New ideas for health policy in France, Canada, and Britain

Nine New ideas for health policy in France, Canada, and Britain

Victor G. Rodwin

There is a widely shared belief among American policy-makers that a national program providing for universal entitlement to health care, in the United States, would result in runaway costs.(1) In response to this presumptive wisdom, nations that entitle all of their residents to a high level of medical care and simultaneously spend less than the United States, are often held up as exemplars. Canada's system of national health insurance (NHI) is the most celebrated example.(2) French NHI, a prototype of western European continental health systems, is another case in point. Britain's National Health Service (NHS), although typically considered a "painful prescription" for the United States (Aaron and Schwartz 1984), assures first dollar coverage for basic health services to its entire population and spends the smallest share of its gross domestic product (GDP) on health care expenditures (Table 9.1).

All of these countries have produced sonic of the leading physicians and hospitals in the world. judging by various measures of health status, they are in the same league or better than the United States. In Britain, life expectancy at 60 - when medical care may have an important impact - is lower than in the United States. But in the United States over 15 per cent of the population remains uninsured for health care services while spending, as a per cent of GDP, surpasses that of all industrially advanced nations (OECD 1985).(3)

Table 9.1 Health care expenditures arid health stains
	Health expenditures(1984)as % of GDP	Life expectancy (1980) at birth at age 60 males females males females				Infant mortality 2(1981)
France	9.1	70.1	78.3	17.2	22.3	.89
Canada	8.4	71.0	79.0	18.0	23.0	.85
Britain'	5.9	711.2	75.9	115.9	20.5	1.02
United States	10.7	69.6	76.7	17.2	22.4	1.09
Sources: Data on health expenditures arc from Schieber and Poullier 1986; data on life expectancy and infant mortality arc from OECD 1985, Tables F.1 and F.2: 131.
Notes: 1. All data are for the United Kingdom. 2. Infant mortality is expressed in death-rates of infants below 1 year per 100 live births.

Virtually no one in Canada or ill western Europe - not even the fiercest critic - would want to import or even emulate the American system of financing and organizing health care. But ill spite of this prevailing view, a number of fashionable American ideas, most importantly the concept of a health maintenance organization (HMO), have drifted north to Canada and across the Atlantic to Europe. These ideas are hardly popular. They are simmering and they represent a potentially creative response to a number of present concerns in France, Canada and Britain. Although all three of these countries, especially Canada and Britain, have eliminated financial barriers to care, policymakers still face three festering problems.

Economists, for example, emphasize that cost containment should not be confused with allocative efficiency in the use of health care resources. They point to the possibilities of obtaining more value for the money spent oil health care in France, Canada, and Britain, as well as in the United States.(4) This applies not only with regard to improving health status (Cochrane 1972) but also with respect to altering input mixes in the provision of health services - taking advantage of costclIfiective treatment settings, e.g. ambulatory surgery, and personnel, e.g. nurse practitioners.

Public health and medical care analysts criticize the lack of continuity of care between primary, secondary, and tertiary levels. Although health planners in France, Canada, and Britain have called for redistributing resources away from hospitals to community-based ambulatory care services and public health programs, the allocation of resources within health regions has been notoriously biased in favor of the more costly technology based medical care at the apex of the regional hierarchy (Rodwin 1984).(5) The consequence of this allocational pattern has been to weaken institutional capability for delivering primary care services. This has exacerbated the separation between primary, secondary, and tertiary levels of care thus making it difficult for providers to assure that the right patient receives the right kind of care, in the right place and for the right reason.

Consumers have noted the inflexibility of bureaucratic decision-making procedures and the absence of opportunities for exercising for what Hirschman calls "voice," in most health care organizations. Indeed, the problem of control and how it should be shared between consumers, providers, managers, and payers is at the center of all criticisms levelled against the current structure of health care delivery in France, Canada, and Britain (Rodwin 1987). In all of these systems, decisions about what medical services to provide, how and where they should be provided, by whom and how often, are separated from the responsibility for financing medical care.

In the context of these problems - inefficiency in the allocation of health care resources, lack of continuity between levels of care, and the absence of voice in most health care organizations - the concept of an HMO, in combination with elements of market competition, has a certain intellectual appeal. Since an HMO is, by definition, both an insurer and a provider of health services, it establishes a link between the financing and provision of health services. Since it is financed on the basis of prepaid capitation payments, its managers have an explicit budget as well as a clearly defined clientele (population at risk). Moreover, since an HMO is responsible - on a contractual basis - for providing a broad range of primary, secondary, and tertiary level services to its enrolled population, it has powerful incentives to provide these services in a cost-effective manner while simultaneously maintaining quality SO as to minimize the risk of disenrollment.

