Mental Health Consult

No-suicide contracts: Can they work?


An elderly, retired, married African American man sought psychiatric treatment for depression and suicidal thoughts. He had a detailed, lethal suicide plan, but he had not taken any steps to carry it out.

Dr. Caroline Roberts

Dr. Caroline Roberts

He met DSM-5 criteria for a major depressive episode, and he described a lifelong history of recurrent depressions as well as hypomanic episodes. He was diagnosed with bipolar II disorder, and he began weekly therapy, as well as medication. Despite several static and dynamic suicide risk factors, the psychiatrist also noted that he was help seeking and future oriented. He seemed transparent during his initial appointments. He did not have access to lethal means and welcomed the psychiatrist to communicate openly with his spouse.

The patient had never attempted suicide, there was no family history of suicide, and there was no psychosis or substance use disorder present. He was able to commit to reaching out to the psychiatrist, his spouse, or emergency personnel in the case of worsening suicidal thoughts or imminent suicidal action. He remained in the outpatient setting. His suicidal ideation faded and depression receded as psychotherapy and pharmacotherapy continued.


Depression and suicidal ideation are ubiquitous in the practice of psychiatry. Psychiatrists draw from an array of assessment and management tools when this common clinical challenge arises. Among these tools is the no-suicide contract (NSC). The NSC goes by many names, including the no-harm contract and suicide prevention contract.1 It is a promise, verbal or written, from the patient to not attempt suicide and to tell a loved one or psychiatric provider instead.2 The verbal exchange between the patient and therapist described in the case fits the widely accepted clinical definition of an NSC. The contents and implementation of NSCs vary greatly; no standard approach is taught in psychiatric training.3 The American Psychiatric Association has warned against over-reliance on them, emphasizing that they have not been proven effective. It advises that NSCs should not be used independently of other tools or outside well-established patient-provider relationships.4 A 2007 review of the literature on NSCs concluded that there were no data to support their effectiveness and some data that they might even cause harm.5

The origin of the NSC

The NSC is fairly young and its foundation arguably weak. Its evolution has been traced back to a study published in 1973 by Robert C. Drye, MD, and associates on the effectiveness of a questionnaire for the assessment of suicide risk.6 The questionnaire centered on the patient’s reaction to the statement, “No matter what happens, I will not kill myself, accidently or on purpose, at any time.” The authors placed special emphasis on the words “I will,” which they felt to be a stronger indicator of commitment to safety than “I promise.” The authors thought the latter statement sounded like a child’s empty reply to a demanding parent. The authors reported a 100% success rate with “approximately 600 patients” across geographic regions and clinical settings.7 The study group is not further described, and that the authors contend that the intervention had “complete effectiveness in evaluating suicide risk” should give pause to anyone aiming to practice evidence-based psychiatry.

The theoretical basis of the NSC has been presumed by others to be based, in part, on the principles of transactional analysis. Specifically, the suicidal patient is seen as occupying the child ego state, and the NSC is seen as a means of moving the patient into the less problematic adult ego state. It has been argued, however, that an NSC can achieve exactly the opposite. The contract can pit the patient against the clinician, entrenching the patient deeper into the child ego and, therefore, suicidal state.8

Michael Craig Miller, MD, and associates proposed other psychological reasons why NSCs may be counterproductive. They write, “Psychological pitfalls abound, and any of them may contribute to a contract being thoughtless, unrealistic, irrelevant, cynical, punitive, or coercive.”9 They postulated that the NSC grew out of and assumes the same shared decision-making inherent in any therapeutic contract – and they argue that this assumption is flawed given the legal power clinicians have over suicidal patients. While acknowledging this problematic power differential, the authors go on to urge clinicians to aim for shared decision-making and a shared burden of risk when discussing treatment with suicidal patients.

Possible NSC common factors

Psychiatry, like the rest of medicine, is increasingly practiced in an evidence-based manner. The NSC should not be excluded from this movement. To this end, a recently published, randomized study of 97 active duty Army personnel seeking emergency behavioral health evaluation compared the effectiveness of NSCs and with an alternative intervention, the crisis response plan (CRP). The CRP was chosen because it had been suggested by the Joint Commision as an alternative to the NSC, although it also has little evidence supporting its use.10

The NSC and CRP interventions of the Army study were very similar. Both included suicide risk assessment, supportive listening, provision of crisis resources, and referral to treatment. In addition, the NSC intervention included asking whether the patients could keep themselves safe at home. The CRP intervention included collaboration with the patient to identify warning signs of crisis, self-management skills, and support persons. A seemingly small but interesting difference between the two interventions was which member of the dyad, patient or clinician, created a written record of the discussion. In the NSC group, the assessor did the writing, while in the CRP group, the patient controlled the pen.

The results of the study were intriguing. Suicidal ideation declined faster in the CRP arm. Participants in the CRP arm were 76% less likely to attempt suicide over 6 months, although this effect decreased and lost statistical significance when controlling for baseline severity of suicidal ideation. Despite those promising data, the only completed suicide was in the CRP arm.

The authors compared the makeup of the CRP intervention with key components of dialectical behavior therapy (DBT). They pointed to a 2015 study by Marsha Linehan, PhD, and associates that sought to identify the active ingredients of DBT. The Linehan study indicated that attending to warning signs and using self-management tools and social supports contributed more to the success of DBT than the individual therapy component. Interestingly, these were the same features that set the CRP intervention apart from the NSC in the Army study. Perhaps these are the common factors of effective counseling of suicidal patients.

Indeed, these factors seem to harken back to the NSC as originally envisioned by the late Dr. Drye – a patient-driven collaboration. Dr. Drye and associates wrote: “This approach developed out of our belief that the only therapeutic contracts likely to lead to change are those developed by the patient himself, for which he will assume responsibility.” While the data presented by Dr. Drye and associates were weak, the theory behind their NSC – patient commitment – seems solid. Commitment strategies, which grew out of social psychology, are effective and heavily used in DBT, including to decrease suicidal behaviors.11


Next Article: