Clinical Loyalties and the Social Purposes of Medicine

Today, more than ever, physicians practice amid conflicting expectations and pressures from patients and third parties. Health plans and hospitals employ economic leverage to an unprecedented degree to influence clinical decisions. Myriad government and private actors award entitlements and opportunities based on medical judgment. Courts call upon physicians to opine on issues affecting litigants' rights and responsibilities. More generally, health professionals face growing pressure to serve ends that fit awkwardly with the ideal of fidelity to patients.

This article reviews the conflicts between social purposes and fidelity to patients that vex clinical practice. It then considers the principal legal and ethical paradigms by which these conflicts have been managed. These paradigms are inadequate, I contend, because they either deny or unsuccessfully finesse the reality of contradiction between society's expectations of medicine. Neither medical ethics nor law can resolve this contradiction; rather, we must mediate it, case by case, in myriad clinical circumstances.

"In every house where I come," the Hippocratic Oath promises, "I will enter only for the good of my patients."¹ Yet the use of clinical methods and judgment for purposes other than patient well-being has a long history. For centuries, physicians have treated wounded soldiers to return them to combat, given medical opinions to legal authorities, and used clinical techniques for religious and cultural purposes. In the face of scarcity or threats to public health, they have compromised patients' individual interests for the benefit of others.

Uses of clinical skill for purposes other than promotion of patients' individual interests can be grouped into 3 categories: (1) pursuit of public health aims, (2) furtherance of non–health-related social ends, and (3) ascription of rights, responsibility, and opportunity based on health status. These categories overlap, and each encompasses diverse clinical activities, some of which have achieved wider ethical approval than others.

By clinical activities, I refer to uses of medical skill in work with individual human subjects. This encompasses diagnosis and treatment of patients, evaluations for courts and other third parties, and use of medical techniques on human subjects for purposes other than patient care. It excludes uses of biomedical knowledge that do not engage physicians personally with human subjects, eg, crafting of government policies. It also excludes conduct toward others that does not entail medical skill—eg, behavior toward coworkers and service in administrative and political roles. The ethical and legal challenges these activities can present lie beyond this article's scope.

Physicians often incorporate public health concerns into their diagnostic, therapeutic, and other clinical decisions. In so doing, they sometimes sacrifice patients' individual interests to advance community-wide ends, typically without triggering ethical objections. Vaccination is illustrative. When immunization rates are high and the incidence of the feared disease is minimal, vaccination of any one individual may hold more risk (of adverse effects) than benefit for that person. The justification for vaccination in such circumstances lies in its long-term collective benefit: high immunization rates prevent epidemics.^{2 - 3} Other examples include choosing antibiotics with an eye toward slowing the evolution of resistant bacterial strains⁴ and reporting contagious disease to public authorities. Medical research with human subjects represents a variation on this theme, to the extent that subjects bear risk for the benefit of future patients.

Clinical skills also serve social purposes unrelated to health, including military readiness, criminal justice, control of violence, and religious observance. Some interventions, eg, treatment that returns wounded troops to combat, yield health benefits that in turn further nonmedical ends. Other measures, eg, male circumcision, serve nonmedical purposes without yielding clear health benefits⁵ or even, as with execution by intravenous injection,⁶ by inflicting injury. Uses of clinical methods to serve purposes unrelated to health tend to generate more controversy than does service of public health ends, even when the non–health-related purposes at issue are widely seen as legitimate. Objections to medication of prisoners to quell physical resistance^{7 - 8} and to restore competence to stand trial⁹ or be executed ^{10 - 11} are illustrative.

Personal rights, responsibility, and opportunity often turn on health status. As arbiters of health status, physicians thus frequently act as gatekeepers when these are at stake. Psychiatric opinion in insanity defense cases is a classic example. Clinical evaluation can also be decisive in civil law matters, including competence to make contracts¹² and wills,¹³ damages in personal injury cases,¹⁴ and disability insurance claims.¹⁵ When medical assessment of this sort yields favorable results for the person being evaluated (eg, access to desired employment or insurance benefits), such gatekeeping tends not to trigger controversy; indeed, it is often seen in positive moral terms. Clinical gatekeeping that yields unfavorable outcomes for evaluees (eg, criminal punishment) more often evokes objections.¹⁶

The rise of managed care has created new gatekeeping. On behalf of health care payers, physicians now regulate patients' access to expensive medical services.¹⁷ When this gatekeeping is performed by a physician remote from the bedside—eg, a utilization manager uninvolved in a patient's care—it does not present the problem of conflicting clinical obligations that is the focus of this article. Lacking a clinical relationship with the patient, the utilization manager is not burdened by the expectations of fidelity that characterize the patient's relations with caretakers. A different picture prevails when payers ask physicians who attend patients to act simultaneously as guardians of an insurance pool. Clinical gatekeeping of this sort, typically performed by primary care physicians with financial and other incentives to withhold services,^{18 - 19} presents the problem of conflicting purposes starkly.

Managed health plans defend such gatekeeping as necessary to rein in unnecessary clinical spending. Indeed, evidence suggests that putting physicians at risk for the cost of their decisions through capitated payment, income withholding, and performance bonuses restrains medical spending more effectively than do other methods used by payers.^{20 - 21} Plan managers say their quality control efforts prevent skimping on needed care.²² But appraisal of medical necessity is an uncertain enterprise, fraught with ignorance about the comparative efficacy of clinical options and veiled conflict over the balancing of benefits and costs.

This new form of gatekeeping explicitly challenges the premise that physicians should act toward their patients as faithful fiduciaries. Yet potent moral arguments, tied to the social advantages of medical spending restraint, have been marshaled to justify it.²³ Indeed, some leading health policy commentators have portrayed the professional ideal of exclusive devotion to individual patients as an outmoded concept, ripe for replacement by a model in which physicians cooperate with health plan managers to make medical practice more frugal.^{24 - 25}

As the foregoing review indicates, clinical role conflict pervades medicine. Conflicts of obligation have occasioned myriad legal and ethical disputes—over duties to patients vs the public health, the justice system, insurers, and other social institutions.

Courts, commentators, and professional bodies have typically considered these disputes in a piecemeal fashion, without relying on overarching conceptions of professional duty in the face of conflicting expectations. Yet the outlines of 3 prevailing paradigms for addressing such conflict can be discerned: (1) insistence on undivided commitment to patient well-being as the sole ethical guidepost in conflict situations, (2) dismissal of patient welfare as ethically irrelevant when physicians act clinically on behalf of third parties, and (3) reliance on consent to justify actions or arrangements that depart from the principle of undivided commitment to patients.

In some commentary, undivided commitment to patients is the sole guidepost for clinicians who face conflicting expectations. This ideal is variously rendered in the Hippocratic Oath, the maxim "first, do no harm,"²⁶ and the bioethics principles of beneficence and nonmaleficence.²⁷ Denunciations of physician involvement in human rights abuses²⁸ and capital punishment²⁹ commonly invoke this ideal. Some critics of cost-conscious gatekeeping by treating physicians hold that denial of potentially beneficial care is unethical per se because it violates this ideal.^{30 - 31} Along similar lines, courts have suggested that physicians owe their patients a duty of undivided loyalty and vigorous advocacy when insurers try to avoid paying for care that physicians deem necessary.³²

This absolutist approach is open to the charge that it takes a penurious view of the medical profession's public responsibility. Few would argue with the view that physicians should steer clear of complicity in torture or other human rights abuses. But the premise that physicians should not use their skills on behalf of such legitimate social purposes as public safety, justice, and control of health spending is at odds with robust conceptions of the profession's social responsibility.

A classic response to this concern is the claim that the public interest is best served by undiluted professional commitment to individuals' well-being.³³ This answer appeals as an escape from the problem of conflict between public service and private duty, and it is a corollary of the more general, politically appealing idea that the public interest is best advanced by individuals' pursuit of their personal aims. But this claim is unproven and untestable. A more modest response is that clinical ethics lacks means for balancing duties to patients and society. Ethical debate about clinical aims, Alan Stone has argued, reflects a "dialectic" of obligations: "help your patient" and "do no harm."³⁴ Beyond these patient-oriented confines, Stone suggests, medical ethics has nothing to say.

Yet the ubiquity of clinical work that furthers social ends belies the proposition that all such activity is beyond the pale. To be sure, the frequency of a practice does not, in itself, show that it is ethical. Yet, as the above review of social purposes suggests, medical judgment and skill have enormous moral import beyond the bedside. Medicine's myriad social applications reflect moral beliefs about responsibility and opportunity, individual rights, and preservation of public health and safety. A medical ethics decoupled from public morality would be at best insensitive and at worst arrogant. This suggests the need for a broader ethical dialectic—one that encompasses medicine's social functions as well as its therapeutic and caring purposes.

The failure of traditional medical ethics to encompass this broader dialectic has encouraged reliance on a sharply different paradigm. This paradigm, more often tacitly followed than openly stated, justifies clinical work for third parties on the ground that this work does not create a patient-physician relationship. Absent this relationship, the paradigm holds, clinical work for third parties is not medical practice. The subjects of such clinical work are not patients; thus, the possibility of harm (or benefit) to them is ethically irrelevant.

Assertions along these lines have appeared in varying contexts, some notorious in the recent history of medicine.^{35 - 36} Physicians who determined detainees' "fitness" for torture under authoritarian regimes in the 1970s and 1980s maintained that their work served state campaigns against subversion and thus should not be judged by the ethics of patient-physician relations.¹¹ Likewise, US physicians who performed nontherapeutic radiation experiments in the 1950s privately justified their actions in national security terms, eschewing patient-oriented ethics as a subsidiary concern.³⁷

Until recently, open reliance on such arguments was deemed unseemly. Forensic psychiatry is a case in point. Attempts to specify the ethics of forensic practice long incorporated the premise that examining clinicians should look out for evaluees' well-being. To be sure, understandings of this duty varied. Some authorities saw it as central to the forensic examiner's role,³⁸ while others held that forensic practitioners serve the justice system first and their evaluees second.³⁹ Some said medical testimony should advance evaluees' legal interests,³⁸ while others saw duty to examinees as limited to referral for treatment.⁴⁰ But the proposition that psychiatrists can entirely disregard evaluees' welfare remained beyond the bounds of respectability.

In the 1990s, however, the claim that consequent harm is ethically irrelevant when physicians serve the state or other third parties began to appear in mainstream ethics discourse. For several years, forensic psychiatrists have urged the American Psychiatric Association (APA) and the American Medical Association's Council on Ethical and Judicial Affairs (CEJA) to adopt this stance toward forensic evaluation.^{41 - 42} Thus far, neither the APA nor CEJA has done so overtly. But a 1995 CEJA report on the ethics of psychiatry on death row⁴³ approvingly referenced an article asserting that "the forensic psychiatrist in truth does not act as a physician" and that concern about consequent harm is thus ethically "mistaken" in forensic work.⁴⁴

Attempts to put clinical work for third parties beyond the regulatory reach of medical ethics authorities are engaging courts and legislatures. In 1994, an Illinois judge ordered the state's medical disciplinary board to proceed against several physicians who allegedly assisted an execution by lethal injection,⁴⁵ in violation of ethical proscriptions.^{6 ,29} Within weeks, the Illinois legislature passed a law stating that physician attendance at executions does not constitute medical practice and is thus beyond the disciplinary board's jurisdiction.⁴⁶

Managed health plans have sought similar immunity for their prior-approval programs. Disciplinary complaints against gatekeeping physicians have been filed in several states by both patients and physicians disturbed by denials of coverage. Health plans and physician defendants have responsed that coverage determinations are not medical decisions and thus lie outside the jurisdiction granted to medical boards by state medical practice acts.⁴⁷

The claim that assessment of medical need on health plans' behalf is not medical practice will probably play a central role on another legal front. Under current law, people who obtain health insurance via the workplace (ie, most Americans) cannot recover from their health plans in state or federal courts for harm deemed to arise from negligent administration of benefits.⁴⁸ However, health plans (and gatekeeping physicians) can be held liable under state law for negligent medical practice.⁴⁹ This dichotomy sets the stage for litigation over whether myriad gatekeeping programs devised by health plans constitute the practice of medicine.

Efforts to put clinical work beyond the regulatory reach of medical ethics and law disclaim the problem of consequential harm by characterizing it as irrelevant to the ethics and legality of the professional conduct at issue. This strategy involves large risks. To begin with, immunization of spheres of clinical work from the ethics and law of medical practice is not readily subject to analytic constraint. If, for example, consequent harm is ethically irrelevant when psychiatrists do forensic evaluations, why should harm to human subjects matter when physicians perform risky nontherapeutic research or execute convicts by lethal injection? Absent such distinctions, the claim that an activity lies beyond the reach of medical ethics expands to encompass all that clinicians might do on behalf of third parties.

The moral tensions of forensic practice have inspired an attempt to draw such lines. Forensic psychiatrists who claim that consequent harm is ethically irrelevant to their work distinguish between administering the law and other social purposes. But this argument covers more than its proponents would like. It encompasses, for example, medical methods of punishment, including lethal injection. More generally, if the social value of administering the law warrants denying the ethical import of harm, can't the same be said of social benefits that derive from clinical work on behalf of any third party? Collective benefits may well bear on the ethics of such work. But the logic of social benefit cannot justify the claim that consequent harm to individuals should not count as a cost.

One might finesse this problem by positing that consequent harm is ethically irrelevant if lawfully inflicted. This would allow condemnation of some troubling activities, eg, assessment of fitness for torture (a breach of international law⁵⁰ ), while permitting such practices as forensic evaluation and cost-conscious gatekeeping by clinical caretakers. Unreflective reliance on law is, however, deeply problematic. It abdicates the task of explaining why law should determine the ethical import of consequent harm. Moreover, it ignores an especially troublesome aspect of such harm from a medical perspective. To the extent that a medical hand in doing harm is evident, public confidence in medicine as a caring enterprise is put at risk.¹¹

This risk animates the profession's objections to involvement in such high-profile lawful impositions of harm as execution by intravenous injection.²⁹ If, as I argue below, public trust in physicians as caring professionals is a social good, then this risk constitutes a social cost (and an ethical problem), even when the harm at issue is lawful. This cost may at times be worth bearing to achieve some social benefit. The ethical irrelevancy thesis neglects the need to strike such balances.

A more robust response to clinical role conflict looks to informed consent by clinical subjects to legitimize medical activities that benefit third parties at subjects' expense. This approach emerged in response to revelations of abuse of research subjects during and after World War II.^{51 - 55} Over the past 20 years, courts and commentators have eyed consent as the solution to an expanding range of role conflict problems. Prior consent has been invoked to legitimate forensic and other evaluations performed for nontherapeutic purposes,^{56 - 57} disclosure of confidential clinical data (eg, homicidal ideation) to prevent harm to others,⁵⁸ and release of medical records to health care payers for cost-control purposes.⁵⁹

More recently, proponents of competition between managed health plans as a way to contain medical costs have argued that subscribing to a plan constitutes consent to its subsequent rationing decisions.⁶⁰ This ex ante consent, they assert, authorizes plan physicians to withhold potentially beneficial care based on cost without saying so at the bedside. Thus, in a sense, the logic of consent has come full circle, from prophylaxis against clinical actions contrary to patients' interests to justification of professional authority to weigh the interests of others over those of the patient.

This plasticity reflects a problem rarely addressed in bioethics commentary. Simply put, the question of whether an act (eg, joining a health maintenance organization [HMO]) constitutes informed, autonomous consent to the consequences requires an underlying judgment about the moral acceptability of the actor's situation. By itself, the logic of consent is indeterminate. It draws its content from multiple, often unstated moral premises about the acceptability of limits on an actor's options and the adequacy of the actor's understanding of each option's consequences.⁶¹

Absent such moral content, human actions can be characterized in either voluntary or deterministic terms.^{62 - 64} The armed robber's proposition, "your money or your life," is a classic example. It presents the listener with a choice between 2 actions (and their consequences). Our belief that this proposition coerces reflects our sense that the options posed are morally repugnant. Analogously, the physician who asks a person to agree to forensic evaluation poses a choice, as does the firm that offers its employees membership in a health plan. Whether the person who assents to forensic examination thereby consents voluntarily to the legal consequences is a morally laden question. Its resolution rests on prior judgments about the legitimacy of the choice situation and the needed level of knowledge about potential legal outcomes. Likewise, whether a woman who joins an HMO thereby consents to her physician's subsequent refusal to approve a mammogram cannot be answered without prior moral judgments about the adequacy of her health insurance options and the minimum acceptable level of subscriber knowledge about the HMO's coverage policies.

Mainstream bioethics theory, that invoked by courts and government commissions, has yet to make a place for these underlying moral judgments. Instead, leading bioethics authorities tend to present the question of whether a person has consented autonomously to something as an empirical inquiry.^{27 ,37 ,65} When there is consensus on the underlying moral issues, this analytic incompleteness is not a problem at the bedside. The law of informed consent, for example, has empowered patients in therapeutic settings.⁶⁶ But this accomplishment rests upon rough agreement on supporting moral judgments—about the necessary degree of disclosure and the proper limits on physicians' attempts to persuade.^{61 ,63}

With regard to many of the role conflicts that vex US medicine, we are nowhere near to agreement on the moral questions that underlie claims of prior consent. Since mainstream bioethics theory does not admit the presence of such questions, it lacks analytic tools for limiting the plasticity of the logic of consent. Proponents of the legitimacy of any number of clinical activities that serve third parties can find words or acts to characterize as prior consent. This analytic failure should hardly be surprising, since the key underlying questions, which concern the moral force of sundry reasons for departing from the ideal of undivided loyalty, are the very quandaries that claims of prior consent purport to avoid.

The prevailing paradigms for coping with clinical role conflict either deny or unsuccessfully finesse the moral problems such conflict presents. The proposition that physicians should never stray from undivided commitment to patient welfare is belied by the social, economic, and cultural import of health status and by the ubiquity of clinical work that serves nonclinical ends. The notion that some clinical work on behalf of third parties lies beyond the scope of traditional medical ethics is both difficult to contain and neglectful of the requisites for bedside trust. The paradigm of consent allows tacitly for the striking of different balances in different contexts between the interests of clinical subjects and other parties. By itself, however, the logic of consent is unrevealing about how these balances should be struck.

Can we do better? The need for myriad contextual moral judgments precludes resolution by a simple, elegant algorithm. We would make more progress, I submit, by reconceiving the work of constructing an ethics of role conflict in more modest terms, as an ongoing effort to mediate, case by case, between clinical fidelity and medicine's social purposes. We could thereby come to appreciate conflict between medicine's caring and social purposes as in need of ongoing management, rather than as a problem to be solved once and for all. Such an approach should be sensitive to both the moral import of medicine's social purposes and the ethical significance of fidelity to patients. A detailed account of how these concerns might be mediated in diverse clinical contexts is beyond this article's scope, but I shall suggest a general framework.

First, we should recognize the ethical value of clinical work on behalf of socially desirable ends. We cannot appraise this worth precisely, case by case, but we can value such work in rough proportion to our understanding of the social benefit it yields.

For example, when society links a right (or responsibility) to medical criteria, the ethical weight we give, on the positive side, to clinical assessment of a subject's qualifications for this right (or responsibility) should roughly correspond to the moral significance we attach to reliance on medical criteria. Thus, psychiatric examination bearing on criminal responsibility should be valued in proportion to the moral import we ascribe to mental state, understood medically, as a basis for determining culpability. Analogously, the ethical value we confer on cost-conscious gatekeeping by primary care physicians should be tied to the moral weight we give to controlling health care costs.

