The Tarasoff Rule: The Implications of Interstate Variation and Gaps in Professional Training

Recent cases of violence committed by persons who have received psychiatric treatment have revived important questions about therapist–client confidentiality and the circumstances in which that confidentiality can or must be broken (1). More specifically, what is the status of training and clinicians’ knowledge about their responsibilities when a patient communicates a threat of imminent violence? These questions stem from the Tarasoff v. Regents of the University of California rulings (2, 3) issued in the 1970s. Prosenjit Poddar, a graduate student at the University of California (UC) Berkeley, told his therapist that he wanted to obtain a gun and kill the object of his infatuation, Tatiana Tarasoff. The therapist informed the campus police about the threat but neither the police nor the therapist warned Tarasoff directly. Poddar then murdered Tarasoff when she returned to campus from summer vacation, an event that occurred two months after Poddar broke off contact with his therapist.

Tarasoff’s parents sued the campus police and the UC Regents for failing to warn their daughter, and in Tarasoff I (2), the California Supreme Court stated that therapists have a duty to warn others who are in foreseeable danger from the therapists’ patients. The duty to warn enunciated in the first ruling was expanded in an opinion issued after a rehearing of the case, Tarasoff II (3), wherein the court stated that the therapist has a duty to “use reasonable care to protect the intended victim against such danger (4).” In the hands of the California court, the duty to warn a potential victim evolved into a duty to protect, a more expansive obligation that may include warning the potential victim, telling the police, and taking “whatever steps are reasonably necessary under the circumstances.” In the three decades since the Tarasoff rulings, the duties to warn and protect have been applied, refined, and revised in many states’ case law and statutes, including those of the state of California itself.

In this review, we raise and examine three sets of questions about Tarasoff duties, all aimed at stimulating more empirical, legal, and conceptual scholarship on these neglected topics. First, what key ethics related and legal questions does substantial interstate variation in duties to warn or protect raise? What ethics principles should guide therapists in states with legal leeway in how to discharge the duty as they handle potential threats? Second, to what extent do mental health professionals understand the details of their given state’s codification (or lack thereof) of Tarasoff-type duties, and to what extent are other relevant professionals (law enforcement officers, university officials, and lawyers involved in risk mitigation) aware of how to assist therapists in discharging their duty? Finally, how do courts typically rule in cases surrounding the duty to warn or protect, and what reasoning do courts use? We conclude with a brief discussion of gaps in existing conceptual and empirical work in each of these three areas.

The Policy and Legal Landscape

There is no blanket federal duty to warn or protect; instead, there is substantial state-by-state variation in whether and how the duties are defined and codified. Each state falls into one of four general categories, elucidated in a 2012 review by Griffin Edwards of state statutes and case law (Table 1):

• 1) A duty to warn or protect is mandated and codified in a legislative statute (23 states);

• 2) a duty to warn or protect is not codified in a statute but is present in the common law supported by precedent (10 states);

• 3) a duty to warn or protect is neither codified in statute nor present in state case law, but states permit a breach of confidentiality in the therapeutic relationship if a threat is present (11 states);

• 4) there are neither statutes nor case law offering guidance on the issue. (6 states) [Ref. 5].

Other surveys of state statutes have given slightly different breakdowns of states that fall into each category owing to different methodologies that have been used for classifying states. For instance, in a 2002 review, Herbert et al. (6) classified 27 states as having a mandatory duty to warn in statute or common law (compared with Edwards’ 33 states), 9 states as having a permissive duty (as opposed to Edwards’ 11 states), and 13 states as having no statutes or case law offering guidance (as opposed to Edwards’ 6 states). These discrepancies stem from different standards for when a state has proposed a duty through common law: Edwards’ survey suggested that if a court states that it is prepared to rule in favor of Tarasoff, should the appropriate fact pattern be present, a state (e.g., Alabama and Georgia) has established a common law duty, whereas Herbert et al. classified these states as offering no legal guidance. This discrepancy in survey findings highlights the complexity created by interstate variation, a complexity that has been reviewed extensively else where (5, 7, 8). We therefore ask, what are the legal and ethics-related implications of such significant between-state variation in the duty to protect or warn?

