January 1, 1999

It was gratifying to read the article "Experts Share Advice on Reducing Risk When Treating Potentially Violent Patients" in the October 2 issue for its useful tips on violence management and risk management. It was proportionately frustrating to have those same experts make the common mistake about the 1976 California legal decision Tarasoff v. Regents, a decision described in the article as holding that psychiatrists had a "duty to warn" potential victims of danger from patients. This is simply wrong: the 1976 decision (the second and final version) defined a duty to take reasonable steps to protect potential victims.

This correction is not mere nitpicking. Warning victims breaches confidentiality without materially affecting the safety of victims, whose options, once warned, are quite limited; warnings, moreover, may precipitate preemptive violence rather than preventing it. In contrast, the clinician's focus on protecting victims permits attention to the best clinical response to clear and present danger: voluntary hospitalization, with involuntary commitment as the fallback position.

Why do so many experts, including forensic psychiatrists who should know better, make this same mistake repeatedly? My own hypothesis is that-when the earlier, superceded Tarasoff decision in 1974 proposed a duty to warn (and was greeted by a storm of protest from mental health organizations and others)-the idea of warning someone outside the therapy dyad was so arrestingly nonclinical a notion that it became stuck in the collective temporal lobe of psychiatry to such a degree that even experts still get it wrong.

Finally, the article omits an important approach to decreasing violence: involving the potential victim in the assessment and management of the dangerous situation by such means as phone calls, meeting in a protected setting, and the like, as described by Wulsin and colleagues in the May 1983 issue of the American Journal of Psychiatry. Interventions of this type offer hope of decreasing the actual danger in the situation, in addition to protecting against liability. Clinicians should consider this approach as well as the others suggested in the article.

Thomas G. Gutheil, M.D.
Program in Psychiatry and the Law
Mass. Mental Health Center
Harvard Medical School
Cambridge, Mass.

PRMS Responds

We appreciate the opportunity to respond to Dr. Thomas Gutheil's letter regarding our article on working with violent patients. We have reread both our article and the opinion in Tarasoff v. The Regents of the University of California, 17 Cal.3d 425, 551 P.2d 334 (1976). We find that Dr. Gutheil's comment regarding our treatment of Tarasoff is not entirely accurate.

Dr. Gutheil contends that we described Tarasoff as "holding that psychiatrists had a 'duty to warn' potential victims of danger from patients." We did not say this. We stated that "[the] Tarasoff court held that once a psychologist knows that a client poses a danger of violence to another, the psychologist has a duty to exercise reasonable care to protect the foreseeable victim of that danger. Discharging the duty may mean warning the potential victim . . . ." This is paraphrased from the Tarasoff opinion itself, the second and final version to which Dr. Gutheil refers. Id. at 431. We then went on to explain that various subsequent courts have broadened the duty to include psychiatrists and have specified an affirmative duty to contact the foreseeable victim.

We believe that Dr. Gutheil's comment stems from our statement that Tarasoff "established the precedent for a 'duty to warn.'" We support this statement. While the court formally recognized a broad "duty to protect" and never explicitly mandated contacting the foreseeable victim, the concept of warning permeates the opinion and is the principal focus of the court's decision-making process. Indeed, although the wording of the court's holding uses the inexact and undefined phrase "reasonable care to protect," 17 Cal.3d at 450, a thorough reading of the opinion and the subsequent case Thompson v. County of Alameda, 27 Cal.3d 741, 614 P.2d 728 ( 1980), reveals that "reasonable care to protect" refers primarily to the concept of warning.

The only cause of action recognized by the court in Tarasoff was titled "Failure to Warn on a Dangerous Patient," and that cause of action was allowed to be amended specifically to include the psychologist's failure to warn the victim. 17 Cal.3d at 450. The court's reasoning regarding this particular matter concludes that "[i]n this risk-infected society we can hardly tolerate the further exposure to danger that would result from a concealed knowledge of the therapist that his patient was lethal. If the exercise of reasonable care to protect the threatened victim requires the therapist to warn the endangered party or those who can reasonably be expected to notify him [or her], we see no sufficient societal interest that would protect and justify the concealment. The containment of such risks lies in the public interest. For the foregoing reasons, we find that plaintiffs' complaints can be amended to state a cause of action against defendants . . . , for breach of duty to exercise reasonable care to protect [the victim]." Id. at 442. In other words, failure to warn may breach the therapist's duty to protect. No other examples of breaches are given.

