Article Text

Download PDFPDF

Are physicians obligated always to act in the patient's best interests?
  1. David Wendler
  1. Correspondence to David Wendler, Department of Bioethics, NIH Clinical Center, Building 10, Room 1C118, Bethesda, MD, USA 20892-1156; dwendler{at}


The principle that physicians should always act in the best interests of the present patient is widely endorsed. At the same time, and often within the same document, it is recognised that there are appropriate exceptions to this principle. Unfortunately, little, if any, guidance is provided regarding which exceptions are appropriate and how they should be handled. These circumstances might be tenable if the appropriate exceptions were rare. Yet, evaluation of the literature reveals that there are numerous exceptions, several of which pervade clinical medicine. This situation leaves physicians without adequate guidance on when to allow exceptions and how to address them, increasing the chances for unfairness in practice. The present article considers the range of exceptions, illustrates how the lack of guidance poses ethical concern and describes an alternative account of physician obligations to address this concern.

  • Obligations
  • Fairness
  • Best Interests
  • Risks
  • philosophy
  • education for healthcare professionals

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The image of the physician at the bedside, dedicated to the present patient, has been consecrated in art and literature. Guidelines translate this image into the principle that physicians are obligated to act in the best interests of the present patient.1 2 The American Medical Association directs physicians to regard as ‘paramount’ the interests of the patient they are treating.3 The International Code of Medical Ethics instructs physicians to give their patients ‘complete loyalty’,4 while the American College of Physicians holds that physicians' ‘primary commitment must always be to the patient's welfare and best interests’.5 Many commentators agree. Marcia Angell argues that physicians' ‘sole obligation’ is the care of the present patient,6 while Hellman and Hellman instruct physicians ‘always to act in the patient's best interests’.7

Commentators and guidelines also recognise individual exceptions to this ‘traditional’ view of physician obligations, sometimes within the same document. The World Medical Association's Medical Ethics Manual cites the statement in the International Code of Medical Ethics that ‘A physician shall owe his/her patients complete loyalty’.8 The ethics manual then goes on to note, ‘physicians may in exceptional situations have to place the interests of others above those of the patient.’ It might be possible to sustain both views if the exceptions to the traditional view were rare, perhaps limited to war and national emergencies. In fact, as a recent article notes, the extent to which physicians should consider the interests of individuals other than the patient in front of them has been ‘greatly underestimated and has not attracted significant scholarly analysis’.9 This claim is supported by a review of the literature that reveals at least 27 exceptions (table 1), a number of which pervade medical practice.

Table 1

Exceptions to the traditional view of physician obligations

The present situation, then, is one in which the traditional view is widely endorsed and exceptions to this view are widely acknowledged, yet little systematic guidance is provided regarding which exceptions are appropriate and how they should be handled.10 This situation is troubling. It leaves physicians without adequate guidance on how they should take into account the interests of others when providing patient care. It also leaves policy makers without the resources necessary to evaluate proposals for new practices and policies that conflict with the traditional view. How are policy makers to determine whether such proposals should be adopted when they do not have the means to evaluate which exceptions are appropriate and which are inappropriate? To address these concerns, we need to recognise and accept the pervasiveness of appropriate exceptions to the traditional view and develop an alternative understanding of physician obligations, one that protects the present patient without ignoring the interests of others.

Exceptions to the traditional view

Widespread endorsement of the traditional view notwithstanding, individual exceptions have been widely discussed and often endorsed. One debate considers the extent to which physicians should ‘weigh patient welfare against competing considerations of economics and of justice’.11–13 Other debates evaluate possible exceptions for military medicine14 15 and prison medicine.16 To develop an estimate for the number, range and type of exceptions, I conducted an informal survey of the literature. This survey identified 27 exceptions that are widely regarded as appropriate (table 1).

The range of exceptions highlights the fact that physicians often have competing and legitimate claims on them. Many of the exceptions can be traced to the simple fact that physicians have limited time but many patients. As a result, they cannot always promote the present patient's interests without undermining the medical care of their other patients. The traditional view, if followed strictly, would prohibit physicians from responding to emergencies whenever doing so set back the interests of the present patient.

