Printer Friendly

Health care expenditure and other data.

Health care expenditure and other data

This international compendium from the Organization for Economic Cooperation and Development Secretariat contains available data for the 24 Member countries for 1960 through 1987. The Federal Republic of Germany is listed as Germany. Tables are presented for expenditure on health, health care pricing trends, social protection and public participation, utilization of medical services and available personnel resources, selected variations in common medical care practice, selected health status indicators, and demographic and general economic background data.

There may be some inconsistencies between the data presented in the articles in this issue and those in this data compendium. The articles were based on earlier versions of the Health Data File. As a result, revisions to the data in a number of countries, including Canada, France, Germany, and Italy, are not reflected in the articles. However, the changes are relatively minor and do not affect the results presented in the articles in any substantive way.

Health Data File: Overview and methodology

Reliable information is essential for rational policymaking. Unfortunately, in many countries, health policies have been framed without the help of systems of national health accounts. Even now, the public debates on health policy in many of the industrialized countries are conducted in terms of partial information (i.e., the division between the private and public shares in the financing of health care is based on hazy concepts and the health accounting framework is typically fragmentary). Moreover, even where these shortcomings do not apply, the dearth of outcome indicators considerably hampers analyses of the overall performance of health systems as well as the effects of particular policy interventions.

Although the simple availability of reliable information does not necessarily lead to appropriate clinical or service delivery decisions, in most areas of public policy, reliable information is an essential ingredient in improving the policymaking process. Few policies, particularly those pertaining to social programs, are framed in an international context. Mainly this is because of the intrinsically national nature of social systems and the dearth of valid, internationally comparable data.

Statistical systems are regularly modified to adapt to the requirements of a changing environment, incorporate new data sources, monitor a wider territory, and facilitate interpenetration with other, related statistical data. Such changes are typically generated by a domestic process; rarely are they triggered at the international level. A community of countries may, however, decide to embark on a new generation of concepts and methods to elaborate, for example, national income and product accounts or employment statistics. The publication Measuring Health Care 1960-1983: Expenditure, Costs and Performance by the Organization for Economic Cooperation and Development (OECD) in 1985 triggered such a process of reevaluation of accepted health information and accounting standards in many countries. A sizable number of statistical series have been substantially revised during the past decade to correspond to more exacting standards. For example, in their 1988 national review of progress toward the World Health Organization (WHO) objective "Health for All by the Year 2000," the Danish authorities stressed the importance they attach to the OECD statistical work (and its impact on their welfare-state development) "whose compilation and processing of data, especially concerning health economics and health activities, is considered as a necessary supplement to the WHO activities in these fields." However, the process of revision of accounting standards in OECD countries is far from finished and is likely to be continued in the early 1990s.

Three main approaches(1) to compiling and processing official data at the international level can be identified: * An international (or related) agency fully finances a

survey and, thus, controls its variables. Identical

questions, common concepts and definitions, and

single processing are employed. * Agencies from various countries determine that an

area is worth a cooperative effort, work together to

harmonize concepts and definitions, and supply the

corresponding time series to a single compilation unit. * An analyst attempts to "massage" data from various

countries, using as closely comparable units as can be

obtained from the information readily accessible.

The first approach is the surest way to attain fuller cross-national comparability. However, it is only rarely used because it is the costliest in money terms. The second is the most classic method, but it requires years to implement because numerous meetings are necessary to agree on boundaries, classification principles, accounting units, compilation lead-times, etc. The third approach is the least precise because judgments are made by the compiler only. However, with this approach, several years are saved and costly experts' meetings are avoided. Although the learning curve inherent in this approach leads to considerably improved data in later editions, the enhancement process is constrained by the ability of a single investigator to perceive the intricacies of complex systems pertaining to many countries and by inherent difficulties in gaining access to the building blocks of the statistical series to be compared.

Sooner or later, the second approach becomes unavoidable. However, the OECD Health Data File belongs to the third group. Thus, it should be viewed with a knowledge of the weaknesses inherent in such an approach. Its current strength lies in the lack of an alternative.

Data systems are intellectual constructs. As such, they are designed to summarize behavioral relationships that are empirically complex and often not well understood: Statistical systems are thus regularly challenged by their users (and even internally by their producers) for inadequacies in the reductionism employed. Therefore, an international compilation cannot be superior to its national constituent parts.

Because data originate from several sources in each country, and those sources are not necessarily consistent with one another, the Health Data File is fraught with international inconsistencies, gaps in coverage, and definitional heterogeneity. However, in international health comparisons, the tradeoff of precision for timely accessibility of the data is a difficult but necessary one.

The OECD Health Data File is designed to facilitate the identification of trends rather than detailed policy prescriptions, provided that the series exhibited is qualified and the data are used with a certain level of caution. Space limitations preclude adding numerous footnotes to the tables, and the qualifications stated in this overview only briefly address sources and methods. More detailed information on sources and methods that would provide a better understanding of the data being compared are presented in Measuring Health Care.

The 67 tables presented in this compendium are a subset of the more than 400 tables being worked on at the OECD as part of the Health Data File. They reflect the October 1989 status of the process initiated with Measuring Health Care in 1985. Great care has been taken to present as many continuous time series as possible, but breaks are unavoidable. The cost to statistical offices of revising past data in all series after a benchmark revision is often prohibitive.

As is usual with data from statistical systems, the estimates for the last year (mostly 1987) are provisional, and those for the 1 or 2 previous years (usually 1986 and 1985) are only semifinal. Aside from methodological revisions, quantitative information, on which macroeconomic estimates are based, is usually incomplete at the time of first compilation and is supplemented by "projections" of the recent past.

Finally, although most of the data come from national data bases to which the OECD Secretariat has had generous access, modifications have been required to improve cross-national comparability. Sometimes, the information supplied here differs from the customary national presentation of health trends. The cross-national presentation does not automatically entail a superiority of one set of data over the other. These data are published under the authority of the Secretary General of the OECD. Thus, national authorities are not formally committed to changes made to published national series in order to enhance cross-national comparability.

The first hurdle to overcome in developing these tables was a boundary problem. In most countries, health care is monitored by a ministry of health. However, ancillary activities, such as armed forces health services, school health services, industrial medicine, family planning centers, and old-age homes with varying degrees of medical services are treated in a number of ways by the agencies responsible for their administration. To date, recognition of the boundary problem has led to important convergences in individual countries' data reporting, thus increasing comparability across countries. However, significant differences among countries do not yet permit full homogeneity. Denmark, for example, does not classify "residential" beds provided under the auspices of its Social Affairs Ministry as inpatient medical care beds. Sweden, on the other hand, considers "nursing" beds to be part of inpatient medical care institutions. Dealing with ambulatory care (outpatient clinics, office-based medicine, etc.) results in similar problems. Chiropractors are licensed professionals in some countries and treated as quasi-healers in others.

In a number of cases, the statistical subsystems (for example, expenditure and manpower data) of different agencies have heterogeneous definitions, with detailed, underlying series inaccessible in the published sources. This factor alone suffices to explain important variations between successive versions of the data, such as that included in the 1985 version of Measuring Health Care and that included here. A still greater difficulty than accessing data is that of consistency, even in each country's own definitions. Moreover, in some cases, the definitions used in a country's most quoted reference vary slightly almost annually because the time series shown varies in response to the policy emphasis of each particular year. The most common pitfall consists in using parameters with identical names as if they were comparable. Although similar terms are used, the meanings may differ because they may not be based on formal prior agreement on concepts, definitions, estimating methodologies, etc.

In a comparative international data base, the problems encountered at the national level are magnified. The omission of data on pharmaceuticals distributed through mail-order houses would have no effect on French or Spanish health outlays, but it would create a sizable gap in the U.S. data. Over-the-counter pharmaceutical sales are not directly observed in many countries. Therefore, their estimation requires an apparent-consumption approach established on the basis of production data minus exports plus imports plus net changes in inventories (stocks). Such an approach obviously leads to misclassification of the containers of medicines that can be found in cabinets in many homes that have been purchased but not consumed.

