A Journal of Philosophy, Applied to the Real World

Motives and Markets in Health Care

Department of Philosophy, University of Wisconsin-Madison


The truth about health care policy lies between two exaggerated views: a market view in which individuals purchase their own health care from profit maximizing health-care firms and a control view in which costs are controlled by regulations limiting which treatments health insurance will pay for. This essay suggests a way to avoid on the one hand the suffering, unfairness, and abandonment of solidarity entailed by the market view and, on the other hand, to diminish the inflexibility and inefficiency of the control view. It suggests that the way to mitigate these problems is to recognize the malleability of motivation and the range of factors, in addition to financial incentives, that may influence the behavior of patients and especially physicians.


A health-care system aims to promote the health of members of its target population and thereby to contribute to their opportunities, autonomy, and well-being. Because resources are limited and have other valuable uses, the satisfaction of health needs must be efficient, and the health-care system must also act fairly and in a way that respects the individuals it serves, often at times when they are extremely vulnerable. How should citizens arrange the details of the health care system so as to accomplish all this? Health-care and health-insurance administrators, caregivers, and members of the population at large must be motivated to undertake actions that promote autonomy, fairness, efficiency, and, of course, health.

In this paper I am going to consider the problems that arise in determining what choices should be open to patients and health-care providers (more specifically physicians) and how they should be motivated to make the best choices. These questions are interdependent. If patients and providers cannot be motivated to make good choices, then there is reason not to give them choices. In this essay, I shall focus exclusively on incentives for patients and health-care providers (mainly doctors). I shall not consider questions concerning incentives for those who are neither patients nor doctors or incentives for purchasing insurance, and I shall have only a little bit to say about motivating people to choose healthy activities or habits. Nor shall I have anything to say about nudges (Thaler and Sunstein 2008; Schwartz 2011)—that is, structuring choice situations to take advantage of flaws in human deliberation. My focus is on rational decision-making by doctors and patients concerning specifically medical treatment and prevention of disease and disability and on the institutional design that determines what decisions they can make, what incentives influence their decision-making, and what determines their responsiveness to incentives.1

The first three sections in this essay examine two views concerning how to organize a health care system, which I call respectively, “the control view” and “the market view.” The market and control views are extreme positions, neither of which can be implemented without some admixture of the other. Rather than arguing for a specific compromise, I shall in sections 4 and 5 sketch a “third way.”

The market view assimilates patients to consumers and physicians to profit-seeking firms or employees in such firms, and it relies on economic incentives to allocate health care and to control costs. Libertarians espouse this view, and it is implicit in some Republican-Party proposals concerning health-care reform. The control view, in contrast, allocates resources by relying on constraints on physician and patient choices, mainly in the form of insurance provisions and best-practice guidelines. No actual health-care system entirely excludes individual choice, and none leaves everything to the market. The fundamental contrast between the two views lies in the favored method of allocating health-care resources. The market view relies on physician and patient choices subject respectively to profit seeking and budget constraints, while the control view limits costs via insurance regulations directing doctors and patients to make efficient health-care choices. As I shall argue, factors other than market choices or health-care regulations are and should be at work.

The control view

Many health care needs are difficult to predict and expensive to satisfy. Centralized provision of health care that buffers individuals from these risks is accordingly an attractive policy, which is realized in most affluent countries. But public spirit will not induce patients to economize, and if health care is free there is little else. If health care is subsidized but not free, individuals will have a self-interested as well as a public-spirited reason to economize, though as subsidies increase, an individual’s financial stake in economizing diminishes. Individuals who are ill and who seek medical care are mainly concerned about the probable outcomes and burdens of alternative treatments. With respect to the outcome, an individual cares about what health states the treatment is expected to lead to and the consequences of those health states for the individual’s opportunities, autonomy, and well-being, and for family and friends. With respect to the burdens of treatment, the individual cares about the discomfort, disability, and distress treatment involves, its expense, and its consequences for family and friends. From the patient’s perspective, the social costs of treatment alternatives and the burdens a patient places on the health-care system and the public purse are usually of little direct concern, though there is no reason to suppose that individuals have no concern at all about the social burden they may be imposing. Patients most of all want the best treatment, taking into account the parameters sketched above; and to get that treatment they want health-care providers who are motivated to serve their interests and who are competent to do so.

Because health care decisions are complicated, patients need expert advice, and because so many different factors bear on the interests of patients, they need doctors who are willing and able to get to know their aspirations, preferences and circumstances well and who will provide them with relevant and accurate information and counseling about the alternatives and their consequences.2 The ideal physician from the patient’s perspective is something like a caring, patient and selfless big sister, who is also a master of the relevant medical science. Which feasible institutions will generate health care providers that approximate this ideal?3 Licensing, regulation, standards, remuneration, and sanctions of all sorts shape the behavior of physicians. But a great deal that goes on in the doctor’s office is beyond the reach of regulation.

