Commentary: So the Pendulum Swings—Making Sense of the Duty to Protect ======================================================================== * Patrick K. Fox ## Abstract Psychiatry has been struggling for nearly 40 years to make sense of the duty to protect. The great jurisdictional disparity as to what constitutes the duty has been a significant contributing factor. The American Psychiatric Association (APA) released the Model Statute in 1987 to establish a framework to guide legislators and courts toward consensus, to some effect. In response to case law and statutory requirements in most states, psychiatric practice has incorporated the assessment of risk to third parties by patients as an essential element of psychiatric assessment and care. Although court cases shortly after the *Tarasoff* decision expanded the scope and breadth of the duty to protect, in recent years there appears to have been a shift toward a more narrow interpretation as to what conditions must exist to find a defendant psychiatrist guilty of failing to exercise the duty properly. The threshold for the duty to warn or protect often rests precariously beside the criteria permitting an exception to confidentiality, placing the psychiatrist in a tenuous position. If appellate verdicts continue to find for the defendant psychiatrist in cases claiming a breach of the duty to protect, it could have an impact on how psychiatrists assess and manage threats made by patients toward third parties. In the 35 years that have passed since the *Tarasoff II* decision by the California Supreme Court, much has been written about the implications of this decision for the practice of psychiatry and the physician-patient relationship. The *Tarasoff* duty is a confusing metonym, at times used to describe a duty to warn, while at other times referring to a duty to protect. This confusion has been present since the case's inception, as the original 1974 decision's lack of clarity with respect to the nature and breadth of the duty led to a subsequent decision by the same court two years later. While 37 states have a *Tarasoff*-like duty, the diversity of included criteria with respect to imminence, identifiability of victim, level of threat, and person to whom the threat is conveyed creates confusion as to what, exactly, one is referring. Comparing how verdicts are rendered across the nation with respect to a duty to protect is like comparing apples and oranges. I commend Soulier, Maislen, and Beck1 for their extensive and enlightening work in this important and daunting area of psychiatric research. In 1987, the American Psychiatric Association's Council on Psychiatry and Law developed a Model Statute intended to aid local branches of the APA in their efforts to clarify the scope and breadth of the psychiatrist's duty to protect.2 It was hoped that by adopting a uniform set of criteria for when a physician has a duty to prevent harm to others, a reasonably prudent physician standard would be promulgated. While some states did incorporate aspects of the Model Statute, the duty-to-protect landscape in this country remained a patchwork of statutory or case-based, permissive, or affirmative duties, addressing various degrees of foreseeability of harm to potential victims who may or may not be readily identifiable. The duty to protect has remained in a persistent state of flux for nearly four decades. While the pendulum may recently have swung in favor of physician-defendants, the lack of consistency across jurisdictions as to what the duty to protect entails and when it applies makes it difficult for clinicians to employ a methodological approach to specific clinical situations. Physicians must still consult with colleagues or legal counsel to determine whether the facts of the clinical situation with which they are confronted requires a warning or compels further measures to protect members of the public. This is particularly important in permissive, rather than instructive duty to warn or protect states and in those states in which the duty to protect is established through case law rather than statute. A novel and somewhat radical concept in the mid-1970s that created quite a stir within the psychiatric community, the duty to warn or protect has become a central aspect of patient care. In the years immediately following the *Tarasoff* decision, commentary and debate focused primarily on the chilling effect that the decision and its resultant duties would have on the psychiatrist-patient relationship and confidentiality in particular. Despite the concerns of leaders in the field, subsequent research demonstrated that the effects of a duty to warn on the psychotherapist-patient relationship were minimal and that patients accepted that confidentiality is not absolute but subject to notable, justifiable exceptions.3 Other physicians at the time were worried about unnecessary hospitalization of patients as a means of limiting physician liability, although