The definitional problem is not confined to expenditure series, as students of the quality of medicine also experience. In his article in this issue, Enthoven cites the paucity of hospital records in some European countries that he visited. Simple breakdowns, such as the distinction made between admissions and readmissions, are not available in many countries. Yet, this is an important variable in considering whether the pressure to lower the length of hospital stays leads to the early discharge of frail patients who soon require new spells of inpatient care. The recording approaches of major hospital subsystems in some countries differ, so the lags in accessibility of complete data sets can be fairly long.

The medical evidence of death records also can rest on cultural differences. For instance, one OECD country may record little mortality from bronchitis, simply because this classification is not used in its medical schools; deaths from "other upper respiratory diseases" are correspondingly more frequent. International comparisons thus require the aggregation of two or more subcategories.

A single issue of the Health Care Financing Review cannot encompass a full description of each of 24 countries' statistical sources and methods and their possible reconciliation. Only brief discussions are provided in the following sections.

National accounting concept

The expenditure series (Tables 1-9) initially was conceived as a national accounting (income and product accounts) series meant to satisfy two criteria: quasi-comprehensiveness of the underlying identity and adherence to rigorous economic classification principles. The identity ensures that all costs incurred for a stated purpose or function are added up. The components are private consumption of medical care--hospital care, physician services, pharmaceuticals, therapeutic appliances, other insurable benefits (except those related to sickness benefits paid in cash, which belong to an income maintenance function)--and general government outlays on health care, including public health preventive services, administration and regulation, etc.

The qualification "quasi" is used for several reasons. For one, there is a grey area related to those services that contribute to the health status of the population but are classified differently in various countries: under agriculture, under health, or elsewhere. Because detailed information is not readily accessible, it has not been possible to reallocate these services systematically. Whenever detailed information was available and permitted it, the reallocation has been affected for armed forces health services, prison health services, school health services, and publicly funded medical research and development, which are treated as auxiliaries to defense, justice, education, or science support outlays in the national accounts. In the national accounts, expenditure for health on private business premises is treated as an "intermediate" outlay, although the same services rendered to civil servants are final expenditure. Eye tests for airline pilots and similar examinations are necessary for rendering services such as safely flying passengers, but mostly the primary purposes of these services are preventive screening, which implies that they should be added to the other "final" consumption outlays. This has been done to the extent possible.

Also, gross capital outlays have been added to the aggregate outlays. In some countries, such as France, a depreciation allowance is included in the pricing (for example, the price of hospital services) and has to be deducted to avoid double counting; this was not always possible. However, the range of errors and omissions has been amply reduced over time. The estimated size of the gap or the inclusion of unwanted categories of expenditure, errors, and omissions may be in the range of 96 to 102 percent of the desired amount, much less for many countries.

Another major problem is that certain national accounting systems still are rudimentary. In public debates in a few countries, such as Belgium, the OECD estimates have been criticized as being on the low side. Almost never is it suggested that OECD estimates are overstatements. However, in France, a 1987 benchmark reevaluation led to a substantial downward revision, which is still only partly documented. Underreporting in the basic national systems from which the data are extracted regularly leads to small reevaluations of the estimates. The publication of the tables in this issue of the Health Care Financing Review is likely to generate further changes in ongoing accounting methodologies. Countries, such as Switzerland, that had no health accounts have established a working party to create such a system. Tentative satellite accounts have been developed in Finland and Norway. The Health Care Financing Administration in the U.S. Department of Health and Human Services publishes detailed National Health Accounts data (which, with only a 1-percent difference, nearly correspond to the National Income and Product Account Health Outlays data, compiled by the Bureau of Economic Analysis in the U.S. Department of Commerce). Health and Welfare Canada publishes data according to the same accounting rules as those used in the United States. Details published in the Dutch accounts or made available by the Central Bureau of Statistics suggest that the boundaries and economic classification principles used in the Netherlands are close to those adopted in North America. In France, Germany, and the United Kingdom, the accounting system has gained in reliability over time. However, the international compilation for countries is still crude. It is hoped that greater cross-comparability will be reached in the 1990s through international harmonization of boundary concepts and classification principles.