Physicians also face conflicts between their current patients and other patients. Perhaps the most noteworthy instance of this class of exception arises in the context of physician training. Attending physicians take steps to minimise the risks that training new physicians poses to patients. Nonetheless, putting care in the hands of physician trainees, necessary for training a new generation, sometimes poses increased risks,17 even with careful supervision.18 Placement of a central line poses more risks in the hands of a physician who is performing the procedure for the first time compared with a physician who has placed many central lines in the past year. It follows, on the traditional view, that experienced physicians should never instruct trainees to learn this procedure by performing it on the present patient.

Other exceptions to the traditional view trace to the fact that physicians happen to be people who have non-professional claims on them, including claims arising from their status as friends and family members. Physicians should not leave surgery to avoid missing a golf outing. But physicians do get to retire, take vacations and have days off, even when doing so is not in the interests of their present patients. The fact that the legitimate claims on physicians cannot always be reconciled highlights the shortcomings of the traditional view and the need for systematic guidance.

Costs of ignoring the exceptions

Given the importance of protecting the present patient, one might insist that physicians should adopt the traditional view and ignore the exceptions. This response, while understandable, poses significant costs. First, physicians who ignore the interests of others would be blocked from many current practices vital to promoting health, including reporting public health dangers, responding to cardiac arrests and training new physicians, whenever these practices conflict with the interests of the present patient to any extent.

Second, insisting on the traditional view implies that proposed new practices and policies that involve balancing the interests of the present patient against the interests of others are necessarily unethical. For example, it has been argued that managed care19–21 and randomised clinical trials22 are unethical simply on the grounds that they conflict with the obligation to ‘do what is best for his or her patients’.23 This is not to argue that managed care and randomised trials are ethical. The point is that pretending there are no exceptions to the traditional view precludes accurate analysis of proposed new practices and policies that involve physicians balancing the interests of their patients against the interests of others.

Third, failure to develop an alternative to the traditional view forces physicians who recognise the appropriateness of some exceptions to balance the competing claims on them without adequate guidance, increasing the chances for mistakes and unfairness. This is a serious concern in a number of areas, including physician training. In the absence of adequate guidance, powerful patients are less likely to have trainees learn on them, leaving the risks of training new physicians to the poor and powerless.24 This result need not be overt nor intentional. Physicians who are forced to make ad hoc judgements may unintentionally end up implementing unfair patterns of treatment.

Abiding by principles, living with exceptions

Essentially all rules have exceptions.25 The right to free speech does not imply that individuals may divulge state secrets or yell ‘fire’ in crowded theatres. Chaos is averted, not by ignoring the exceptions, but by developing guidelines to manage them. The same approach applies in our relationships. Parents, like physicians, have more power and knowledge than their children. And parenting has the goal of promoting the interests of the child.26 Yet, as courts note, the principle that parents should promote the best interests of their children does not imply that they should ignore ‘the interests of others’.27 Commentators agree, pointing out that the best-interests standard with respect to children does not and cannot mean that parents should do everything possible to promote the interests of an individual child no matter what the costs or impact on others.28 Parents are not required to give their every waking moment to their children, even if doing so would maximally promote the children's interests.29

The fact that we should respond to the claims of third parties does not destroy our relationships nor lead us to neglect those for whom we care and whom we care for. It does not lead to parents ignoring the needs of their children. Instead, this fact helps to define the nature of our relationships by clarifying the extent of our obligations to significant others. The parent–child relationship is defined, in part, by the extent to which parents are obligated to promote their child's interests and the situations in which they properly respond to others in need, even when doing so conflicts to some extent with their child's interests.

What implications do the appropriate exceptions to the traditional view of physician obligations have for our understanding of when physicians should respond to the interests of individuals other than the present patient? Benatar and Upshur recently have noted the need to develop an alternative to the traditional view, one that recognises the appropriate exceptions.8 They argue that the exceptions can be accommodated by appeal to a public health ethic which maintains that physicians have obligations to the common good. This approach allows the authors to address some of the more important exceptions to the traditional view, especially those that arise in the context of military medicine.

To develop this alternative requires a positive account of physicians' obligations to the present patient. This addition is needed to block the conclusion that physicians have no more an obligation to the present patient than to anyone else. After all, the fact that one patient happened to make an earlier appointment does not imply that this patient is more important than another nor that this patient's interests are more significant. The priorities of public health alone, such as the social good and proportionality, do not provide a basis for the claim that the interests of the present patient, while not always determinative, have more weight.

The extreme view that physicians are obligated to promote the interests of all individuals, without any special attention to the interests of the present patient, seems to approximate the obligations of police officers.30 Police officers do not have a greater obligation to take care of the individual in front of them. Instead, they decide who has the greatest need that can be addressed and attend to that individual first. This approach makes sense because the police–public relationship tends to involve discrete encounters with groups of individuals.

Effective medical care, in contrast, requires ongoing and personalised relationships. It requires physicians to engage with the whole patient in front of them, as a unique individual, with a particular history and needs and preferences. On this view, the present patient has a stronger claim than others on the physician not because the patient is more important, but because the patient is part of an important relationship. This implies that the obligations physicians have to the present patient are not, strictly speaking, a result of the fact that they are physicians. Obligations of that sort would imply that physicians must take care of all those they encounter in their lives.

The present suggestion is that the obligations to the present patient result from a particular kind of relationship, the clinical relationships into which physicians enter. It is when physicians enter such relationships that the obligations to care for the present patient arise. And the extent of the obligations within the relationship is determined by its nature. There are other moral claims on physicians, including ones derived from the fact that physicians are involved in other significant relationships that make legitimate, and sometimes conflicting, claims on them.

One way to recognise these competing claims, without ignoring the needs of the present patient, is to regard physicians' obligation to promote the interests of the present patient as ‘pro tanto’ rather than strict.31 Pro tanto obligations, unlike strict obligations, can be overridden, even in the normal course of events, by considerations of greater weight. Parents, for example, have a pro tanto rather than a strict obligation to promote their children's best interests. Parents' obligations to their children do not preclude parents from driving an ill neighbour to the hospital, even when doing so conflicts slightly with their children's interests.

The view that physicians have a pro tanto obligation to promote the interests of the present patient has the virtue of allowing physicians to balance the interests of the present patient with the interests of others. Given the complex web of relationships into which physicians enter, it is unlikely that any algorithm will be able to determine precisely when physicians should respond to the interests of others. What we need, then, is a method to ‘to mediate, case by case, between clinical fidelity (to the present patient) and medicine's social purposes’.10 Fortunately, guidelines have been developed already for several of the exceptions. For example, many jurisdictions have guidelines that obligate physicians to report communicable diseases, even when doing so poses confidentiality risks to the present patient. Analysis of these guidelines suggests a preliminary framework that can be used to develop guidance for the other exceptions (table 2).

Table 2

Framework to manage the exceptions

A framework to manage exceptions

To ensure physicians are able to focus on providing care, and to block undermining of the interests of the present patient, an oversight authority should be established to develop guidelines on when it is appropriate for physicians to act contrary to the interests of the present patient. Given that these guidelines are intended to protect the physician–patient relationship, professional societies may be best situated to develop and oversee them. As in the case of distributing scarce resources, the development of guidelines for the exceptions should include ‘public engagement and involvement of relevant stakeholders’.32 33 These guidelines should be developed on the basis of at least the five considerations discussed below.

Physicians should act contrary to the interests of the present patient only when there is a compelling justification. For example, the literature on military medicine largely agrees that physicians may act contrary to the interests of the present patient when otherwise ‘society's interests would be significantly sacrificed’.34 While taking into account societal interests, guidelines must protect the physician–patient relationship. To consider one implication of this principle, guidelines should provide specific guidance that physicians implement, rather than requiring physicians to decide unilaterally when the interests of others should outweigh the interests of the present patient. Existing guidelines on reporting communicable diseases specify which diseases to report, limiting physician judgement to the medical question of whether the present patient has one of the listed diseases.

Physicians should minimise risks by compromising the present patient's interests only to the extent necessary. For example, it has been suggested that physicians may prescribe less effective medications, but only when they offer significant savings and ‘the individual patient is not likely to lose much’.35 Remaining risks should be equitably distributed. Since changes in physician schedules for coverage sometimes are contrary to the interests of the present patient, it is important to ensure that schedules are not tailored to the rich and powerful.