If doctors prioritize the interests of their patients, then it is hard for them to control costs. The control view accordingly assigns the responsibility for controlling health-care costs to those who regulate the health-care system, including, crucially, those who specify health-insurance coverage. They set the rules determining which treatments and services caregivers can provide. Treatment choices that would use too many resources are ruled out, typically by requiring patients to pay for them out of pocket, though some treatments can be excluded by law or by recommendations of professional groups. For example, the British National Health Service will not pay for drugs that are not sufficiently cost-effective. The U.S. is (in theory) an exception with its untenable commitment to provide every effective treatment, regardless of cost, to the old and the poor.

With sufficient constraints concerning general policies and health-insurance reimbursements, patients and caregivers can be freed from any responsibility for economizing. Within the constraints, patients can think about their health care exclusively from a personal perspective, and doctors can (in principle) be caring, patient and selfless experts devoted entirely to the interests of their patients.

This strategy of universal coverage has considerable advantages. It simplifies the decisions that patients and caregivers have to make, and in principle it provides the same health care to everyone except the few who are wealthy enough to pay out of pocket for uncovered treatments. It enhances the confidence that patients can place in their doctors. It is a wonderful thing to have confidence that your doctor will do everything that he or she can for your welfare—provided, of course, that this confidence is justified.

On the other hand, this method of economizing restricts the range of options available to physicians and patients. General rules about which treatments can be provided are blunt instruments. Individuals differ in their conditions, circumstances, interests, and responses to treatments. Any practicable scheme of resource allocation that restricts the choices of doctors and patients will be insensitive to many relevant differences among patients. For example, a costly test whose results would be of little importance to one patient may relieve an overwhelming anxiety in another. Insurance company or government regulations preempt decisions by patients and recommendations by physicians, and these regulations will inevitably be faulty.

The market view stated and refined

These drawbacks to employing centralized control to determine practices and control costs motivate the market view. According to the market view, individuals, with the advice of health-care providers, decide on their own treatment, subject to the constraints imposed by their budgets. Since individuals have to bear the costs of their decisions, no more will be spent on health care than the members of the population want to spend, and their choices control costs without any centralized rationing. Given their budgets, individuals ration for themselves: Just as individuals decide on what to eat and on how much of their budgets to spend on food, so individuals should decide on what health care to purchase.

Any realistic formulation of the market view requires several important qualifications. Health care, unlike commodities such as DVDs or services such as hair styling cannot be left entirely to the market. There are four reasons why. These do not imply that the market view is untenable. They show instead that some qualifications and some refinements are needed. The first difference between health care and hair styling is that health care, like food, has become a basic need. Just as most people are not prepared to abandon their fellow citizens to starve, so they are not prepared to let them die for lack of simple medical treatment. Defenders of the market view consequently call for subsidies or vouchers for those who cannot otherwise afford to meet their needs.

Second, there are health-related public goods, such as the monitoring of potential epidemics, that cannot be left to the market. The development of medical knowledge is also an important public good.

A third well-known complication with health-care markets lies in the asymmetries of information concerning appropriate care in the doctor-patient relationship. The protections that markets provide for customers at the supermarket or mall, which consist in the threat that customers will take their business elsewhere, are inadequate in medical markets, which expose individuals to risks of exploitation and grave harm. Licensing and regulation of health-care providers may be needed, though some defenders of the market view may argue that that the gravity of the threats will call into existence private agencies to address these threats by gathering (and selling) information concerning the quality of health-care providers.

The fourth distinctive feature of health care markets derives from the fact that, unlike the need for food or shelter or less pressing goods, individuals often find themselves with unanticipated expensive and compelling health-care needs. Even if it were possible to save for such contingencies, it would be inefficient if people had to do so. Any tenable version of the market view envisions that individuals who cannot afford to self-insure—a large majority of the population—will purchase health insurance.

The fact that a large portion of health care must be paid via insurance brings with it two difficult problems, which, to varying extents, affect insurance generally. The first of these is adverse selection: Since individuals know more about their health than insurance companies do, insurance policies will be a better deal for those with private information that their need for health care is greater than average. For example, insurance plans covering pregnancy and childbirth will attract women who are pregnant or who plan on becoming pregnant. Those buying insurance will thus tend to incur higher medical expenses than the average person, and premiums must be high enough to cover these costs. Higher premiums will in turn discourage healthier people from purchasing insurance, shrinking the pool of the insured and forcing still higher premiums. Unless insurance companies counteract this adverse selection, health insurance markets collapse. Counteracting adverse selection requires that private insurance policies for individuals exclude those with pre-existing conditions and refuse to cover conditions that people can choose, such as pregnancy. But our collective benevolence precludes abandoning those without insurance. Private health insurance must thus be supplemented with some sort of public provision for those who cannot insure themselves.

Health insurance also creates problems of moral hazard. Once insurance is in place, people face weaker incentives to economize on health care and, to a lesser extent, to care for themselves. Markets for health care and health insurance can counteract moral hazard problems by deductibles, co-pays, or coverage limits. Regardless of the effectiveness of these measures, there are limits both to the amount that individuals are willing to spend for health insurance and to the amount of taxes they are willing to pay to provide government subsidies for the poor or for government subsidized insurance for those with pre-existing conditions. So there will be limits on the amount that is spent on health care. Defenders of the market view can conclude that, provided that government does not gum up the works, cost containment is automatic.

