"Magical Thinking," Suicide, and Malpractice Litigation

Harold Bursztajn, MD; Thomas G. Gutheil, MD; Archie Brodsky; and Evelynne L. Swagerty, JD
Bull Am Acad Psychiatry Law, Vol. 16, No. 4,1988

Prospective clinical assessment of suicidality differs significantly from that used retrospectively in malpractice litigation. In the latter context, the judge or jury may be susceptible to hindsight reasoning and a disproportionate emphasis on the specific method of suicide, exaggerating its foreseeability and "magically" linking the means of death to the treating clinician, especially in the case of suicide by an overdose of prescribed medication. Such magical thinking, moreover, is rooted in the clinical context of suicide: The errors of reasoning observed in the courtroom exhibit striking parallels with the mind-set of the suicidal patient. An understanding of these dynamics suggests appropriate precautions for the clinician and thus contributes to the prevention both of suicide and of malpractice litigation.

The authors acknowledge their indebtedness to Victoria Alexander for significant editorial contributions to the preparation of this manuscript. Drs. Bursztajn and Gutheil, Ms. Swagerty, and Mr. Brodsky are affiliated with the Program in Psychiatry and the Law, Massachusetts Mental Health Center, 74 Fenwood Rd., Boston, MA 02115. Address reprint requests to Dr. Gutheil.

In considering malpractice liability for a patient's suicide, the courts as a rule are appropriately respectful of the uncertainties inherent in the assessment of suicide risk. [1] In practice, the determination of liability is guided by two questions: "Was the clinician's evaluation at the time sufficiently thorough to assess the patient's suicidality? If so, were adequate preventive measures taken, given the level of assessed risk?" Nevertheless a central problem in all negligence cases, including malpractice, is that the determination is made in retrospect, with knowledge of the outcome. Thus the judge or jury may be susceptible to reasoning by hindsight ("Of course the suicide was predictable; he killed himself, didn't he?") This kind of hindsight reasoning (which is one form of magical thinking as we refer to it here) [2] obscures the prospective uncertainty of the outcome, exaggerates its foreseeability, and scants the possible therapeutic benefits of less restrictive treatment options.

A hallmark of magical thinking in the aftermath of suicide is a disproportionate emphasis on the specific means of self-destruction, particularly when connected with the treating physician. Thus the physician (even when not actually negligent) may be a more likely target for accusations of negligence if a patient overdoses with medications prescribed by that physician than if other means of suicide are used. Yet magical thinking is much more than simply an artifact of litigation, for it is deeply rooted in the ecology of suicide. In this article we shall explore both the clinical and judicial manifestations of magical thinking in the context of suicide. By understanding the errors of perception and logic that may characterize suicidality as well as litigation in the wake of suicide, psychiatrists can equip themselves to perform better suicide assessments and help prevent malpractice liability by reducing the risks both of the tragic outcome itself and of judicial misinterpretation after the fact.

Magical Thinking in Malpractice Litigation

Unexamined, erroneous assumptions about causality and responsibility that typically underlie malpractice actions based on suicide are illustrated in the following vignette:

A middle-aged woman was hospitalized for depression with some suicidal ideation. Treated with antidepressant medication, she showed consistent improvement in mood over the course of several weeks in the hospital. At the end of that period, the patient prepared for her discharge. Since she was planning to spend some time out of state in an area where pharmacies were scarce, she asked for a month's supply of her antidepressant medication. Her physician, recalling that the patient had been seriously suicidal when she was off the medication and noting that she showed no current evidence of depression, complied with this request. Two days after discharge, the patient took her entire month's supply of medication and died. The patient's husband sued her physician for malpractice, claiming that the patient should not have been given such a large amount of medication, because if taken all at once, it represented a lethal dose.

The source of possible error here lies in an inappropriate emphasis on the specific means used for suicide. The outcome would have been the same if the patient had hanged herself, slit her wrists, or jumped from a window—any of which she would have been free to do outside the hospital. Had she used any of these other means instead of the medication, the retrospective assessment of liability might have focused more appropriately on the question of whether she was safe to leave the hospital at the time of her discharge.

In cases such as this, however, the need for a multifaceted evaluation tends to be obscured by an "if only" style of wishful thinking: "If only this one thing had [or had not] been done, the deceased would be alive today." In the exclusive attention given to the specific means of death the broader context of the psychiatrist's assessment of suicidality is lost.

