Harm is a central concept both in the practice of medicine and in ethics. Hence, it is no surprise that in bioethics harm plays a prominent role. The proper goal of medicine is to prevent, alleviate, or eliminate harm to patients that result from disease or injury. Moreover, some medical interventions themselves have a serious potential to cause additional (iatrogenic) harm—for instance, pharmacological side effects or even death from surgery.
General prohibitions against inflicting harm on others supposedly belong to the principles of any moral code, and an attitude of non-malevolence (taking care that others do not suffer harm) is widely regarded as a core virtue and as a decisive source of moral motivation. Beyond this, bioethics is concerned with harm-related judgments, obligations, prohibitions, and problems. The harm at stake is most often harm to patients. Such debates as those about professional duties toward patients, about matters of resource allocation, or about the limits of patient self-determination all deal in part with actual or potential suffering, dysfunction, pain, or death of patients. Some problems, however, relate to potential harm to third parties. For example, HIV-positive patients risk infecting uninformed sexual partners and pregnant women who consume drugs risk harm to their unborn children. Other ethical questions deal with harm to health professionals themselves—when, for instance, a physician faces treating a contagious patient under substantial personal risk. And finally, various arguments in bioethics address the possibility of long-term social harm resulting from certain permissive practices (the so-called slippery slope argument). For example, critics of prenatal selection against severe genetic diseases predict shrinking social solidarity with the handicapped and with their justified claims to social support.
While its central role in bioethics thus cannot be disputed, harm remains a vague and contested concept that in and of itself does not provide much moral guidance. What counts as harm varies greatly, as do the scope and relative importance of the prescriptions not to inflict, to prevent, or to remove harm.
An instance of harm may be assessed with reference to kind, degree, and duration. Risk assessment, not considered here, also includes the probability of harm's occurrence. According to the Oxford English Dictionary, harm is "evil (physical or otherwise) as done to or suffered by some person or thing; hurt, injury, damage, mischief." As far as harm is relevant to moral deliberation, however, this broad concept must be restricted.
First, harm should be understood as person- (or animal-) regarding, that is, as consisting of events or states of affairs that are negative for someone—as expressed in Joel Feinberg's definition of harms as "setbacks to interest" (p.31). As long as the sticky question of what counts as interests remains open, this concept of harm is still neutral to various ethical positions. Problems start with determining who counts as a bearer of interests; for instance, do embryos (as potential persons), the deceased or permanently unconscious (as former persons), or animals bear interests? These issues, although obviously important for evaluating abortion, transplantation, decisions to end treatment, or animal protection, will not be pursued here.
Secondly, ethics in general ethics and bioethics in particular have to restrict their focus on those instances of harm that are in some way or other linked to human action. It would not make sense to morally deliberate about ineluctable evils, deplorable though they may be. Rather, harm is ethically relevant only if it occurs or persists in consequence to human agency, be it by action or omission, from intention or negligence, but not from unavoidable ignorance. Thus, what counts as harm with relevance to bioethics is context-relative: harm is contingent upon professional knowledge and medico-technical progress.
Thirdly, bioethics, reflecting both ordinary moral and non-moral language usage, commonly differentiates between harm on the one hand and mere loss or lack of benefit on the other. Harm is not simply conceptually complementary to benefit (interest satisfaction), but it also represents a significant disservice to its victim. Along the scale of interest satisfaction, there are numerous positions of submaximal satisfaction (disbenefits) that it seems inappropriate to call harms. There is thus an asymmetry between harm and benefit in the sense that harm pertains exclusively to the basics of well-being. It may be wrong to prevent someone from obtaining a luxury good, but nevertheless, its consequence does not qualify as harm. Another argument elucidating this asymmetry emphasizes that harm has or leads to distinct phenomenal qualities of bodily or psychological painfulness and suffering, which is by no means true for all instances of lacking benefit (e.g., Noddings). Moreover, pity for someone's experience of harm is a motivation distinct from other forms of benevolence (e.g., Sidgwick).