The market view models patients as consumers, with a preference ranking over different bundles of commodities and services and an overall budget constraint, and it models health-care providers as firms that seek to maximize their net financial returns or as employees of such firms. Although conceiving of patients as consumers imposes a particular structure on their wants, it is compatible with almost the whole range of things that patients are concerned about. Regarding providers as profit-maximizing firms or as employees of such firms, on the other hand, drastically limits the factors that influence their behavior. It is a misleading idealization.

Even with (a) vouchers for those who are poor (in order to address the fact that health care is a basic need), (b) licensing and regulation (to mitigate asymmetries in information concerning treatments), (c) provisions for public health and medical research, and (d) health insurance supplemented with aid for those who are excluded from the insurance market, leaving health care to the market remains problematic. Egalitarians will be repelled by the deep inequalities that result, and the actual inefficiency of private health care insurance in the United States compared to systems of public health insurance or provision in other wealthy countries should make one suspicious of the theoretical arguments in defense of the efficiency of markets.

Defense of a refined market view

Yet, as we have seen, planning and control have serious problems, too. Insurance regulations concerning what treatments will or will not be paid for are bound to be inflexible and insensitive to the differences in individual needs and preferences. The fact that markets protect individual choice makes them appealing. Provided that government policies address the pressing problems of fairness caused by inequalities in wealth and adverse selection within the health-insurance market, the market view might appear to be the better alternative for controlling costs, because it relies on individual choice rather than bureaucratic fiat to allocate resources.

Consider, for example, routine mammograms to screen for breast cancer and routine PSA tests for prostate cancer. The tests are individually inexpensive, but they are used so widely that they are cumulatively costly (and they lead to expensive follow-up tests and treatments). Their benefits are variable and dubious. Moreover, their use and hence their costs can be anticipated. Whether or not these tests are advisable, unlike an x-ray for someone who may have broken a bone, routine mammograms and PSA tests are not called for by an unanticipated condition. They are as expected as a regular oil change in one’s car, for which no one buys insurance.

Whether it is advisable for particular women to have routine mammograms or whether it is sensible for particular men to have routine PSA tests depends on their medical history and how much weight they place on lowering the risk of death as compared to incurring the risks and side-effects of diagnostic procedures and treatment. The tests are inexpensive enough that individuals can purchase them for themselves, but the interference of government and interest groups in the form of “best-practice” guidelines and insurance rules encourage choices that may be inappropriate for many individuals.

Permitting individuals to decide for themselves (in consultation with their doctors) has the potential of economizing more flexibly and with fewer limitations on individual freedom. Moreover, defenders of the market view argue that if individuals paid for these tests out of pocket, competition would drive their costs down.4 If one makes financial costs irrelevant to individual decision making by requiring that insurance policies cover the tests—as states or the Federal Government do—individuals have no financial incentive to economize, and the commitment to reimburse individuals for these tests encourages individuals to have them. Requiring coverage for these tests may also persuade patients that these tests are beneficial: Why else would reimbursement be required?

Government is not the only culprit. With respect to the PSA test, the American Cancer Society has issued the following statement:

The American Cancer Society does not support routine testing for prostate cancer at this time because we believe proper pretest guidance and education is necessary. Doctors and other clinicians should provide information on the potential risks and benefits of PSA testing to appropriate patients, allowing them to make an informed decision on testing.5

Yet the American Cancer Society also supports requiring insurance policies to cover the test on the grounds that, “Prostate cancer screening should not be prevented because of the reimbursement limitations of health insurance plans.” Unless the American Cancer Society maintains that no medical procedure that is potentially beneficial to anyone should be precluded by reimbursement limits, some explanation is needed for why PSA tests are especially beneficial. Moreover, almost every man in the U.S. can afford a $30 PSA test, and special provision can be made for the few who are so poor that $30 is a significant barrier to obtaining the test. For those who are not indigent, if the test is not worth $30 to them and there is no compelling reason of social policy to provide the test, why should insurance policies (and hence, ultimately, everyone purchasing insurance) be required to pay? With a stronger commitment to markets, the meddling of non-governmental organizations such as American Cancer Society would be less consequential.

The following table helps clarify what forms of payment are most efficient with respect to which procedures:

Efficient Financing


Low cost

High cost

Expected procedures

Direct payment


Unexpected procedures

Direct payment


If costs are low, it is efficient for people to self-insure and save the administrative insurance overhead. If costs are high but can be anticipated, insurance is less efficient than saving or borrowing, again because of overhead costs. Concerns about fairness call for subsidies for those who cannot afford the procedures. Insurance is the most efficient way to finance health care only when costs are high and unexpected. Within a market model, the only role for government (apart from prosecuting fraud) is to provide the subsidies or vouchers demanded by charitable concern for those who are ill and cannot pay and by fairness concerns. Left to themselves, people will not purchase insurance when they can more efficiently self-insure, and health care will be more efficient.