We propose that magical thinking in the determination of negligence for suicide has the following distinctive characteristics: a conception of foreseaability in absolute, yes-or-no terms as opposed to the more probabilistic, risk-benefit reasoning in which the clinician must engage; the use of hindsight [3] rather than reconstruction of the conditions under which the clinician exercised foresight; failure to acknowledge the uncertainty that surrounds clinical judgment; [4] unicausal rather than multicausal explanation for the suicide; a perception of the physician as active agent, the patient as passive victim; disproportionate weight given to the specific means of death as a factor in determining whether the suicide could have been foreseen or prevented; and a perception of the means of death as if it were exclusively under the physician's control.

The trial court's reasoning in Hirsh v. State of New York [5] demonstrates the oversimplifications of magical thinking. In this case a state hospital was charged with negligence in the death of a 38-year-old man, diagnosed as manic depressive, depressed type, who overdosed on Seconal. He was noted to be suicidal at the time of admission and had attempted suicide twice (once by hanging and once by overdose) several weeks earlier at another hospital. The source of the Seconal was never determined.

The trial court held that "the State violated simple rules which should have been followed so that [the patient] could not have taken his own life." Specifically, the court stated that the hospital staff should have inspected the patient's person and bed so that any unauthorized possession of medications could not possibly have gone undetected:

The decedent would have committed suicide only in either of two ways. Either be had the Seconal in his clothing, in which case the hospital should have found and removed it, or he procured it in the hospital from the dispensary or storeroom or from a person in the hospital. In either of these cases, the State would have been negligent ... [emphasis added].

All the elements of magical thinking—certainty, hindsight, unicausality, foreseeability, and a disproportionate emphasis on the means of suicide—are present in this statement. The host of possible paths to suicide is reduced to the single issue of the patient's access to the instrument actually used.

Although the patient was known to be at high risk for suicide, the trial court focused exclusively on the instrument of death and the patient's access to it. The court thus assumed, first, that prevention of the patient's suicide was merely a matter of taking sufficient precautions (however restrictive) and, second, that had the hospital staff prevented the patient's access to the pills, the patient would still be alive. The appellate court took a different view and reversed the judgment of negligence:

The State could not have provided an employee to watch every move by this unfortunate man during 24 hours of the day. We are not persuaded that it is evidence of negligence that he was not repeatedly wakened and his bed searched during the night. If institutions for the mentally ill are required to take all of the precautions contended for in this case, and are to be held liable for such debate mistakes in judgment, patients would be kept in strait jackets or some other form of strict confinement which would hardly be conducive to recovery ... An ingenious patient harboring a steady purpose to take his own life cannot always be thwarted.

This realistic assessment reflects an awareness of uncertainty, of the need to balance risks against benefits, and of the patient's actions as an independent cause contributing to the outcome.

When the question of negligence in the wake of suicide is reduced to "how the patient got those pills" or "why that window was left open," the actual complexity of the circumstances leading to suicide is obscured. The underlying assumption about cause and effect is magical in that it is more certain, simplistic, and symbolically based than is the reality of clinical practice. Through this type of magical thinking, the links between the death and its instrument, and between the instrument and its "dispenser" (the physician or hospital) may offer a tempting psychological shortcut to the fixing of blame. We suggest that this process may influence the outcome of negligence litigation in an inappropriate manner.

The determination of negligence in this area should best focus on two questions: the adequacy of the clinician's assessment of suicidality and the adequacy of the precautions taken to prevent a foreseeable suicide attempt. In addressing these questions, the courts have recognized that neither the prediction nor the prevention of suicide can be accomplished with certainty. Moreover, the fact that risks are inherent in therapeutic measures does not itself prove negligence when harms result; instead, the risks of treatment (including suicide itself) must be weighed against the potential benefits. [1]

Suicidality and Magical Thinking

A court's attribution of special meaning to the particular means of suicide may coincide with its special meaning for the patient. For example, the choice of pills prescribed by the physician may reflect the patient's experience of the therapeutic relationship. From a psychodynamic standpoint, Abraham's model of predisposition to melancholia involving the oral phase of development may explain the depressed suicidal patient's susceptibility to using the physician's medication for an overdose. [6,7] Clinicians know that for such patients oral medications often have powerful symbolic meaning (e.g., as the mother's poisoned milk).

Beyond the choice of a particular means of suicide, there is another, perhaps deeper correspondence between magical thinking as manifested in suicide litigation and in suicidality itself. The flaws of reasoning that may distort the retrospective judicial assessment of suicidality have clinical parallels in the reasoning of suicidal patients. For example, the absolute, black-and-white reasoning that may characterize litigation after suicide is analogous to the dichotomous thinking (ideal life versus death) of the suicidal person. [8] Judgment from hindsight is observed in the pressed person's guilt and regret over losses ("I should have done" such and such). Failure to acknowledge uncertainty is mirrored by the deterministic hopelessness characteristic of depression—the certainty that one's mood will never change. [9] Moreover, the tendency to rely on a unicausal explanation is also characteristic of the depressed person's tendency to see either the self (internal causation) or the world (external causation) as the sole cause of difficulties. [9] The judicial perception of the patient as powerless and the physician as omnipotent may thus be identical with the patient's own perception.

