■ Introduction
1OECD countries are spending a considerable amount of money on their health systems. In 2003, the latest year available for international comparisons, health spending reached a peak of 8.8%, on average, of their gross domestic product (GDP). In the United States, health spending reached 15.0% of GDP and in France, spending as a share of the economy was above the OECD average at 10.1%.
2A combination of factors lies behind the continued growth in health spending. Advances in medical technologies, population ageing and rising public expectations are among them. Over the past decade, health spending has outpaced economic growth. This is in part the result of economic slowdown, especially in the early 2000s. However, health spending growth was exceeding economic growth even before the slowdown in the economy. During the 1990s, the difference between health spending growth and economic growth was over one percentage point on average among OECD countries. And, in 2002, health spending resumed its fast growth in several OECD countries, giving rise to policy concerns about the sustainability of health systems financing. The fact that the large majority of health spending comes from public sources in most OECD countries has increased pressures on already stretched public budgets.
3The question of the right or appropriate level of health care spending does not however have easy answers at hand. Over the past two decades, OECD countries have implemented a variety of measures to contain costs and improve value for money in health systems. These have included, for example, downscaling of the hospital sector and decreasing the number of health personnel, reducing households’ entitlement to publicly financed goods and services, and increasing cost sharing on, for example, prescription drugs and dental care. However, strong control on public spending on health might impact on other policy goals, including shortages in nurses and doctors and prolonged waiting times for elective surgery.
This article [1] presents data on health care spending for the 30 OECD countries using OECD Health Data [2] as its source. It draws attention to the main trends and reviews cross-country differences. The article first discusses the contribution of health care expenditure to countries’ gross domestic product, showing how increasing health expenditure and the slowdown in economic growth have contributed to the growing share of health in the economy. The article then analyses how health care expenditure is financed, and finally it reviews how it is spent on different types of medical goods and services. It should be noted that it focuses solely on examining the level and differences in health spending and does not discuss the important links between health status and health expenditure.
■ Health expenditure represents a sizable and growing share of the economy
4OECD health systems accounted for an average of 8.8% of GDP in 2003. There are, however, substantial variations across countries. The United States spent as much as 15% of its GDP on health, followed by Switzerland with 11.5% and Germany with 11.1%. Conversely, countries such as Slovakia, Poland, Mexico and Korea [3] spent less than half the level of the United States on health. French spending on the health system was above the OECD average at 10.1% (Figure 1).
Health expenditure as a percentage of GDP, 2003
Health expenditure as a percentage of GDP, 2003
5The share of GDP devoted to health has increased over the past few decades. Considering the 29 countries for which comparable historical series are available, this share increased from 7.1% in 1990 to 8.8% in 2003 (Table 1). The United States has consistently been the largest spender since 1980. France has also been above the OECD average over the same period. The increase in spending as a share of GDP for France has generally been in line with the OECD average, compared to a few other OECD countries, such as the Czech Republic, Turkey and Portugal, which have seen more significant change (Figure 1, right).
Total expenditure on health, share of gross domestic product (GDP), 1970 to 2003
Total expenditure on health, share of gross domestic product (GDP), 1970 to 2003
6The growing importance of health spending in the economy of OECD members is in part a consequence of the slower pace of GDP growth compared to health spending growth. The difference in such growth rates has been remarkable in certain periods. Between 1990 and 2002, on average, across 29 OECD countries, annual per capita real [4] health expenditure grew more than two percentage points faster than the growth in real income per capita (Table 2). In 2003, health care spending growth continued to outpace economic growth in many OECD countries. In the United States, for example, the share of health expenditure in GDP increased from 14.6% to 15% of GDP – as health expenditure grew more than twice as fast as GDP growth. In 2003, in France, health spending grew at 4.6% in comparison to the whole economy which saw zero growth.
Growth of per-capita health expenditure compared to GDP growth, 1990 to 2003
7However, the rate of health spending growth has not been uniform over the years (Table 3). For example, the decade of the 1990s can be broken down into three sub-periods:
8– Between 1990 and 1992, health spending in OECD countries grew rapidly.