The other cause of action addressed by the court, titled "Failure to Detain a Dangerous Patient," was barred by governmental immunity and, thus, became moot. Id. at 450. Because of this, the court did not answer the question of whether the defendants also had a duty to commit the patient in order to protect the victim. The correlate of this decision is that even therapists who are otherwise protected by governmental immunity may still have a duty to warn in order to protect. The duty to warn and the duty to detain, both means of protecting, can be viewed, and judged, separately.

In Thompson the court clarified its previous ruling and explicitly limited it to identifiable victims. There is no "general duty to warn," 27 Cal.3d at 752, it states. Rather, in Tarasoff the court had concluded that "the therapist has a duty to warn either 'the endangered party or those who can reasonably be expected to notify him . . . .'" Id. Once again, the court's only example of "reasonable care to protect" is warning, referred to specifically as a "duty to warn." As in Tarasoff, no other examples are given.

In technical terms, Tarasoff established that, in California, a treater, in this case a psychologist, owes a legal duty to a third party who is not a patient (i.e., the foreseeable victim). This duty involves exercising "reasonable care to protect" the foreseeable victim, for example, warning the victim. Failure to exercise "reasonable care to protect" (i.e., breaching the standard of care) can result in the treater being held liable for damages caused by the patient. This can be described as a "duty to protect." The "reasonable care" component of the theory may include acts other than warning, but the court does not specify them. Clearly, then, reasonable interpretation and practical application of Tarasoff can be limited for purposes of discussion to a "duty to warn."

We are not alone in our assessment of Tarasoff. Numerous courts and academics throughout the country have identified Tarasoff as the foundation and the precedent for a "duty to warn." They usually refer to a "duty to warn" or a "duty to warn or otherwise protect." Even when they refer to a "duty to protect," they speak almost exclusively about the concept of warning. See, for example, Brady v. Hopper, 570 F.Supp. 1333 (D. Colo. 1983); Chrite v. United States, 564 F.Supp. 341 (E.D. Mich. 1983); Leedy v. Hartnett, 510 F.Supp. 1125 (M.D. Pa. 1981); Douglas M. McIntosh, Carmen Y. Cartaya, Psychotherapist as Clairvoyant: Failing to Predict and Warn 59 Defense Counsel Journal 569 (1992); Sherrie A. Wolfe, The Scope of a Psychiatrist's Duty to Third Persons: The Protective Privilege Ends Where the Public Peril Begins 59 Notre Dame L. Rev. 770 (1984). For a listing of additional supporting cases and law review articles, refer to the end of this letter on the Psychiatric News' Web site at .

Dr. Gutheil maintains that his comment is not nitpicking, because a focus on protecting the victim is more constructive clinically than a focus on warning the victim and thus violating confidentiality. We recognize the distinction. Warning is but one method available to a psychiatrist clinically when faced with protecting a potential victim. However, in many jurisdictions either the duty to warn is mandated by state statute or the failure to warn exposes the psychiatrist to potential liability in court. It is absolutely true that unauthorized disclosure violates confidentiality and may threaten the therapeutic relationship. It is precisely because of this conflict that the dilemma is so powerful, resonates so loudly. The "duty to warn" is an imperfect resolution, and it is based on public policy, not on professional expertise. It is ultimately a society-based decision, not a clinical decision.

Our article was a risk-management article, and we chose to discuss two of the most troublesome risk management-related issues involved with treating potentially dangerous patients: confidentiality and the duty to warn. We did not believe a thorough analysis of Tarasoff and its legacy would have been appropriate given the circumstances.

Jacqueline M. Melonas, R.N., M.S., J.D.
Marynell Hinton, M.A.

The writers are staff attorneys at PRMS, the administrator of the APA-sponsored Professional Liability Insurance Program.