There are two complications. First, low cost expenditures can add up, and when there are many low-cost expenditures, it is misleading to place them in the first column in the table above. What is relevant is not the cost of individual procedures or treatments but total costs to patients. Second, insurance is also a means by which society can pool risks and influence the distribution of health care. Health insurance can help insulate the distribution of health care from inequalities in wealth. In addition, in the absence of insurance, those in bad health, whether this is anticipated or not, would face higher expenses than those in good health; and as a matter of justice and social solidarity, we may not want those who are already unfortunate with respect to their health also to suffer financially. One way to mitigate the financial consequences of poor health is, of course, insurance. Efficiency is not the only consideration, and insurance for anticipated high cost procedures may be called for.

It is possible to control costs entirely by government regulations that specify what services and treatments doctors can provide, just as it is possible to run an economy by government fiat. But central control generally does not work very well. It limits freedom and creates inefficiencies. Market economies work much better than centrally controlled economies, because markets roughly align individual interests with social benefit and make far better use of information. Yet the “curses” of our benevolence, our social solidarity, and our concerns with distributive justice place such serious constraints on market allocation of health care, that the case for markets sketched in this section is far from decisive.

Superseding the market-control dichotomy: Motivating doctors

Given the asymmetries in information concerning appropriate treatment, the stakes in the decisions, and the emergency settings in which decisions must often be taken, what maximizes doctors’ incomes will often differ from what serves the interests of patients. On the other side of the relationship, in a market system, patients facing difficult medical decisions are forced to consider costs. This complicates the decision making of patients (and the advising of physicians, particularly when there are multiple insurance plans with different coverage limits). Moreover, charging patients introduces morally troubling inequalities. If treatment were free and the remuneration of physicians did not depend on which treatments they provide, most of the troubling inequalities in treatment could be avoided, medical decision making would be simplified, and it might be possible to align the interests of patients and physicians more closely. But costs must be controlled. Are we stuck with relying on markets or control or on some combination of the two?

Any way to mitigate the need for control must motivate doctors and patients to make choices that benefit patients and control costs. If these motivating factors are financial costs and rewards, then we’re back to the market view. What makes possible a third way is the recognition that there are other incentives in addition to financial incentives. This third way will not eschew all control and all financial incentives, but it will rely heavily on non-financial incentives that are ignored by the market and control views. Let us consider, in very general terms, the factors that can push doctors and patients to make choices that efficiently promote health, welfare, autonomy and fairness.

Let us begin by considering the incentives that influence the choices doctors make. The next section considers what motivates patients. Julian Legrand (2006) distinguishes two exaggeratedly simple characterizations of the motivations of care providers. Either they are “knaves,” who are motivated by self-interest, or “knights” who are altruistic. Economists typically model everyone as knaves, concerned exclusively with their own interests, and one of the drawbacks of economics is the extent to which it legitimizes and encourages the motivations that it pretends are already typical (Frank, Gilovich and Regan 1993). The market view usually follows economists in taking both patients and providers to be self-interested. Although Legrand argues persuasively that this crude distinction between knights and knaves clarifies disputes between social-democratic and neo-liberal policy makers, this distinction is (as Legrand recognizes) not particularly useful in understanding what motivates health-care providers. There are four reasons:

First, although Legrand distinguishes “knights” from “knaves,” the important distinction is really between knavish and knightly actions. A cut-throat bond trader may be a pussycat at home with the kids. Shaping choices in particular contexts need not require shaping characters.

Second, many different motivations count as self-interested, and many different motivations count as altruistic. Dr. Smith, who is consumed by vanity and cares about little besides the admiration of patients and colleagues, is as self-interested as Dr. Jones, who cares only about her annual income; but Smith and Jones are unlikely to treat their patients in the same way. If Smith succeeds in his objective of securing the admiration of his patients, he will be regarded as altruistic (as well as skilled), and his treatment of his patients resembles the treatment that a genuinely altruistic doctor would provide. Jones might simulate a deep concern for her patients as a way to boost her fees, but she is likely to appear more self-interested than Smith. Conversely both doctors who are exclusively concerned with their own patients’ health and those who are also concerned about conserving resources for other patients may be “knights.” But their objectives and behavior differ.

Third, many important motivational factors are neither self-interested nor altruistic. Consider, for example, trustworthiness or reciprocal altruism. Whether a doctor keeps her promises is a separate question from whether she is self-interested or altruistic.

Finally, motivations are mixed. Many things motivate doctors. Some of these are self-interested, others altruistic, and still others neither self-interested nor altruistic.

Although I do not have sociological evidence to support the following claim, it is not controversial to maintain that the following four considerations are among the factors that motivate doctors: (1) their patient’s interests, (2) devotion to public service (including the promotion of medical knowledge), (3) self-interest, and (4) legal and normative constraints. What constitutes the self-interest of doctors varies, but the main self-interested considerations include remuneration, job satisfaction, status, colleagues’ respect, and avoiding malpractice suits. I shall not attempt to rank the considerations, because their relative importance is bound to vary among doctors. Some doctors care more about their patients or about the good opinion of others than other doctors. In any case, it is a serious mistake to assume that doctors are motivated only by their pecuniary interests, both because it is false and because it legitimizes and encourages such motivation.