9– Mounting concerns over such fast growth resulted in several OECD countries implementing reforms and measures to cut health expenditure or reduce the pace of growth (Docteur and Oxley, 2003; Mossialos and Le Grand, 1999). This led to a slowdown in the rate of growth of health spending, which was 2.6% between 1992 and 1997, on average.
10– Health expenditure started to rise rapidly again at the end of the 1990s, and this continued into the new millennium.
Average real growth of per capita health expenditure, 1980 – 2003
11In the United States, while insurers’ efforts to manage care had met with some success during the 1990s, a combination of consumers’ and providers’ backslash against the most restrictive forms of managed care has led insurers to loosen their interventions in this area with a resulting renewal in rapid cost growth. In some countries, such as the United Kingdom, Canada and Ireland, health spending growth was the outcome of policies specifically aimed to increase the level of public spending on health. Tight budgetary controls implemented in these countries during previous years had brought about constraints in capacity and surgical activity. Preoccupations over unpopular waiting times for elective surgery have encouraged policy makers in the United Kingdom to release the tight grip on expenditure growth (Hurst and Siciliani, 2003).
12Moreover, health expenditure growth has not been uniform across countries (Figure 2).
Growth of health expenditure relative to gross domestic product, 1980 – 2003
Growth of health expenditure relative to gross domestic product, 1980 – 2003
13For example, over the past two decades it has been faster than the OECD average in some of the countries with lower levels of income and health expenditure per capita, such as Korea, Ireland and Portugal. [5] Growth rates have also been above average in some high-income countries, for example in Australia and the United Kingdom.
14On the other hand, Germany, Switzerland and France had seen growth rates of health spending below the OECD average, yet health spending remained above GDP growth rates, resulting in an increase of health spending as a share of GDP over the years.
15Conversely, fast GDP growth in Finland, Canada, and Italy brought about a stabilisation, or a reduction in the share of GDP devoted to health.
16Some countries, whose share of health expenditure in GDP was relatively low in 1990, experienced high growth rates of health expenditure compared to GDP. Yet this does not mean that countries with a higher share of health spending in GDP have necessarily experienced slower growth in health spending (Figure 3).
17Germany, Switzerland, and the United States had both high health expenditure as a share of GDP in 1990 and fast growth in health spending relative to GDP. This suggests that a high share of health spending in GDP need not necessarily result in slower growth in total health expenditure.
■ Per capita health spending in OECD countries differs to large extent
18The United States also ranks first in per capita health spending. In 2003, the United States spent 5,635 US$ PPP per capita on health care, compared to less than 1,000 US$ PPP for lower spending countries such as Mexico, Korea and Eastern European countries (Figure 4). The United States spends more than double the OECD average of 2,394 US$PPP, [6] compared with about 70% more than the OECD average using the percentage of GDP as an indicator, at 2,903 US$ PPP.
Breakdown of per capita health spending into public and private expenditure, 2003
Breakdown of per capita health spending into public and private expenditure, 2003
19France spends around 20% above the OECD average. There is also wide variation in the breakdown of per capita spending on health into public and private shares across the OECD, although public spending largely exceeds private spending for the majority of countries.
20The level of economic development explains in part the differences in per capita spending across countries. However, the positive association between income per capita and health expenditure per capita, shown in Figure 5, is weaker when clusters of countries with different levels of economic development are considered. Income is surely an important factor behind health expenditure growth, particularly when considering the variation in health spending across countries at different stages of development. However, there is no single, most appropriate level of health spending that fits all countries, especially when considering those at higher stages of development. While richer countries tend to spend more on health, there is considerable variation in spending among countries with comparable incomes. Such variation, and the fact that the income-spending relationship is weaker for richer countries, also suggests that income growth need not necessarily result in spending growth. Other factors, including policy choices about health systems’ structure and their regulation are likely to impact upon health expenditure levels (OECD, 2004a).
Per capita expenditure on health and income per capita, 2003
21Similarly, there are variations in per capita spending on health for countries which have similar shares of their GDP devoted to health (Figure 6).