The use of international classifications other than those proposed by the OECD should be mentioned.(2) An example is a classification based on institutional similarity, such as social security. Estimates from these other classification systems do not include purely private medical outlays, and a fair number of public health outlays included here are also omitted. The expenditure levels estimated by the OECD Secretariat are higher than those

reported elsewhere.

The heterogeneity of sources for the segments of the expenditure series is another cause of gaps in the identity. Total expenditure equals institutional care plus ambulatory care plus pharmaceutical purchases plus therapeutic appliances plus collective services (such as armed forces medical care) plus public health programs plus biomedical research and development outlays plus administrative outlays. However, for most countries, the gaps are not large.

Prices and incomes

Detailed expenditure flows are only one prerequisite of the accounting substrata needed for health policy analyses. To obtain a measure of real resources devoted to health care across countries or over time requires controlling for price trends. Have the resources devoted to health care genuinely decreased in some countries? Have they merely been stabilized in most, as exhibited by the simple ratio of total health expenditure to gross domestic product (GDP) or gross national product (GNP)? The answers provided rest on the evidence collated or estimated by statisticians regarding prices. (Schieber and Poullier briefly address the relationship between economic growth and the rate of consumption of health services in their article in this issue.) The methods of elaboration of the implicit price indexes of GDP and GNP are well documented; those of the numerator (health care consumption) are less so.

In a dynamic economic process, individual prices vary as a reflection of new scarcities, gains in productivity, market power, and a host of other demand and supply factors. If an aggregate price index for medical consumption--GDP deflator or consumer price index (CPI)--is used, specific changes occurring in inpatient care, ambulatory care, pharmaceutical production, therapeutic appliances, and government-supplied services are not captured. The OECD Health Data File includes distinct price measures (shown as 1980 = 100 in the accompanying tables) for each consumption function. The indicators that are accessible are mostly details of private consumption price trends, and these often fail to reflect the impact of changes in real factor costs. Input price measures, however, also are gradually collated.

At a more aggregate level, a new "total" medical consumption price index is presented in Table 13. This replaces the measure of "private" consumption trends previously presented in Measuring Health Care. The ol measure was a fair indicator for Belgium, Switzerland, and the United States, where private consumption constitutes from two-thirds to more than four-fifths of total consumption. However, it was deceptive for other countries. For instance, in Canada, it reflected barely more than pharmaceutical prices. For some countries (Australia, Austria, Denmark, Finland, Greece, Iceland, Italy, Japan, Norway, Sweden, and the United Kingdom) the new measure is a weighted index of private and general government consumption of medical care and health services. For other countries (Belgium, Germany, Ireland, Luxembourg, Spain, Switzerland, and the United States), the price index shown reflects only private consumption; this can be a source of serious bias. For Canada, France, and the Netherlands, total health consumption indexes are calculated by the national authorities, but the underlying methodology is not given along with the published figures.

The Canadian price index is a weighted sum of the following: physician services (i.e., fee-for-service increases in medical care insurance plans, not salaries), which are calculated for each province and tabulated by Health and Welfare Canada; dental care and optometrist services, which are measured by consumer prices; average wages and salaries in hospitals (80 percent) combined with hospital supplies (20 percent), which is information collected by Statistics Canada; and prescribed and nonprescribed drugs, which are two consumer price series collected by Statistics Canada. The hospital index is also applied to other institutions.

The French index appears to be equally divided between a weighted price measure of the consumer price components and a cost index for public institutional care. Information on both wage costs and supplies is obtained in an annual survey conducted since the early 1980s. Consumer prices were the only deflators available for public hospital costs in the earliest years.

The Netherlands annually publishes current and constant price data for a range of inpatient care institutions, several ambulatory services, medical goods, and health administration services. The estimates allow calculation of a measure of total costs since 1980; provisional figures were calculated for the period 1972-80 (the estimation for past years is not yet completed); and consumer price figures were derived from several national accounting versions prior to 1972.