In addressing this question, we can distinguish between three general categories of states: those that mandate some duty to warn or protect (and that often specify whether law enforcement, the victim, or a combination should be “warned”); those that allow therapists to warn by protecting them from liability for breach of confidentiality if they do so, but do not require them to issue a warning; and those that offer no statutory or case law guidance. We highlight the ethics-based and legal implications of this variation for health professionals.

Discharging the duty in states with mandatory statute.

Clinicians in states with statutes or case law that mandate a duty to warn or protect face at least one source of ambiguity: the risk assessment required to determine whether a patient’s threat level is high enough to require the therapist to discharge the duty. In assessing this risk level, clinicians should be aware of state requirements that affect how this threshold is set: whether the victim must be identifiable (most states); whether the threat must be imminent (most states); or whether the threat need not be imminent but must be serious (few states). In addition, even for states that explicitly state that the violence must be “imminent” to impose the duty to warn upon the therapist, clinical commentators often specify different definitions of how the law ought to interpret imminent, ranging from a few days to a few weeks to several months, with state case law also using different notions of what counts as imminent (9). Furthermore, the definition of imminence that is relevant for duty-to-warn cases, where a therapist breaches confidentiality as a result of a threat of imminent violence, may and often does differ from the definition of imminence that is relevant in cases such as civil commitment, where a therapist deprives a patient of liberty as a result of imminent danger to the patient or others (9). This source of confusion has led commentators to recommend that clinicians focus less on the imminence of the threat in Tarasoff cases and more on the patient’s demonstrated capacity to carry out the threat (i.e., whether the patient has a history of violence, whether the patient has experienced situational triggers that have exacerbated violence in the past, and what can be done to intervene) (9, 10).

Once a clinician has decided that a patient poses a threat of violence sufficient to trigger a duty, the limited empirical data that are available on how clinicians actually discharge this duty suggest that many may take liberties with their states’ specific legal guidance. For example, although California state law mandates that clinicians warn potential victims and a law enforcement agency of serious threats of physical violence (11, 12), data from San Francisco suggest that many therapists use involuntary civil commitment for patients who pose a threat (13). This approach may be legally valid if the patient meets the criteria for civil commitment and the therapist determines that the patient no longer poses a threat of imminent violence after the commitment period. Perhaps these therapists pursue commitment because they see it as a more effective means of reducing a patient’s risk than warning a potential victim. Further research should disentangle therapists’ reasons for pursuing a given course of action in states with mandatory statutes to ensure that therapists do not use warnings and civil commitment as substitutes where inappropriate.

Discharging the duty in states with common law suggesting mandatory duty.

In states where a mandatory duty to warn or protect has arisen exclusively from the judiciary, clinicians’ understanding of their specific state’s regulations may be compromised. Arguably, case law is more dynamic than legislative statutes elucidating duties, some of which have remained unchanged over the past 20 years, and the changes in case law may be difficult to interpret, not well publicized, and generally nontransparent. Thus, although clinicians may have heard of Tarasoff and think that they understand its broad implications, case law in their state may offer guidance that differs significantly from the broad implications of the initial Tarasoff case. For instance, after Virginia rejected Tarasoff duties in Nasser v. Parker and before it passed a mandatory Tarasoff statute in 2010, it nonetheless found that allegations that providers had taken responsibility for a psychiatric patient raised questions of fact for a trial, opening up providers to the expense and time of litigation (14). If the case law in a given state is not clear and well publicized, therapists in these states face an additional source of ambiguity in deciding how to handle a violent threat.

Discharging the duty in states with permissive statutes.

Therapists in the second category of states, those with permissive statutes that protect therapists from liability for breach of confidentiality in the case of threats but do not obligate them by statute to warn or protect the potential victim, arguably face more difficulty in determining how to protect potential victims than do therapists in states with mandatory statutes and case law. Therapists in states with permissive statutes have several legally acceptable options when a patient makes a violent threat against an identifiable victim: continue therapy as planned without issuing any warning, change the therapy to contain the threat and protect the potential victim, warn the victim, warn law enforcement, warn the victim and law enforcement, or determine if there are grounds for civil commitment based on the patient’s dangerousness to others.