If one wants to design health-care institutions so as to promote health in a way that is fair, flexible, respects patients’ autonomy, and economizes on resources, the designer needs to know what is wanted of doctors. One view, which exempts doctors from any responsibility to conserve resources, holds that they should provide treatments that are as responsive as possible to the interests of the patients, where the patients themselves, if competent, largely define what is in their interests. But given the large and unavoidable asymmetries in medical knowledge, doctors should not simply sell the permitted services that their patients want to buy. They should also help the patient define what constitute the best outcomes, the least burdensome treatments, and the justifiable claims on insurance benefits (whether public or private). In many cases, patients effectively delegate the choice of treatments to their physicians. In emergencies, there is no realistic alternative, and in many non-emergency decisions, the factors are too complicated for patients to make up their own minds. Rather than offering a menu of available treatments and then carrying out the one the patient independently selects, doctors should use their technical expertise and their detailed knowledge of the particular patient to rank the items on the menu. Doctors must usually defer to the patient’s wishes (or at least to their refusals), but within this limit, they exert a good deal of influence.

To help with the problems of allocating scarce health-care resources, doctors need to economize on the cost of treatments they recommend. When there is no difference in the expected benefit to the patient, the doctor should employ treatments that use fewer resources, and where benefits are slight and costs are large, doctors should recommend the less expensive alternative. However, as is obvious, the obligation to provide the best treatment to patients may conflict with an obligation to conserve resources. When more expensive treatments are better for patients, devotion to the best interest of the patient pushes one way, while economizing on the use of resources pushes the other way. Trading off efficacy, cost, and burden to the patient is difficult, and unless the best interest of the individual patient is given absolute priority, there is bound to be conflict. Different institutional designs will mitigate or aggravate this conflict. The challenge is to design a set of institutions that motivates doctors to make sensible trade-offs between their patient’s interests, which are paramount, and controlling costs.

Let us begin by considering incentives that can motivate doctors simply to serve their patient’s interests. The design of institutions that nurture the needed skills and motivations must begin with the four general motives that physicians already have, even though by means of training physicians and selecting medical students, designers of the health care system can influence the mix of motives they build on. Some of these motivations already push in the right direction and only need to be supported and strengthened. First, doctors often care about their patients, and that care affects their behavior. Some of their patients are friends and neighbors, and even when they are strangers, doctors recognize the grave responsibilities of providing medical care. For their patients’ sake, doctors care about whether treatments are successful and not too burdensome. This intrinsic altruistic concern is grounded in sympathy and benevolence as well as in a commitment to the duties of a physician. It is strengthened and made more effective by closer personal ties between the physician and the patient. A system such as the British National Health Service that encourages (or used to encourage) long-term relationships between patients and general practitioners helps to develop a caring relationship and enables doctors to get to know patients better and make better judgments about what is in their patients’ interests.6 In contrast, a system like that at the University of Wisconsin, where I teach, that provides a financial inducement for changing from one HMO to another as their premiums differ, disrupts doctor-patient relations. Similarly, the change from a system whereby one’s doctor directs one’s hospital care to a system of hospitalists weakens the link between physician and patient and hence the intrinsic incentives motivating doctors to promote their patients’ interests. Although probably more expert with respect to some medical issues, the hospitalist will typically never have met the patient before and will be unlikely to have the same personal relationship to patient or the detailed knowledge that comes with it.

Other ways in which the concerns of physicians for their patients can be strengthened lie in selecting the right individuals to become doctors and nurturing a caring culture among doctors. This entails on the one hand encouraging relatively selfless and caring individuals to become physicians.7 Higher incomes may attract greater talents, but at the same time, they may also attract students who are more interested in their future incomes than in caring for others. Lower incomes, especially if they result from spending more time with patients, might secure better doctors. Selection to medical school might also be made to depend more heavily on character judgments, but I am skeptical the ability of admissions committees to assess applicants’ characters, which will in any case, adapt to the culture and norms of the profession.

Establishing and supporting social norms that encourage doctors to become involved with their patients is crucial to this vision of medicine. Doctors need to be rewarded rather than punished for spending enough time with their patients to get to know their needs and concerns—despite the fact that from a short-run perspective, longer visits may appear to waste resources. Office schedules and technology can be helpful or harmful. For example, the shift to computerized records, which has obvious advantages, may result in greater physician engagement with the computer than with the patient.

With respect to securing physicians’ commitment to their patients and enhancing their knowledge of what is truly in their patient’s interests, beneficent concern for their patients is the best motivator, because it does not rely on the contingent and imperfect coincidence between the other motives and the desired outcome. But it is not a perfect motivator, and too much personal involvement can be dangerous. As the AMA stated in the mid-19th century, “the natural anxiety and solicitude which he [the doctor] experiences at the sickness of a wife, a child, or anyone who by the ties of consanguinity is rendered peculiarly dear to him, tend to obscure his judgment and produce timidity and irresolution in his practice.”8 I doubt, however, that too much beneficence is likely to be a frequent problem. Even with the best institutions, beneficence is limited; and other motivations come into play.

