Abstract
States have responded to the Tarasoff duty to protect by passing statutes in all but 13 states. Such statutes either mandate or permit warning a potential victim. In this study, we analyzed 70 Tarasoff-related cases from a Westlaw-based search between 1985 and 2006. We determined the extent to which clinicians are being held liable for breach of the Tarasoff duty in statutory and nonstatutory states, whether there is language in the statutes that permits warning compared with statutes that mandate warning, and whether recent Tarasoff decisions better reflect the inherent ambiguities in clinical mental health practice. We found 70 appellate cases, and only 6 were plaintiff verdicts. Statutes that mandate warning a victim appear to be the most protective of clinicians. Seven of the 17 remanded cases came from the jurisdictions with permission-to-warn statutes, suggesting that permission rather than a strict mandate to warn may increase the liability for clinicians. Notwithstanding the language of statutes, the protections from Tarasoff are not extended to poor clinical judgment, particularly in the controlled inpatient setting.
The first Tarasoff v. Regents of the University of California1 opinion in 1974 enunciated a duty for psychotherapists to warn others who are foreseeably endangered by their patients. In 1976, the court2 expanded this duty saying, “When a therapist determines … that his patient presents a serious danger of violence to another, he incurs an obligation to use reasonable care to protect the intended victim against such danger” (Ref. 2, p 346). The clinician could warn the victim, hospitalize the patient, or notify authorities, depending on the clinical facts. The legal duty and its implications for clinical practice continue to be a source of concern, even anxiety, for psychiatrists and other mental health professionals. This concern in turn has led to a line of commentary and analysis now more than 30 years old, distinguished by a wide range of opinions and by the intensity with which the opinions are presented.
Clinicians’ fears of being found negligent were fueled by several early Tarasoff decisions that appeared to many psychiatrists to make little or no clinical sense. In Jablonski v. United States,3 a Veterans Administration psychiatrist was found to be liable for the death of a woman who was well aware of the risk she ran by her behavior. In Petersen v. Washington,4 considered to be the first driving case to find Tarasoff liability, a state hospital psychiatrist was found liable when, five days after discharge, a schizophrenic patient caused an auto accident in which the victim suffered a knee injury. The court appeared to give heavy consideration to the fact that the psychiatrist had failed to extend the confinement after the patient had driven recklessly on the hospital grounds, one day before the expiration of the commitment. In Lipari v. Sears Roebuck & Co.,5 the case was settled out of court after the Federal District Court in Nebraska assigned liability following injuries to random victims when a patient fired a shotgun in a crowded nightclub. In Davis v. Lhim,6 a patient who had never been violent or threatened his aunt attacked her. Again, the state hospital psychiatrist was found liable. Finally, in Naidu v. Laird,7 a Delaware state hospital psychiatrist was found liable when his patient caused an auto accident that killed a man, but the accident occurred five and one-half months after the patient was discharged and 500 miles away from the hospital where he was treated. After issuing a Tarasoff warning, one psychiatrist was found liable for breach of confidentiality.8
In response to these and other decisions, psychiatrists and other mental health professionals lobbied for statutes that provided guidelines for the exercise of the duty to protect. Thirty-seven states now have statutes. The laws are all fundamentally similar, but they may differ in specific language. Some are triggered by imminent danger, others by serious danger. Some are triggered by threats to identifiable victims; others are more open-ended—an identifiable victim triggers the duty—but a specific threat is not required. The statutes also differ on what they entail for the clinician. Some are permissive, stating that the clinician may warn; others are prescriptive, stating that the clinician must warn.9 They all include a hold-harmless provision for protection against liability for breach of confidentiality for any good-faith warning or other protective action. These statutes reflect not only the lobbying effort of mental health professionals, but also a heightened appreciation from the public, judges, and legislatures that clinicians face an exceptional dilemma when confronted with patients who make threats toward others.
Commentators have addressed the value of the statutes as social policy as well as for clinical practice, with various conclusions (e.g., Felthous,10 and Beck11). Although the commentators expressed several concerns, no recent study has examined and reported on the extent to which courts are actually finding psychiatrists liable for breach of the duty to protect. A critical question for the profession is whether there are differences in the outcome of cases brought in states with statutes (statutory states) compared with cases brought in states relying entirely on case law.
In this article, we review and report on outcomes of duty-to-protect cases brought in statutory states and in states relying on case law. We first review recent commentary, to illustrate the wide range of concerns and opinions currently held. We then review case law comparing outcomes of cases in statutory and nonstatutory states.