Not to inflict, to prevent, or to remove harm usually takes moral precedence over providing those benefits the lack of which does not count as harm. Such asymmetry between harm and benefit has been traditionally acknowledged (e.g., by John Stuart Mill), but a more systematic focus on harm is a rather recent development of applied ethics, with its eye to more concrete moral rules (a notable exception being Jeremy Bentham's 1789 taxonomy of "pains" by sources, kinds, and circumstances). The improvement of people's well-being being a more or less central goal of any moral code, concrete efforts must first focus on the most important obstacles to well-being, that is, on existing or potential harm.
Understanding harm as a significant setback to some-one's interests already implies that usually it ought to be avoided. In this sense harm is a weak normative concept, carrying a presumption of evaluative negativity. However, not every infliction or non-prevention of harm to another person is, all things considered, necessarily wrong, and in just this sense harm is not a strong normative concept. For instance, not to treat a particular patient in a tragic triage situation may be a deplorable but ethically-justified decision. Likewise, foregoing life-saving surgery on a competent patient because he autonomously decided against it, by no means "wrongs" him, in the sense of violating legitimate moral claims (Feinberg). Where harming thus does not necessarily mean wronging, the same is, of course, also true the other way round. One ought not conflate people's legitimate claims to justice or self-determination with those of not being harmed. Less clear cut is the distinction between harms and offenses, where the latter cause unpleasant, though not harmful, mental states. In the context of medicine, patients might be frustrated, shocked or irritated by inefficient hospital structures or by physicians who behave rudely. Whether such states of offendedness turn into proper harm seems to be but a matter of degree and duration.
Harm and Harm-Referring Duties in Bioethics
Assessing harm and distinguishing it from offenses, minor hurts, or non-harmful instances of lacking benefit requires an analysis of harm's nature and of how to determine its significance. Particularly in the context of healthcare, many instances of harm and potential harm to patients are widely uncontested, namely: severe lack of functioning resulting from bodily or mental disease, enduring pain, substantial suffering, gross disfigurement, or premature death. Another, easily neglected category of possible harm in the context of medical practice is of a psychosocial nature: for example, patients may experience absorbing anxiety, mistrust, alienation, helplessness, loss of self-control, loneliness, or annoyance due to structural and human deficits. In particular, the work of feminist ethicists (e.g., Noddings; Warren) and physician-ethicists (e.g., Cassell; Pellegrino and Thomasma) has created a new awareness of widely neglected kinds of harm to patients that occur in daily medical practice and that can largely be reduced or avoided when caregivers are humane and sympathetic. Even beyond the individual patient-caregiver relationship, general loss of trust in contemporary biomedical institutions and practices, in researchers and clinicians seems to be a prevalent and deeply troubling problem (O'Neill).
Finally, harm may occur as a setback to patients' higher-level "critical interests" in living a life they consider good (Dworkin). Notably, decisions about one's time and manner of dying are likely to relate to such highly personal, critical interests. Focusing on these issues would involve yet another conceptual enlargement of (modern) medical harm.
All of these states or events are setbacks to individuals' interests in basic well-being, and thus univocally considered harmful. In principle, they can be relevant to bioethics whenever they potentially occur or persist as a consequence of intentional behavior, where behavior must be understood in a broad sense. Hence, ethically-relevant harm can result from both omission and commission, from individual or collective acts, from a patient's own decision or from someone else's.
Harm can be intended, merely foreseen, or accepted as a lesser evil when compared to the consequences of all available alternatives, and it can be intended with regard to an identified or to a statistical addressee. Take a patient's premature death, to illustrate the broad variance of agents and victims and of causal and intentional modes under which ethically relevant-harm can occur in medicine. This premature death could, for instance, be the consequence of: a physician's decision to stop life-saving treatment, a negligently wrong treatment, an unfortunate research intervention, the patient's own decision against further treatment, a rationing policy, or a negligent infection from undisclosed sexually transmittable disease.