There are currently so many models of HMOs in the United States that it is unwise to generalize about them. Nevertheless, the evidence based on a large number of stable HMOs in the 1960s and 1970s is persuasive in demonstrating that this form of health care financing and organization can reduce hospital admissions by as much as 40 per cent when compared with conventional fee-for-service practice (Luft 1981).

The idea of introducing HMOs - or similar kinds of health care organizations - into national systems that provide universal entitlement to health care resembles, in many ways, the American experience of encouraging medicare beneficiaries to enroll in federally qualified HMOs or competing medical plans (CMPs). The idea usually involves two reforms. It spurs policymakers to combine regulatory controls with competition on the supply side; and it encourages them to design market incentives for both providers and consumers of health care. In this chapter I examine some new ideas along these lines for France, Canada, and Britain and conclude with an assessment of their viability.

France: les reseaux de soins coordonnes (RSC)

France is noted for combining NHI with fee-for-service private practice in the ambulatory care sector and a mixed hospital sector of which two-thirds of all acute beds are in the public sector, and one-third in the private sector (Rodwin 1981). Physicians in the ambulatory sector and in private hospitals (known as cliniques) are reimbursed on the basis of a negotiated fee schedule. Roughly 15 per cent of all physicians are allowed to set their own fees. And physicians based in public hospitals ? the principal teaching and research institutions - are reimbursed on a part-time or full-time salaried basis. private cliniques are reimbursed on the basis of a negotiated per diem fee. Public hospitals used to be reimbursed on a retrospective cost-based per them fee but they have received prospectively set "global" budgets since 1984.

There are several problems in this system. From a public health point of view, there is inadequate communication between full-time salaried physicians in public hospitals and solo practice physicians working in the community. Although general practitioners in the fee-for-service sector have informal referral networks to specialists and public hospitals, there are no formal institutional relationships which assure continuity of medical care, disease prevention and health promotion services, post-hospital follow-up care, and more generally systematic linkages and referral patterns between primary, secondary, and tertiary level services.

From tile point of view of economic efficiency criteria, there are additional problems in the French health care system. Oil the demand side, two factors encourage consumers to increase their use of medical care services: the uncertainty about the results of treatment and tile presence of insurance coverage. To reduce the risk of misdiagnosis or improper therapy physicians are always tempted to order more diagnostic tests. Since NHI covers most of the cost, there is no incentive - neither for the physician nor for the patient - to balance marginal changes in risk with marginal increases in costs. This results in excessive medical care utilization.

Oil the supply side, fee-for-service reimbursement of physicians has provided incentives for them to increase their volume of services so as to raise their income. Likewise, per diem reimbursement of cliniques and hospitals created incentives to increase patient lengths of stay. The recent imposition of global budgets, in France, has eliminated this problem but they represent a blunt policy tool - one which tends to support the existing allocation of resources within the hospital sector and, possibly, to jeopardize the quality of hospital care. It is relatively easy for a hospital to receive all annual budget to maintain its ongoing activities but extremely difficult to receive additional compensation for higher service levels, institutional innovation or improvements in the quality of care. Even with prospective budgets, hospitals naturally seek to ma ' ximize the level of their annual allocations and to resist budget cutbacks.

In summary, providers under French NHI have no financial incentives to achieve savings while holding quality constant or even improving it. Consumers have few incentives, other than minimal co-payments, to be economical in their use of medical care. And, there are no incentives to move the French system away from hospital-centered services toward new organizational modalities.

Traditional solutions to these problems go in the direction of making patients pay higher co-payments. For example, a 3 dollar daily co-payment charge was recently imposed oil all hospital in-patient stays. Reimbursement for drugs has become more restrictive, particularly for those with more questionable therapeutic effects. Also, the government is allowing more physicians to refuse assignment of their fees and engage in extrabilling. The problem with these proposals is that they focus only on the demand side. They do nothing to promote supply-side efficiency. It is in response to this challenge that a proposal was recently developed to introduce a system of HMOs under French NHI.

In French, the concept of an HMO was translated as a reseau de soins coordonnes (RSC) - a network of coordinated medical services. The proposal, published in the French Review of Social Affairs by two French economists, a French physician and the present author (Launois et al. 1985), is based on six principles:

Preservation of entitlements under NHI

All compulsory pay-roll taxes for NHI remain unchanged. All those covered under French NHI, i.e. 99 per cent of the population, remain covered. The current level of benefits becomes a minimum benefit package under the new plan.

Supply-side modernization through the creation of RSCs

Qualified RSCs - with minimum benefit packages - are retitiired to allow open enrollment. RSCs could be organized by a variety of sponsors. They would promote vertical integration in the health sector and place hospitals, day surgery facilities, physicians, and other health-care professionals at risk for providing cost-effective medical services.

Promotion of integrated medical care

The RSC assumes a contractual responsibility for providing its enrolled population with all health services covered under French NHI. The patient chooses a primary care physician who is in charge of making proper referrals and managing patient

Prepayment on a capitation basis

The RSC receives a pre0-paid capitated monthly fee directly front the beneficiary's NHI fund. This payment is equal to the actuarial cost based on the enrollee's age, sex, and health status. The RSC's annual budget is equal to its annual capitation payment multiplied by the number of its enrollees. Within that constraint, managers have an incentive to minimize costs and maximize patient satisfaction so as to avoid disenrollment.

Marginal shifts in health care financing

Most of the capitated fee is financed directly by the beneficiary's NHI fund. But since, in the aggregate, consumers pay roughly 15 per cent of all health expenditures through co-payments, to make the proposal financially viable there is an additional prepaid contribution by the beneficiary at the time of enrollment. This would be equal to the difference between the capitation fee charged by the RSC and the actuarial cost calculated by the beneficiary's NHI fund. There is no payment at the time of service use, and all enrollees who cannot afford the additional contribution are eligible for a state subsidy.

Competition between RSCs

Enrollment in RSCs is voluntary. This results in three levels of competition. First, between RSCs and traditional NHI. Second, between RSCs themselves. Third, between health care providers to whom RSCs will send their enrollees presumably on the basis of their ability to keep quality high and costs low.

The six principles of this proposal were inspired by Alain Enthoven's (1980) Consumer Choice Health Plan for the United States. But whereas Enthoven's plan is designed to create a new form of NHI for the United States, tile RSC proposal is largely a strategy to promote supply side efficiency within an already existing NHI system. As in the case of competing medical plaits (CMPs) - HMOs for Medicare beneficiaries in the United States - if French beneficiaries choose to enroll in an RSC, they would lose their coverage under traditional NHI. just as CMPs have to accept all Medicare beneficiaries who choose to enroll, all RSCs would have to accept all French NHI beneficiaries who choose to enroll, which could be 99 per cent of the population. Thus the problem of adverse selection is somewhat reduced, although by no means absent.

Canada: publicly financed competition

Under Canadian NHI, although coverage for drugs is far less than in France, there are no co-payments; there is first-dollar coverage for hospital and medical services. Physicians in ambulatory care are paid predominantly on a fee-for-service basis, according to fee schedules negotiated between physicians' associations and provincial governments. In contrast to France, physicians in hospitals are most often paid on a fee-for-service basis, as in the United States.

There are few private for-profit hospitals in Canada such as French cliniques and American proprietary or investor-owned institutions. Most acute care hospitals in Canada are private nonprofit institutions. But their operating expenditures are financed through the NHI system. And most of their capital expenditures ?ire financed by the provincial governments.

In the United States, Canada's health system is typically depicted as a model for NHI (Andreopoulos 1975). Its financing, through a complex shared federal and provincial tax revenue formula, is more progressive than the European NHI systems financed on the basis of payroll taxes. Canada's levels of health Status are high by international standards. And it has achieved notable success in controlling the growth of health-care costs. What, then, are the problems in this system?

From the point of view of health-care providers, there is, above all, a crisis of underfinancing. physicians complain about low fee levels. Hospital administrators complain about draconian control of their budgets. And other health care professionals note that the combination of a physician "surplus" and excessive reliance on physicians prevents an expansion of their roles. Although Robert Evans (1987) contends that Canadian costcontrol policies cannot be shown to have jeopardized the quality of care, providers and administrators, alike, claim that there has been deterioration since the imposition of restrictive prospective budgets.

Leaving aside the issue of quality, the same issues discussed in the context of France are present in Canada, with respect to economic efficiency. Neither the hospital physician nor the patient have an incentive to be economical in their use of health care resources. On the demand side, since patients benefit from what is perceived as "free" tax-financed first-dollar coverage, they have no incentive to choose cost-effective forms of care. For example, in the case of a demand for urgent care, there is no incentive for a patient to use community health centers rather than rush directly to the emergency room.

On the supply side, physicians lack incentives to make efficient use of hospitals which are essentially a free good at their disposal. There are no incentives for altering input mixes to affect practice style (technical efficiency). Nor are there incentives for providers to evaluate service levels and the kinds of therapy performed in relation to improving health status (allocative efficiency). It could be argued that these problems are common to all health systems. But they are especially acute in a system characterized by a bilateral monopoly that tends to support the status quo. On the one hand, providers have strong monopoly power which they use to defend their legitimate interests; on the other, the monopsony power of sole source financing (under Canadian NHI) keeps provider interests in check at the cost of not intervening in the organizational practice of medicine.