These striking parallels between characteristic modes of perception in the legal and clinical contexts of suicide can be explained in two (not necessarily mutually exclusive) ways. First, suicidal patients may exhibit in intensified form some common fallacies to which human intuition and reasoning are susceptible in the face of a difficult, anxiety-provoking decision. [10-12] In addition, suicidal patients may "infect" not only family members and clinicians [13,14] but, secondarily, attorneys, judges, and juries with their pathological style of thinking—in particular, a need to exact retribution, the law of the talion, [15] insofar as suicide may be, in part, a retributive act. Just as the patient's suicide "punishes" the family, so the family may wish to punish the physician.

The following case illustrates the way in which a suicidal patient may act on the basis of magical meanings attached to prescribed medication:

A borderline adolescent with juvenile-onset diabetes took a massive overdose of her own injectable insulin and left the empty bottles— each with the prescribing (i.e., treating) psychiatrist's name on the label—in that psychiatrist's mailbox. Subsequent litigation turned upon the issue of the patient's responsibility for keeping and managing her own insulin while hospitalized for depression and impulsivity. Although she had made previous suicide attempts, some serious, she had not used this method before, nor had she given warning of increased suicidal intent.

In this case the grounds for finding the physician negligent on the basis of the means of suicide would be relatively weak. There remains, however, the troubling fact that the patient chose this particular means as a clear expression of hostility toward the physician. Under such conditions, then, we may infer that some patients, when suicidal, may be more at risk for using a physician-prescribed means of death than other means.

What therapeutic measures can be taken to protect patients who are likely to act out in this way, without denying them the benefits of medications or abandoning the many patients for whom the benefits exceed the risks? At the same time, how can clinicians protect themselves against the retrospective imputation of blame in the event of suicide? The following case exemplifies a comprehensive therapeutic approach in which the risk of the patient's engaging in magical thinking with respect to medications is anticipated and assessed along with other risks and benefits:

A woman with a history of alcoholism, oppression, and mood swings was admitted following a massive, self-administered overdose of insulin prescribed for her diabetic son. By the time of her psychiatric admission the patient no longer exhibited the vegetative symptoms of depression present before her suicide attempt. Instead she became manic during her first two weeks in the hospital. During this period the factors precipitating the overdose were explored. Not diabetic herself the patient had suffered the death of a brother from complications of diabetes two years earlier. At the time of the overdose she felt that her relationship with her son was threatened by his impending marriage. By overdosing on her son's insulin she unconsciously sought to reestablish a close connection with her deceased brother as well as with her son.

The risks and benefits of medication were discussed with the patient in the context of her recent reaction to loss. In her case the usual risks of lithium were exacerbated by an additional risk—namely, that she might overdose, magically to join her dead brother, who had worked for a firm that made lithium. Since she was not depressed at the time, the risks of lithium were judged to outweigh the benefits. During her hospitalization, when the patient became manic, she agreed to a trial of the neuroleptic drug perphenazine, which has a therapy/toxicity index of 100/1, as compared with 3/1 for lithium,[16] and thus has a lower likelihood of life-threatening complications with an overdose.

The patient's mania cleared within days. The remaining four weeks of hospitalization were spent monitoring the effects of the drug and beginning psychotherapy. At the time of discharge the risks and benefits of outpatient medications were reviewed with the patient. Lithium prophylaxis was thought to be contraindicated in part because the patient had become euthymic. Also, she still had not completed the process of grieving. She was assessed as being at risk both for acting on the medication's magical association with her brother and for resuming alcohol use with increased impulsivity. Therefore, with the patient's consent, lithium was not prescribed on discharge. The daily dosage of perphenazine was gradually tapered, and outpatient follow-up was arranged.

Six months later, as she experienced her grief more fully, the patient became depressed and started drinking again. Admitted at her own request, she became manic as she worked through her grief. After her condition had stabilized, she was judged competent to disclose her suicidality (if present) and to address the risks and benefits of lithium. After the magical associations of lithium had been further clarified in psychotherapy, the patient and her physicians together decided that the risks of long-term use of perphenazine now outweighed the risk of an overdose of lithium. Two weeks after admission perphenazine was discontinued and lithium begun. A euthymic response occurred rapidly. During the ensuing four weeks of therapy the competence assessment and risk-benefit discussions were repeated, and it was judged safe to discharge the patient on lithium. One year later, the patient was functioning well and continuing with outpatient therapy.