The first two strategies, where the therapist does not issue a warning, may be more legally risky than the other strategies, since permissive statutes protect a therapist only when he breaches confidentiality and do not necessarily protect a therapist in the following situation: when he does not breach confidentiality, when there are strong reasons for such a breach, and when a victim is hurt. For instance, in Almonte v. New York Medical College (15), a federal district court ruled that a physician training a psychiatric resident studying child psychiatry could have a common law duty to future patients of the resident to warn that the resident had pedophilia, even though Connecticut (whose state law applied) is a permissive-statute state, and so the psychiatrist was not violating a statutory duty.

Although immediately notifying law enforcement and potential victims of a threat may be the safest course of action legally, there are moral reasons supporting other strategies in permissive states. Most simply, the duty to protect is grounded in the moral right of the potential, identifiable victim to avoid preventable death or substantial, debilitating physical injury. The potential victim has a moral claim against the clinician who, by virtue of his relationship with the patient and his knowledge of the patient’s violent intentions, is in a position to try to prevent this serious harm from occurring. Therefore, the moral duty to protect involves a goal for the clinician, to protect the victim, while minimizing the extent to which various interests of the patient are set back in pursuit of this goal, including liberty interests and some degree of privacy in clinical communication. This moral basis serves as the foundation for legal duties to protect, which specify cases where legal requirements of clinician–patient confidentiality are removed to allow a clinician to try to avert harm to an identifiable victim.

We argue that a clinician can best show respect for patient rights while protecting potential victims by first alerting a patient that he is considering a notification and discussing whether the statements or gestures indicating violent intent were made sincerely. For instance, a patient may express a violent fantasy in a moment of anger, and the therapist might explore the extent to which the patient plans to act on the fantasy and whether he has the means to do so. The law bolsters this approach to assessing the sincerity of violent fantasies by generally holding that a patient’s violent fantasies do not necessarily impose a duty to warn upon the therapist, unless the patient has a history of violence or other risk factors (10). The therapist should then indicate that if the threat is sincere, he will notify the authorities and the potential victim. These steps should be combined with the continuation of intensive counseling.

Data support this strategy of working collaboratively with a patient: therapists who talk with patients first about the need to report a threat, as opposed to unilaterally notifying law enforcement and a potential victim with no disclosure to the patient, report that the warning either had a positive or no impact on the therapeutic relationship. Warnings that were not discussed beforehand had negative ramifications for this relationship (16). Therefore, respecting patients through disclosure about possible warnings and discussion about the threat has both intrinsic value in supporting patient interests and instrumental value in minimizing damage to the therapeutic relationship. In turn, an intact therapeutic relationship may better alleviate a patient’s future violent intentions than a damaged relationship or one where the patient breaks off all contact. Of course, there may be contexts in which it is difficult or impossible to discuss the warning with the patient (for example, if the therapist decides that the threat warrants a warning after the end of a therapeutic session and the patient has broken off all contact) but in general, therapists should attempt discussion to show respect to the patient and his trust in the therapist.

Discharging the duty in states with no statutory or common law guidance.

Therapists in the third category of states (those without any statutory or common law guidance) face a more fraught set of options than therapists in states with either a mandatory or permissive duty, since therapists in this third category have neither an established affirmative duty to warn or protect nor a legally established justification to breach confidentiality in exercising such a warning. As a result, therapists in these states are open to significant legal risk no matter how they proceed: if they breach confidentiality, a patient could sue, and if they fail to breach confidentiality and a victim is harmed, the victim could sue. That legal risk could be reduced by the passage of legislation, exemplified by the American Psychiatric Association’s (APA) Model Statute or another state’s statute that clarifies therapists’ duties regarding potentially violent patients.

Variations within a state in interpreting the duty to warn or protect.

We have noted that the passage of a statute clarifying a therapist’s duty to warn or protect reduces the legal risk a therapist faces, but the existence of a statute by no means eliminates this risk, because of intrastate conflicts and inconsistencies between statutes and judicial rulings, which highlight that a therapist’s risk is by no means eliminated by the existence of a statute. A few examples: the Arizona Supreme Court found that Arizona’s Tarasoff statute did not shield therapists from exposure to civil damages because such a limitation would violate the Arizona Constitution (17). The Utah Supreme Court extended therapists’ duties beyond what Utah’s original Tarasoff statute prescribed, holding them liable in cases where they should know of a threat as well as in cases where they have actual knowledge (18). The Supreme Court of New Hampshire held that New Hampshire’s Tarasoff statute does not pre-empt common law duties to warn, and so medical professionals can be subjected to common-law tort claims, even if they do what the Tarasoff statute mandates (19). Courts in several states have issued verdicts seemingly in tension with Tarasoff statutes without citing or distinguishing the statutes (20, 21).