Stoddard (1985) has characterized the problems of the Canadian health system as "financing without organization." In his view, Canadian provinces "adopted a 'pay the bills' philosophy, in which decisions about service provision ? which services, in what amounts, produced how, by whom, and where ? were viewed as the legitimate domain of physicians and hospital administrators" (Stoddard 1985: 3). The result of this policy is that provincial governments were concerned about maintaining a good relationship with providers. This concert) has not avoided tough negotiations and occasional confrontations. But there has been no effort to devise new forms of medical-care practice, e.g. HMOs or new institutions to handle the growing burden of long-term care for the elderly. The side effect of Canadian NHI has been to support the separation of hospital and ambulatory care and to reinforce traditional organizational structures.

As in France, or the United States, there are, in essence, two strategies for managing the Canadian health system and making adjustments. The first involves greater regulation oil the supply side - even stronger controls oil hospital spending, more rationing of medical technology, more hospital closures and mergers and eventual prohibition of extra billing. The second involves increased reliance upon market forces oil tile demand side various forms of user charges such as co-payments and deductibles now advocated as forms of privatization. Neither strategy is likely to succeed on its own. The former will control health?care expenditures in the short run but it fails to affect practice styles. Its effectiveness runs the risk of exacerbating confrontation between providers and the state and jeopardizing health care needs. The latter deals with only part of the problem - the demand side - and neglects the issue of supply side efficiency. It provides no mechanism by which consumer decisions can generate signals to providers to adopt efficient practice styles. Moreover, it is likely to raise the level of total (public and private) expenditures.

Due to the deficiencies which may occur if each strategy is followed independently, Stoddard (1983) has devised an innovative proposal for the province of Ontario, one that relies oil the use of market forces while maintaining the full benefits of a compulsory and universal NHI program. His proposal, which lie calls "publicly financed competition," rests oil four principles:

Creation of three payment modalities on the supply side

Physicians would have the choice of practicing in solo or group practice in the fee-for-service modality, or accepting a capitation fee per person enrolled in their practices, or accepting salary payment in return for working in community health centers organized by the public sector. Fees in the fee-for-service modality mould correspond to the current fee schedule and extra-billing would be allowed to continue. The capitation rate would be based on the average cost of insured services per patient across all three payment modalities. Salaries as well as staffing, programs, and service mix in the community health centers Would be set by Ministry of Health planners.

Financing of NHI is unchanged

All citizens would pay for health care through the tax system as they currently do.

Choice of primary care provider

All citizens would continue to choose a primary care provider but they would have to commit themselves to one modality of selected primary care for a specified period of time. The NHI program would no longer cover services not sought from or approved by the primary care provider. All services used by each patient over the course of the year would be charged to the appropriate payment modality.

Calculation of premium for each payment modality

At the end of each enrollment period, the premium for each modality would be adjusted. based on its total costs. The least costly modality would then become the baseline which would be fully covered under the Ontario Health Insurance Plan. Patients enrolled in the two more costly modalities would have to pay the difference between the baseline and the higher premium.

Although these principles are not as elaborately developed as the French RSC model, they are equally provocative and present a serious challenge to the status quo. Since the relative premiums of the three modalities are calculated on the basis of the average per capita cost including utilization, there would be powerful incentives to reduce such utilization. Assuming government measures are taken to assure a minimum level of medical care quality across payment modalities, these four principles create a system in which the patient benefits from seeking an efficient provider and the provider benefits by choosing cost-effective styles of practice. The level of health benefits remains the same across the three modalities; access to care would not be restrained by user charges; and adverse selection between payment modalities would be carefully monitored by requiring open enrollment and eventually introducing premium adjustments which would take into account age, sex as well as health status.

Britain: internal markets and HMOs

Britain is the exemplar of a National Health Service. It is financed almost entirely through general revenue taxation and accountable directly to the Department of Health and Social Security (DHSS) and Parliament. Access to health services is free of charge to all British subjects and to all legal residents. But despite the universal entitlement, Britons spend only 5.9 per cent of their GDP on health care - one half of what Americans spend as a percentage of their GDP.

Although the NHS is cherished by most Britons, there are, nevertheless, some serious problems concerning both the equity and efficiency of resource allocation in the health sector. With regard to equity, in 1976 the Resource Allocation Working Party (RAWP) developed a formula for the allocation of NHS funds between regions (DHSS 1976). The formula represents one of the most far-reaching attempts to allocate health care funds because it incorporates regional differences in measures of health status. Slow progress is now being made in redistributing the aggregate NHS budget along the lines of RAWP, but substantial inequities still remain both from the point of view of spatial distribution as well as from the point Of View Of social class (Townsend and Davidson 1982).