The final Australian private consumption expenditure deflator is a weighted measure of doctors' and dentists' fees for a range of services paid for by consumers; hospital services (weighted the same as the general government index, described next); and hospital and medical fund administrative expenses (fixed wage for the salary component; CPI trends for meals, transportation, and telephone services; and producer prices for paper and printing). For the general government index, wages and salaries for hospital, medical, and ancillary staff are weighted by fixed-weight wage and salary rates, and supplies are weighted by the corresponding input prices.

Finland also uses separate deflators for intermediate consumption and each component of value added, summing up the costs. The consumption of fixed capital is based on a capital stock model. Private health services are mainly based on social security schedules for paying patient bills, the physician fees and laboratory billing serving as control measures for a range of services. Dental services are evaluated mainly by employment data and CPI-monitored fees.

Luxembourg uses consumer price data for dental care, inpatient bed days, laboratory tests, and pharmaceutical products when evaluating market goods and services. Government health consumption is not specifically deflated.

Spain deflates nonmarket services through a weighted index of wages and supplies, with detailed calculations based on previous year = 100. (This index is not published and is not included in the estimates shown here.) Private medical consumption is deflated by the relevant CPI components.

The United States deflates physician, dentist, and other professional services and care in for-profit hospitals by the relevant CPI components. The hospital room component is used for care in for-profit hospitals. Nonprofit hospital and nursing home prices are composite input prices established by the Health Care Financing Administration. Drug preparation and sundries are deflated by the relevant medical care commodities CPI components, and ophthalmic and orthopedic appliances are deflated by the relevant eye care CPI components. Medical and hospitalization insurance is deflated by the weighted average of physician and hospital services based on estimated benefits composition; it is not separated from income loss insurance and workers' compensation insurance in the published details, the latter two being deflated by the total CPI index. Government purchases are deflated by the appropriate wage and supplies indexes. (This information is not published and is not included in the estimates shown here.) Fixed investment is also deflated by a range of relevant measures for medical and surgical instruments, hospital furniture, hospital construction, etc., including separate measures for the relevant merchandise exports and imports.

Only a limited amount of input cost data is available, so the series shown does not, in principle, include adjustments for productivity. The headings used for Tables 10-13 reflect common acceptance of pricing trends in health economics, but strictly defined they are conventional deflators rather than input price indexes. The distinction relates to the level of observation. Similar services may have two prices. For example, physicians may be committed by law to apply one level of fees to insured patients but may have some freedom to apply another price to private patients. Where two or more sets of prices and price trends exist for given commodities or services, the level of observation does not permit differentiation. The statistician is compelled to make decisions on the basis of partial knowledge, which is only gradually incremented. The data presented reflect a provisional state of the art that is believed to be superior to use of CPI, GDP, or GNP aggregate price trends. The widespread use of separate indexes for wage components and supply components suggests that, notwithstanding the extreme variety in the basic parameters observed (e.g., in inpatient care institutions), the underlying deflation approach is reasonably comparable across countries.

Social protection

The documentation of differences across countries in the level of social protection is essential to the understanding of variations in the private-public mix in the financing of health services. Typically, only descriptive documentation is accessible. The OECD Health Data File constitutes an attempt to translate such descriptive documentation into a series of indicators reflecting the extent to which social preferences have moved countries toward coverage of the population by a public scheme and toward substituting collective for individual financing of medical benefits.

The indicators of public coverage constitute an entitlement index. In countries with universal access to publicly funded health services, the measure is straightforward. For all others, an interpretation is required of institutions and regulations that, at times, are fairly detailed. An index of social security coverage only is insufficient. In the Netherlands and Spain, for example, specific schemes--run by entities that are legally private--exist for specific groups of public employees. Since the schemes are compulsory for the groups concerned, the outlays have been treated as public, and so should the potential beneficiaries. Various systems offer gradients of coverage. For instance, in the Netherlands, a universal tax-funded scheme for long-term hospitalization has been run since 1968, but one-quarter of the population has not been subject to compulsory insurance for shorter hospitalization and other medical benefits. Ireland has a three-tier system under which, in mid-1987 (but not in the first two decades monitored), public hospital accommodation was accessible to all. Category III of this system pays for physician care. The population with voluntary health insurance, or VHI (private coverage) is estimated to have increased from 17.3 percent in 1976 to 29.3 percent in 1987. The benefits of VHI are mainly in the ambulatory care and pharmaceutical functions (plus physician services in hospitals for category I). However, a measure of the covered population is not simply the difference between 100 and the percent of population with VHI, because a special entitlement program for pharmaceutical benefits exists, notably for chronic care.

The estimates shown in Tables 15-17 are, for most countries, best estimates from the OECD Secretariat. These are likely to be revised as more knowledge is gained of the 24 health systems under study. Several of these systems, in turn, comprise distinct subsystems. The appropriate index for a country cannot be a simple weighted average for that country, as each subsystem exhibits a number of exemptions and special cases.

An estimate of the average level of public cost sharing is still more difficult to establish. The nomenclatures of services accessible without charge (full reimbursement), those for which there is nominal cost sharing (partial reimbursement), and those for which gratuitous supply or reimbursement is denied (such as some cosmetic surgery, services supplied by physicians working outside a social contract framework, and medicines listed as nonreimbursable) sometimes comprise several hundred entries, and no detailed spending data are available. In addition, personal status affects entitlement and cost sharing. Pensioners are exempt or pay lower rates in some countries, such as Belgium; children are entitled to free dental care in several countries of Northern Europe; war victims are entitled to zero cost sharing in France; the unemployed and/or other underprivileged groups have access to specific programs in some countries, such as the United Kingdom and the United States.

No OECD country publishes an index of cost sharing. The generosity index, as it is sometimes labeled, is not a simple ratio of public-to-total expenditure. In the nearly three decades spanned by the observations, notable changes have occurred. Most countries have liberalized access to or are partly subsidizing reproductive services (family planning, prenatal and postnatal care, abortion, etc.) Older social care services aimed at the killer diseases of the period before World War II (tuberculosis, poliomyelitis, etc.) have been displaced by programs aimed at a new killer--acquired immunodeficiency syndrome (AIDS)--and by a range of treatment options made possible by advancing medical technology (such as treatment of end stage renal disease and the prevention of congenital malformations). In the opposite direction, the development of thousands of new medicines has forced many governments to moderate or even slightly reduce the level of pharmaceutical benefits. In actuality, this range of trends is still ill-translated in Tables 18 and 19. However, data from these tables can be used to find signals about turning points (years during which public schemes increased or decreased benefit levels) and to generate awareness about the typically modest size of patient cost-sharing requirements in most OECD countries. In many public debates, the public effect of cost-sharing techniques is overstated. However imperfect, the indexes produced serve a didactic purpose.

Medical care use and personnel

Health planning has long been conducted in terms of ratios, such as hospital beds per 1,000 population, with better endowed neighboring countries serving as targets. Underutilized beds contribute little to raising the health status of populations, and the differences in countries' occupancy rates are startling. The OECD Health Data File has thus been designed as a measure of usage in addition to a measure of actual inputs. Many previous international listings included licensed but retired or otherwise inactive physicians. Efforts to standardize definitions so as to include similar inpatient institutions or active medical personnel have yielded only partial results to date. Further progress is expected in subsequent versions of the Health Data File.

Time series are more reliable than cross-sections, although this is not always the case. For example, the data on inpatient care beds in Norway matches the expenditures reported in Tables 3 and 4 only from 1980; earlier data seriously underreport the input into the Norwegian health system.

Tables 20-30 also differ from other classic presentations in attempting to portray an annual average (obtained from a daily census or other recording system) or a midyear position. Retired medical and paramedical personnel and graduates who are not practicing are not counted.

Efforts to standardize concepts cannot always overcome cultural and institutional differences. For instance, in some countries, physicians are allowed to charge their patients for diagnostic and prescriptive advice given by phone; elsewhere, they are not. The frequency of doctor consultations may be correlated with payment methods (as discussed by Sandier in this issue). The trends shown in Table 23 are probably more reliable than cross-country comparisons.