Total health expenditure per capita and as a percentage of GDP, 2003
Total health expenditure per capita and as a percentage of GDP, 2003
22For example, both Greece and France spent around 10% of GDP on health in 2003, but health expenditure per capita in Greece was more than 30% lower than France. Variations in spending levels are due to a combination of cross-country differences in prices and volumes of health care. For example, while spending in the United States is substantially higher than any other OECD countries, it provides comparable volumes of care to other OECD countries (Docteur et al., 2003). A combination of factors explains higher spending in the United States, ranging from higher administrative costs linked to its fragmented financing system, to greater market power of providers over third-party payers than occurs in other OECD systems, particularly those where governments exercise monopsonic or collective bargaining power over providers (Reinhardt et al., 2004). Reliance on competition among insurers and providers in the United States also results in physicians trying to attract patients through intensity of treatment, amenities and other aspects of quality, while hospitals also attract patients and physicians by similar mechanisms (McClellan et al., 2002).
■ Public funding continues to be the main source of financing health expenditure
23Health expenditure can be financed through different combinations of public sources (general government revenues and social health insurance) and private sources (out-of-pocket payments and private health insurance). Private health insurance includes both employer sponsored schemes and individually purchased policies. Out-of-pocket payments include both direct household payments for purchasing medical goods and services as well as cost sharing on coverage schemes. Out-of-pocket payments represent health funding that is not reimbursed by either a public or a private thirdparty programme.
24OECD governments have adopted three different approaches to ensure broad population coverage of health expenditure.
25• A first group of countries has achieved universal or near universal cover through national public health coverage, be it a tax-funded system or a social health insurance programme (e.g., the Nordic, Mediterranean – including Italy – and Eastern European countries, Canada, Australia, New Zealand, Korea, and Japan).
26• A second group of countries has promoted basic public coverage only for specific population groups, leaving others with the option to buy private health insurance, self-insure or be uninsured (e.g., the Netherlands and Germany). In the case of the United States, only those who are poor
27or elderly are entitled to public coverage, while the majority of the population is reliant upon private funding sources.
28• A third approach, represented by Switzerland, [7] has been to ensure universal coverage by compelling the entire population to buy basic health insurance. Only three OECD countries have not attained universal health insurance: the United States, where about 14% of the population was uninsured in 2001; Mexico, where about half of the population is not part of the social security system; and Turkey, where public coverage was only available for two thirds of the population by the end of the 1990s (OECD, 2004b).
29The mix of schemes used to promote population coverage affects funding patterns of health systems. Yet even in countries which have achieved universal coverage there can be significant funding from private sources, depending on how large are service gaps within public systems, the importance of cost sharing, and the role and size of private health insurance markets.
30For example, Korea has achieved universal national health coverage through a social health insurance system, yet private sources account for nearly half of total health financing (OECD, 2003c).
31On the other hand, while only about one quarter of the US population is insured through public programmes, the United States has one of the highest levels of public spending per capita in the OECD area, and public spending as a share of GDP is comparable to other OECD countries with larger share of public financing (Docteur et al., 2003).
32Public financing is the dominant form of health financing in all OECD countries apart from Korea, Mexico and the United States (Figure 7). It accounted, on average, for 72% of total health care expenditure across OECD in 2003. [8] The public share was more than 80% in several countries, including the Czech Republic and Slovakia, Northern European countries, Japan, and the United Kingdom. France was positioned slightly above the OECD average, with a public share of 76% of total health expenditure.
Private, out-of pocket payments represented 19.4% of total health expenditure in 2002, [9] or about three quarters of total private financing. (In Figures 7 and 8, OECD countries are ranked by increasing share of out-ofpocket spending). Cross-country variability is considerable. Out-of-pocket
Health expenditure by source of funding, 2003
Health expenditure by source of funding, 2003
33payments account for more than 30% of total health expenditure in Switzerland, Korea and Mexico (and around half of total financing for the latter two countries). For most other countries with available data, it varies between 10% and 30% of total health expenditure. In France, household expenditure represents 10% of total health finances. With exceptions on both ends of the spectrum, countries with lower per capita income have a larger share of out-of-pocket spending, and vice versa. Out-of-pocket payments finance diverse health goods and services across OECD countries. Pharmaceuticals are one major component in most countries, while other components include, with variation across countries, dental care, longterm care, and outpatient services.