With regard to efficiency, the problems are even more severe because NHS resources are extremely scarce by international standards. Since there is less slack, the marginal costs of inefficiency are higher than in western Europe or the United States. And since the NHS faces the same demands as other systems to make available new technology and to care for an increasingly aged population, British policy-makers recognize that they must pursue innovations that improve efficiency. But there are numerous institutional obstacles in the way.

The tri-partite structure of the NHS is, itself, a major source of inefficiency. Regional health authorities (RHAs) are responsible for allocating budgets to hospitals in their regions. Hospitalbased "consultants" are paid on a salaried basis with distinguished clinicians receiving "merit awards" and all consultants have the right to see a limited number of private fee-paying patients in "pay beds." Outside the RHA budget are family practitioner committees (FPCs) responsible for remunerating general practitioners (GPs), ophthalmologists, dentists, and pharmacists. The GPs are reimbursed on a capitation basis with additional remuneration coming from special "practice allowances" and fee-for-service payment for specific services, e.g. night visits and immunizations. Separate from both the RHAs and the FPCs are the local authorities (LAs) that are responsible for the provision of social services, public health services, and certain community nursing services.

Such an institutional framework creates perverse incentives to shift borderline patients from GPs to hospital consultants, to the community, and back to the hospital. GPs, for example, have no incentive to minimize costs and can impose costs on RHAs by referring patients to hospital consultants or for diagnostic services. NHS managers can shift costs from the NHS to social security by sending elderly hospitalized patients to private nursing homes. And, consultants can shift costs back on to the patient by keeping long waiting lists thereby increasing demand for their private services. As in France and Canada, neither the patient nor the physician in Britain bear the costs of the decisions they make; it is the taxpayer who pays the bill.

Three recent strategies, all of them inadequate, have attempted to deal with this problem (Maynard 1986b). The first came promptly with the arrival of the Thatcher government. After cautious attempts to denationalize the NHS by promoting a shift toward NHI and privatization, the Conservative government backed off when they realized that such an approach would not merely provoke strong political opposition but also increase public expenditure and, therefore, conflict with their budgetary objectives (McLachlan and Maynard 1982). Instead, the strategy was narrowed in favor of encouraging competition and market incentives in limited areas. To begin with, the government allowed a slight increase of private beds in NHS hospitals. In addition, it introduced tax incentives to encourage the purchase of private health insurance and the growth of charitable contributions. Also the government encouraged local authorities to raise money through the sale of surplus property and to contract out to the private sector such services as laundry, cleaning, and catering

The second response was the Griffiths Report, which resulted in yet another reorganization in the long history of administrative reform within the NHS. Mr Griffiths, the former director of a large British department store chain, introduced the concept of a general manager at the department (DHSS), regional, district, and unit levels. This individual is now presumably responsible for the efficient use of the budget of each level of the NHS. The problem, however, is that the tri-partite structure of the system remains unchanged; and the general managers have very little information about least-cost strategies (across the tripartite structure) for generating improvements in health status.

The third and most recent response to the problem of improving efficiency has been to reduce the drug bill. Since April 1985 the government has limited the list of reimbursable drugs and reduced the pharmaceutical industry's rate of return. These measures will help contain the costs of the only openended budget within the NITS. But there is no evidence that they will have any impact on the efficiency of health care expenditures.

The more innovative efficiency-improving ideas have been developed by Enthoven and Maynard. They concern the promotion of "internal markets" and HMOs within the existing system of entitlements provided under the NHS. The essence of these ideas is to create financial incentives for each district to provide its residents with the best medical care possible, even if it has to purchase services outside of its boundaries. The aim is to maximize the benefits of health service expenditures, as measured by some measure of health status, e.g. qualityadjusted life years (QALYs); or to minimize the costs of sustaining a given level of QALYs. It sounds entirely theoretical but cost-effectiveness studies can produce empirical results. Recent findings indicate that the cost of a QALY of hemodialysis in a hospital is fourteen times that of a corollary artery by-pass graft and more than fifteen times that of a hip replacement (Torrance 1984; Williams 1985).

Short of allocating the entire NHS budget so as to maximize QALYs, there are a number of efficiency-improving measures that could be taken in the short run. For example, to avoid long queues for elective surgery in some regions and excess capacity in others, incentives could be devised to reward those regions receiving what the British call "cross-boundary flows." Or to persuade GPs to prescribe economically, a system could be devised to allow GPs to share in the savings. Beyond these examples of internal markets, Enthoven and Maynard have proposed variations of all HMO Plan for the NHS.

