Psychiatric Malpractice Grand Rounds: The Tarasoff Dilemma

September 27, 2019

An expert provides sage advice to avoid a malpractice lawsuit, even in the face of potentially tragic outcomes.

Forensic psychiatrist and Editor in Chief Emeritus of Psychiatric Times, James L. Knoll IV, MD, weighs in on a duel case report: To Invoke or Not to Invoke: Tarasoff Is the Question, by Elliott B. Martin Jr, MD.


Thou shalt not stand idly by the blood of thy neighbor. -Leviticus 16

Psychiatrist Elliott Martin, MD, graciously provides us with two rich clinical scenarios and notes the understandable confusion, anxiety, and ethical conundrum that may arise when the Tarasoff conflict is “invoked.” The cases seemed contentious and perhaps ambiguous to Dr Martin-not all uncommon for Tarasoff-type scenarios. In his case examples, Dr Martin refers to the Tarasoff duty as a duty to warn, and so let us take a moment to clarify this often misunderstood concept.

It has been my experience that confusion persists regarding the meaning and use of the terms duty to warn and duty to protect. This may be partly due to the fact that there were two Tarasoff decisions. The first Tarasoff decision in 1974 created a duty to warn in California and was based on the special relationship between therapist and patient.1 This first decision was unprecedented, and quite upsetting, to therapists due to its legal compromise of patient confidentiality. An oft quoted line from this first Tarasoff decision made it clear that the Court was concerned with social policy: “The protective privilege ends where the public peril begins.”

In “Tarasoff II,” the California Supreme Court reheard the case, noting plaintiffs’ argument that therapists failed to exercise reasonable care to protect Tatiana Tarasoff.2 Although the police were warned, no other steps were taken such as detaining Poddar or warning Tatiana of the danger. In its 1976 ruling, the Court replaced duty to warn with a duty to protect.2 The famous quote from Tarasoff II, which was adapted by many states across the country, made the change clear: “When a therapist determines, or should determine, that his patient presents a serious danger of violence to another, he incurs an obligation to use reasonable care to protect the intended victim from danger.”

Initially, there was concern that this exception to confidentiality would have a disastrous effect on psychiatric practice, despite the fact that most therapists had embraced such a duty before the Tarasoff ruling.3 Over time, it became clear that the concerns about the potential loss of confidentiality did not have an adverse impact on psychiatric practice.4 In 2013, after decades of misunderstanding, California passed legislation which unambiguously established a sole duty to protect.5 The California statute removed all references to duty to warn, and provides a “definitive clarification.”6

The persisting confusion among clinical psychiatrists about the duty to protect is understandable given that: 1) there are no widely accepted clinical guidelines to follow; 2) the duty varies by state; and 3) individual states’ duties have changed over time according to case law and/or state statute.7

Although Tarasoff applied only in California, the ruling had a national reverberation. The duty to protect articulated in Tarasoff was subsequently interpreted more broadly by other courts throughout the US. One of the broadest interpretations occurred in the 1980 case of Lipari v. Sears, Roebuck & Co.8 This case involved a VA patient who shot strangers in a crowded nightclub, without ever threatening a specific person, and one month after terminating psychiatric treatment. The court rejected the Tarasoff limitation to an identified victim, imposing not only a duty on therapists to predict violence, but also a duty to protect unidentified victims in the general public.

Similar cases in the wake of Tarasoff eventually led to strong objection to such legal expectations. A remarkable example of this was the case of Naidu v. Laird, which further expanded the duty to unidentified victims and unintentional harm.10 The case involved a patient with schizophrenia who killed another man in a motor vehicle crash. The patient’s psychiatric history included violent behavior, ramming a police car with his automobile, and driving off the road at high speed. The Supreme Court of Delaware held that five and a half months after a hospital discharge was not too long a period to support a finding of negligence when a psychiatrist was found liable for failing to foresee a patient’s potential to act violently towards the general public. Two decades after Tarasoff, courts around the country began to reflect ambivalence about the extension of the duty to protect.

As a result, some case law shifted to require that the threat be clearly made, and that the duty extended only to reasonably foreseeable victims-not to the general public. Many states subsequently adopted statutes known as “Tarasoff-limiting statutes,” which gave specific criteria (ie, a credible threat made against an identifiable victim). Duty to protect statutes have been passed in all but 13 states.10

When faced with a potential duty to protect in clinical scenario, the first step is to become familiar with the specific Tarasoff duty in one’s state, as well as any evolving case law which may create nuances in how the duty should be carried out. The precise language specifying the duty in Dr Martin’s jurisdiction will be critical. For the sake of discussion, let us note that states with mandatory duty to protect statutes contain language that can often be distilled down to two criteria: 1) an explicit, realistic threat, 2) against an identifiable person.11

Turning to Dr Martin’s first case, it was decided that there was an ethical and clinical duty, despite the lack of a legal duty to breach confidentiality. Dr Martin goes the distance-handling an extremely difficult meeting with his patient’s mother. His empathic interaction and warning to her were clearly performed in good faith. Such interactions, particularly when well documented, are extremely likely to be viewed favorably by a jury in the event of a lawsuit. Further, one must imagine there may have been a greater potential for a lawsuit had the mother not been warned in such a compassionate, patient manner.