initial studies did not find this to be the case.4 ## Ethics of the Duty to Protect In response to the expectation by society and the courts that psychiatrists have an obligation to assess and manage their patients’ risk to others, practitioners have come to view the duty to protect and the management of patient risk to others as they would other essential elements of care, such as the taking of a patient history and diagnostic formulation.5 I can scarcely conceive of a psychiatric interview in which the patient's risk to self or others is not addressed. For while an adverse clinical outcome in which a patient harms another may not constitute a breach of a *Tarasoff*-like duty, it may qualify for a claim of general negligence or failure to comply with an appropriate standard of psychiatric care. Aside from the imposed duty to protect through statutes and case law, one can argue that psychiatrists have an ethics-based obligation to protect third parties from patients who pose a risk. This duty can be viewed as an extension of the principle of nonmaleficence, in which the psychiatrist takes reasonable steps to protect the patient from the adverse consequences that can result from acting on his violent impulses. For example, if a patient presents to a psychiatrist's office appearing visibly distressed, stating that he is “going to lose it” and is uncertain as to how he may act, sound clinical judgment suggests that the psychiatrist consider further evaluation of the patient and the threat that he poses to society and himself generally, whether or not a legally required duty to protect has been satisfied. In both permissive and compulsory states, the ability to warn third parties of a dangerous patient is typically framed as an exception to confidentiality. The high threshold established by most states for when a duty to warn is triggered by requiring imminence, expressed threat, and identifiable victim serves to limit civil liability for the psychiatrist, but can have additional, disadvantageous clinical consequences. For instance, how should the psychiatrist proceed in clinical situations that satisfy some, but not all of the criteria necessary to trigger a duty to warn or protect? ### Case Example 1 A young man with delusional disorder, erotomanic type, was admitted to a psychiatric facility for restoration to competency to stand trial related to charges of stalking, harassment, violation of a restraining order, and criminal trespass. These charges stemmed from several instances in which he approached a minor female at her school, home, and place of employment in an effort to convey to her the depth and magnitude of their love for one another. A restraining order had been filed by the parents of the same female after the patient approached the family as they dined in a restaurant to ask for their daughter's hand in marriage. The patient had also sent the young female dozens of nonthreatening e-mails in which he expressed delight in their undying love for one another. The patient was found not competent and not restorable by the criminal court on the basis of his refusal to accept psychotropic medications and the court's opinion that the charges did not satisfy *Sell* criteria.6 The patient was subsequently civilly committed and later petitioned the probate court for discharge after a three-year period of hospitalization in which his condition did not improve and mail and phone restrictions had been in place to prevent unwanted communication with the victim or her family. After the presentation of evidence at the probate hearing, the judge decided that the patient no longer required a hospital level of care and he was discharged. Connecticut does not compel, but instead permits, disclosure of a threat of harm to third parties and has not established a statutory or case-based duty to protect. The relevant statutory exemption to confidentiality reads, “Communications or records may be disclosed when the psychiatrist determines that there is a substantial risk of imminent physical injury by the patient to himself or others….”7 The statute does not require that a specific threat be conveyed by the patient to the therapist. In the described case example, while it was likely that the patient would resume his efforts to contact and perhaps approach his female victim, he never made an attempt to or expressed a desire to harm her physically. Strict adherence to the statute would not permit disclosure based on a failure to meet the “substantial risk of imminent physical injury” standard, despite awareness that the patient's potential to approach the minor female and her family remained high. ### Case Example 2 A man was brought to the hospital for psychiatric treatment following a domestic disturbance at his home. The police reported that they had frequently responded to the home due to recurrent domestic disputes. On arrival at the emergency room, he was intoxicated and appeared manic. He stated that he