In Enthoven's plan, which he considers a form of "market socialism," a district continues to receive a RAWP-based per capita revenue and capital allocation and remain responsible for providing health services to its resident population (Enthoven 1985). In contrast to the present system, however, it receives additional compensation for services provided to residents from other districts and it controls referrals to providers outside its district. fit short, the district controls all budgets within the tripartite structure and purchases health services from the most cost-effective sources outside its borders. In effect, it operates as an HMO. Consultants and GPs enter into a variety of contractual arrangements with district authorities and district authorities are free to enroll consumers near the borders of a neighboring district.

fit Maynard's plan the GP functions as a client budget holder (Maynard 1985). All Britons receive a voucher front the NHS which entities them to sign up with a GP of their choice. The voucher generates a per capita payment to the GP in return for the provision of comprehensive health care for a year, after which the patient can choose another GP. The GP is responsible not merely for providing primary care but also for purchasing hospital services front public or private hospitals.

Both plans would provoke rapid reorganization of the health sector in Britain. The Enthoven plan would shift power to district managers - far more than they now exercise following the Griffiths reforms. The Maynard plait would shift power to GPs who would need to hire managers to assist with HMO formation. Needless to say, the details of these plans require a great deal more study. But even at such a level of generality, what is most interesting is the extent to which they resemble new ideas in France and Canada.

HMOs and universal entitlement: the promise and potential pitfalls

Ideas about introducing HMOs and elements of market competition into national health systems with universal entitlement hold promise because they point to the possibilities of combining some of the best features in the United States, Canada, and western Europe. The French plait for RSCs, the Canadian proposal for publicly financed competition, and the ideas about internal markets and HMOs in Britain focus on combining the supply-side efficiency embodied in a well-managed HMO with the financial security of a universal NHI system. To the extent that such ideas can be made to work in practice, they would probably provide more realistic models for the United States than the present structure of health care financing and organization in either France, Canada, or Britain.

But are these new ideas for health policy feasible in either the United States or France, in Canada or Britain? It would be naive to, conclude without adding some cautionary observations.

The proposals we have examined rest on two important assumptions: first, that competition between health care organizations will increase efficiency in the allocation of resources; and second, that health care providers can be motivated to change their behavior in response to financial incentives. The first assumption fails to circumvent a fundamental characteristic of health care markets - "informational assynimetry." The prevailing uncertainty about the effectiveness of various forms of medical care and the inability of consumers to assess quality makes them likely to turn to physicians for advice. Economists have shown that in markets characterized by agent-principal relationships in which buyers and sellers are unequally informed, competition does not necessarily lead to efficiency (Arrow 1963). In traditional, fee-for-service medicine, financed by third-party payers, physicians are likely to err on the side of overutilization. In HMO-type settings, due to prepayment, financial incentives are reversed and there is a risk of under-utilization. HMOs may increase competition between providers more than traditional indemnity coverage, but given the special characteristics of the health sector, it is impossible to draw inferences oil the basis of economic theory about the impact of increased competition on welfare (Weisbrod 1983).

The second assumption fails to acknowledge that health care providers do not behave like profit-maximizing firms. Only a small fraction of hospitals in France, Canada, Britain as well as the United States, are proprietary institutions. In the main, they are public and non-profit organizations with powerful missions and community allegiances. As for physicians, although much of their behavior, particularly in France, Canada, and the United States, has an entrepreneurial character, they are, nevertheless, members of a highly reputed profession and have consequently internalized a powerful set of values and norms. The extent to which financial incentives will influence the behavior of health care providers is bounded by the psychological, cultural. and institutional context within which they work (Brown 1981).

If, in deference to realism, we relax these two assumptions, it is important to note that the combination of HMOs and universal entitlement betrays a number of potential pitfalls.

First, efforts to promote competition between RSCs in France, the three payment modalities in Canada, and districts or GI) client budget holders in Britain, may result in competition over attributes other than price and quality. The theory of monopolistic competition suggests that a system of competing health care organizations would lead to product differentiation. Competition may well be focused on features other than delivering medical care, for example, amenities, marketing, and advertising. Also, there is a risk of collusion between competing health care organizations, which may result in providers demanding government regulation to maintain their market share.

Second, efforts to promote competition create incentives for providers to engage in risk selection. This would result in health risks or expected medical care costs being distributed unevenly among RSCs, payment modalities, or client budget holders. Of course, in elaborating the operational details of all these proposals, attempts would be made to identify the health risks of all beneficiaries based on criteria such as age, sex, residence, and perhaps even health status and certain socio-demographic characteristics. Nevertheless, even if health care organizations are compensated for beneficiaries with higher health risks, whatever system of risk rating is used, studies based on the experience of Medicare's competing medical plans (CMPs) suggest that the possibilities for risk selection are abundant (Eggers 1980).

Third, efforts to change physician behavior by confronting them with new financial incentives are likely to place physicians in the uncomfortable position of choosing between their ethical obligation to do the most for their patients, their natural inclination to pursue their own interests and organizational constraints, which encourage them to contain costs. Such a position is bound to erode doctor-patient relationships with no assurance of efficiency improvements in the allocation of health resources.