In this case the patient's alcoholism, her prior use of insulin to establish a self-destructive bond with her brother and son, and her brother's involvement in the manufacture of lithium alerted her therapists to the heightened significance that lithium might have for her. In planning her treatment, they were able to anticipate and avoid the possibility that the patient might act on these special meanings, with potentially destructive consequences.

The magical meaning that medications may have for a suicidal patient does not justify blaming a suicide retrospectively on the single act of prescribing; that is, the patient's magical thinking does not validate magical thinking on the part of attorney, judge, or jury. When the patient has a specific, clinically relevant rationale for choosing a means of suicide, however, the clinical and legal question is whether that rationale has been properly assessed. Thus, the alleged negligence of a physician would lie not in the act of prescribing but in the failure to consider the risk of magical thinking on the part of the patient and to address it with the patient.

For the physician the clinical and legal pitfall would be to allow precautions related to a particular means of suicide to obscure the ever-present need for a broad and careful assessment of suicidality, followed by appropriate standard precautions, irrespective of a particular means of self-harm. For example, the practice of prescribing small doses of medication, which some clinicians favor, may or may not impress a judge or jury with the clinician's foresight, but it is unlikely to deter patients' intent on self-destruction, [17] inasmuch as they can easily save small weekly doses until a lethal supply is at hand, obtain medications from more than one source, or use drugs prescribed for someone else. Moreover, many other means of suicide are available, even in the clothing worn to the physician's office. By relying on limited doses of medication, the clinician may naively assume that no other preventive measures are necessary.

The use of frequent small prescriptions may yet be therapeutic insofar as it communicates the therapist's concern for the patient's well-being or cognizance of the danger of impulsivity. If used without the patient's active participation and consent, however, this practice may instead communicate a patronizing distrust of the patient or a defensive posture on the part of the physician, which in turn may invite dangerous regression. [17] When the clinician acts without communication, the patient is invited to do the same: the physician's "impulsivity" (i.e., acting without prior discussion) may thus mirror and amplify the patient's impulsivity.

The law does not require that every psychiatrist be psychodynamically sophisticated. However, all clinicians should be aware of the special dynamic meanings that medications (and other oral substances and images) have for depressed patients. Indeed, the prudent clinician will consider the potential for magical thinking on the part of the patient and in legal proceedings. We propose that the best precaution against magical thinking in both the clinical and legal contexts is a carefully documented risk-benefit analysis shared with the patient through the informed consent procedure. [18] Discussion with the patient should cover the risks and benefits of hospitalization versus discharge and, secondarily, the risks and benefits of drug prescription, as we have described.

Risks arising from the seemingly magical potency of medications can be explored by asking the patient directly about feelings of hopelessness and suicidal intent. [19] When the patient's vision of an intolerable situation is brought into the open, it can be examined critically rather than accepted fatalistically. Moreover, explicit engagement with the patient's suicidal ideation enables patient and therapist to form an alliance around safeguarding the patient's life as the first priority.

Thus, when prescribing for the suicidal patient, the clinician should ask the patient to consider—along with the other risks of medication—the risks of magical thinking in connection with the drugs. By treating these risks as though they were properties of the medication and by using counterprojective techniques, [20] the clinician may encourage the patient to express uncomfortable emotions. For example, the clinician might say, "We've noticed that one side effect of this medicine is that it can tempt people to take too much at once and harm themselves. You know, people have all kinds of ideas about pills..." Special care should govern this exploration if the patient has a history of drug or alcohol abuse, has previously taken an overdose of prescribed medications, or exhibits personal or family dynamics in which drugs have special meaning.

The patient's responses to such inquiries should be critically assessed. A patient who insists, "I'll never do it," may be unwilling to face the actual risk involved. On the other hand, one who admits, "Look, if I really want to do it, I can always find another way," shows some ego strength by at least disavowing magical thinking. By the same token, the risk of the patient's withholding disclosure of suicidal intent cannot be dismissed. Therefore, we recommend extending the assessment of the patient's competence to weigh the risks and benefits of leaving the hospital or taking prescribed medications. That assessment should include an evaluation of the patient's competence to disclose suicidal thoughts or fears, if they are present [18,19,21] A patient whose denial of suicidal intent is delusional or grossly incompatible with the clinical reality may be regarded as potentially incompetent to make an honest disclosure of suicidality and thus as presenting a substantial risk of suicide.