34Out-of-pocket payments represent the most regressive form of financing health care costs, and can pose access barriers to low income groups, particularly when they represent a significant portion of households’ disposable income. The incidence of out-of-pocket spending in total household consumption increased in most OECD countries over the past decade. Only in a few countries – Denmark, France, Korea and the United States – did the share remain constant or even decrease (Huber and Orosz, 2003).
35Private health insurance (PHI) accounts for only 6.1% of total health funds [10] on average and about a quarter of private health funds (Figure 8). Its relative importance in financing total health expenditure varies between the high of 36% in the United States and the zero of several OECD countries with no PHI markets. There is no link between the importance of PHI in financing total health expenditure and the level of economic development of a country.
36Cross-country variation in the share of private health insurance in total health spending is due both to differences in the size of the population covered by private policies and to the heterogeneity of benefits provided by PHI insurers across the OECD area:
37• PHI represents the sole form of cover for a large fraction of the population in the United States, where public programmes cover only the poor (Medicaid) and the elderly (Medicare).
• PHI plays a similar role for smaller population groups in the Netherlands (where one third of the population – those with upper income levels – is not eligible to social health insurance) and in Germany (where a tenth of the population opts out of sickness funds insurance).
Breakdown of public financing, 2002
Breakdown of public financing, 2002
38• PHI also remains a sizable funding source (above 10% of total health expenditure) in Canada, where it pays for prescription drugs and dental care which are not publicly reimbursed, and France, where it covers cost sharing in the social security system.
39• PHI also duplicates universal coverage in Australia, Ireland, New Zealand, Italy and the United Kingdom, among others, thereby offering a parallel, private alternative to universal public systems (Colombo and Tapay, 2004; OECD, 2004b).
The structure of health financing has been by and large stable over time. Nonetheless, there have been some fluctuations; for example, public health financing increased during the 1970s, then stabilised and declined slightly during the 1990s. In France, the public share of health expenditure has remained stable at around 76 – 77%. In general, there seems to be some convergence in the share of public spending across the OECD area over time. Countries whose level of public spending on health was above the OECD average saw a decline in the public share over time, and the opposite pattern occurred for countries with public spending below the OECD average (Figure 9). Out-of-pocket payments as a share of total health expenditure have slightly increased over time.
Changes in the public share of health expenditure, 1990-2003
Changes in the public share of health expenditure, 1990-2003
■ Spending on health by type of care: the growing role of pharmaceuticals
40OECD countries differ substantially in their levels of spending for different types of care (Figure 10). For the 24 countries for which data are available, inpatient care represented an average of 37% of current health expenditure in 2003, compared to 34% for ambulatory care and 22% for medical goods. [11] While prevention and public health are increasingly considered key to improvements in health outcomes, OECD countries do not allocate many resources to keep people from becoming ill. On average, OECD countries spent only 2.9% of total health expenditure on organised public and private prevention programmes (Figure 11). France spent around 2.5% of total current health expenditure in 2003. [12]
Current expenditure by medical services, medical goods and collective care, 2003
Current expenditure by medical services, medical goods and collective care, 2003
Expenditure on prevention and public health as a share of current health expenditure, 2003
41Inpatient care represents over 50% of total current health expenditure in Iceland and Denmark, and nearly half in Switzerland. It accounts for less than 30% in Canada, the United States and Spain.
42Medical goods and services are over a quarter of total health expenditure in some Eastern European countries, for example in Slovakia (46%), Hungary (33%) and the Czech Republic (27%). The United States spends only about 14% of total health expenditures on medical goods. Despite its low share, the absolute level of per-capita spending on pharmaceuticals in the United States is the largest in the OECD area (Figure 12)
Per capita expenditure on pharmaceuticals and share of total health expenditure, 2003
Per capita expenditure on pharmaceuticals and share of total health expenditure, 2003
43In 2003, France spent 38% of its current health spending on inpatient care, 28% on ambulatory care, and 25% on medical goods and services, not significantly different from the OECD average.