Variations in the purpose of invoking the duty to warn or protect.

In addition to interstate and intrastate variation in the duty to warn or protect, therapists also face variability in the purpose for which a state’s case law and/or statute will be invoked in legal cases. A victim’s family may prosecute a therapist for failing to discharge the duty appropriately, or a therapist may use duty to warn or protect statutes as a legal shield if prosecuted for a breach of confidentiality by a client whom the therapist reported. In other words, the statutes can be a liability if a therapist inappropriately does nothing and a client commits a violent act, or the statutes can be a form of protection if a therapist appropriately does something and a client sues for breach of confidentiality. The diversity of uses of these statutes highlights why therapists and other covered mental health professionals should have a nuanced understanding of their state’s regulation or case law. The next section reviews whether therapists and allied health and legal professionals achieve this level of understanding.

Who Knows What?

Research conducted in the first 20 years after the Tarasoff ruling found that although many therapists were aware of the case, there was substantial misinformation regarding the extent of a given state’s law and whether it required therapists merely to warn authorities about a potentially harmful patient or instead to take steps to protect the potential victim (22). However, the past decade’s increase in mandated ethics and forensic education as part of graduate training or continuing education requirements may make these early studies of therapist misinformation outdated and inaccurate.

Training for psychologists and psychiatrists.

tarasoff-related training is given to clinical psychologists and psychiatrists in both formal and informal settings, with formal training related to the duties likely to fall under the heading of ethics or forensic education. For psychologists, the American Psychological Association requires training in ethics as a part of accredited PhD programs, and their most recent survey of continuing education standards found that 26 states require between 1.5 and 6 hours of training in ethics annually for licensure maintenance (23, 24).

While future psychiatrists are in medical school, they receive substantial training in ethics. A 2008 survey found that the average medical school required 35.6 hours of bioethics instruction, yet this training may be nonspecific to the situations that mental health professionals face (25). Once they choose the specialty of psychiatry, psychiatrists receive training in Tarasoff-related duties from two sources during residency: formal training and response to informal, on-the-job events. For formal training, one survey of training content in psychiatric residency found that more than 88 percent of programs taught about the duty to warn, with more than 10 percent of programs not responding and only 0.7 percent saying that they did not teach about the duty (26). This formal training is supplemented by informal, on-the-job situations that prompt the resident to discharge a duty to warn or protect. One survey of second-through fourth-year residents in psychiatry found that 22 of the 46 residents surveyed had issued a Tarasoff warning at some point during their training (27). Therefore, for both clinical psychologists and psychiatrists, topics related to the duty to warn or protect appear to be well covered through both formal education and hands-on experiences. However, questions remain about the prevalence and depth of education about a crucial step in the process of potentially discharging a duty to warn or protect: risk assessment to determine the threat level of the patient. A 1997 survey of psychiatric residents found that a third received no training in assessing and managing a patient’s risk of violence and another third described their training as inadequate (28); a 1990 survey of psychologists reported a mean of 3.3 hours and a median of 0 hours of formal training in risk assessment (29). Although training in risk assessment may have increased in quantity since the time of these surveys and although some argue that actuarial tools for violence risk assessment such as the Violence Risk Appraisal Guide (VRAG) are more accurate predictors of violence than unstructured clinical judgment (30), clinicians still need training on how to translate scores on the measures into complex decisions in the clinical setting (31). The education of mental health professionals in Tarasoff-related duties not only should outline the scope of these duties but should also teach practical risk assessment and management techniques that clinicians need for appropriate assessment and handling of a threat. For instance, therapists should learn to make a distinction between a patient who makes a threat by communicating an intent to harm versus one who poses a threat (i.e., has planning and building capacity for a violent act) (32). Training should emphasize that, although the law focuses on therapists issuing warnings after violent communication, a therapist’s ethics-related obligations should extend to patients who pose a legitimate threat to identifiable others.

Psychologists’ knowledge of the duty.

Unfortunately, simply quantifying the amount of formal and hands-on training related to Tarasoff may offer too rosy a picture of the extent to which psychiatrists and psychologists understand the duties mandated or permitted by their specific states. In one study, Pabian et al. (33), found that 63 percent of the psychologists surveyed had completed a graduate course in ethics, and 87 percent of those had received instructions on the Tarasoff ruling and their responsibilities for dangerous patients. Most clinicians appeared confident that this training adequately informed them about their duties, with only 10 percent expressing uncertainty about their legal duties surrounding potentially violent patients.

However, despite such confidence, 76 percent of the psychologists surveyed were incorrect in selecting the statement that best described their given state’s duty-to-protect law. The errors stemmed from a misunderstanding of the circumstances that trigger a duty to protect, with many therapists incorrectly thinking that the duty extends beyond the bounds of imminent danger in their state and a misunderstanding of the level of threat that triggers a duty to protect, with 41 percent of therapists incorrectly believing that they should warn the victim and law enforcement when the likelihood that patient will follow through on the threat appeared low or uncertain (33). Perhaps owing to this inflated understanding of the circumstances and threat level that trigger a duty to protect, the therapists reported experiencing a scenario requiring duty to warn or protect fairly frequently, approximately once every two years. The results of this survey suggest that although psychologists are aware of Tarasoff and receive graduate ethics training, many are confused or misinformed about the specifics of their states’ statutes and common law on the duty to warn or protect. The gap in knowledge that is understandable, given the complexity of the regulations, but nevertheless is a gap that training and continuing education programs should work to remedy.

Other health professionals’ knowledge of the duty.

The analysis by Pabian et al. (33) was limited to psychologists, most of whom practiced in outpatient settings. It does not include psychiatrists. Furthermore, at least 20 states seem to extend the duty to physicians who are not necessarily credentialed in psychiatry as long as the physician purports to offer mental health treatment (Table 2). We are not aware of data that assess psychiatrists’ knowledge of their state’s duty to warn or protect statutes and case law or the knowledge of nonpsychiatrist physicians, social workers, nurse practitioners, and other health professionals.

Further research on allied health professionals’ understanding of Tarasoff is needed, as behavioral health care increasingly relies on nonpsychiatrist and nonpsychologist providers. Health care reform should catalyze the move toward integrated care in which primary care physicians, nurse practitioners, health counselors, and other nonspecialty providers may serve as a frequent point of contact for persons with psychiatric disorders (34). Although some provider types (e.g., psychologists and psychiatrists) are covered by Tarasoff-related duties in most states (Table 2), other provider types (e.g., nonpsychiatrist physicians) are covered in only a subset of states. Health professionals not explicitly covered by their states’ duty to warn or protect statutes lack clear statutory protection from liability if they breach confidentiality to report a Tarasoff-type threat. They are in a risky situation resembling that of health professionals in states that altogether lack duty to warn or protect statutes.

Other parties’ knowledge of the duty.

Research on other potential parties in the equation (i.e., law enforcement professionals or institutional legal and risk management offices who may be informed of a patient’s potential dangerousness) is also scarce. One 1998–1999 survey of Michigan and North Carolina police officers found that only 3 percent were familiar with the Tarasoff ruling and only 24 percent reported that their stations had specific policies on when and how to warn potential victims of a violent patient, despite the fact that 45 percent of stations had received at least one Tarasoff-related warning from a mental health professional (35). No empirical scholarship of which we are aware examines how the legal and risk management offices of institutions such as universities or integrated health care systems advise therapists on how to handle a Tarasoff-related warning.

The Narrowing of the Duties

Whereas much of the early commentary surrounding Tarasoff consisted of dire proclamations about the damaging effect the rulings would have on psychotherapy, with commentators arguing that the therapeutic relationship would be irremediably compromised by the ensuing regulations, recent court cases illustrate that therapists are very rarely held liable. Soulier et al. (36), in an analysis of 70 appellate cases from 1985 to 2006, found that 46 were decided in favor of the mental health professional, 6 were decided in favor of the plaintiff (although only 4 of these used Tarasoff statutes), and 17 were returned to trial courts for further litigation. Mental health professionals were exonerated on the following bases: no imminent threat was communicated to a therapist about an identifiable victim; the victim was already aware of the danger; or the therapist warned the victim, but the victim took actions that went against the warning.