When a social end or activity is undesirable, any clinical contribution toward its pursuit should be weighted negatively. Medical assessment of prisoners' tolerance for torture⁶⁷ is an example. When a social purpose is a subject of moral controversy, spillover of this controversy into debate over the ethics of clinical work that serves the purpose at issue is inevitable. We cannot expect medical ethics to resolve questions of public morality that society as a whole has not put to rest.

Though medicine's social purposes merit moral recognition, fidelity holds a privileged place in clinical ethics. Mediation between clinical fidelity and medicine's social purposes should take account of what we sacrifice, in ethical terms, when physicians fail to keep faith with their clinical subjects. Bioethics commentary and judicial opinions generally present fidelity to patients in instrumental terms, as a precondition for patient trust. Such accounts tally the medical benefits of trust⁶⁸ —patient willingness to share sensitive information and comply with recommended treatment,⁶⁹ as well as confidence in physicians' clinical explanations.⁷⁰

This understanding of clinical fidelity suggests a mediating strategy that seeks to balance the public benefits of clinical work on behalf of social purposes against the therapeutic costs of breaches of clinical trust. Yet this instrumentalist picture is incomplete. It neglects, even denies, the affective experience of the victim of infidelity—the person who trusts and is betrayed and is thus intimately wounded by another person.⁷¹ Quite apart from any adverse health consequences, the physician's infidelity violates another, often when the other is most vulnerable.

The ethic of clinical fidelity safeguards an intimate sphere of personality against such violation.⁷² As concrete individuals, we crave this safety without regard for its justifiability in cost-benefit terms. Disembodied consideration of the medical benefits of clinical trust fails to take account of this. Balancing, from a disinterested perspective, between the public benefits of medical work that serves social ends and the therapeutic cost of clinical breach of trust would likewise fail to capture all that is lost when physicians break faith at the bedside for social purposes.

To be sure, construction of ethical rules requires that the passions of real people be viewed from some distance⁷³ ; otherwise, collective interests would be without weight. But the idea of an ethics of professional role entails a shortening of this distance, tailored to situations of special human vulnerability.⁷² From this reduced distance, collective concerns retain importance, but keeping faith with concrete persons merits special regard.

This suggests a vantage point—somewhere between the disinterested cost-benefit perspective and private expectations of loyalty and fears of betrayal—for mediation between the ethic of clinical fidelity and the moral weight of medicine's social purposes. From such a vantage point, these social purposes and the instrumental benefits of patient trust should both "count," but protection against betrayal is a privileged matter. Put differently, if my metaphorical mediator were a real person, he or she would be self-consciously biased toward people's expectations of fidelity.

This approach cannot mechanically resolve role conflict questions, but it can frame discussion to take account of the public and private values most at stake. It can also channel the working through of inevitable disagreements—about the moral import of particular social purposes and the medical and other consequences of particular departures from clinical fidelity.

This approach presents a threshold question: Does the physician conduct at issue occur within a clinical relationship? Absent personal engagement between physician and clinical subject, medical work that serves social purposes does not implicate the fidelity concerns considered herein. Thus, for example, a physician who performs preauthorization review for an insurer without personally evaluating patients does not break faith with them, although this practice may raise other issues. The same is the case for physicians who give advice to public health or law enforcement agencies that does not derive from a particular clinical relationship.

The existence of a clinical relationship introduces interpersonal fidelity concerns. When the relationship is transient and the subject knows its nontherapeutic purpose, these concerns are typically small enough to be outweighed by the accompanying social benefits. Clinical examinations to assess eligibility for disability benefits are illustrative. Yet even in such cases, the potential for breach of faith is nontrivial. Examinees may be better informed cognitively than affectively about an evaluator's nontherapeutic purpose. Culturally shaped expectations of professional benevolence, encouraged by a clinician's empathic manner, may engender feelings of trust and, later, violation.¹¹

To a point, we must tolerate such ethical awkwardness, so long as health status remains a basis for bestowing rights and opportunities. But it may at times be too great, even when personal engagement is fleeting. Objections to physician-assisted executions are illustrative. However brief (and unempathic) the physician's attendance, whether to supervise the lethal injection or merely to pronounce death, complicity in the punitive taking of life constitutes a singular and dramatic form of abandonment.⁷⁴ Not only does this abandonment break faith with the condemned; its high public profile invites wider skepticism about the profession's capacity to keep faith.⁶