To emphasize it once more: it seems hard to find even one bioethical problem that does not somehow involve aspects of harm to patients or, less frequently, to health professionals or third parties. In all these matters, however, dissent arises when it comes to the comparative evaluation of a particular harm's negativity; in setting standards for professional, social, or personal responsibilities for people's health, and corresponding duties; and in the assessment of distinguishing harm from mere lack of benefit in healthcare.
With regard to duties, some scholars in ethics formulate a distinct duty of nonmaleficence, expressing a prohibition on actions with foreseeable harmful effects. Others, however, include this prohibition as part of a duty of beneficence. This, and whether such obligation is construed as a prohibition on causing net harm to someone (such that, say, shooting a murderer to save the lives of his three victims would not count as maleficent), or on harming itself (the shooting would be maleficent, though perhaps justified), is a question of terminological and classificatory preference. The duty of nonmaleficence is still indeterminate under any of these descriptions, not only because they reintroduce the problems of harm assessment but also because they are silent about permissible limits and trade-offs.
Recognizing a distinct principle of nonmaleficence is fairly common in medical (in contrast to general) ethics. It is meant to guide actions by caregivers in those situations that are most likely to produce harm. However, depending on both formal tailoring of concepts and on normative perspectives, there exist formal and substantial differences among bioethical perspectives in what is understood as nonmaleficence. Tom Beauchamp and James Childress, for instance, turn to the four duties of beneficence originally distinguished by William Frankena in 1973. Frankena's classification of duties is based on a distinction between harm and benefit and on the action's causal mode:
- not to inflict harm;
- to prevent harm;
- to remove harm; and
- to promote good.
Beauchamp and Childress modify Frankena by subsuming the first duty under nonmaleficence and leaving the last three duties under beneficence. Their distinction between the two duties of nonmaleficence and beneficence thus corresponds to the difference between negative and positive duties (i.e., duties of omission versus duties of commission), again depending on aspects of causality. Beauchamp and Childress do not, however, take this classification as such to be normatively decisive; rather, they intend to capture ordinary language usage, mirroring the empirical fact that noninfliction of harm often is achievable at lower cost to the agent than is obeying positive duties. It is in this generalized sense that the obligation of nonmaleficence frequently has priority over beneficence.
Along these lines, Allen Buchanan and Dan Brock have suggested that appeals to nonmaleficence in medicine be understood as specific reminders: in Hippocratic times, not to forget that some treatments were only burdensome and not beneficial; in contemporary times, to correct "for professional biases toward over-treatment of non-communicating patients in conditions of great risk or profound uncertainty" (Buchanan and Brock, p. 256). These reminders pay attention to medicine's increasing potential not only to benefit patients but also to inflict iatrogenic harm upon them (Sharpe and Faden).
The duty of nonmaleficence may conflict with the autonomy of patients who request treatment that physicians consider harmful (e.g., unjustified surgery, futile chemotherapy, or drugs). With an eye to precisely this conflict, H. Tristram Engelhardt, Jr., understands the duty of nonmaleficence as a justification to limit patients' self-determination.
Problems with Harm in Medicine
As to the more precise nature of harm and to the scope of harm-related duties, bioethics inherits some of the controversies of general ethical theory. A crucial question is to what extent there are objective criteria for identifying and evaluating harm. If such criteria could be found, this might, for example, justify a physician's overriding a patient's own "harmful" preferences. Or, such criteria could be adduced in surrogate decision making for noncommunicating patients, as well as in matters of allocative justice (where it becomes crucial to evaluate medical interventions in terms of their comparative tendencies to avoid or alleviate net harm).
The issue of determining criteria is linked to the objectivity/subjectivity debate concerning people's well-being and ability to live a good life (see Griffin), and the setbacks to these. With most experts agreeing that there is an irreducible plurality of harms, the subjectivist view takes harm to be a significant setback only to actual wants or desires, possibly after procedural safeguards have been met. Here, for instance, a patient's death due to intentional non-resuscitation would be harmful only if the patient, when informed and asked, would opt for treatment. In the objectivist view, harm is a significant setback also to interests that are want-independent, but related to ideals of a good life. Here, death due to non-resuscitation could be harmful to a patient, regardless of whether he or she wants it.