Responding to the needs of others typically will expose the present patient to minor risks, such as prescribing a cheaper antibiotic that poses a slightly increased risk of minor nausea. These cases do not raise serious ethical concern. In contrast, cases in which the present patient may be exposed to greater risks do raise serious concern and should be subject to additional protections. In the context of clinical research, for example, adults may be exposed to greater than minor risks for the benefit of others, but only when they consent.36

This preliminary framework should be relatively straightforward and intuitively plausible. Much of the hard work lies in implementing it. Minimising and equitably distributing risks makes sense once one attends to the concern. The problem is that failure to address the issue explicitly leads to the potential for inappropriate considerations to influence practice and policy. Applying the framework to an actual case will help to clarify this point and, hopefully, will provide the basis for future work on modifying it.

Example of physician training

The importance of training new physicians provides a compelling justification for the practice of learning by doing, even when care by trainees poses somewhat increased risks to the present patient (see table 2).29 30 Physician training reveals that consistent endorsement of the traditional view would undermine patients' interests, including those of the present patient. It is in the interests of all patients to have a sufficient number of trained physicians. The need to train new physicians is, of course, widely recognised. Unfortunately, the influence of the traditional view has led to a situation in which physician training is often conducted in a way that fails to balance appropriately the clinical interests of the present patient with the interests of all patients in having a sufficient pool of well-trained physicians.

Some institutions and physicians ignore the fact that training new physicians can pose increased risks to the present patient. Many other institutions recognise this fact but, perhaps influenced by the traditional view, fail to provide systematic guidelines to train new physicians in a way that balances appropriately the relevant interests. In the absence of systematic guidance, physician training often is carried out in an ad hoc manner, with the attending physician deciding when, and on which patients, trainees will learn.

This approach raises several concerns. Even when physicians attempt to conduct training in as equitable a manner as possible, the absence of guidance raises concern that training will end up being performed on less privileged patients. For example, patients who are more savvy and aware of the risks posed by trainees are more likely to insist on being treated by someone with experience. And attending physicians may unconsciously make decisions about when trainees are involved in ways that favour the privileged and powerful. Insistence on the traditional view will only exacerbate these concerns, leading physicians to ignore rather than distribute equitably the burdens and risks associated with physician training.

To ensure that some patients and groups are not exploited, professional societies should develop guidelines for training new physicians that reflect input from the relevant stakeholders. To minimise risks, these guidelines should require appropriate supervision and mandate that trainees begin with the least complicated cases. To protect the physician–patient relationship, experienced physicians should not be forced to decide on a case-by-case basis which patients are seen by a trainee. Instead, trainees could be assigned to patients at the institution level, providing a way also to ensure that the risks of physician training are distributed equitably. Finally, when supervision is not sufficient to reduce the risks to a negligible level, as may be the case with some invasive procedures, guidelines might require explicit consent and consider ways to compensate those who agree to be seen by a trainee.37


The widely endorsed ‘traditional’ view that physicians are obligated always to act in the best interests of the present patient is intended to protect patients' interests, yet evaluation of the literature reveals that there are many appropriate exceptions to the traditional view. It follows that insisting on the traditional view will undermine medical care in the long run, increase the potential for unfairness and leave physicians without adequate guidance. To address these concerns requires an understanding of physician obligations that recognises that physicians are subject to legitimate, and sometimes conflicting, claims. One way in which these claims might be recognised, while still protecting the present patient, would be to regard physicians as having a pro tanto rather than a strict obligation to promote the interests of the present patient. This analysis suggests a preliminary framework to develop guidelines for physicians on when and how they should balance the interests of the present patient with the interests of others. Future work will be needed to evaluate this framework and to use it to develop guidelines for the cases in which physicians need to balance conflicting claims on them.


The author would like to acknowledge the valuable contributions of Annette Rid MD, Marion Danis MD, Steve Pearson MD, Franklin Miller PhD, Seema Shah JD and Ezekiel Emanuel MD, PhD.



  • Funding This work was funded by the NIH Clinical Center, USA.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Other content recommended for you