These competence assessments may also help prevent magical thinking in the judicial setting. For example, in one malpractice case involving an overdose with prescribed medication, the court decided that because the patient had been judged competent to live outside the hospital, the suicidal action itself, rather than medical negligence, was the proximate cause of the patient's death. (Indeed, this was the traditional legal view before the 1940s.) Therefore, the suicide was not a foreseeable consequence of the physician's decision to prescribe medication. [22] If this approach is more widely followed, so that the standard of care for the suicidal patient includes a careful assessment of competence in decision making, the freedom of patients will be safeguarded, and clinicians will be freed (at least in a court of law) from the expectation that they be clinical mind readers who exercise total control of their patients' decisions.


We have suggested a clinical approach that is consistent with both the legal and clinical imperative to give full weight of attention to the patient's autonomy while remaining alert to the danger of suicidality. Such careful consideration is the best available antidote to the human tendency (in clinician, patient, judge, and jury alike) to think magically, thereby reducing a complex web of causation and influence to a single cause. We propose that the clinician use the informed consent dialogue creatively, undertaking a thorough, documented assessment of the patient's competence to weigh the risks and benefits of decreased restrictions, of prescribed medications and their psychodynamic meanings, and of disclosure of suicidal intent. Such assessments may well reduce both the likelihood of suicidal actions by patients and the inappropriate use of hindsight in malpractice litigation.

  1. Gutheil TG, Bursztajn H, Hamm RM, et al; Subjective data and suicide assessment in the light of recent legal developments: I. Malpractice prevention and the use of subjective data. Int J Law Psychiatry 6:317-29,1983.
  2. Shweder RA: Likeness and likelihood in everyday thought: magical thinking and everyday judgments about personality, in Thinking: Readings in Cognitive Science, Edited by Johnson-Laird PN, Wason PC Cambridge, Cambridge University Press, 1977, pp 446-7
  3. Fischhoff B: Hindsight/foresight: the effect of outcome knowledge on judgment under uncertainty. J Exp Psychol 1:288-99, 1975
  4. Bursztajn H, Feinbloom RI, Hamm RM, Brodsky A: Medical Choices, Medical Chances: How Patients, Families, and Physicians Can Cope with Uncertainty. New York, Delacorte/Lawrence. 1981
  5. Hirsh v. State of New York, 183 N.Y.S. 2d 175 (1959), 202 N.Y.S. 2d 296 (1960)
  6. Abraham K: The first pregenital stage of the libido (1916), in Selected Papers on Psychoanalysis. New York, Basic. 1960, pp 248-79
  7. O'Brien JP: Increase in suicide attempts by drug ingestion. Arch Gen Psychiatry 34:1165-9, 1977
  8. Neuringer C, Lettieri DJ: Suicidal Women: Their Thinking and Feeling Patterns. New York, Gardner, 1982
  9. Beck AT: Cognitive Therapy and Emotional Disorders. New York, International Universities, 1976
  10. Allison GT: Essence of Decision: Explaining the Cuban Missile Crisis. Boston, Little, Brown & Co. 1971
  11. Janis IL, Mann L: Decision Making. New York, Free, 1977
  12. O'Keefe DL: Stolen lightning: The Social Theory of Magic New York, Vintage, 1983
  13. Havens LL: The anatomy of a suicide. N Engl J Med 272:401-6, 1965
  14. Buie DH. Empathy: its nature and limitations. J Am Psychoanal Assoc 29:281-307, 1981
  15. Racker H. The meanings and uses of countertransference. Psychoanal Q 26:303-57, 1957
  16. Baldessarini RJ: Chemotherapy in Psychiatry (rev). Cambridge, MA, Harvard University Press, 1985, p 35-6
  17. Gutheil TG, Appelbaum PS: Clinical Handbook of Psychiatry and the Law. New York, McGraw Hill, 1982
  18. Bursztajn H, Gutheil TG, Hamm RM, et al: Subjective data and suicide assessment in the light of recent legal developments: II. Clinical uses of legal standards in the interpretation of subjective data. Int J Law Psychiatry 6:331-50, 1983
  19. Shein HM, Stone AA: Psychotherapy designed to detect and treat suicidal potential. Am J Psychiatry 125:1247-51, 1969
  20. Gutheil TG, Bursztajn H, Brodsky A: The multi-dimensional assessment of dangerousness: competence assessment in patient case and liability prevention. Bull Am Acad Psychiatry Law 14:123-9,1986
  21. Havens LL: Explorations in the uses of language in psychotherapy: counterprojective statements. Contemp Psychoanal 16:53-67, 1980
  22. Speer v. United States, 512 F. Supp. 670 (N.D. Tex. l98l)