Ethics and the Triage Model in Managed Care Hospital Psychiatry

by Harold J. Bursztajn, M.D., Thomas G. Gutheil, M.D., and Archie Brodsky

One manifestation of the health care cost-consciousness and substantial control that managed care organizations exercise over clinical practice (Schreter et al., 1997) has been the increasing use of the concept of triage as a rationale for clinical decisions. In its original battlefield context, triage (a French word meaning sorting or choice selection) has been defined as follows: The evaluation and classification of casualties for purposes of treatment and evacuation. It is based on the principle of accomplishing the greatest good for the greatest number of wounded and injured men in the special circumstances of warfare at a particular time. The decision which must be made concerns the need for resuscitation, the need for emergency surgery and the futility of surgery because of the intrinsic lethality of the wound. Sorting also involves the establishment of priorities for treatment and evacuation (U. S. Department of Defense, 1975).

Triage commonly entails dividing those seeking treatment into three categories:

  1. those who are so well off that they do not need treatment or can wait to be treated;
  2. those who are so severely wounded that they have little hope of survival even with treatment; and
  3. those whose chances of survival would be significantly improved by timely treatment (U. S. Department of Defense, 1975).

Traditionally associated with military and civilian disaster medicine, triage has become an established and recommended clinical procedure in hospital emergency rooms and other general medical settings (Vickery, 1975; Rund and Rausch, 1981). As such, it has had considerable appeal for understaffed, underfunded mental health centers where staff members perceive themselves to be, and commonly are, embattled and besieged by sheer numbers of patients. Triage has been seen by some observers to be helpful in mental health settings (Edelwich and Brodsky, 1980).

However, the clinical and ethical pitfalls of the application of the triage model are perhaps even greater in psychiatry than in general medicine, since clinicians' and patients' emotional reactions to one another play a vital and much discussed role in the treatment of mental illness. For this very reason, psychiatry, with its awareness of such interpersonal dynamics, is equipped to critically evaluate the impact of this form of decision making and its potential pitfalls in other areas of medicine, where interpersonal dynamics have real, if less well understood, effects on illness and treatment (Bursztajn et al., 1981). Here we shall examine how group process, evoked hostility and patients' repetition compulsions can contribute to the misuse of triage in inpatient psychiatry, and how such misuse can be observed and prevented.

Triage as a Decision Rule: Ethical Implications

Triage represents a modification of the simple decision rule: the greater the need, the greater the resources to be applied. Triage reasoning follows this rule up to a point-the point where a patient's ability to survive (or to benefit from treatment) becomes seriously open to question. Once that critical threshold is reached, the triage principle calls for the application of fewer rather than more resources to the care of that patient. In triage situations, a patient in extremis might be denied treatment on the grounds that a disproportionate expenditure of limited resources to save one patient would cause more damage to other patients (in the aggregate).

In a sudden disaster, where there are many simultaneous casualties and the resources at hand are obviously limited, the need for such tragic choices is incontestable (Calabresi and Bobbitt, 1978); catastrophic circumstances justify judiciously callous choices. However, to invoke the triage model to justify individual treatment decisions in a stable clinical setting-under stressful but not catastrophic circumstances-such as a mental health facility, raises, serious ethical questions (O'Donnell, 1960; Lucas, 1975; Childress, 1978; Winslow, 1982).

In the first place, the judgment that a patient is beyond help is inevitably a probabilistic one. Like any other clinical judgment, it is a question not of absolute certainty, but of a greater or lesser degree of certainty (Bursztajn et al., 1981). Triage decisions are made on the basis of such implicitly probabilistic judgments; for example, that there is a 90%, 95% or 99% probability that a patient "cannot survive" or "cannot improve." To balance such estimates against the probabilities of saving other patients, as well as to compare the value of the lives saved or enhanced in each case, is a delicate matter. Moreover, while the anticipated outcome (i.e., nonsurvival or nonimprovement) is not absolutely certain to occur, the assumption that it will occur may become a self-fulfilling prophecy.

The decision to treat a patient as hopeless may (for whatever combination of organic and psychological reasons) close the gap from, say, a 95% to a 100% probability of a tragic outcome. Just as in Paris before World War I, the word triage was used to mean the sorting out and throwing away of wilted or dead produce (Rund and Rausch, 1981). Today there is cause for concern that the labeling of people as dead (or its equivalent) will leave them fit only for discard. The study of the psychology of survival in the Holocaust confirms that people who are labeled as dead either by themselves or by others tend to engage in a powerful form of socially self-fulfilling prophecies (O'Keefe, 1982; Bettelheim, 1960; Des Pres, 1976).