Fourth, all of the above potential pitfalls suggest that the new ideas for health policy, which we have examined, would create extraordinary possibilities for "gaining the system." For example, to skim healthy young patients and keep away frail elderly patients, a French RSC or the Canadian payment modality based in community health centers might decide to invest in exceptional amenities for a new birthing center and I. under-service" geriatric cases. Or in Britain, GP client budget holders might make referrals with few limitations for young healthy patients but drastically restrict them far elderly people. Suppose the Gil makes too few referrals and lowers the drug bill too much in order to appropriate a larger share of the savings. Would there be sanctions? Who would monitor the system?

That there are possibilities for gaming tile system has led Alain Enthoven (1986) to recognize that consumers could not negotiate effectively oil their own. They would need "sponsors" to "manage the demand side (and) to make the market achieve desirable results." But this fact should not detract front tile promise held by the idea of combining HMOs and universal entitlement. It merely exposes the illusion that competing HMOs could somehow operate as all alternative to strong government regulation. Clearly in a system of competing HMOs under NHI or within all NHS, sponsors would demand vigilant government regulation.

Periodic open enrollment;
Standard benefit packages with minimum specified benefits;
Standardized information disclosure by all competing health care organizations of data oil utilization trends, per capita costs (including premiums and out-of-pocket expenses), hospital mortality and patient characteristics (including health status);
Spot checks oil the veracity of the disclosed information;
Monitoring of quality.

How would such a system of "regulated" or "managed" competition compare with existing forms of more centralized regulation and budget control in France, Canada, and Britain? Would it succeed in producing efficiency improvements, greater continuity of medical care, and more flexibility in decision Making procedures? The evidence, at this point, is too fragmentary to serve as any reliable guide for policy. Nor is it ever likely to persuade analysts who are predisposed to accept the competition/regulation dichotomy in health policy.(6) But the curious melange of competition and regulation implied by these ideas for health policy in France, Canada, and Britain, do suggest one conclusive proposition. Whatever reforms are pursued in the health sector, there will always be a number of underlying tensions - between the patient's desire to take extra precautions and mobilize a maximum amount of attention. irrespective of costs, and a collective desire to contain costs; between a clinician's inclination to err on the side of overprovision, at the margin, and an HMO's or a government's rationale for making decisions on the basis of statistical averages; and between an HMO's or a government's persistent attempts to measure medical care activities and performance and the formidable difficulties of perceiving and measuring results.

Acknowledgements

I would like to thank james Knickinan, who originally suggested that I write this paper and subsequently commented extensively on its content. I am also grateful for Lorraine Mead's diligent secretarial assistance.

Notes

An often cited empirical basis for this presumption is a study by J. Newhouse, C. Phelps and W. Schwartz, "Policy options and the impact Of national health insurance," New England Journal of Medicine (1974) 290, 24: 1345-58.
See e.g. Andreopoulos (1975) and the more recent work of Robert Evans (1985).
Estimates of the uninsured range from 15 per cent to 20 per cent of the population. In 1984 the Current Population Survey estimated that 35.1 million people, 17.1 per cent of the population under 65, were without insurance. The percentage increases if one broadens the definition to include the underinsured and otherwise medically disadvantaged. See M.B. Sulvetta and K. Swartz (1986) The Uninsured and Uncompensated Care: A Chartbook, Washington, DC: National Health Policy Forum George Washington University, June.
The literature in health economics is abundant with examples of efificiency-improving changes in patterns of health care organization. For classic statements on this theme, see Fuchs (1975) and Enthoven (1978). For examples in France, Giraud and Launois (1985); in Canada, Evans and Robinson (1980); in Britain, Abel-Smith (1976) and Maynard (I 986b).
Daniel Fox (1986) traces the evolution of this pattern - what lie calls "hierarchical regionalism" - in an historical study of health policy developments in Britain and the United States.
The misleading nature of this dichotomy has been well analyzed by Luft (1985) and Schramm (1986), among others.

References

Aaron, H. and Schwartz, W. (1984) The Painful Prescription Rationing Hospital Care, Washington, DC: Brookings Institutions

Abel-Smith, B. (1976) Value for Money in Health Services, London: Heinemann.

Andreopoulos, S. (ed.) (1975) National Health Insurance: Can We Learn from Canada? New York: Wiley.

Arrow, K. (1963) "Uncertainty and the economics of medical care," American Economic Review 53, 5: 941-73.

Brown, L. (1981) "Competition and health cost containment," Milbank Memorial Fund Quarterly 59, 2.

Cochrane. A.L. (1972) Effectiveness and Efficiency, London: Nuffield Provincial Hospitals Trust.