The important practical questions we examined are first, to what extent are courts finding defendant psychiatrists and others liable for breach of the duty to protect, and second, whether there is any difference in outcomes of cases in states with and without statutes.
Recent Commentary
Several studies have been conducted regarding the meaning for clinicians of the change in language from Tarasoff I to Tarasoff II (Herbert12 and Felthous10). Herbert's close analysis of the Tarasoff cases concludes that the 1976 Tarasoff II opinion extinguishes the 1974 Tarasoff I. Herbert was quite precise in his analysis of the change in the black letter of the law, but was dubious about whether the change meant much, if anything, for clinical practice, noting that the change from “duty to warn” to “duty to warn or other-wise protect” “is a distinction with little practical difference” (Ref. 12, p 418).
Felthous10 appeared to agree with Herbert. He noted that although Tarasoff II states a duty to protect, “the real change brought about by the Tarasoff decision was that one of the protective measures … was warning the intended victim” (Ref. 10, p 338). Weinstock et al.13 understood the matter differently. They considered that the Tarasoff court intended a duty to protect, which was distinct from a duty to warn, but which was “often misinterpreted to be a duty to warn” (Ref. 13, p 523).
Commentators have also addressed inconsistencies in how courts have analyzed and decided these cases and the variation from state to state in how the statutes codify the duty to protect. Kachigian and Felthous14 analyzed the case law in 23 statutory states. They found a wide variety of ways in which post-statute case law related to the statutes. Some courts failed to reference the statute, others briefly referenced it, and others that referenced it varied in holding whether the statute did or did not create a duty on the facts at issue. They note that therapists value clarification and consistency from the courts in how to understand the duty and that the effect of statutes on later cases was not consistent at all. They concluded that the statutes had been of doubtful value in providing the hoped-for clarity in defining the duty to protect.
Other commentators have reached different conclusions. Mossman15 raised several concerns. At the most fundamental level, he was critical that in articulating the duty, the Tarasoff court relied on a probabilistic theory of violence prediction that was fundamentally unsound. Like other commentators, he was also concerned that the Tarasoff decision had been unclear on when the duty was triggered and on how clinicians should meet it. In contrast to Kachigian and Felthous,14 he ultimately concluded, using the Ohio statute as an exemplar, that statutes had provided an adequate framework within which clinicians could exercise their clinical judgment. Geske,16 using the Indiana statute as his specific example, also concluded that the statutes are adequate to the purpose for which they are intended.
For Felthous,10 specific differences in language from case to case or statute to statute become an important source of uncertainty in understanding how therapists are expected to exercise clinical judgment in different jurisdictions. In his discussion, he illustrated how these differences in language can lead to contrasting assessments of the extent to which the clinician is required to exercise clinical judgment as distinct from being guided by the black letter of the statute. He also articulated a potential new basis for warning victims. He suggested that there may be an unexpressed (presumably clinical) ethic that views a patient's threat as “an essentially aggressive act per se” (Ref. 10, p 344). The victim is already victimized by the threat, whether it is made directly to the victim or not. “From this perspective the victim has a right to be warned of the threat regardless of how great or little the risk of violence in fact is” (Ref. 10, p 344). We disagree, and we address this point in our discussion.
Other authors have examined variation in the facts on which courts have found that a duty exists, noting differences that are important to clinicians as to what facts trigger a duty (Felthous10 and Walcott et al.17). Some courts have found a duty only after the patient has made an actual threat. Other courts have held that duty arises solely from a history of the patient's violent behavior, in the absence of any direct threat to the victim, whether made to the therapist or to anyone else.18 Some courts have found that the outpatient therapist has sufficient control of the patient to trigger a duty to protect; other courts have held that the degree of control is not sufficient to trigger a duty.
A recent case involving a threat conveyed not by the patient but by a family member, led to two suits that have been the subject of considerable discussion. In Ewing v. Goldstein19 the patient's father warned his son's outpatient therapist, Dr. Goldstein, that his patient had threatened his ex-girlfriend and her new boyfriend. Goldstein advised hospitalizing the patient who was then admitted voluntarily. The following day, the inpatient psychiatrist, Dr. Levinson, planned to discharge the patient as not suicidal, over the objections of Goldstein. The discharge occurred as planned, and one day later the patient killed the new boyfriend Keith Ewing, and then himself. Levinson settled and so was not a party to the suit that followed.
Ewing's survivors sued Goldstein. The appellate court, in a decision considered by many legal and clinical observers to be simply wrong, found that the California statute mandates a duty to warn and that Goldstein, having failed to warn Ewing, was potentially liable. In the second case, Ewing v. Northridge Hospital Medical Center,18 the Ewings sued the hospital social worker who had done the initial evaluation, claiming again a failure to warn. Again, the appellate court found that the therapist had a duty to warn and that her failure to do so created a triable issue.
The Ewing decisions led to several critical law review articles (e.g., Edwards20 and Smith21). “In the Ewing decisions, the courts extended a statute and statutory definitions beyond their ordinary meanings” (Ref. 20, p 183). Edwards went on to criticize the decision for requiring therapists to breach confidentiality on the basis of deadly information given by family members. Such hearsay information forces a psychotherapist to appraise the accuracy of the information, and potentially act in an inappropriate manner should the information be highly inaccurate. She noted also that the court did not define family members and that the definition of family has itself undergone great recent change.
Smith was particularly outraged by the Ewing decisions, but her analysis suffers from several misconceptions. She begins with a section headed “The Treatment of Dangerous Patients” (Ref. 21, p 294), as if dangerousness were a fact rather than a matter of clinical judgment. She also ignores the notion that the patient's right to confidentiality does not survive the clinician's concern for potential violence to self or others. She writes, “therapists are legally prohibited from communicating with third parties, including family members, regarding an adult patient's treatment absent the patient's consent” (Ref. 21, p 310).
Smith is concerned about the threat to patient confidentiality that Ewing posed, and she proposed a novel solution: when in doubt, commit the patient. If a therapist is somehow confronted with a communication of a threat relayed by a family member that he or she believes is genuine, the option of temporary civil commitment under the LPS [Lanterman, Petris, and Short] Act should be considered as a more appropriate and already existing alternative to expanding the criteria that trigger the duty to warn [Ref. 21, pp 313–14].
Smith appears to believe that a patient is less damaged by being locked up for 72 hours than by having confidentiality broken. In her view, involuntary commitment provides: … added protection for both the patient and the therapist… . LPS has the safeguard of further evaluation of the patient, which allows the therapist to assess the credibility and accuracy of the family member's communication before deciding to break confidentiality, instead of being compelled to warn based on the communication alone [Ref. 21, p 317].
We agree that commitment is often necessary in cases of reasonable danger in which a clinician cannot divine all the facts, but it would be an egregious violation of basic civil rights simply to commit all patients without judgment.
The more important consequence of Ewing was a concerted and ultimately successful effort by California psychiatrists and others to lobby for a change in the law. Weinstock et al.13 wrote a careful, balanced account of the whole experience, and it was Weinstock himself who led the state psychiatric effort to revise the California statute. The revised statute makes clear that warning is but one of several possible actions that fulfill the duty to protect in California. This statute, § 43.92 of the Civil Code,22 was amended and became law on January 1, 2007.
At the most basic level, commentators have been deeply divided on whether the legal duty the courts have articulated is good social policy or is mistaken. In 1976, Stone23 argued that the duty to warn represented a potential disaster for the psychotherapeutic professions because it would require therapists to breach confidentiality. This breach would in turn prevent patients from trusting therapists, essentially ending any possibility of a therapeutic alliance. Later, in a 1984 publication, Stone24 expressed reservations about the Tarasoff duty, but acknowledged that the practice of psychotherapy had not been as seriously damaged as he and others had initially feared.
Stone modified his position, but others are not convinced. Herbert thinks that the duty to warn is an unwarranted imposition on therapists who cannot predict violence, and the obligation to do so has bad consequences: “Patients are … walking law suits under Tarasoff” (Ref. 12, p 421) and “indisputably it has weakened psychotherapy” (Ref. 12, p 421). “Tarasoff does real harm” (Ref. 12, p 422). “Whether a patient means particular words as a threat … simply cannot be assayed” (Ref. 12, p 422).
By contrast with Herbert, Beck11 has seen the duty to protect as constructive. Walcott et al.,17 were critical of early Tarasoff decisions, but argued that, over time, the developing case law appeared to represent less of a threat to clinical practice than had earlier decisions. Similarly, in Almost a Revolution, Paul Appelbaum concluded: In principle, the duty to protect is difficult to reject, especially for members of professions dedicated to assisting others in need … and with courts and legislatures muting the more problematic forms that the duty has taken, the duty to protect now looks like a reasonable step in the evolution of mental health law [Ref. 25, p 103].
None of the articles reviewed thus far reported on the extent to which psychiatrists and other mental health professionals currently are being found liable for breach of the Tarasoff duty to protect. The points raised by the various commentators are important, but as a practical matter, clinicians must know whether and to what extent they are being held liable for breach of the legal duty in statutory and nonstatutory states.