The fundamental distinction between "want regard" and "ideal regard" (as a difference between subjective versus objective concepts of interest) was introduced by Brian Barry in political philosophy. In that area, lack of autonomy in forming one's wants is less obviously a danger than it is in medicine, where patients can so easily be ill informed, manipulated, or otherwise incompetent when forming their preferences. Therefore, at least certain procedural safeguards—such as standards of informed consent—are not inconsistent with "want regard" in medicine. Other safeguards, like elevating standards for patient competence to a level commensurate with the expected harm that would result from acting in accordance with patient choice (e.g., Buchanan and Brock), arguably cross over into "ideal regard." In any case, there is room for hybrid positions between the extremes of pure want regard and ideal regard. Consider forcing a Jehovah's Witness to be transfused with blood. Justifying this by reference to the patient's presumed objective interest in the preservation of his life falls under ideal regard. Arguing that the patient would want the transfusion if she were not bound to her irrational belief system puts harm assessment by want regard under some ideal-regarding constraint.
A common argument in favor of taking harm as an objective concept stresses the broad consensus in what "rational persons desire to avoid for themselves" (Culver and Gert, p. 70). Reference to the obvious consensus about the desirability of avoiding disease, disability, pain, premature death, and suffering, presupposed in daily medical work, is familiar from the debate over concepts of disease (Culver and Gert). To concur on this point does not imply acknowledging universal standards for all sorts of harm. Rather, pain, disability, and premature death are seen as universal harms simply in being setbacks to very basic interests, the satisfaction of which is instrumental to practically all conceptions of a good life. It would, of course, not come as a surprise to find this true for many kinds of harm, in contrast to mere lack of benefits.
Even more serious problems with defining medical harm hinge on the need to compare two instances of harm, such as those from alternative treatment courses or from alternative resource distributions. Such ranking judgments are needed on kinds of harm (for example, pain versus addiction; premature death versus disfigurement; disease versus a restricted lifestyle) and on how much, when, and for how long harm is to be accepted, and for what purpose. At least implicit comparative evaluations of risks of harm and benefit are involved in virtually any treatment decision or medical indication (Veatch). Here, more fundamental disagreement starts: Some authors emphasize the great variability in comparative harm assessment, pointing to its relatedness to the context of each patient's irreducibly personal or parochial conception of a good life (e.g., Engelhardt; Veatch). This position has nurtured so-called autonomy-centered bioethics, which considers the assessment of harms and benefits to be the patients' business only. In contrast to this position, other scholars want to keep at least some objective ground for evaluations: medical interventions should, according to them, be determined futile not by patients, but by professional standards whenever they appear to be disproportionately harmful and thus "not reasonable" (Brody); or these scholars see interpersonal variability in ranking harm—though it exists—as not predominant and therefore not ruling out a beneficence-centered bioethics.
Other fundamental problems relate to the legitimate scope and relative importance of the obligations to prevent or to remove harm. First, some such actions, although morally laudable, are not required of the agent because they pose undue burdens or risks for him. For example, a therapist need not risk his own death in treating a violent patient. But how far do these agent prerogatives go? And how are they determined and justified? Secondly, harm preventing or removing actions sometimes ought to give way to other overriding duties (e.g., the duty to remove still greater harm from another or to respect patient self-determination). However, there are many different views as to what counts as overriding duty. Between the two extremes—understanding nonmaleficence as the trivially indeterminate principle "avoid harm (whatever that is) unless it is outweighed" or having as many specified duties as there are different normative theories—attempts have been made to give a more specific meaning to nonmaleficence without leaving the middle ground of broader consensus.