44Several reasons explain variation in the distribution of spending by type of care. These include, among others, innovation in medical technology, government reforms to improve efficiency and changes in the way resources are allocated – including for example payment reforms (Docteur and Oxley, 2003). Cross-country differences in prices for different factors of production also influence such variation (Huber and Orosz, 2003).
45Pharmaceuticals represent a particularly important – and rapidly growing – area of health care spending. In 2003, pharmaceutical expenditure accounted for 17.5% of total health spending for the OECD area. [13] Yet variation is substantial, as this item accounts for less than 10% of total health spending in Denmark and 38.5% in Slovakia. In France, it was above the OECD average, standing at 20.9%.
46Variation between countries in pharmaceutical spending also reflects large variations in unit prices, and volumes prescribed and consumed. The United States is the largest spender. In 2003, this country spent as much as 728 US$ PPP per capita on pharmaceuticals. France is the second highest spender on this item in the OECD area, with per capita 606 US$ PPP in 2003, followed by Canada and Italy. There is not necessarily a link between levels of per capita spending and the importance of pharmaceuticals in total health expenditure. In fact, the share spent on pharmaceuticals can vary greatly for countries with similar health spending per capita.
For example, Denmark and France have roughly the same spending per capita on health. Yet, Denmark spent 10% of total expenditure on pharmaceuticals in 2003, while France spent 21%.
Growth in pharmaceutical spending has been notable in the past 10 years. Considering the period between 1997 and 2003 (Figure 13), spending on pharmaceuticals grew, on average, 27% faster per year than total health expenditure across the OECD, and 76% faster in France. This has resulted in an increasingly large share of pharmaceuticals in total health spending over the period. For example, drugs expenditure rose from 18% to 21% of total health expenditure in France. Several reasons explain such fast growth rate, ranging from the production and adoption of new drugs, particularly expensive ones, and the substitution of certain surgical treatments with drugs.
Real annual growth in pharmaceutical spending and total health expenditure, 1997 – 2003
Real annual growth in pharmaceutical spending and total health expenditure, 1997 – 2003
47Financing of pharmaceuticals comes to a larger extent from private sources than is the case for other types of health services. For example, in 2003, private expenditure on pharmaceuticals represented 79% of total pharmaceutical spending in the United States, 62% in Canada and 33% in France. On average, public expenditure in OECD countries represented around 60% of total spending on pharmaceuticals in 2003. By way of contrast, for those countries where this information is available, spending on inpatient care was, on average, 85% publicly financed. This difference is due both to higher levels of cost sharing for pharmaceuticals than for other categories of care, and to the fact that a significant portion of spending on pharmaceuticals is not covered under public insurance schemes. OECD countries have intervened in different ways to contain the fast rate of growth in pharmaceutical spending. Tools adopted range from price controls to measures intended to control the demand for pharmaceutics or influence prescription patterns. Reference prices combine measures of the first and of the second type: they are based on higher patient cost sharing for individuals using high-cost products rather than less expensive generic drugs. Other measures include restricting reimbursement to drugs of proven therapeutic values and delisting other comfort drugs. Several OECD nations have increased cost sharing, as has been the case of Italy, and introduced tight controls over the type and number of drugs publicly reimbursed (Mossialos and Le Grand, 1999; Jacobzone, 2000; Docteur and Oxley, 2003).
■ Conclusions
48Health care spending is rising, and the high cost of health systems is once again on the policy makers’ agenda. In France, health spending as a share of GDP is above the OECD average, and expenditure growth has been faster than GDP growth over the past decade. Yet cost-containment policies, as implemented in several OECD countries during the 1990s, are not necessarily the solution to addressing the continuing challenges posed by the financial sustainability of health systems. In fact, successful interventions to ration health sector capacity and volumes of care in some OECD countries have resulted in health services that no longer match population expectations, as in the case of long waiting times for elective surgery.
49Concerns over health expenditure are often not well founded. As a country becomes richer, its inhabitants will necessarily want to put more resources into improving their own health, especially in view of population aging and rising expectations for more and better care linked to new healthrelated technological advances. In fact, there is nothing intrinsically wrong with high levels of health spending, if that is what people want, even if the result is lower overall consumption of other goods and services.
50There are nonetheless at least two sets of concerns that are legitimately associated with rising health care spending. First, most health spending is still publicly funded. In France, public sources of funding health care represent 76% of total health expenditure, which lies slightly above the OECD average. Increases in health expenditure obviously raise concerns for policy makers struggling with overstretched public budgets. Second, and more importantly, much of the spending on health systems is not at an optimal level of efficiency and quality (OECD, 2004a). There is growing evidence of shortcomings – ranging from provision of inappropriate
51services to underuse of needed services and errors in health-care delivery – that result in unnecessary deaths, disability and poor health, often adding to costs as well.
Rather than containing health expenditure growth, which is liable to result in an undesirable rift between the population’s expectations and the ability of health systems to provide, it is on improving the efficiency and value for money from health spending growth that policy makers’ efforts should concentrate for the future. In association with this, measuring, analysing and monitoring the effectiveness of health spending are indispensable to reduce a certain lack of transparency in the quality and output of health systems. International comparative analyses can help on both grounds, by providing a benchmark for assessing the performance of national health systems, and by providing a toolkit of measures that countries can implement to improve the efficiency of their health system.
Notes
-
[*]
Francesca Colombo, health economist, and David Morgan, statistician, in the Health Division of the OECD.
-
[1]
The views expressed in this article are those of the authors and do not necessarily reflect the position of the OECD. The authors would like to thank Eva Orosz and Elizabeth Docteur from the OECD Health Division for their invaluable comments on this article.
-
[2]
OECD Health Data 2005 is the latest available annual collection of data on the health systems of the 30 industrialised OECD member countries. Expenditure data are collected according to the main aggregates of health spending and financing as well as more detailed data according to function and provision of health care. Countries are at different stages of being able to provide data at the various detailed levels which may affect comparability. More information on the various definitional issues can be found at: www.oecd.org/health/healthdata
-
[3]
In Korea, health spending as a share of GDP was 5.9% in 2001.
-
[4]
Using GDP deflator.
-
[5]
This brought these countries closer to the OECD average of health spending as a share of GDP, as shown in Table 1.
-
[6]
The average is for 2003. It excludes Slovakia. For Australia, Japan and the United Kingdom, data are available for 2002 (OECD, 2005).
-
[7]
This will also be the case of the Netherlands after the implementation of recently approved reforms to transform the current system, where individuals above an income threshold are not eligible to sickness fund coverage, into a uniform private health insurance system, mandatory for the entire population (Dutch Ministry of Health, Welfare and Sport, 2002).
-
[8]
Excludes Belgium and Slovakia, for which data are not available. (OECD, 2005).
-
[9]
The average excludes Belgium, Greece, Portugal, Sweden, Turkey and the United Kingdom, for which data are not available. For Australia, Japan and Korea, data are available for 2002 only (OECD, 2005).
-
[10]
The average excludes Belgium, Greece, Portugal, Sweden, Turkey and the United Kingdom, for which data are not available. For Australia, Japan and Korea, data are available for 2001 only. A number of OECD countries have no or insignificant PHI markets, whose contribution was estimated as nil (OECD, 2005).
-
[11]
The average excludes Belgium, Greece, Ireland, New Zealand, Portugal, and the United Kingdom. Medical goods include pharmaceuticals and other medical non-durables, and therapeutic appliances and other medical durables (OECD, 2005).
-
[12]
In should be noted that spending on prevention and public health is less comparable between countries. For example, in France as in other countries, a large share of expenditure on prevention may be included under spending on curative and rehabilitative care.
-
[13]
The average does not include Poland and Turkey, as data are not available.