While the court and forensic specialists were understandably focused on whether the criteria were met for a legal duty, Dr Martin opted to address his ethical and clinical duties. This was an acceptable option, particularly in jurisdictions that have a permissive, as opposed to a mandatory, duty to protect. Yet another option, depending upon the clinical circumstances of a case, might be for the psychiatrist to collaborate with the patient to carry out the warning.12 This would involve a clinical interaction in which it is explained to the patient that the psychiatrist wishes to act in the patient’s best interest (which clearly involves the patient not being arrested for a violent crime). It would be ideal for the psychiatrist and patient to make the warning together, if the patient gives permission and is willing to cooperate. In cases where the patient refuses to cooperate despite a therapeutically communicated explanation, and the patient also does not meet civil commitment criteria (eg, the threat flows from antisocial personality disorder), the psychiatrist may then inform the patient that a warning will be made in his or her best interests and per legal duty.

In his second case, Dr Martin describes what the stalking literature refers to as an Intimacy Seeking stalker who suffers from an erotomanic delusion.13 Making matters more complicated is the presence of a bizarre delusion (believing he was impregnated), as well as the patient’s strong desire for revenge towards his “love/hate object.” This latter element of the case adds significant risk and may be considered a “high risk psychotic phenomenon” according to the Stalking Risk Profile-a structured professional judgment instrument intended to assist in the assessment of the type and level of risk in stalking cases.14,15

Other risk factors per the stalking literature that Dr Martin describes include the patient’s persistent and elevated level of anger and strong paranoid ideas. Indeed, Intimacy Seeking stalkers with erotomanic delusions tend to be impervious to the threat of criminal sanctions and may even regard court appearances and imprisonment as the “price of true love.” In cases of erotomanic stalkers who delusionally believe a relationship already exists with the victim, there is the risk that stalkers may become jealous due to misperceptions of “infidelity” on the part of the victim. Though the overall risk is lower than stalkers of former intimate partners (Rejected Type stalker), erotomanic stalkers can occasionally be responsible for extreme violence.16

Dr Martin gives clear data for us to consider when we turn towards an analysis of the duty to protect scenario. For this second case, we find clear threats (both verbal and written) made against a readily identifiable victim (a classmate in his school). Having confirmed that the two relevant criteria are met, the psychiatrist then has a number of intervention options to consider, depending upon the clinical scenario and statute. Note that we have not yet reached a need to carry out the duty to protect by breaching confidentiality. Indeed, the duty to protect does not inexorably demand a call to police and/or the identified victim. First, a careful consideration of the available options should take place. These options most often include:

• Hospitalization (or escort to a hospital emergency department for evaluation)

• Notifying police

• Warning the potential victim

• Increasing the frequency of outpatient appointments

Breaching confidentiality may be viewed as a last option after all other therapeutic options have been exhausted.17 This will always remain, in part, a clinical judgment call. However, one helpful guide is that confidentiality should be breached only if reasonable clinical efforts seem unlikely to provide adequate resolution, and harm to the victim remains foreseeable. When all reasonable options are untenable, consider that “trust”-not absolute confidentiality-is the foundation of the therapeutic alliance. Providing necessary protection “where self-control breaks down is not a breach of trust when it is not deceptive.”18

Therefore, circumstances permitting, the psychiatrist should inform the patient about the decision to breach confidentiality. Psychiatrists can take comfort in knowing that they have little basis “to fear being sued successfully for a bad outcome if the clinical practice has been reasonable” and when the psychiatrist’s actions were motivated by concerns over the welfare of the patient and threatened third parties.19 Potential liability is further reduced when psychiatrists contemporaneously document their risk assessment and clinical rationale.

Contemporaneous documentation provides the most believable evidence that the psychiatrist was diligent in gathering facts prior to exercising clinical judgment. Thus, psychiatrists should document what options were considered, and the clinical basis for rejecting or proceeding with a particular option. Consultation with a psychiatric colleague, as well as hospital legal counsel, should be considered in difficult cases. Dr Martin does precisely this, making liability very unlikely even if harm should occur to a third party.

Returning to the details of Dr Martin’s second case, a strong argument could be made that the duty to protect was eliminated or became nonexistent (at the time of discharge) due to good psychiatric inpatient treatment. The patient’s return of “insight,” clinical improvement, and willingness to adhere to treatment were all important risk-reducing factors. Although it was not specified, I am interpreting Dr Martin’s statement that “we had no reason to keep him any longer” to mean that the patient was no longer endorsing an explicit, realistic threat against the identifiable third party.

While such cases leave psychiatrists with many lingering concerns, I would frame the situation differently than saying we “turf” such cases or make “(creative) assumptions.” Rather, I would suggest that we appropriately and skillfully handle such cases in hopes that the inpatient team will be able to resolve the situation with good psychiatric care. In this manner, the “baton” of the duty is carefully handed forward until there is either a resolution of the danger, or the duty must be carried out in accordance with the relevant laws. Note that none of this implies that the mental health professionals working with the patient are not allowed to communicate the relevant risk issues with each other (with the patient’s consent).

Nor does any of the foregoing imply that the duty will never have to be revisited in the future, should the patient’s clinical condition worsen and the elements of duty to protect resurface. Finally, I would add that I am unaware of any law or policy in which a potential victim cannot be told the name of the patient who has threatened harm in cases where it is decided that confidentiality must be breached. Indeed, failing to do so seems likely to enhance the level of fear in the third party, as well as defeat the purpose of the warning since the third party would not know who they need to avoid or seek protection from.

Concluding remarks

Psychiatrists’ duty to protect in the context of a patient’s 1) realistic threats toward 2) identifiable third parties is a well-established exception to patient confidentiality. The duty has foundations in clinical ethics and was acknowledged even prior to the time that the Tarasoff case established a legal duty. Psychiatrists should be familiar with the duty to protect laws in their own state. When a potential duty to protect scenario arises, it should be first addressed as a clinical issue, and an array of options considered prior to breaching confidentiality. It is quite possible that clinical interventions may eliminate the need to violate confidentiality entirely. The fact that the duty is, perhaps temporarily, eliminated should not prevent psychiatrists from remaining alert to its reoccurrence, communicating relevant risks to other treatment providers and documenting a reasonable risk management plan. Even in the event of a tragic outcome and lawsuit, “judges and juries are likely to be more impressed by psychiatrists trying to do the most protective thing for patients as opposed to merely protecting themselves.”20


The author would like to acknowledge the late Douglas Mossman, MD, friend, colleague, brilliant scholar, mathematician, philosopher, and musician.

Dr Knoll is Director of Forensic Psychiatry and Professor of Psychiatry at SUNY Upstate Medical University in Syracuse, NY. He is Editor in Chief Emeritus of Psychiatric Times (2010 to 2014).


1. Tarasoff v. Regents of University of California, 529 P.2d 553 (Cal. 1974).

2. Tarasoff v. Regents of University of California, 551 P.2d 334 (Cal. 1976).

3. Slovenko R. Confidentiality and testimonial privilege. In: Rosner R (Ed). Principles and Practice of Forensic Psychiatry. New York, NY: Oxford University Press, Inc., 2003.

4. Buckner F, Firestone M. “Where the Public Peril Begins”: 25 Years After Tarasoff. J Legal Med. 2000; 21:2

5. Cal. Civ. Code § 43.92 (2013)

6. Weinstock R, Bonnici D, Seroussi A, Leong G: No Duty to Warn in California: Now Unambiguously Solely a Duty to Protect. J Am Acad Psychiatry Law. 2014;42:101-108.

7. Knoll J. The Psychiatrist’s Duty to Protect. CNS Spectrums. 2015;1-8.

8. Lipari v. Sears, 497 F.Supp. 185 (D. Neb. 1980)

9. Naidu v. Laird, 539 A.2d 1064 (Del. 1988)

10. Soulier M, Maislen A, Beck J. Status of the psychiatric duty to protect, circa 2006. J Am Acad Psych Law. 2010;38:457-473.

11. Mossman D. Critique of Pure Risk Assessment Or, Kant Meets Tarasoff. Univ Cincinnati Law Rev. 2006:24;523-609.

12. Phillip J. Resnick, MD, personal communication, September 3, 2019

13. Knoll J, Resnick P. Stalking Intervention. Curr Psychiatry. 2007;6:31-38.

14. MacKenzie R, McEwan T, Pathe M, et al. Stalking Risk Profile: Guidelines for the Assessment and Management of Stalkers. Victoria, Australia: StalkInc. Pty Ltd & Centre for Forensic Behavioural Science; 2009.

15. McEwan TE, Shea DE, Daffern M, et al. The Reliability and Predictive Validity of the Stalking Risk Profile. Assessment. 2018;25:259-276.

16. Mullen PE, Pathé M, Purcell R. Stalkers and Their Victims. Cambridge: Cambridge University Press; 2000.

17. Appelbaum P, Gutheil T. Clinical Handbook of Psychiatry and the Law. Philadelphia, PA: Lippincott Williams & Wilkins; 2007.

18. Slovenko R. Psychotherapy and Confidentiality, 24 Cleve St. L. Rev. 1975:375, 395.

19. Soulier M, Maislen A, Beck J. Status of the Psychiatric Duty to Protect, Circa 2006. J Am Acad Psychiatry Law. 2010; 38: 457-473.

20. Weinstock R, Bonnici D, Seroussi A, Leong GB. No duty to warn in California: now unambiguously solely a duty to protect. J Am Acad Psychiatry Law. 2014; 42:101-108.