A Psychiatrist’s Reaction to a Patient’s Suicide

As the overwhelmingly most common cause of death from psychiatric disorders, suicide has been the topic of many books, research articles, and philosophical treatises (1–3). In the literature on suicide, among the least commonly discussed topics is the reaction of mental health professionals when one of their patients in treatment commits suicide. This relative silence is especially noteworthy given that a substantial proportion of the more than 30,000 individuals who commit suicide yearly in the United States have been in treatment with a mental health professional, and many more have had recent contact with a primary care physician (4).

In the following case discussion, interspersed with a review of the sparse relevant literature, the responses of one clinician to a patient’s suicide are presented.

CASE PRESENTATION

Within a week of finishing his residency, Dr. G received his first referral to his private practice. Paul, a man in his early 20s, had been referred for treatment of depression by Dr. G’s former residency director/mentor. Paul had a history of dysthymia (although the term was not yet in use) and recurrent major depression, with a history of one suicide attempt several years before. When first seen, he was moderately depressed and expressed minimal suicidal ideation. Over the next 6 months of treatment, he was seen in psychotherapy once to twice weekly and was treated with a series of antidepressants, to which he either had an inadequate response or developed intolerable side effects. Six months after beginning treatment, a close relative of his, who lived in the same city and who was a major source of support, told Paul that he was moving to another city. Immediately, Paul became far more suicidal, announcing that he would eventually kill himself but that he wouldn’t do so immediately or in the near future because he did not want his parents, who lived in the Midwest, to suffer the pain of shipping his body back home. He said that he would wait until he returned to his hometown and then kill himself. Dr. G and Paul discussed hospitalization, but after being convinced that Paul had no imminent suicidal intent, Dr. G agreed to defer hospitalization, since Paul promised to call if he felt worse. Because he was busy at work and could not schedule an office visit for the next few days, Paul agreed to call Dr. G 2 days later (on a Friday) to establish whether any further intervention would be needed before the weekend. When Paul had not called by 3:00 p.m. on Friday, Dr. G began phoning him. Over the next few hours, between sessions with other patients, Dr. G phoned Paul’s workplace and his home and eventually managed to reach the one friend of Paul’s whose name he knew. When it became clear that no one had seen Paul that day, Dr. G called the local police and asked them to go to Paul’s apartment. They did so and saw his lifeless body through a window. The groceries he had bought that day were still on his kitchen table. He had overdosed on all the antidepressants he had in the apartment (tricyclics, which he had been taking at the time of his death, and monoamine oxidase inhibitors, which he had taken previously). He left no suicide note. The police called Dr. G, who then phoned Paul’s family to inform them of the tragedy.

Six months out of residency, Dr. G felt overwhelmed and numbed. For the first few days after Paul’s suicide, Dr. G walked through his life, doing all of his usual professional and personal activities. No one noticed any difference in his behavior. Internally, though, he felt as if he were acting, consciously trying to imitate his own customary behaviors. Although this sense of being depersonalized, which he had never experienced before, faded after a few days, these feelings would return for brief periods of time later, typically following some event that triggered anxiety related to the suicide.

Along with the numbed feelings, Dr. G began to feel an overwhelming sense of shame and embarrassment. While objectively knowing that it was not true, Dr. G felt as if he stood out as the only psychiatrist among his colleagues and friends who had ever had a patient commit suicide. (Because Dr. G would not discuss this event with anyone other than one close friend who, in fact, had never had a patient kill himself, he did not subject this distortion to verification.) He was afraid that if other psychiatrists knew about this event, the consequences would be severe. He feared that others would stop referring depressed patients to him (and mood disorders were his specialty), that his beginning academic career would be aborted, and that he would become the object of scorn within the professional community. Linked to these fears was a parallel terror of being sued and, consequently, of being publicly humiliated. (He was too young and inexperienced to realize that most malpractice suits are actually handled relatively quietly within large professional communities.) Interspersed with these feelings of shame and anxiety, concerns of self-protection for his career emerged; he began to brood about the documentation in his patient’s medical chart. Had he sufficiently documented his decision to not hospitalize Paul during their last appointment? How would the chart appear to expert witnesses or a jury? He negotiated with himself that if another patient committed suicide within the next year, he would simply move to another city and restart his psychiatric practice there, assuming that his current community would never accept him as competent if two suicides in his practice occurred within a relatively short period of time.