Courts appeared to rule in favor of the victims only in marked cases of negligence by the mental health professional or institution: in Almonte (15), a psychiatrist/patient who was being seen by another psychiatrist admitted his sexual attraction to children, but was recommended for a child psychiatry fellowship where he raped a child. In other cases, inpatient psychiatric treatment was terminated against medical advice or because the patient’s insurance coverage ran out (36).

Ultimately, we conclude that the courts converge on three themes of the duty to warn or protect: hold therapists liable only in obvious cases of negligence that result in harm to a victim; recognize when a state has a permissive statute, rather than an obligatory one; and do not hold therapists liable for violence that occurs well after the termination of therapy (37). We note that these narrow interpretations of therapist liability stand in contrast to earlier rulings that ignored the language of the statutes and interpreted broad liability, such as cases in which therapists were held liable for motor vehicle accidents that occurred months after termination of therapy.

The data sample of Soulier et al. (36) was limited, in that it included only appellate cases and therefore did not include verdicts in state trial courts that are not appealed or cases settled before trial. To try to examine verdicts from state trial courts, they reviewed records from an APA-sponsored insurance company that provides coverage to psychiatrists as defendants in various legal cases, a review that should have identified trial verdicts that are not appealed. Their review turned up little evidence of money spent on helping defend psychiatrists in Tarasoff-type cases. Cases settled before coming to trial were not captured by either method, and perhaps therapists accused of improper discharging of Tarasoff duties are more likely to settle with the plaintiff rather than go to court. Yet if one accepts the appellate decisions as fairly representative of how courts now rule in Tarasoff-type cases, it seems that in contrast to early cases, courts now resist the notion that a therapist has a duty to protect the general public and even appear increasingly likely to reject the notion that an outpatient therapist’s relationship with the client grants the therapist sufficient control to warn or protect potential victims (36, 38). Therapists may still encounter the time and distress of litigation, but it appears that therapists who choose to defend themselves in court rather than settle the case are increasingly less likely to be held liable for a patient’s violence except in cases of marked negligence.

Conclusion

Interstate variation in the duty to warn or protect raises normative questions about how this variability may impede mental health professionals’ knowledge of their duties. Inadequate knowledge not only exposes therapists to legal risks, but also may impede a therapist’s ability to fulfill an identifiable victim’s moral claim to be warned about or protected from substantial harm. When legal scholars have difficulty parsing the reasoning behind various Tarasoff-related rulings, it seems unreasonable to expect mental health care professionals and law enforcement officers to discharge these duties correctly without increased guidance and support.

Our review also highlights the need for more empirical, legal, and conceptual scholarship in multiple areas related to various parties’ understanding of the duty to warn or protect. First, researchers should move beyond documenting gaps in knowledge about state-specific Tarasoff regulations to investigate the forms of education most helpful for remedying these gaps. Second, research should examine the motivations of therapists who practice in states that mandate notifying law enforcement but choose to take other measures to reduce the threat of the patient. Are these therapists unaware of their state’s specific legal suggestions for discharging the duty, or are they aware but deliberately choosing to address the threat through means other than those suggested in the states, and why? Third, existing empirical research has focused largely on therapists in states with mandatory statutes. Future research should examine how therapists in states with permissive statutes weigh their various legally acceptable options.

Fourth, researchers should attend to health care’s increasing emphasis on nonspecialist mental health care by investigating nonpsychiatrist/nonpsychologist health providers’ understanding of the duties imposed by certain states. Fifth, research should focus on the knowledge and training of other parties who are often practically involved in Tarasoff–related situations: the police, institutional legal and risk management offices, university professors, and others. Now that most commentators recognize Tarasoff duties as acceptable aspects of therapist–client relationships, these areas of future scholarship can examine the nuances of these duties, how the array of professionals responsible for performing these duties are trained in essential skills, and how they discharge the duties in institutional and integrated care contexts.