How to Handle Pluralist Harm Assessment
Undeniably, different people have very different notions of what medical harm would be for themselves or for others. Autonomy-centered bioethics has seen its task as spelling out procedures to foster a "morality of mutual respect" (Engelhardt) and patients' self-determination. This approach leads to particular concern for informed consent, policies for advance directives, substituted judgment, and so on. A contrasting approach urges that instead of inviting radical individualism in assessing medical harm, we redetermine medicine's substantive goals. Daniel Callahan, for example, argues that such individualism results in net harm to all by consuming too many resources for marginal benefits and setting wrong priorities in our lives. Stressing the importance of expectations and cultural presumptions in determining what individuals view as harm, Callahan hopes to find arguments acceptable to the whole of society—in favor, for instance, of decreasing individual expectations for life-prolonging treatment in old age.
Other authors concur that individualistic harm assessment is the wrong paradigm for medicine: "Moral atomism" is viewed as impoverishing medical practice socially and morally, that is, as giving up grounds on which a sense of community and good decision making should develop (Pellegrino and Thomasma). Others see "moral atomism" as leading to a waste of physicians' power to assist patients in pursuing their goals (Brody, p. 50), or as leading to paralysis in crucial policy questions, such as how to determine the best treatment interests of incompetent patients (Emanuel). Ezekiel Emanuel opts for communitarian healthcare settings, where groups of patients and physicians shape medicine according to their shared assessment of harms and benefits; others are confident that the consensus on harm in the context of medicine is substantial (Pellegrino and Thomasma; Cassell; Brody). They see the main problem in "the view that the physician respects autonomy by taking a negative, hands-off stance" (Brody, p. 50), which they argue ought to be given up in favor of assisting patients, in a critical and trustworthy manner, to assess harms and benefits.
Prominent Controversies on Medical Harm
A prohibition on killing is often taken to be the most important negative duty of nonmaleficence, death being a major harm for most people. Generally, the same is true for the medical context, with the contested exception of assistance in dying. Proponents of active voluntary euthanasia for terminally-ill patients are not only prepared to give priority to patient self-determination in these situations, but would not even consider the resulting death a harm and its intentional provision maleficent—rather to the contrary. Controversies over these issues across many cultures result from different views on the allegedly harmful or benefiting nature of a patient's death from assistance—be it by active killing, by withholding or withdrawing life support, assisted suicide, or indirect euthanasia. Those who insist on normative differences between these various forms of assistance often give normative weight to the involved causal or intentional differences. A prominent instance of such an argumentation is the controversial Roman Catholic doctrine of double effect, according to which, for example, indirect euthanasia can be justified in spite of the death that may result, since the latter is not intended but merely foreseen as a by-product of beneficent painkilling. Other opponents of aid in dying argue with the social harm than could be expected from one or several of these practices once, established as legitimate option for the terminally ill (the slippery slope argument).
Yet another debate centering on the concept of harm-concerns cases of sexually-active patients who carry a sexually-transmittable virus (e.g., HIV) and refuse to inform their partners. Legal prescriptions aside, bioethicists are divided as to whether the treating physician, who cannot convince his patient to the contrary, has a duty to inform those at risk. Obviously the obligation to prevent harm to others conflicts with the professional obligation to confidentiality; violating confidentiality might also lessen the general trust in physicians' patient advocacy.
The heated controversies on prenatal diagnosis, gene therapy, and wrongful birth and wrongful life issues focus on possible harm to future children or their parents, but also on those who are living with genetic handicaps. Once again, bioethicists dissent on what to identify as harm, how to evaluate its negativity, and how to balance related duties against other ethical obligations.
In summary, there is a remarkable tension between harm's undisputed importance in bioethics and the numerous different ways in which it comes to be conceptualized and evaluated, thus mirroring the plurality of existing ethical approaches.
bettina schoene-seifert (1995)
revised by author
SEE ALSO: Animal Welfare; Bioterrorism; Buddhism, Bioethics in; Circumcision; Competence; Death; Death, Definition and Determination of; Double Effect, Principle or Doctrine of; Environmental Ethics; Ethics; Harmful Substances, Legal Control of;Holocaust; Homicide; Human Rights; Infanticide; Injury and Injury Control; International Health; Law and Morality; Malpractice; Mistakes, Medical; Moral Status; Pain and Suffering; Paternalism; Psychiatry, Abuses of; Research, Unethical; Smoking; Utilitarianism and Bioethics; Warfare
Barry, Brian. 1965 (reprint 1990). Political Argument: A Reissue with a New Introduction. Berkeley: University of California Press.