If doctors are to play a role in cost containment, they cannot be concerned exclusively about their own patients. Legal regulation, best-practice guidelines, and social sanctions are ways in which doctors can be motivated to economize while at the same time serving as additional ways that doctors can be motivated to take to heart the interests of their patients. The social expectations of doctors coupled with the status they enjoy create incentives to provide good and socially responsible care even when doctors have no benevolent concern for their patients and no generalized altruistic concern about cost containment or population health. These social expectations are enforced, as they should be, by formal legal and social sanctions. Carelessness and neglect are punished with malpractice suits, revocations of hospital privileges, and state disciplinary actions. There are also informal sanctions, which may be severe.9 Losing the respect of one’s colleagues is a serious penalty.

Normative constraints backed by legal or social sanctions, unlike benevolent concern for patients, are not always a good thing. They can lead to behavior that is harmful to patients and wasteful of resources. Malpractice suits enforce norms of acceptable treatment, but at a cost of significant resources devoted to litigation and to defensive medicine. Replacing malpractice suits with an alternative way of compensating patients who have been harmed would probably save resources, but it might also weaken incentives for good treatment. Designing a set of social norms and sanctions that will direct physicians to serve their patients and to conserve resources is a complicated matter, involving sophisticated social modeling, experimentation and, especially, a realistic appreciation of what motivates doctors.

Next we come to financial incentives, which are obviously of great importance. If doctors’ incomes rise with more tests and procedures, patients are likely to be subjected to more tests and procedures.10 If physicians have to bear some of the cost of the tests and procedures they prescribe, patients will be subjected to fewer. It might seem that the ideal compensation scheme would break any connection between remuneration and the choice of treatments and procedures. But the question is subtle, because testing and treatment have non-monetary costs and benefits. One way to counteract the failure to economize, which is an inevitable consequence of motivating doctors to serve their patients and removing costs constraints, is to give doctors a financial stake in performing fewer tests or a limited budget that can be used to purchase treatments for their patients (Hausman and Legrand 1999). Getting the financial incentives right is very hard. Financial incentives like those that are common in the U.S., which give doctors a stake in more expensive treatment, are costly and are among the factors that render the U.S. health care system so wasteful (Gawande 2009). On the other hand, financial incentives to do less, while lowering costs, could be harmful to patients both directly and by undermining trust in the doctor’s motivations.11

That brings us to the last of the motivating factors, which I described vaguely as devotion to public service. I have in mind the fact that medicine is a calling not just to help those specific individuals who happen to be one’s patients but to relieve the burdens of disease generally, insofar as it lies within the capacities of a single physician to do so. The commitment to public service includes the obligation to help people in emergencies, regardless of whose patients they may be, to volunteer in clinics to treat strangers, to share observations and insights that may help other physicians to improve their practice, and to participate in studies and experiments designed to improve medical knowledge. The strength of these motives varies widely, and they depend heavily on institutional details. A medical system that requires all doctors to carry out extensive public service early in their careers (as is the case in South Africa), may discourage more self-interested individuals from entering the field and help to make salient general health needs. A medical system that leads medical students to incur large debts, in contrast, is likely to lead to a greater concern with earnings. The extent to which doctors are committed to public service affects public attitudes, which in turn strengthen or undermine the commitment.

The commitment to public service is one basis upon which to motivate doctors to economize in the use of health-related resources. Doctors can help to control health care costs, and in some regards, relying on doctors is preferable to relying on insurance regulations. Doctors have much better information about the needs of individual patients and they can be more flexible. But it is difficult to construct institutions in which doctors are able and motivated to make responsible compromises between the interests of specific patients and the interests of some wider population

Motivating patients to make a public health care system efficient

It is, I believe, impossible to rely on patients for any significant cost control, while at the same time meeting the demands of benevolence and fairness. Conservatives and defenders of the market model disagree. They believe that health-care costs can and should be controlled by the choices medical consumers make among insurance policies and among services paid for out of pocket. For example, eliminating reimbursements for expected inexpensive procedures such as mammograms or PSA tests and requiring that patients pay more out of pocket for their health care leads people to economize and reduces costs. But this a priori claim is apparently refuted by the apparently greater efficiency of state-run systems. It seems that the administrative and transactions costs of private insurance systems cancel out the savings that result from individual economizing.

Efficiency is, moreover, not the only consideration. If relying on individuals to economize leads to unfair or harmful states of affairs, then some other way of controlling costs needs to be found. In many cases there are moral reasons, including paternalist considerations, for discouraging economizing and instead subsidizing expected and low-cost procedures or requiring that insurance policies reimburse individuals for those expenses.12

Defenders of greater reliance on economizing by medical consumers as the means to control costs emphasize the moral hazard implicit in free or subsidized access to health care, which leads to overuse and to carelessness in personal care, both of which increase costs. They are right to point out the serious moral hazard problems health insurance creates, but private insurance is no less subject to these problems than state-supported health care.13 If insurance policies pay for treatments and drugs, which may be extremely costly,14 patients will choose the more efficacious or less burdensome treatment, regardless of the cost. Deductibles will not help much. Co-pays will help, but if patients only have to pay one-fifth or one-tenth of the total, their influence may be small; and if one makes them larger, then one creates serious inequities. It is moreover implausible to believe that an ethic of individual responsibility condemning taking more than one’s fair share of the social resources devoted to health and welfare will do much to control costs, even though such an ethic, encouraged by the advice of physicians can make some contribution. Ultimately, much of the burden of controlling cost will rest on insurers who will cover only some of the potentially efficacious treatments.