When, by contrast, patients are themselves committed to the social purpose served by medical intervention, clinical fidelity is not compromised, even if the intervention endangers patients' physical or psychological well-being. The wounded soldier who enthusiastically rejoins his fighting unit after treatment that sacrifices long-term efficacy to achieve a speedy result is not a victim of breach of faith. The physician's duty of fidelity is to the patient as a person, not to a professional or other conception of well-being at odds with the patient's life aims. On the other hand, to the extent that a person's assent to a nontherapeutic purpose stems from external pressure, breach of clinical fidelity is a concern. Genital cutting by medical personnel of African and Middle Eastern women under community pressure to consent⁷⁵ illustrates this poignantly.

The more powerful the message of fidelity conveyed within a clinical relationship, the more compelling a social purpose should be to justify departure from the ethic of undivided loyalty. Health plans that make primary care physicians into gatekeepers, with strong incentives to deny access to beneficial care, pose a special problem in this regard. The promise of fidelity signaled by sustained, empathic engagement is painfully at odds with the reality of a double agenda.

Disclosure of this dual role, by either plan officials or clinical caretakers, hardly dispenses with the problem of infidelity. Expectations engendered by bedside empathy persist despite cognitive input to the contrary, especially when anxious patients yearn to believe. From this standpoint, cost-control methods that align caretakers' and payers' interests are more troubling than schemes that empower clinicians to act as patient advocates, at arm's length from plan managers. The arguable efficiency advantages of the former¹⁹ may be outweighed by corrosive effects on patient-physician relations.

These concerns bear greatly on the legal controversies that have marked the rise of managed care. They weigh in favor of regulation designed to limit physicians' financial risk to a small fraction of income, and they merit judicial recognition of professional duties of loyalty and patient advocacy vis-à-vis health plans.³² Remarkably, US health law has not yet widely recognized a duty of loyalty, distinct from professional obligations with respect to quality and confidentiality. Capitation schemes, penalties, and other incentives to withhold (or provide) services should prompt judicial suspicion when clinical loyalty is at issue. Physicians' ownership interests in laboratories and other facilities to which they refer patients should trigger similar judicial concern.⁷⁶

In addition, courts should not enforce provisions in contracts between physicians and health plans that bar disclosure of uncovered treatment options or financial arrangements between caretakers and payers. Moreover, judges should approach litigation over health plan deselection of physicians with an eye toward the potential use of deselection to align physicians with health plans' financial interests.

We cannot reconcile our public and our deeply personal expectations of medicine. Medical ethics can at best manage the tension between them. The approach I have urged here concedes the inevitability of conflict between medicine's social purposes and the ideal of clinical fidelity. This approach can help to frame a discourse about appropriate limits on clinical work that serves third parties, public purposes, or both. This discourse ought to acknowledge that medicine can legitimately serve such purposes, but it should affirm the special role of interpersonal fidelity in clinical work.

Understandings of the duties implied by clinical fidelity are bound to differ and to change over time. But at the core lies faithfulness to persons, not adherence to disembodied principles. Assessment of the social costs and benefits of breaches of fidelity should be done from a perspective sympathetic to the experiences of trust and betrayal.

Beyond the benefits considered herein, vigorous defense of clinical fidelity could revitalize the damaged self-conception and morale of the medical profession.⁷⁷ For a generation, the profession's much-cherished autonomy has been under attack, as both an instrument of paternalism and a cover for social welfare–reducing restraint of trade. To no small degree, such criticism has been merited. Yet the erosion of physician independence in the 1990s has both dispirited practitioners and rendered them less able to keep faith with patients.

By making a stand for clinical fidelity, medicine in the United States could reanimate its sense of purpose and promote those facets of professional autonomy likely to contribute to the common good. In so doing, physicians might also contribute modestly to reducing the insecurity and incivility in the United States fed by repeated collective experiences of breach of faith.