Bentham, Jeremy. 1789 (reprint 1963). An Introduction to the Principles of Morals and Legislation. New York: Hafner.
Buchanan, Allen E., and Brock, Dan W. 1989. Deciding for Others: The Ethics of Surrogate Decisionmaking. New York: Cambridge University Press.
Callahan, Daniel. 1991. What Kind of Life: The Limits of Medical Progress. New York: Simon & Schuster.
Cassell, Eric J. 1991. The Nature of Suffering: And the Goals of Medicine. New York: Oxford University Press.
Culver, Charles M., and Gert, Bernard. 1982. Philosophy in Medicine: Conceptual and Ethical Issues in Medicine and Psychiatry. New York: Oxford University Press.
Dworkin, Ronald. 1993. Life's Dominion: An Argument about Abortion, Euthanasia, and Individual Freedom. New York: Knopf.
Emanuel, Ezekiel J. 1991. The Ends of Human Life: Medical Ethics in a Liberal Polity. Cambridge, MA: Harvard University Press.
Engelhardt, H. Tristram, Jr. 1996. The Foundations of Bioethics, 2nd edition. New York: Oxford University Press.
Feinberg, Joel. 1984. Harm to Others. New York: Oxford University Press.
Frankena, William K. 1973. Ethics, 2nd edition. Englewood Cliffs, NJ: Prentice-Hall.
Griffin, James. 1986. Well-Being: Its Meaning, Measurement and Moral Importance. Oxford: Clarendon.
Mill, John Stuart. 1863 (reprint 1979). Utilitarianism. Indianapolis, IN: Hackett.
Noddings, Nel. 1989. Women and Evil. Berkeley: University of California Press.
O'Neill, Onora. 2002. Autonomy and Trust in Bioethics. New York: Cambridge University Press.
Pellegrino, Edmund D., and Thomasma, David C. 1988. For the Patient's Good: The Restoration of Beneficence in Health Care. New York: Oxford University Press.
Sharpe, Virginia A., and Faden, Alan I. 1998. Medical Harm: Historical, Conceptual, and Ethical Dimensions of Iatrogenic Illness. New York: Cambridge University Press.
Sidgwick, Henry. 1966. The Methods of Ethics, 7th edition. New York: Dover.
Veatch, Robert M. 1991. The Patient-Physician Relation: The Patient as Partner, Part 2. Bloomington: Indiana University Press.
Warren, Virginia L. 1992. "Feminist Directions in Medical Ethics." In Feminist Perspectives in Medical Ethics, ed. Helen Bequaert Holmes and Laura M. Purdy. Bloomington: Indiana University Press.
harm / härm/ • n. physical injury, esp. that which is deliberately inflicted: it's fine as long as no one is inflicting harm on anyone else. ∎ material damage: it's unlikely to do much harm to the engine. ∎ actual or potential ill effect or danger: I can't see any harm in it. • v. [tr.] physically injure: the villains didn't harm him. ∎ damage the health of: smoking when pregnant can harm your baby. ∎ have an adverse effect on: this could harm his Olympic prospects. PHRASES: come to no harm be unhurt or undamaged. do more harm than good inadvertently make a situation worse rather than better.do (someone) no harm used to indicate that a situation or action will not hurt someone, whether or not it will provide any benefit: the diet of milk and zwieback certainly did him no harm. mean no harm not intend to cause damage or insult: this was cruel, but they meant no harm by it. no harm done used to reassure someone that what they have done has caused no real damage. out of harm's way in a safe place.