To address this question we review case law in states with statutes and, for comparison, we also review over a comparable time period case law in states without statutes. The purpose of this comparison is to evaluate whether there appears to be a different risk of being found negligent in states with statutes as contrasted to the risk in those states without.
Methods
A Westlaw search of cases and secondary materials citing Tarasoff from 1985 through 2006 yielded 612 cases and 99 articles discussing the effects of this landmark decision. Shepardizing and a hand search supplemented the electronic search. To track and analyze the most recent evolution in Tarasoff doctrine, we conducted a detailed study of all of the cases from 1985 through 2006. We found 102 appellate cases in which a mental health professional or institution was sued for breach of the duty to warn or protect. We excluded from our data 32 cases that concerned testimonial privilege or the disclosure of privileged information from a clinician to a third party. We abstracted the records and recorded data from the remaining 70 cases. Data abstracted includes the legal theory under which the case was brought; the identity of the defendant or defendants; patient characteristics including history of violence, possible motivation, diagnosis, and relationship to the victim; the extent of the injury; the victim's prior awareness of the threat; and the legal outcome of the cases.
This method tracks state appellate cases, both published and unpublished, and it also tracks federal appellate and federal trial cases. It does not track state trial court cases that are not appealed or cases that settled before trial. To gather relevant data on plaintiff verdicts in these latter types of cases, we obtained insurance company payout data from the APA-endorsed Professional Risk Management Services and from an insurance company that insures many Massachusetts psychiatrists. Plaintiffs’ verdicts that are not appealed or cases that settle with favorable outcomes for plaintiffs should be reflected in insurance company payout data. These records will reflect trial court verdicts for plaintiffs, but not for defendants.
Results
There were 139 defendants in these 70 cases, including 42 psychiatrists, 8 other physicians, and 20 other mental health providers including psychologists, nurses, therapists, psychiatric specialists, and counselors. Forty-six facilities and the organization that owned or was otherwise responsible for the facility were named. Also named were 2 patients, 11 state or Federal government entities, and 2 insurers. Eight defendants classified as other included two drug companies, two executive directors, one inn, one shelter, one undefined individual, and one owner of the vehicle.
There were 92 allegations in the 70 cases: 39 negligence, 37 failure to warn or protect, 12 negligent release, 4 medical malpractice, and 2 missing data. In only 12 cases was a failure to warn or protect the plaintiff's sole allegation. Most cases also claimed either negligence or negligent release.
Forty-seven cases were decided in favor of the defendants, 6 were decided for the plaintiffs, and 17 were returned to the trial court for further litigation. Of the six plaintiffs’ verdicts, four were based on violation of the duty to protect, and two on breach of a common law duty. Table 1 summarizes the verdicts of 70 Tarasoff cases in 28 states that mandate an affirmative duty to warn, 10 jurisdictions (including the District of Columbia) that permit a duty to warn, and 13 states that do not have Tarasoff statutes.
Verdicts in 70 Tarasoff Cases
Table 1 shows that there were more plaintiff verdicts in states without statutes (1/5, 20%) than in states with statutes (5/65, 7.69%). This result is consistent with our expectation, but is not a statistically significant finding (χ2 = 0.14, p = .9057).
The table also reveals an unexpected finding: the likelihood of cases being remanded to the lower court was higher in permissive states (7/14 cases, 50%) than in either mandatory states (10/51 cases, 19.6%) or states with no statutes (1/5 cases, 20%). This is a dramatic difference that trends toward a significant relationship (χ2 = 5.623, p = .0601).
In the following sections, we present the results of our analyses of individual cases. The results are organized by verdict, within verdict by the legal status of the duty within the jurisdiction (statute mandating an affirmative duty to protect, permissive duty statute, or no statute) and finally by the legal theory on which the case was decided.
Defendant verdicts were returned under a variety of legal theories, including no Tarasoff duty in this jurisdiction; Tarasoff duty, but not on the facts of this case; and defendants statutorily immune. Within each category, we present at least one case example, trying to select those that illustrate some legal or clinical point that has general applicability.
Defendant Verdicts
Thirty-seven Cases from 28 States With Statutes That Mandate an Affirmative Duty
No Communicated Threat to the Therapist Against a Specific Victim
The most common reason a court in a state with an affirmative duty Tarasoff statute did not find a duty is that the patient had not previously communicated to the therapist a threat against an identified victim. The 28 cases, courts, and accompanying state statutes are shown in Table 2. We further categorized the cases according to the patient-victim relationship.
Verdict or Summary Judgment for the Defendant Because There Was No Communicated Threat to the Therapist Against a Specific Victim
Victim Well Known to the Patient, but no Communicated Threat to the Therapist.