would kill his wife and their two young children if his wife ever left him. When asked whether he had reason to suspect that his wife may have been planning to leave him, he stated that he did not, but added, “If I can't be with them, no one will.” He was held in the psychiatric emergency room for observation, and as his intoxication subsided, he appeared less angry and in better emotional control. On re-examination, the psychiatrist inquired about the statements that the patient had made earlier in which he expressed his intent to kill his family if his wife left him. He restated his intent to kill his family under those circumstances, but refused to allow the psychiatrist to phone his wife or other collateral sources, citing confidentiality and his right to privacy. He also demanded to be discharged, emphasizing that he was merely stating what actions he would take in the event that his wife filed for divorce, adding that he remained committed to their relationship and hoped that they could resolve their perpetual disagreements. This vignette illustrates a scenario in which the psychiatrist had concern for the safety of the patient's wife and children, although the concern was predicated on meeting a specific condition. The psychiatrist had no information to ascertain the probability that the patient's wife would leave him, other than his assertion that their relationship was salvageable. Permission to disclose information is not authorized for failure to meet the imminence criterion. While the psychiatrist could have phoned the patient's wife and inquired whether she intended to divorce her husband without revealing his identity and role or the nature of the phone call, this approach was impractical. ## A Precarious Balancing Act In this circumstance, the psychiatrist may not be liable for failing to warn or protect the wife in the event that she or her children are harmed, but the hold-harmless clause for making the warning may not apply, as the specific conditions necessary to justify disclosure may not have been met, exposing the psychiatrist to a breach-of-confidentiality complaint. By limiting the distinctions between the criteria necessary to convey and gather the information necessary to aid in clinical risk management with the threshold necessary to invoke a duty to warn or protect, the clinician is forced to walk a razor-thin line. Psychiatrists may resolve this tension by limiting their assessment of threat. If a state requires that a specific threat about an identifiable individual be conveyed, the psychiatrist may adopt a see-no-evil, hear-no-evil, speak-no-evil approach. While this approach may reduce the psychiatrist's anxiety about potential exposure to civil liability for breach of a duty to protect, it is not consistent with sound clinical judgment and seems antithetical to an appropriate standard of care. While the results of the research of Soulier *et al*.,1 who found that there has been a shift in verdicts in duty-to-protect cases in recent years in favor of defendants, is encouraging, it is only modestly so. As they have stated in their article, their research did not include state trial court cases that had not been appealed, or those that had settled before trial.1 This limitation in sampling may substantially underestimate the rate at which complaints are made against psychiatrists for a breach of the duty to protect, as in many cases such complaints are resolved without a trial, and fewer still are appealed after trial. In many instances, physicians are covered under the malpractice policy of the institution for which they are employed. In such cases, the hospital may pursue settlement, even though the merits of the case are questionable, in an effort to limit potential financial liability and reduce litigation costs. Furthermore, the researchers’ evidence suggests that malpractice insurers such as the one endorsed by the APA do not categorize claims based on a failure to exercise properly a duty to protect.1 It is unlikely that the APA-endorsed insurance company has never encountered a situation in which one of its insured has had an action brought against it for failure to exercise this duty. It is more likely that duty-to-protect cases are categorized under a broader liability category, such as general negligence. ## Defensive Medicine Although malpractice tort claim verdicts in favor of plaintiffs are relatively low, fear of civil litigation continues to affect the manner in which patient care is delivered. In 2008, the American Medical Association conducted the Physician Practice Information Survey. Relying on data compiled by the Physician Insurers Association of America, the AMA survey found that 65 percent of claims were dropped, dismissed, or withdrawn; 25.7 percent were settled; 4.5 percent were decided by alternative dispute mechanisms; and 5 percent were resolved by trial, with the defendant prevailing in 90 percent of the tried cases.8 Despite the low rate