Definitional differences contribute to the intercountry differences in pharmaceutical consumption observable in Table 24. Hospital outlays on pharmaceuticals are excluded. In some countries, medicines delivered in outpatient wards are also excluded. Physicians in rural areas of some countries are allowed to dispense medicine, a source of minor data distortion. In Japan, one of the largest medicine consumers of all OECD countries, physicians in urban areas can dispense medicine as well. The number of units of medicine reported in Table 24 reflects only residual sales in pharmacies. In most countries, prescriptions refer to the number of items listed by physicians on each prescription form. Typically, more than three items constitute one prescription form for French general practitioners. Records also vary across the spectrum of countries with respect to renewable items for chronic treatments: Should six boxes of beta-blockers prescribed for a hypertensive patient over one-half year be entered as one prescription or six? In compiling international data, one lacks the necessary details to ensure consistent treatment across countries.

Medical practice variations

The literature on the impact of cultural and other factors on medical consumption in adjacent catchment areas and on the appropriateness of care in similar socioeconomic conditions has become sizable. (In this issue, McPherson adds to the literature survey.) Smaller efforts have been devoted to highlighting similarities across countries, whether tied to cultural factors or to prevailing socioeconomic incentives. Some facets of these large area variations are illustrated in Tables 31-49.

In Tables 31-37, mean length-of-stay is shown by discharge category using the 18 classes of the International Classification of Diseases (ICD) adopted under the auspices of the World Health Organization and used by all OECD countries. However, several countries do not yet record all their hospital admissions exhaustively and systematically, as indicated in the Enthoven article in this issue. In principle, the surveys on which these tables are based cover the entire nation for which data are reported. However, data for Australia relates only to the most populous State (New South Wales), and data for the United Kingdom come from a one-tenth sample based on a British classification developed by the Office of Population and Census. (This classification is also used in Ireland.) By and large, the figures reported for the 1970s are based on the Eighth Revision of the ICD. Use of the Ninth Revision spread only slowly across the spectrum of countries, so the figures for the 1980s are partly Eighth Revision and partly Ninth Revision. This factor is even more important when considering the three-digit entries in Tables 38-44. In some countries, maternity data are not aggregated with general hospital data, introducing some distortion. The extent to which psychiatric care is provided in hospitals differs, too. The entries for mental disorders in Table 31 are the least comparable, and this affects the overall total. The occasional reallocation of services inside a nation's hospital structure influences total length-of-stay; this is the case with Sweden beginning in 1984. The data for Switzerland are for a hospital federation covering more than four-fifths of all beds. The data for Austria and Germany are only for patients belonging to the largest health insurance schemes. The average length-of-stay figures for New Zealand are weighted for private and public institutions.

These tables also reflect conceptual differences related to definitions of personnel per bed. Data on all staff employed by hospitals (Table 45) are not completely comparable because the extent of subcontracting differs among countries. The definitions of nurses also vary (Table 46). Although there are persistent conceptual differences, considerable efforts are brought to bear to reduce the comparability gap that occurs when national series are simply collated. By and large, the trends are consistent, and no amount of statistical massaging would crush the important intercountry variations observed.

Health status indicators

Although some progress has been achieved, OECD countries do not monitor the outputs of their health systems satisfactorily because appropriate indicators are few. International organizations have limited their collection of data to a small set of "standard" indicators. The data on life expectancy at birth and at various ages (Tables 50-57) and infant and perinatal mortality rates (Tables 58 and 59) originate from national yearbooks and are sometimes supplemented by demographic journals. Mortality rates (Table 60) are supplied from the annual OECD publication Labour Force Statistics.

Part of the progress in a longer life span in the 1950s and 1960s is attributable to a sharp drop in infant and perinatal mortality. New diagnostic techniques to cope with high-risk pregnancies and the advent of sophisticated gynecology-obstretical centers to cope with difficult deliveries have been essential factors in accelerating the downward trend in perinatal mortality and, by limiting complications at birth, have contributed to the improved well-being of both mothers and babies. The death rates for infectious diseases have dropped in OECD countries by about nine-tenths in the past half-century. However, in macro-policy terms, these measures are usually held to be of limited value. They may reflect advances in medical knowledge and the diffusion of certain medical procedures, but they do not lend themselves to the measurement of the outcome of public health programs and policy.

Demographic and economic background

International comparisons require ratios or the use of a common numeraire to deal with differences in population sizes and currency units. In this data compendium, Tables 1-9 and selected other entries are supplied in their original national currency expression and without reduction for population sizes. The required denominators or multipliers are published, though not always at hand when needed for analytical purposes. General economic and demographic background data are provided in Tables 61-67. The principal source for these data is Volume 1 of National Accounts 1960-1987, published by OECD in 1989. This source was occasionally supplemented by the underlying corresponding national publications. The OECD National Accounts and accompanying demographic data reflect guidelines that have been gradually established by the statistics profession in a harmonization process started in the early 1950s.

Differences observed in, for example, hospital beds per 1,000 population or pharmaceutical outlays per person when using national sources or the OECD data are not necessarily attributable to the numerator but may be caused by the denominator. Although an effort has been made to "massage" the numerator entries, the harmonization process of working from details to publish the most comparable aggregates is still in its infancy, whereas the body of principles on which the denominators are based is considerable.

Population figures are mid-year estimates. International agencies have opted to publish gross domestic product (GDP) estimates, not gross national product (GNP) or gross national expenditure (GNE) estimates; the difference--net factor income from abroad--is typically not very large and is not the reason why OECD GDP and domestic GNP figures are often at great variance. Canada, the United Kingdom, and the United States, to mention three large countries, adopted estimation methods in the 1940s which have not been fully incorporated in the standardized national accounts of the United Nations, the OECD, and (with some specific developments) the Statistical Office of the European Community. The standardized approach is reported here, the data being calculated by the concerned national statistical offices and meeting with their official approval.

The implicit GDP price indices, shown as 1980 = 100, are originally established in each country's own base year prices, then rebased at a subaggregate level by the OECD Secretariat whenever the base was not 1980. Several countries have various bases for, say, 10-year periods. The subperiods are linked by a chain index.

Typically, past international comparisons converted national currencies at exchange rates (Table 66). The estimates shown are an annual average of daily observations. The vagaries of foreign exchange markets lead to grievous underestimates or overestimates of the variables analyzed.

Purchasing power parities are presented in Table 67. Indices representing the average prices in each country relative to the average international prices to purchase the same market basket of goods and services were calculated for most OECD countries for 1980 and 1985. A regression method was used to determine 1975 data. Partial estimates also existed for 1950, 1955, and 1970. From these base years, and with national accounting details, the series has been retropolated and extended beyond 1985.

(1)A more elaborate discussion of these approaches can be found in "OECD experiences with the initiation and coordination of health indicator systems, with special emphasis on interinstitutional coordination and comparability" in Indicators and Trends in Health and Health Care, D. Schwefel, ed., Berlin-Heidelberg, Federal Republic of Germany, Springer Verlag, 1987. (2)A quantitative expression of the differences among classification systems can be found in Structural Problems of Social Security Today and Tomorrow, Proceedings of the European Institute of Social Security, Leuven, Belgium, ACCO, 1988, Table A-1.
COPYRIGHT 1989 U.S. Department of Health and Human Services
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:International Comparison of Health Care Financing and Delivery: Data and Perspectives; Compendium
Author:Poullier, Jean-Pierre
Publication:Health Care Financing Review
Date:Jan 1, 1989
Previous Article:What can Americans learn from Europeans?
Next Article:Longitudinal patterns of Medicare use by cause of death.

Related Articles
Overview of international comparisons of health care expenditures.
Cost containment in Europe.
Revisions to the National Health Accounts and methodology.
The burden of health care costs: business, households, and governments.
The health care systems of British Columbia and Washington state: learning from the neighbors.
U.S. health expenditure performance: an international comparison and data update.
National health accounts: lessons from the U.S. experience.
Administrative costs in selected industrialized countries.
Measuring the relationship between income and NHEs.
Business, households, and governments: health spending, 1991.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters