 In a gate-keeper system, patients must see a general practitioner and get a referral before seeing a specialist.19 The idea is that a doctor is a better judge of when a patient needs further specialized treatment than the patient himself. 20 The gate-keeping physician—a lower-cost general practitioner—stands as a barrier for patients who may otherwise rush out and consume the higher-cost services of a specialized doctor. This method of cutting costs is somewhat controversial; gate-keepers themselves complain about the pressure to keep costs low by reducing referrals and the inherent inconsistency between that pressure and their desire as doctors to make decisions about their patients’ care without regard for economic considerations.21 The countries that employ “some gate-keeping” may use financial incentives to encourage patients to get referrals, while those with no gate-keeping offer no incentives and do not make referrals a requirement for seeking specialized care.22 The rough comparisons seen in Figures 3–7 above appear to support these physicians’ concerns; in almost every case, those countries without gate-keeping measures achieve better health outcomes than those with gate-keeping. In fact, Groups 5 and 6—those with the strictest
JESSICA R. ADOLINO & CHARLES H. BLAKE, COMPARING PUBLIC POLICIES: ISSUES AND CHOICES IN SIX INDUSTRIALIZED COUNTRIES at 212 (CQ Press 2010). 19 Id. 20 Id. 21 Id. 22 V. Paris et al., “Health Systems Institutional Characteristics: A Survey of 29 OECD Countries,” OECD Health Working Papers, No. 50, (OECD Publishing 2010) at 42.
gate-keeping measures—perform worse than Group 2—systems with only some gate-keeping— in almost every health outcome. However, the difference in performance for these groups could also be due to the level of public control over the health care system; Groups 5 and 6 have mostly public insurance and provision of services, Group 2 has private insurance available beyond the basic coverage. This theory is explored in the previous subsection. Another interesting contrast is that between Group 2 and Group 3. One might expect these two categories to perform similarly, given that both encompass countries with public insurance for basic coverage. Indeed, intuition may suggest that a system (like that of Group 2) that allows for more private insurance beyond the basic coverage would lead to better health outcomes than one (like Group 3) that is more restrictive of supplementary private coverage. The contrary results may be an indication of the powerful negative effect of gate-keeping. The gate-keeping attribute could not be included in the description for Group 1, because the defining characteristic for that group is exclusive public provision of insurance, and the countries differ in their use of gate-keeping. Supplementary data, however, show that all of the countries in this group employ at least some gate-keeping. 23 Considering Group 1’s poor showing in nearly every metric, this fact lends further support to the notion that gate-keeping is harmful to a population’s health.
C. Expenditure Any country designing a system for health care must take cost into consideration. A system that allows each person unlimited health care, free of charge, would certainly achieve the best health outcomes; with no rationing, there would be far fewer preventable deaths.
See V. Paris et al., “Health Systems Institutional Characteristics: A Survey of 29 OECD Countries,” OECD Health Working Papers, No. 50, (OECD Publishing 2010) at 42.
Unfortunately, such a system is not currently viable anywhere in the world, which makes the expenditure comparisons noteworthy. As with health outcomes, Groups 3 and 4 perform well, and Groups 1 and 2 perform poorly. Groups 5 and 6, notably, also beat 1 and 2. Again, it should be noted that the United States skews the Group 1 results somewhat. However, even with the United States’ data omitted, Group 1 still performs worst: Figure 10. Expenditure on health with the United States omitted.
Expenditure on Health
12 10 8 6 4 2 0 1 2 3 4 5 6
The high average for Group 1 cannot be blamed only on the United States; it is a result of the consistent high costs seen across Group 1 countries. It seems unlikely to be mere coincidence that the three most expensive systems of the twenty-four—the United States, Switzerland, and Germany—all fall in Group 1. This observation could have important implications for reform. If systems that rely so exclusively on market mechanisms for insurance and care are consistently the most expensive— and, as we have seen, generally perform poorly—then reform should tend toward increasing government involvement in health care. The low costs of Group 5 and 6 systems support this idea.
Of equal importance is the status of Group 3 as the lowest-cost system of insurance and health care delivery. As described in the previous and following subsections, Group 3 is a top performer across health outcome metrics. Such success in outcomes is striking on its own, but when coupled with the extremely low costs of the Group 3 system, it presents interesting possibilities for cost-effective reform. Even acknowledging the rudimentary nature of this analysis, Group 3 countries appear to be doing something right, at least. At first glance, public insurance with no gate-keeping—the key characteristics of Group 3—might seem to be a budgetbalancer’s nightmare. Somehow, though, these countries manage to keep costs down, and with overall excellent results in health outcomes. This is in slight contrast to Groups 5 and 6, which also have low costs but do not tend to succeed as much in health outcomes as Group 3.