DHSS (1976) Sharing Resources for Health it: England: Report of the Resource Allocation Working Party, London: HMSO.

Eggers, P. (1980) "Risk differential between beneficiaries enrolled and not enrolled in an HMO, Health Care Financing Review 4: 55-73.

Enthoven, A.C. (1978) "Cutting costs without cutting quality of care," New England Journal of Medicine 298, 22: 1,224-38.

_. (1980) Health Plan: The Only Practical Solution to the Soaring Cost of Medical Care, Reading, Mass: Addison-Wesley.

_. (1985) "Reflections on improving efficiency in the National Health Service." occasional paper, London: Nuffield Provincial Hospitals Trust.

_. (1986) "Managed competition in health care and the unfurished agenda," Health Care Financing Review, annual supplement.

Evans, R.G. (1985). "Illusions of necessity: evading responsibility for choice in health care," Journal of Health Policy, Politics and Law 10, 3, Fall: 431-67.

_. (1987) "Holding the line: the Canadian experience with global budgeting," in M. Berthod-Wurmser, V. Rodwin, Y. Soutey and J.C. Henrard (eds) Systeme de sante, pouvoirs public et financeurs: qui controle quoi? Paris: Documentation Franqaise.

Evans, R.G. and Robinson, G. (1980) "Surgical day care: measurements of the economic payoff," Canadian Medical Association Journal 123: 873-80.

Fox, D.M. (1986) Health Policies, Health Politics: The British and American Experience 1911-1965, Princeton, NJ: Princeton University Press.

Fuchs, V. (1975) Who Shall Live? New York: Basic Books.

Giraud, P. and Launois, R. (1985) Les Reseaux de soins, medecine de demain, Paris: Economica.

Hirschman, A. (1970) Exit, Voice and Loyalty, Cambridge, Mass: Harvard University Press.

Launois, R., Majitoni-d'Intignano, B., Stephan, J.C., and Rodwin. V.G. (1985). "Les reseaux de soins coordonnes (RSC): propositions pour une reforme profonde du systeme de sante " Revue Franqaise des Affaires Sociales I (January-March): 37-61.

Luft, H.S. (1981) Health Maintenance Organizations: Dimensions of Performance, New York: Wiley.

_.(1985) "Competition and regulation," Medical Care 23 (May): 383

McLachlan, G. and Maynard. A. (eds) (1982) The Public/Private Mix for Health: The Relevance and Effects of Change, London: Nuffield Provincial Hospitals Trust.

Maynard, A. (1985) "HMOs in the NITS? Health Summary 2, February: 5.

_. (1986a) Annual Report opt the National Health Service, York: Centre for Health Economics.

_.(1986b) "Financing the UK National Health Services, Health Policy 16: 329-40.

it OECD (Organization for Economic Cooperation and Development) (1985) Measuring Health Care, 1960-1983, Paris: OECD.

Rodwin, V.G. (1981)"The marriage of national health insurance and La Medecine Liberale in France: a costly union," Milbank Memorial Fund Quarterly 59: 16-43

_. (1984) The Health Planning Predicament: France, Quebec, England and the United States, Berkeley, Calif.: University of California Press.

_.(1987) "Le controle des pouvoirs publics et des payeurs: comparaisons internationales," in Systeme de Sante, pouvoirs publics et financeurs: qui controle quoi? M. Berthod-Wurmser, V. Rodwin, Y. Soutey and J.C. I Henrard (eds), Paris: Documentation Franqaise.

Schicber, G. and Poullier, JAI. (1986) "International health care spending," Health Affairs, Fall.

Schramm, C. (1986) "Revisiting the competition/regulation debate it) health care cost containment," Inquiry 23: 236-42.

Stoddard, G. (1983) "Publicly financed competition in Canadian health care delivery: a viable alternative to increased regulation?" in Proceedings 4) the Conference on Health in the '80s and '90s and Its Impact on Health Sciences Education. Toronto: Council of Ontario Universities.

_.(1985) "Rationalizing the health care system," in T. Courchene. 1). Conklin and G. Cook (eds) Ottawa and the Provinces: The Distribution of Money and Power, Toronto: Ontario Economic Council Conference.

Torrance, G.W. (1984) "Health status measurement for economic appraisal," paper presented at the Aberdeen meeting on Health Economics.

Townsend, P. and Davidson, N. (eds) (1982) Inequalities in Health: The Black Report, Harmondsworth: Penguin.

Weisbrod, B. (1983) "Competition in Health Care: A Cautionary View," in J. Meyer (ed.) Market Reforms in Health Care, Washington, DC: American Enterprise Institute.

Williams, A. (1985) "Economics of coronary artery bypass grafting," British Medical Journal 291, August 3: 326-9.