As part of the endless reworking and replaying of the last session with Paul and the day of his suicide, Dr. G began to wonder about clues he may have missed that might have indicated Paul’s suicidal intent. On the day of the suicide, for example, someone had called in to Dr. G’s answering machine and hung up. (This was, in fact, a relatively common occurrence.) After the suicide, Dr. G became convinced that Paul  had placed the call just before he overdosed. For the next few months, ­anytime his answering tape recorded a   hang-up, Dr. G quickly thought through his case list, wondering which of his suicidal/borderline patients had called and was now making a suicide attempt. His ruminations on the last session with Paul led to a pattern of awakening in the middle of the night, obsessing about the meaning of remarks or gestures made by one of his ongoing borderline patients at the end of the previous session. Was Bob’s closing remark at the end of today’s session, “Yeah, I guess I’ll see you next Thursday,” said a bit more sullenly than usual? Was he hinting that he was not going to return next week because he would kill himself before then? Dr. G struggled to not call these fragile borderline patients between sessions, aware that he would be calling to reassure himself rather than for appropriate therapeutic concerns.

Even with patients who did not exhibit significant suicidal ideation, Dr. G altered his pattern of assessing such thoughts. Without being consciously aware of the shift, he began to ask patients who were not suicidal about their suicidal thoughts; the questions were asked in a mechanical, rote manner, and the patients’ negative responses reassured him that he was at least asking the right questions.

Dr. G also felt deeply saddened by Paul’s death. He had been fond of Paul, who was in fact not a difficult, exasperating borderline patient, and because of a number of similarities in ethnic and family background, he identified with him. His sadness at the loss of Paul was mixed with another type of grief. He was painfully aware that his hopes and expectations about what a clinical career in psychiatry would be like were permanently changed. He had uncovered one of his primitive fantasies—that with his skill, empathy, and good training, he would make a positive difference in all his patients’ lives. That Paul could kill himself despite Dr. G’s best efforts and judgments made the practice and outcome of psychiatric treatment far less certain. Dr. G mourned the loss of his fantasy. For the first time in his professional life, he wondered whether he had the psychological toughness to tolerate clinical work with sick patients over the decades to come.

The small literature addressing therapists’ reactions to patients’ suicide ­describes several typical reactions. Among the first are disbelief/denial followed by depersonalization, shame, guilt, and the finding of omens (5, 6). As an example of denial, one psychiatrist asked about the chance of recovery for a patient who was described to him as brain-dead after a self-inflicted gunshot wound. In the present case, Dr. G seemed not to have had a great deal of denial or disbelief. Instead, his predominant initial response was de­personalization, characterized by a feeling of numbness and unreality, while he maintained his usual behavior patterns.

Shame and guilt are common and predominant responses following a patient’s suicide. As described in other accounts, guilt has led psychiatrists to offer false confessions. At a review after a depressed patient committed suicide, one psychiatrist admitted to having prescribed an inadequately low dose of an antidepressant, which then led to a discussion that was critical of his care. In reality, he had prescribed a much higher and more adequate dose, but in response to his guilt at the outcome, he had falsely remembered his own treatment, making it seem more inadequate than it was (5). Dr. G’s fear of public humiliation and of potential litigation (a real threat) typifies therapists’ reactions to patients’ suicide. (No lawsuit was filed in Paul’s case.) Dr. G’s concern with “the way the chart looks” is also common, especially in academic medical centers, where the importance of adequate documentation is particularly emphasized.

Dr. G’s fantasy of moving to another city in response to his shame and fear of retribution by his community is also not unique. As described in a case report (7) unknown to Dr. G at the time, another psychiatrist, after a patient’s suicide, noted, “It reached the point in my fantasy where I was having to leave town, shunned like a leper for the terrible act I had committed.”

The searching for omens, exemplified by Dr. G’s looking for clues that might have alerted him to Paul’s incipient suicide, reflects the need to maintain one’s illusion of control in an unpredictable world in order to bind anxiety. The fact that practitioners and entire institutions alter treatment practices on the basis of a single suicide sometimes reflects the desperation to create new procedures that will allegedly protect against the unwanted event. Some changes in clinical practice as a consequence of a patient’s suicide are, of course, appropriate and reflect improvements in care. Other changes, however, more likely reflect magical thinking. Dr. G’s ritualistic inquiries about suicidal ideation in patients for whom this was clinically not necessary allowed him to feel that he was establishing more effective control of suicidal feelings and behavior in his practice, although these inquiries were not used in a clinically wise or particularly helpful manner. In institutional settings, changing hospital-wide policies and procedures on the basis of a single untoward event is common. To illustrate, a hospitalized patient killed himself after returning from a court hearing in which his petition to be released from an involuntary admission was denied. Thereafter, the ward leaders established a new policy requiring that after similar court hearings, patients would be obliged to stay in the dayroom for the rest of that day, ignoring the many hundreds of other such court hearings that did not lead to suicidal behavior.

Dr. G’s grief at Paul’s death is understandable. Less obvious, although well described in the literature, was his need to mourn the loss of grandiosity about his abilities to positively affect every one of his patients’ lives. This loss is generally considered to be maximally felt by young psychiatrists, either trainees or recent graduates, who have a patient who commits suicide.

Over the next weeks and months after Paul’s suicide, Dr. G continued to feel shame and continued to look for omens regarding potential suicide in his ongoing patients. Although he continued to function normally, internally he was anxious and perseverative. He could not stop replaying the scenes of the day of Paul’s suicide, obsessing about the number of times he had tried to call Paul both at home and at work. Whenever the telephone rang at his home in the evening, his heart rate jumped, and he was (transiently) convinced that the police were calling to inform him that another of his patients had been found dead from an overdose of medications that he had prescribed.

Mixed with Dr. G’s fear that another patient might commit suicide were his growing angry feelings toward Paul, which he did not share with any of his colleagues. He was angry that Paul had not waited until returning home to the Midwest before killing himself; this would have relieved Dr. G of the responsibility of being the treating clinician at the time of his death. Although acknowledging the selfishness of these feelings, Dr. G was angry with Paul for having caused him such pain, anxiety, and guilt. Most of all, however, he was angry with Paul for having killed himself with medications that he had prescribed. To Dr. G, this felt especially unfair—to have a patient commit suicide by using the personal healing ministrations of the doctor, the actual means of his healing. He became painfully aware of the fact that as the medical director of a mood disorders clinic, he was treating more than 100 patients with unipolar and bipolar disorder, virtually all of whom (in the early 1980s) had enough medication at home to kill themselves. His concerns about these possibilities were sufficiently intense that, unconsciously, he transmitted this awareness to some of his patients. For example, one of his patients was a nurse with borderline personality disorder who was taking tricyclic antidepressants and who frequently flirted with low-dose insulin injections as a self-destructive gesture. She eventually became so exasperated with Dr. G’s persistent questions about overdosing—which she had never done or even threatened to do—that she told him, “Look, I can’t promise that I won’t kill myself, but I promise that if I do it, it won’t be with your pills. So, leave me alone already!”

Dr. G’s anxiety, middle-of-the-night awakenings, intrusive thoughts of the day of the suicide, startle response to phone calls, and the previously described depersonalization are all symptomatic elements of a posttraumatic stress syndrome. Other accounts of psychiatrists’ responses to patients’ suicide include their developing hallucinations, accident proneness, and suicidal ideation. In one study, within 6 months after a patient’s suicide, 57% of psychiatrists reported intrusive thoughts and avoidant behaviors at a level comparable to that of clinical populations (8). The two emotions most fraught with internal conflict for the psychiatrist after a patient’s suicide are anger and relief. Dr. G’s anger is fairly typical of responses described elsewhere. Often, the anger is projected at others—the family of the patient, a supervisor if the treating psychiatrist is a resident, nurses in an inpatient facility, and so on. (Projected anger as a method of protecting against grief is thought to be a frequent factor in families filing lawsuits after patients’ suicide.) Relief after suicide results from not having to continue to struggle with a patient’s suicidal ideation and threats. Such relief most commonly occurs after the death of chronically suicidal patients, who tend to exhaust even the most experienced clinicians. Since Paul was not chronically suicidal, Dr. G experienced no relief at all; however, relief definitely accompanied Dr. G’s responses to the suicide of a chronically suicidal patient many years later.

Ultimately, Dr. G wanted to contact colleagues about the suicide, partly to confess and ask for acceptance and partly to simply feel less alone. When he discussed Paul’s suicide with a good friend who was an internist, the friend made a small joke and changed the subject. He talked about the suicide with a few colleagues from his residency, but as he expected, none of them had yet experienced a patient’s suicide. They were sympathetic, but since they had not had similar experiences, their responses were not particularly helpful to Dr. G. In some respect, Dr. G began to feel as if he were a member of a very elite club. Although most feelings associated with this club were deeply painful, he also recognized feeling a sense of specialness, of having gone through an experience that only those who had also had the experience could really understand.

Aside from his attempts to discuss his feelings about Paul’s suicide with colleagues and in his individual therapy, Dr. G made appointments with two former teachers. First, Dr. G went to see his former residency director and mentor, the person who had originally referred Paul to him. Dr. G saw this as going to confess to a father figure, and he hoped to achieve some absolution for his failure. He believed that confession would allow a proper chastisement but that he would avoid total rejection by this important figure. Within himself, he was unclear to what extent he was simply seeking support and/or a more even-handed discussion about how he had handled Paul’s case. Had he made errors that might have contributed to the outcome? The intensity of his feelings made objective thought about the issue almost impossible. His second appointment was with a former supervisor who was a very senior psychiatrist and an expert in treating chronically suicidal patients. Dr. G wanted advice on how to cope with his anxieties about clinical practice and how to survive for the next few decades in this unpredictable, dangerous field.

As might be expected, Dr. G’s mentor was supportive and did not castigate him. However, his mentor acknowledged that he himself had never had a patient who committed suicide while under his care, which made Dr. G again feel as if he were “special” in both the positive and negative senses. Then, he and his former supervisor talked about how to cope with the stresses and uncertainties of clinical work. Dr. G followed the advice to not accept new patients who were known to have significant suicidal potential until he felt he had more substantially “recovered” from Paul’s death. Years later, Dr. G recalled both meetings as being exceedingly helpful although, except for the advice about being more selective with referrals, he remembered almost nothing of the content of either discussion.

Among the most common concerns after a patient’s suicide is the response of colleagues. Dr. G’s responses are typical, vacillating between fears of being shunned and feeling special. Some authors have described the experience of having a patient commit suicide as a rite of passage; “to survive it is testimony to one’s hardiness, endurance, and being a ‘real physician’ ” (5). At the same time, the feelings of aloneness can be debilitating and provoke the seeking out of colleagues, as Dr. G did. Unfortunately, the reaction of his internist colleague is common (although not universal) among nonpsychiatric physicians. Since most physicians have some patients who die while under their care, they do not immediately perceive the different quality of a patient’s suicide for a psychiatrist.

In training programs and psychiatric hospitals, a “psychological autopsy” or suicide review is a standard procedure. In these conferences, the treating psychiatrist presents and discusses the case with other staff members with the goals of evaluating issues of the quality of care and of learning from the tragedy. The usefulness of such a conference for the treating psychiatrist is unclear. At least one study found that the psychological autopsy compounded rather than alleviated the treating psychiatrist’s doubt (9). Too often, psychological autopsies evolve into public shaming. Finding a middle ground between mindless support and excessive critical thinking is surprisingly difficult.

Over the next 10 years, as Dr. G continued to specialize in caring for patients with treatment-resistant depression, he had other patients under his care who committed suicide. As his career evolved and he became more established within his community, a patient’s suicide would not provoke the same fears of a failed career, although many of the other psychological reactions that he experienced after Paul’s death would reemerge, albeit in less intense form. The suicide of one patient who was in psychotherapy with another psychiatrist (who had seen the patient immediately before the suicide), but for whom Dr. G was the ­attending psychopharmacologist, provoked many of the now-familiar feelings of horror, self-doubt, and intrusive images, although the sense of guilt was less intense.

A number of variables will shape the impact of a patient’s suicide on the psychiatrist. Of these, the two main factors are the relationship between the patient and the psychiatrist and the individual psychiatrist’s psychological makeup, including personality features and the developmental stage of his or her career.

It is generally assumed that the more intense the relationship between patient and psychiatrist, the more difficult will be the psychological reaction to the patient’s suicide. Aside from the sheer duration of the therapeutic relationship, the nature and interpersonal intensity of the treatment—i.e., whether the modality was entirely psychopharmacological management, with less frequent and shorter visits, or some form of psychotherapy—seems especially relevant. However, rather deep relationships can develop in the course of psychopharmacological management over long periods of time, which might then predict more intense psychological reactions in the event of a suicide. Another important aspect of the patient-psychiatrist relationship is the extent to which the patient’s treatment is shared with other professionals. Both Dr. G’s experience and common sense suggest that situations in which the treating psychiatrist is the sole provider of care, such as individual solo practice, are associated with the greatest sense of responsibility, loss, and psychological distress in case of a patient’s suicide. A split treatment situation, in which the patient is in therapy with another mental health professional while the psychiatrist provides medical treatment, may buffer the sense of responsibility and guilt. This was especially clear to Dr. G when the patient for whom he was only part of the treatment team killed herself. The fact that another psychiatrist had seen the patient in psychotherapy a few times between her last appointment with Dr. G and her death significantly blunted his sense of guilt, although it did little to minimize the other psychological responses.