If the government explicitly denies coverage for certain treatments, rather than permitting the anonymous insurance market coupled with individual choice and financial constraints to deny coverage, complaint, criticism and controversy is inevitable. But in any feasible market system, this criticism will not disappear: only it will be directed toward private insurance companies rather than the government. On the one hand, hiding behind private insurers has huge political advantages. On the other hand, if a public authority rather than private insurers dictates coverage limits, then they can be defended as serving the public good rather than private profit.

Proponents of the market view might argue that coverage limits in privately purchased health insurance reflect individual economizing, since they result from individual choices among alternative insurance policies rather than rationing decisions by insurance companies. What prevents individuals from receiving more expensive treatments are alledgedly their own decisions concerning which insurance plan to purchase. So rather than doctors or some third party telling patients what treatments they can have, patients decide for themselves (subject to a budget constraint) when they are choosing among insurance policies.

This view is implausible. It ignores limitations on patients’ time, knowledge, and effort. Although people no doubt know that policies have different costs and that they differ in which treatments they cover and in their co-pays or deductibles, few can make fine-grained choices among policies that differ with respect to whether they cover a specific treatment. (And given problems of adverse selection, it is questionable whether insurance markets would function at all if individuals had such choices.) It is implausible to maintain, as some defenders of market views do, that choices among insurance plans (as opposed to a choice of treatments, hospitals, and physicians) should be an important objective of health-care policy.


While limiting health insurance mainly to unexpected and expensive treatments increases efficiency, only small savings can be achieved by increasing the financial incentives on individual patients to economize on the use of health care resources, unless one is prepared to broach the limits imposed by our benevolence, solidarity and sense of justice. The incentives that matter most are those that the health-care system imposes on health-care providers, and many of these are not financial. Of course, financial incentives matter, but they may be harmful rather than beneficial, and benevolence, solidarity and equity limit the extent to which we can rely on the market. The non-financial incentives I have discussed lessen the need for direct control or for financial incentives and market allocation, but they do not eliminate these needs. To control costs equitably requires, unfortunately, explicit rationing decisions on the part of insurance administrators or regulators, whether public or private. This leads to a measure of inflexibility and inefficiency, which we have to live with. How damaging that inflexibility and how great the inefficiency depends on our success in motivating patients to acquiesce in the regulatory limits they face and in the medical advice they receive and especially in motivating doctors and other health-care professionals to act in the interests of their patients and in the public interest.

Acknowledgements: I would like to thank Lydia duBois for her research assistance, and I would like to thank Norman Fost, J. Paul Kelleher, and an audience of bioethicists at the National Institute of Health for helpful criticisms and suggestions.


Abelson, Reed and Julie Cresswell. 2012. “Hospital Chain Inquiry Cited Unnecessary Cardiac Work.” New York Times, August 6, 2012.

Cuffel, Brian, Joan Bloom, Neal Wallace, Jaclyn Hausman, and Teh-wei Hu. 2002. “Two-Year Outcomes of Fee-for-Service and Capitated Medicaid Programs for People with Severe Mental Illness.” Health Service Research 37: 341-359.

Douglas, Sharon. 2009. “Consider Ethics, Patient Rights Before Treating Your Immediate Family.” Amednews.Com. Ethics Forum. Posted April 13, 2009. http://www.ama-assn.org/amednews/2009/04/13/prca0413.htm. Accessed August 23, 2012.

Frank, Robert, Thomas Gilovich, and Dennis Regan. 1993. “Does Studying Economics Inhibit Cooperation?” Journal of Economic Perspectives 7: 159-72.

Gawande, Atul. 2009. “The Cost Conundrum: What a Texas town Can Teach Us About Health care.” The New Yorker, June 1, 2009. http://www.newyorker.com/reporting/2009/06/01/090601fa_fact_gawande

——— 2012. “Big Med.” The NewYorker, August 13, 2012 http://www.newyorker.com/reporting/2012/08/13/120813fa_fact_gawande

Glazier, Richard, Julie Klein-Geltink, Alexander Kopp, Lyn Sibley. 2009. “Capitation and Enhanced Fee-for-Service Models for Primary Care Reform: A Population-Based Evaluation.” Canadian Medical Association Journal 180: E72-81.

Grant, Ruth. 2012. Strings Attached: Untangling the Ethics of Incentives. New York: Russell Sage Foundation.

Hausman, Daniel and Julian LeGrand. 1999. “Incentives and Health Policy: Primary and Secondary Care in the British National Health Service.” Social Science and Medicine 49:1299-1307.

LeGrand, Julian. 2006. Motivation, Agency, and Public Policy: Of Knights and Knaves, Pawns and Queens. Oxford: Oxford University Press.

Schwartz, Peter. 2011. “Questioning the Quantitative Imperative: Decision Aids, Prevention, and the Ethics of Disclosure.” Hastings Center Report 41(2): 30-9.

Thaler, Richard and Cass Sunstein. 2008. Nudge. New Haven: Yale University Press.

1. Unlike Ruth Grant, who uses the term “incentive” narrowly “to mean an extrinsic benefit deliberately designed to alter behavior,” I shall use the term “incentive” in a broad sense that includes any “of the factors that influence our choices or motivate action” (Grant 2012, pp. 38-9).

2. There are, of course, alternative institutional arrangements in which doctors have more limited roles, but I shall have nothing to say about them in this paper.

3. This view may be old fashioned and from the perspective of the new industrialists of medical care (Gawande 2012), it may appear outdated. In their view, the future of medical care lies in diagnostic and therapeutic technology delivered by teams of anonymous experts. For example, at a large recent gathering of a chapter of Health Occupations Students of America, a hospital administrator gave a keynote address listing ten reasons for seeking a career in health services. Two thirds of the reasons consisted of factors such as high salaries, job security, portability, and flexibility, social status and so forth. Only at the end—almost as an afterthought—did the speaker mention that health care workers help to make people healthier, and there was no mention at all of the satisfactions and challenges attached to comforting people in need, allaying their anxieties, and helping them to come to terms with changes in their lives or their approaching deaths.

4. Although mammograms can be had for $75, charges in some hospitals are nearly $900. Defenders of the market view argue that if individuals paid for these tests out of pocket, these disparities would diminish and the average cost would be under $100. See http://clearhealthcosts.com/2011/08/25/the-cost-of-a-mammogram-a-boston-new-york-rivalry/ (accessed July 23, 2012). The cost of PSA tests varies six-fold from $22 to more than $120. See http://health.costhelper.com/psa-testing.html (accessed July 23, 2012). Another source (http://www.joepaduda.com/archives/002181.html—accessed July 23, 2012) puts the cost at $70-$200 (plus the cost of an office visit). Home PSA testing kits are available on the web for $30. Since insurance companies have strong incentives to limit reimbursement rates and both the knowledge and the capacity to do so, it is by no means obvious that market prices for these tests would be lower. Direct payment by individuals does, however, avoid the administrative overhead required by third-party payment.

5.http://www.cancer.org/Cancer/ProstateCancer/MoreInformation/ProstateCancerEarlyDetection/prostate-cancer-early-detection-state-screening-coverage-efforts (accessed July 19, 2012).

6. Personal ties are also encouraged by small practices, but these are burdensome for physicians and, as the technology of primary care expands, they are also uneconomical.

7. See footnote 4 above.

8. Quoted in Douglas 2009. When the close ties between doctor and patient are reciprocal, which is likely if they exist at all, the autonomy and informed consent of the patient may also be threatened.

9. For example, a front-page article in the New York Times reports an investigation by Medicare of a suspiciously large number of heart catheterizations by doctors at HCA hospitals in Florida. (“Hospital Chain Inquiry Cited Unnecessary Cardiac Work” By Reed Abelson and Julie Creswell, August 6, 2012). The unfavorable publicity (with the doctors’ names and pictures included) and the potential legal proceedings provide a powerful (though apparently not always decisive) incentive against such behavior.

10. See for example, Atul Gawande, “The Cost Conundrum: What a Texas town can teach us about health care.” The New Yorker, June 1, 2009. http://www.newyorker.com/reporting/2009/06/01/090601fa_fact_gawande

11. The evidence seems to be that shifting from fee-for-service payments to capitation fees is not in fact harmful to patients. See for example Cuffel et al. 2002 and Glazier et al. 2009.

12. One would think, for example, that those concerned to eliminate abortion would support the requirement that health insurance reimburse women for the cost of birth-control pills. According to an Associated Press article in the New York Times (October 4, 2012), providing free contraceptives to a sample of 9,000 teens in St. Louis reduced the rate of both unintended pregnancies and abortions to less than half the national averages.

13. I doubt moreover that the most serious problems of moral hazard can be cured with high deductible policies, because the most serious problems do not lie with too many visits to primary care physicians or diminished self-care. No doubt people with insurance see their physicians more often than they would if they had to pay. But the trouble, anxiety, and waiting involved in seeing a physician constitute strong non-financial incentives against frivolous visits to the doctor, and the risks of broken bones, obesity, or cancer give people good reasons to avoid dangerous and unhealthy behavior, even if they are confident that medicine can lessen the bad effects. Although the availability of anti-retroviral drugs has probably led people to take fewer precautions against contracting AIDS, how many people decide not to wear seat belts because they have comprehensive health insurance?

14. With different patent policies the costs of drugs and appliances could be much lower. But the political power of the pharmaceuticals companies probably takes this possibility off the table.