of successful plaintiff claims, physicians remain concerned about the arbitrary nature of malpractice lawsuits and verdicts, which is often discordant with the quality of care delivered. Harvard researchers have recently found that medical liability accounted for $55.6 billion (2.4% of total health care costs) in 2008, with $45.6 billion spent on defensive medicine. While psychiatrists scored low (51.4) on a malpractice concerns scale relative to other specialties in medicine, litigation concerns were still significant.9 No one enjoys being sued, and even suits that are dismissed or withdrawn can have a profound and lasting effect on the health care provider. Psychiatrists are likely to continue to practice in a manner that reduces their liability with respect to a breach of the duty to protect, whether or not the complaint is apt to prevail at trial. ## Conclusions While the legal concept of a psychiatrist's duty to protect third parties from dangerous patients has been significantly influenced by the California Supreme Court's decision in *Tarasoff II*, it has not been universally recognized or adopted. The decision's impact on the practice of psychiatry was immediate and profound, and the subsequent court decisions that further expanded the scope and breadth of the concept required further refinement of clinical practice to ensure that it satisfied the current legal requirements. Psychiatry has come to accept the premise that some form of duty to protect others from harm exists, whether statutorily defined on the basis of case law or a common law duty. Our profession generally recognizes that assessment of our patients’ risk to others constitutes an integral part of diagnostic interviewing and case formulation and may be further justified on the basis of an ethics-related obligation to our patients to do no harm. The recent data that demonstrate a shift by the appellate courts to defining more narrowly the situations in which a breach of the duty to protect may be found reduces the potential instances for which a psychiatrist may be liable. This shift in verdicts for duty-to-protect cases from plaintiffs to defendants may reflect the cultural changes brought about over the past two decades with respect to the patient's rights and autonomy, as well as the concerted efforts of those within the psychiatric community who have worked arduously to explain to lawmakers the complexity of risk assessment and our profession's limited capacity to predict human behavior accurately. It is as yet unknown whether psychiatrists will alter the manner in which they assess their patients’ threat to others in light of the waning risk of tort liability. The extent to which the assessment of a patient's risk to others has been inculcated into psychiatric practice suggests that psychiatrists will continue to provide for the safety of their patients and others whom they reasonably conclude may be in peril. ## Footnotes * The author acknowledges fiscal support from Connecticut's Department of Mental Health and Addiction Services. * Disclosures of financial or other potential conflicts of interest: None. * American Academy of Psychiatry and the Law ## References 1. Soulier M, Maislen A, Beck J: Status of the psychiatric duty to protect, circa 2006. J Am Acad Psychiatry Law 38:457–73, 2010 [Abstract/FREE Full Text](http://jaapl.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NToiamFhcGwiO3M6NToicmVzaWQiO3M6ODoiMzgvNC80NTciO3M6NDoiYXRvbSI7czoyMDoiL2phYXBsLzM4LzQvNDc0LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 2. Buckner F, Firestone M: Where the public peril begins: 25 years after Tarasoff. J Legal Med 21:187–222, 2000 [CrossRef](http://jaapl.org/lookup/external-ref?access_num=10.1080/01947640050074698&link_type=DOI) [Web of Science](http://jaapl.org/lookup/external-ref?access_num=000087917300002&link_type=ISI) 3. Beck J: When the patient threatens violence: an empirical study of clinical practices after Tarasoff. Bull Am Acad Psychiatry Law 10:188–9, 1982 4. Appelbaum P: The new preventive detention: psychiatry's problematic responsibility for the control of violence. Am J Psychiatry 145:779–85, 1988 [PubMed](http://jaapl.org/lookup/external-ref?access_num=3381921&link_type=MED&atom=%2Fjaapl%2F38%2F4%2F474.atom) [Web of Science](http://jaapl.org/lookup/external-ref?access_num=A1988P544800001&link_type=ISI) 5. Wise TP: Where the public peril begins: a survey of psychotherapists to determine the effects of Tarasoff. Stan L Rev 31:165–90, 1978 6. Sell v. United States, 539 U.S. 166 (2003) 7. Conn. Gen. Stat. § 52-146f (1969). 8. Kane CK: Medical Liability Claim Frequency: A 2007–2008 Snapshot of Physicians. Policy Research Perspectives. Chicago: AMA, 2010 9. Carrier E, Reschovsky J, Mello M: Physicians’ fears of malpractice lawsuits are not assuaged by tort reforms. Health Affairs 29:1585–92, 2010 [PubMed](http://jaapl.org/lookup/external-ref?access_num=20820012&link_type=MED&atom=%2Fjaapl%2F38%2F4%2F474.atom)