Psychological factors within the psychiatrist also play vital roles in predicting response to a patient’s suicide. These include an obsessional personality style, a tendency to internalize, and vulnerability to anxiety and depression. The fact that Dr. G had been out of residency for only 6 months at the time of Paul’s suicide was unquestionably a critical factor in accounting for his intense psychological reaction. The suicide of a patient is tragic and painful at any time; however, if the psychiatrist is older and has achieved some sense of competence, a mature professional identity, and some respect within his professional community, the ability to buffer the pain with internal strengths is likely to be greater.

A decade after Paul’s death, Dr. G began to lecture to residents and colleagues on the topic of therapists’ reactions to patients’ suicides. Of the many topics about which he regularly lectured, this talk was always the most difficult to prepare and deliver. Before each lecture, he experienced surges of anxiety and unease. He was gratified that each time he gave the lecture, one or more trainees (and, often, more senior colleagues) would acknowledge similar sets of experiences that they too had kept secret because of their sense of shame.

Three of the most important methods for coping with a patient’s suicide are decreasing the sense of isolation, making efforts at reparative, constructive behavior, and using specific cognitive defenses. Decreasing isolation is best managed in the same manner that we typically suggest to our patients—talking to others one trusts and respects, be they lovers, friends, family, colleagues, or one’s own therapist. For each individual, different sets of people will be most helpful. In Dr. G’s case, talking with his girlfriend/wife-to-be (who was herself a clinical psychologist) and discussions with his two former teachers were the most helpful. The psychological autopsy, when done well, can also diffuse the isolating nature of the experience, assuming that it does not evolve into a critical blame-finding discussion. Sometimes, it is helpful to the psychiatrist (as well as to the patient’s family) to meet with the bereaved family, to grieve together in an atmosphere that is hopefully not dominated by blame and projection. In Paul’s case, the family declined Dr. G’s invitation to meet with them. In other cases, when Dr. G did meet with families of patients who committed suicide, the encounters were exceedingly helpful to him and, he observed, to the family members, who were often unaware of some of their dead relatives’ psychological struggles.

Repair is typically attempted by helping others prepare for or cope with similar experiences. Presenting a case of a patient’s suicide at grand rounds conferences, as Dr. G did, and writing a case report are typical examples. My writing of this Clinical Case Conference with myself as the thinly disguised Dr. G as my alter ego is another example. (Writing in the third person also helped me to distance myself slightly from the affect associated with the writing.) As has been pointed out elsewhere (5), the act of public repair may also reflect other motives, such as the need for public confession or a request for support.

Embracing certain facts and philosophical viewpoints can also help one to cope with a patient’s suicide. These include the fact that suicide is a predictable outcome of major psychiatric disorders, especially depression; that suicide is inevitable when an individual is not ambivalent about it; and that the field is currently unable to accurately predict suicidal risk for any individual (10). As a method of reducing the intensity of his guilt/pain, Dr. G often compared the potential outcome of suicide in a population of severely depressed individuals with the potential outcome of death from terminal cancer in patients treated in oncology practice. In both fields, the deaths reflect the natural history of the disorders and the imperfection of therapeutic modalities more than the failure of individual clinicians. Of course, these statements can also be used inappropriately for defensive purposes, to foster a sense of therapeutic nihilism, or to deny therapeutic responsibilities inherent in being a clinician.

CONCLUSIONS

A substantial proportion of psychiatrists will, at some point in their careers, experience the suicide of one or more of their patients. Overall, the field has been relatively silent about this phenomenon, as exemplified by the sparse literature on the topic and the paucity of formal attention given to it in residency training programs. Yet, except for protracted malpractice suits, patient suicides may be the most psychologically difficult experiences encountered in the life of a psychiatrist. Dr. G’s response to Paul’s suicide—­including a combination of post­traumatic stress disorder symptoms, shame, guilt, anger, isolation, and fears of both litigation and retribution from the psychiatric community—is typical of responses described elsewhere. Having a patient commit suicide early in one’s career may be particularly difficult, as it was for Dr. G, suggesting that training programs should prepare trainees for these tragic events. Strategies that decrease the isolation of the clinician and foster perspectives that reduce self-blame may blunt one’s overdeveloped sense of responsibility and limit one’s fantasies of having ultimate control over patients’ lives.

Received April 8, 1999; accepted June 30, 1999. From the Department of Psychiatry, University of California, Los Angeles. Address reprint requests to Dr. Gitlin, 300 UCLA Medical Plaza, Suite 2200, Los Angeles, CA 90095-6968