Finally, we need to ask whether our moral responsibility toward individual patients extends beyond "accomplishing the greatest good for the greatest number" (United States Department of Defense, 1975). Maximizing the total benefits experienced by all patients is a useful decision rule, but it is not a fully sufficient one in a society that values the rights of minorities, including the smallest minority: the individual. Health care institutions are to be judged not only by how well they serve patients in the aggregate, but also by how well they discharge their special duties on behalf of the worst off, the most unfortunate.

Thus, while acknowledging the fact of limited resources and the need for decision rules that (when necessary) take the limits into account, we must also guard against automatically resorting to triage reasoning. Sensible on its face, the decision rule signified by the word triage does not encompass all of the ethical issues involved in the allocation of treatment resources to patients. Moreover, it is subject to misuse (that is, overuse) in the context of the strong negative feelings characteristically evoked by psychiatric patients requiring inpatient hospitalization, that is, mentally ill patients who have already alienated themselves from one social support system.

Unlikability Versus Pathology

The question of triage takes on added significance in the mental health area in the interpretation of mental illness as a consequence of losing out in the struggle for existence (Sloman, 1981). In this view, a person who has been branded by the family as the cause of its problems often provokes, in a compulsive repetition, the same unintentional scape-goating in other social contexts. Extruded by the family and then by society, the mentally ill patient projects unlikability yet again in the hospital as a means of warding off aloneness and fear of the unfamiliar by recreating familiar (in this case pathological) object ties in the new setting.

The institutionalized mentally ill are thus inherently susceptible to the prejudicial implications of triage. Indeed, unlikability itself can function in practice as a partial definition of mental illness, as shown by findings that the patients found unlikable by resident physicians tend to be judged to be in need of commitment (Hamm et al., 1983); that patients who elicit negative reactions from interviewers are most likely to commit suicide (Motto et al., 1981); and by Havens' own clinical description of the unlikability projected by a suicidal patient (Havens, 1965).

Given the close connection between unlikability and pathology in the patient, the use of triage reasoning by the clinician must always be examined as a possible expression of the powerful negative feelings patients can stir up in clinicians (Groves, 1978). It may also serve as a rationalization for projecting onto the patient what is felt to be unsatisfactory in oneself (Klein, 1975). In analytic terms, these universal negative reactions-an inevitable part of the therapeutic process-are referred to as "countertransference hate" (Maltsberger and Buie, 1974).

Experienced clinicians are familiar with trainees' reactions to deep-seated pathology and chronicity, such as "All the patients on this ward are hopeless," or "This patient is too difficult for me to learn from." Such resistance to full therapeutic engagement with the mentally ill can take many forms, including diversionary preoccupations with therapeutic ideology (Gutheil, 1977) or forensic issues (Gutheil, 1979). Premature resort to triage can be another mechanism of this normal resistance, and not just for the neophyte, since it provides an acceptable outlet for uncomfortable, guilt-arousing feelings.

The following case study exemplifies these dynamics:

Case 1: In a state-funded mental health center that has undergone reorganization so as to decrease the number of available inpatient beds, the case of a chronically ill, recalcitrant, assaultive patient is presented at an ethics case conference. A resident in psychiatry who is involved in the case asks whether, in the light of the limited resources now available in the hospital, this patient should continue to be treated at the expense of other patients. The patient is a 29-year-old man who has been hospitalized for eight months. He has a sister who, after a prolonged period of hospitalization, has been progressing well outside the hospital for 10 years.

In the interview conducted at the conference, the patient's pathology is traced to a pattern of interaction within the family. The patient's parents were survivors of the Holocaust. His mother, preoccupied with his identical twin with meningocele, gave the patient little attention. His father, identifying with the Nazi oppressors, would beat him. The patient in turn identifies with the angry, sadistic father and assaults people (including other patients in the hospital). He is driven by his identification with his parents' "survival guilt" with reference to the Holocaust (Russell, 1980; Chodoff, 1980), together with his own particular survival guilt brought on by having avoided the trauma of his defective twin.

In seeking to expel the patient, staff members defend as best they can against the helplessness they feel in the patient's presence by introjecting his projections and becoming the oppressive father that the patient's survival guilt demands. In feeling overwhelmed with the patient, the staff takes over the role of the mother who, herself overwhelmed, directs her feelings not at the sick twin or at her explosive, unpredictable husband, but at the physically healthy twin, the patient.

As a result, the staff declares the patient to be treatment-resistant without ever having tried several potentially applicable treatment modalities. These might include intensive psychotherapy and a focus on the therapeutic alliance and limit-setting in the milieu, as well as more vigorous psychopharmacological intervention on an individual level and a review of the pathological dynamics in the hospital milieu on an interpersonal level. Instead, the staff moves prematurely, albeit understandably, to a concern with the patient's "environmental impact." The issue of scarcity of resources is thus raised in connection with a patient whose parents did indeed live under conditions of scarcity.

In addition to the patient's repetition compulsion and the hostile feelings it evokes in clinicians, a case such as this can also be understood in terms of group dynamics, such as those observed in group psychotherapy (Yalom, 1975; Day, 1981). Attempts to expel an individual by a group are common and occur in the face of perceived scarcity, as when a group becomes too large for intimacy. The individual chosen to be expelled may be one who is close at hand, one who is unable to resist expulsion or one who has a distinguishing feature which thereby becomes obnoxious to the group. When a patient is "triaged out" of a hospital setting on insufficient grounds, such dynamics may be found to operate among both patients and staff.

Common Variations

In order to avoid the misuse of triage, clinicians should be familiar with the different ways in which evoked hostility (countertransference) acts in combination with patient pathology and group process. The dynamics leading to the inappropriate application of the triage model may originate with the patient or with the staff or may represent an interaction between the two.

In the patient-centered variation, a known or unknown pathology defeats all possible modes of treatment, as in the following example:

Case 2: A burly male patient has an uncontrollable seizure disorder not responsive to any known anticonvulsant. A major symptom of this disorder is sudden episodes of violence, after which the patient is post-ictal (i.e., in an altered state of consciousness that follows seizures) and amnesic regarding his actions during the seizure. Staff and other patients are extremely frightened of this man, particularly because the usual measures of controlling medications, restraint and seclusion are ineffective. When he is transferred to a higher security setting, everyone is relieved.

In the staff-centered variation, countertransference is the driving force of the patient's expulsion (Groves, 1978), as in the following case, when a staff member provokes the patient's assault:

Case 3: A tense, withdrawn man is admitted to the ward. One of the newer attendants, unfamiliar with interpersonal issues in working with seriously ill patients, takes an instantaneous dislike to the patient. This feeling is based on a coincidence, namely, that the patient reminds the attendant of a sadistic older brother who used to abuse him viciously over a number of years. The attendant, however, is unaware of this unconscious factor. Nevertheless, he begins to provoke the patient with increasing intensity. Over the next week the patient remains aloof, though increasingly restless. Then, after a several-hour period in which the attendant has roughed him up, taunted and mocked him, and threatened him with various punishments, the patient explodes. In the resulting scuffle, several staff are injured. The patient is transferred to a security setting, while the dominant feeling on the ward is bewilderment over what went wrong.

In the interactive variation, the patient and staff member(s), frightened of each other, engage in mutually escalating provocations, as in the following case:

Case 4: A patient is admitted with a diagnosis of schizophrenia, paranoid type. His chief complaints include the idea that voices are calling him homosexual and prompting homosexual acts. His therapist, a resident who is struggling with his own homosexual impulses, encounters difficulty setting useful limits with the patient. Deprived of needed external controls, the patient becomes more anxious, decompensates further and attacks male staff whom he perceives either as threatening, or attempting to seduce him. After some period of time in this state, he is transferred. The resident's relief contagiously affects the staff who, sensing the problem and feeling intuitively that the issue is a conflicted one, have been hesitant to confront the resident about his poor limit-setting.

These examples form a spectrum of cases which the clinician may experience and should be able to recognize and distinguish from one another.

Mechanisms of Evoked Hostility (Countertransference)

There are numerous ways in which the vicissitudes of inadvertent hostility toward patients (clinical countertransference) can subvert the use of triage by mental health personnel. The forms of countertransference (as seen by the dynamically oriented psychiatrist) that affect triage decisions can, however, be grouped into two main categories: milieu countertransference, whereby the clinician misperceives the clinical setting as one that requires triage decisions; and individual countertransference, whereby the clinician misjudges what triage category a particular patient falls into. Although there is no hard-and-fast line separating milieu countertransference from individual countertransference, the key distinction is between clinicians' reactions to the milieu as a whole and to individual patients.

Milieu Countertransference

Helplessness in the face of chronic illness. Staff members may incorrectly declare a state of limited resources in response to the helplessness, frustration and reduced self-esteem brought on by treating the chronically mentally ill.

Hostility evoked by violent or self-destructive patients. Staff members' anxieties about their own safety and survival in the midst of violent, self-destructive patients may resolve themselves into countertransference (Groves, 1978; Maltsberger and Buie, 1974) which, together with survival guilt (Modell, in press), may motivate an appeal for triage. There is, of course, a place for rational fear when coping with patients who act out violently. At the same time, there is a substantial risk that clinicians will overreact.

Overidentification with patients' feelings. Here the staff internalizes the patients' feelings of helplessness, hopelessness and guilt, and declares a crisis where none exists.

Individual Countertransference

Concentration is on the least difficult patients. Staff members may unconsciously respond to the helplessness and reduced self-esteem brought on by the frustrations of patient care by pouring their therapeutic energies only into those patients who show the quickest response to treatment-in those who, in terms of triage, need treatment the least.

Overidentification with healthier patients. Treatment staff may overidentify with the interests of healthier patients, whom they see as more like themselves (Freud, 1957), rather than with the most seriously ill patients, who serve as distressing reminders of the clinician's own vulnerability to psychic disturbance and even mortality (Modell, in press; White, 1977; Deutsch, 1965; Winnicott, 1965).

Rejection of most difficult patients. The hostility evoked by difficult patients may lead the clinician to deny treatment to those patients on grounds of triage even when they can be helped (Groves, 1978).

Mislabeling of patients as beyond help A patient who causes or appears to cause hopelessness in the milieu (including the staff) may, for that reason, be dismissed as a hopeless case.

Mislabeling of patients as needing continuing help. When resources are indeed scarce and triage is called for, patients who provoke guilt and the consequent reaction formation may be kept in the hospital longer than they need to be, to the detriment of other patients.

Even when the most obvious errors are avoided, it remains difficult to disentangle the effects of countertransference from the legitimate uses of triage. Triage has great appeal for the clinician because it serves as both a prima facie scientific decision-making rule (an ego ideal) and an ethical guide for justifying actions with potentially tragic consequences (a superego demand). Thus, for two of the "best" of motives, clinicians risk premature cognitive closure (Chanowitz and Langer, 1981), as when the statement "we need to triage" is used to close discussion. Such closure prevents the deep thought and reflection needed to ensure that triage is used rationally rather than as rationalization.

Adverse Effects of Triage on Patients and Staff

Even when the shift from individual clinical considerations to environmental impact is made on reasonable grounds, losses are suffered in the process. In focusing on the well-being of the ward or facility as a whole, staff members may inadvertently distance themselves from the individual patient. This distancing reinforces and exacerbates the patient's inability to learn the basic human skill of survival through cooperation. By appearing to be unlikable, the mentally ill patient cuts himself or herself off from the object ties (e.g., family, friends) that would allow for independent-that is, interdependent-survival.

The patient accomplishes this in several ways: by failing to present positive traits with which clinicians can identify, by presenting negative traits with which clinicians would rather not identify and by rejecting the clinician's offer to the patient of the best part of the clinician's self that is presented in a helping relationship. Among the pitfalls of triage, then, is its unwitting use in the service of the patient's pathology.

Triage decisions also have practical consequences for patients which must be considered in an overall cost/benefit analysis. Patients who are declared untreatable and/or a threat to other patients are generally transferred out to less stimulating settings (e.g., the chronic ward). Moreover, discontinuity of care carries with it costs of its own. No matter how scrupulously the referrals and terminations are carried out, rejection and abandonment will be felt. While referrals or transfers are sometimes necessary and may be beneficial to the patient in question as well as to the milieu, they should be undertaken only with a clear awareness of the projected benefits needed to justify these costs (President's Commission, 1983).

Finally, it is not the patient alone who suffers, as the analogy with group therapy suggests (Yalom, 1975). When a therapy group extrudes one of its members, the initial relief experienced by the group dissipates in the face of guilt at the surrender of pathology together with a recognition of the missed opportunity to work through the conflicts reawakened in group members by the "difficult" or "intolerable" member. It seems reasonable to believe that the indiscriminate resort to triage may have similar effects on both the patients and staff of the institutions where it takes place. The same issues are at stake for society as a whole in the current triaging of resources away from health care for the most needy in favor of other public and private agendas (President's Commission, 1983). Ironically, since psychiatric patients are seen as having either intractable or self-remitting illnesses, a skepticism about the efficacy of psychiatric intervention has led to a reduction of resources allocated to mental health care, justified by resort to the triage model itself.


The transplantation of the triage model from wartime to peacetime conditions necessitates reevaluation of its ethical foundations. "Scarce resources" in a psychiatric hospital is not the same thing as in military or civilian disaster medicine. Triage can be a valuable heuristic, but it must be applied in the context of other heuristics if its pitfalls are to be minimized. Ethical heuristics can be useful checks on the misapplication of the triage rule. The best antidote to the uncritical use of triage is a high degree of precision in framing the choices involved. Framing these choices requires self-knowledge and self-scrutiny as to the feelings that difficult patients arouse. Ethical analysis is helpful at this point both as a process for detached empirical description and as a source of principles for evaluating clinical alternatives.

Triage is not in itself a moral imperative, but rather a strategy to serve the larger purpose of healing and reducing suffering in limited resource settings. Moral principles that are relevant to triage include the principles of beneficence, respect for persons and justice. According to the principle of beneficence, we are to avoid harming other persons and to help them where possible. Ordinarily the duty to avoid harm to particular, identifiable individuals is more stringent than the duty to promote the hypothetical or potential well-being of others.

The principle of beneficence suggests that vigilance is needed to ensure that the catch-phrase "limited resources" does not go unchallenged, for example, at case conferences as an all-purpose justification for the avoidance of difficult clinical issues. How are the resources of the ward or facility limited? By what measure? How and to what extent will the continued treatment of a particular patient strain those resources?

It is essential, in other words, to specify concretely how one patient can be experienced as taking the bread out of other patients' mouths and whether the experience has a counterpart in clinical reality. For example, in Case 1 it must be asked whether this experience is instead the product of a reenactment of the patient's survivor guilt, with the milieu staff cast unwittingly in the role of the family members who experience a threat to their survival and therefore react punitively (Russell, 1980).

Similarly, the claim that a patient is untreatable must be subjected to the same critical scrutiny under conditions of triage as in any other clinical situation. In this way, ethical analysis serves self-analysis, and vice-versa. It is to be assumed that countertransference operates in the triage situation as it does in other clinical interactions and must therefore be made accessible to self-analysis and dialogue. If triage is to be used ethically, clinicians must be responsible for identifying and working through the evoked hostility and other disruptive feelings that may arise.

The principle of respect for persons suggests that, although triage is commonly assumed to be equivalent to cost-benefit analysis, it actually involves a more fundamental view of persons and rights. Triage can be usefully understood as a tension between two sets of rights: the right of the patient in question to receive adequate treatment, and the right of other patients not to have therapeutic risks taken on behalf of that patient at their expense. In addition, clinicians who work with violent patients have a right to protect themselves. All too often, however, a careless rush to self-protection carries a higher than necessary price with respect to both the clinician's freedom to treat and the patient's right to treatment.

Respect for persons means that persons are to be treated as ends in themselves rather than as means to the achievement of others' welfare, or simply as units in a cost-benefit equation. Moreover, as the principle of beneficence suggests, the prevention of outright harm done by one patient to others should be seen as a stronger justification for denial of treatment than a concern with the possible depletion of resources available to other patients.

Justice, understood as fairness in the distribution of benefits and burdens within a community, requires that unequal allocation of clinical resources be justified with reference to the welfare of the least advantaged patients. One rule of thumb is to ask how any decision regarding the care of this patient will impact on the resources available not only for the care of the "average" patient, but also for the care of the least fortunate or most difficult patient. Alternatively, one may avoid the pitfalls of overidentification with the healthier patients' interests by asking oneself whether one would still be making the triage decision in question if one were playing a lottery where a possible outcome was to share the identified triage patient's fate (Rawls, 1971).

For triage to be transplanted into the psychiatric setting, attention to its excesses and abuses is a prerequisite for the possibility of its ethical use.

(The authors wish to acknowledge the influence of dialogues with Sissela Bok, Ph.D., and Leston L. Havens, M.D., on the ideas presented. Some of the themes have been introduced in the authors' earlier work including "Use and Misuse of the Triage Principle in the Management of Difficult Psychiatric Inpatients" [unpublished manuscript, 1985] and Kaplan E, Bursztajn HJ, Alexander V et al. Making treatment decisions. In: Gutheil T, Bursztajn HJ, Brodsky A et al. [1991], Decision Making in Psychiatry and the Law, Baltimore: Williams & Wilkins, pp 113-132-Ed.)

The authors gratefully acknowledge the contributions of David Barnard, Ph.D., and Leslie M. Levi.

Drs. Bursztajn and Gutheil and Mr. Brodsky and Ms. Levi are associated with the Program in Psychiatry and the Law, of the department of psychiatry at Massachusetts Mental Health Center.

Dr. Barnard is associated with the Institute for the Medical Humanities at the University of Texas Medical Branch in Galveston.