D. Other Observations This analysis suffers from a number of data and methodological shortcomings, but one in particular is that Group 4 includes only two countries—a relatively small sample size. The original groupings by the OECD put Turkey in Group 4 as well, but Turkey was omitted from this study due to its status as a developing nation. So any conclusions about the virtues of a Group 4 approach could mean either that Group 4 has some uniquely beneficial attributes or simply that Iceland and Sweden have uniquely beneficial attributes. Indeed, the Nordic countries, including Iceland and Sweden, are often lumped together because they share similar traits in a number of social policy areas, including welfare, education, social services, and, of course, health care.24 For this reason, the fact that Group 4 performs well is not as compelling as the consistent good performance of Group 3. The larger sample size of five in Group 3 and the vast cultural,
MARTTI LUJANEN, HOUSING AND HOUSING POLICIES IN THE NORDIC COUNTRIES (2004).
political, and institutional differences between the countries in the group (which also happen to be widely dispersed geographically) seem to refute the notion that the group performed well not because of health policy but because of other shared attributes. Of course, this latter explanation is still viable, but it is less likely here than with Iceland and Sweden—countries that share a vast array of characteristics that may impact public health. Looking at the raw data, one can see that the values for Group 3 are, in general, more tightly clustered than those for other groups (except Group 4, containing only Iceland and Sweden). This holds true even for cancer deaths, 25 the category in which there was the most divergence among countries within each group. Such tight clustering, coupled with the diversity among the countries in the group, tends to support the conclusion that the characteristics of the Group 3 system—public insurance for basic coverage, little private insurance beyond that, and no gate-keeping—have some special and positive effect on health outcomes.
V. CONCLUSION Although, as explained above, this study provides only a basic and incomplete look at health care systems and the outcomes they achieve, the research suggests several opportunities for health care reform as well as new possibilities for further and better research.
A. Avenues for Reform The success of Group 3 is of particular interest in considering potential reforms. Every country wants to keep its health care costs down without sacrificing quality, and these Group 3 countries seem to have found a way to achieve that goal. A Group 3 health system may serve as a model for other nations struggling with rising costs—the United States in particular.
See Figure 7, supra.
The possibility that gate-keeping may have a deleterious effect on health outcomes cannot be ignored in implementing reform. Countries may want to try looking to other costcontainment measures to avoid the possible harm to their populations’ health from gate-keeping. Some nations might balk at the prospect of eliminating one of the most widely used cost-cutting measures, but this analysis indicates that better health outcomes can be achieved without gatekeeping and at lower or equal cost. Reformers may also note the possibility that a middle-of-the-road approach between public and private domination of health care could achieve the most success. And, if a middle ground between public and private provision and insurance is not achievable, this research indicates that a country should favor public provision rather than private, both for better health outcomes and lower cost.
B. Avenues for Further Analysis As data become more readily available for a greater number of countries, the possibilities for research on the relative benefits of particular health system characteristics will grow. Even with the currently available information, however, further study is possible and may provide valuable insights into what works and what doesn’t, what is worth spending money on and what isn’t. As this study progressed, it became clear that the results would change drastically depending on how countries were categorized. If one were to make a comprehensive list of attributes of each nation’s health care systems and then perform a series of similar data comparison on groups of countries for each attribute identified, the results could be very illuminating.
For example, an individual study could categorize countries according to insurance mechanism—employer-based, single payer, health savings account, etc.—and prepare the sideby-side comparisons again. Other possibilities include categorizing countries solely by costcontainment measures, including gate-keeping, doctor or patient incentives, cost-sharing, selective contracting, etc.; by compensation model—fee-for-service, bundled payment, or capitation; or by medical degree or licensing requirements, to name just a few. A robust study analyzing the health outcomes resulting from all of these different metrics individually may reveal which attributes tend to have an impact on outcomes and which do not. Perhaps the most interesting prospect for further research is in trying to understand what makes Group 3 such a success. Reform obviously aims at better health outcomes and lower costs, and Group 3 appears to have found a formula that achieves both. The limits of this study are explained in great detail above; future research might go beyond what was attempted here with particular focus on Group 3. A future study could attempt to isolate the attribute or attributes that are the source of the Group 3 system’s solid performance, or it could seek out other attributes shared by these countries.