In 7 of these 12 cases there was a history of violence toward the victim and/or a credible motive—facts that before the statute often led to plaintiffs’ verdicts. One man killed three members of his ex-girlfriend's family,26 one man shot his girlfriend and her companion,27 two killed their wives,28,29 two killed their mothers,30,31 and one shot his sister and her husband.32 The court's reliance on the statute is well illustrated in Head v. Inova Health Services.30 Alfred Head, age 21, was a voluntary patient at a psychiatric hospital until he was discharged against medical advice. Three days after his release, he killed his mother. Head's father sued the psychiatric nurses, two psychiatrists, and their hospital for medical malpractice. The jury found in favor of the defendants. The Virginia appellate court refused to reconsider the verdict because: In this case, no evidence was adduced that Alfred communicated to any of the Fairfax Hospital nurses or to (his Drs.) “a specific and immediate threat to cause serious bodily injury or death to an identified or readily identifiable person or persons.” Nor was there any evidence that Alfred had “the intent and ability to carry out [any] threat immediately or imminently,” as required by statute before liability may attach [Ref. 30, pp 6,7].
In 5 of the 12 cases, there was a relationship between the patient and the victim, but no past violence or communicated threat.33–37 In Lagow ex. rel. Estate of Welch,33 the court found for the defendants on statutory grounds (no communicated threat to the therapist) but went further. Noting that outpatient treatment had ended 10 months before the killing, the court said that there was no special relationship between the therapist and the patient that could be held to create a common law duty to protect.
In Silvas v. South Bay Community Services,36 after their father killed their mother, the children sued the psychologist who had treated their father. The father's outpatient counseling had ended two months before the killing. The appellate court upheld summary judgment for the defendants, finding that there was insufficient basis to conclude that anything the therapist did or did not do two months earlier was sufficient to have created a triable issue of proximate cause.
In Suzuki v. Eli Lilly & Co.,37 the patient killed his supervisor. Although the patient had expressed anger to his psychiatrist about how he had been treated at work, there had been no communicated threat. Plaintiffs attempted to prevail despite these facts by claiming that the defendants had breached a more general duty to diagnose and treat the patient correctly. The court denied this claim and noted, “whether or not there was negligence in the treatment of [the patient], under the circumstances of this case, [the treating psychiatrist] and [the clinic] owed no duty of care to these plaintiffs” (Ref. 37, p 8).
Victim Not Well Known to the Patient and Communicated No Threat to the Therapist.
In 16 cases, the victim was not well known by the patient, and no threat was communicated. Five of the cases involved a patient in a hospital who injured another patient.38–42 In Sellers v. United States,38 a voluntary patient severely beat another patient at the Veteran's Administration Hospital (VAH). The defense verdict held there was no general duty to the public, because the victim was not readily identifiable. The court added that the patient's voluntary status meant that the VAH did not have a special relationship with him to keep him confined. Two additional cases involved patient violence toward another patient,39,40 and two involved patient violence toward a staff person.41,42 Two cases involved hospitalized patients who were violent toward a random victim in the community while on pass43 and after elopement.44
Four community victim cases involved a motor vehicle injury.45–48 Hartford Insurance v. Manor Inn45 involved a driving injury caused by a patient who escaped from involuntary commitment. The Maryland Court did not reference the Tarasoff statute, but did cite case law (Furr v. Spring Grove State Hospital),49 in which the state did not owe a duty of care to an unidentifiable victim. The remaining random victim cases all involved ambulatory patients who were not known to have threatened the victim.50–52
One random-victim case53 involved a psychiatrist who was treating another psychiatrist. The psychiatrist-patient was a pedophile, and the treating psychiatrist knew that his patient was a pedophile. The pedophile raped a boy whose family then sued five physicians, including the treating psychiatrist. The appellate court upheld the trial court verdict for the treating psychiatrist on grounds that his patient had not communicated to his defendant therapist a threat to an identified victim and that there was no duty to the public at large.
Richter v. Turlo54 is the last of these 28 cases decided on statutory grounds. The report does not provide any facts about the patient-victim relationship. The case is of interest because it is another illustration of a court sharply limiting the duty. In this case, there was no evidence that the patient had ever communicated a threat against the victim to the defendant therapist, but there was evidence that the therapist knew of the threat to the victim from other sources. The plaintiff relied on this in her claim that the therapist had breached a duty to the victim. The Richter court said that the plain language of the statute controls and that the duty is triggered, “if a patient communicates to a mental health professional … (Ref. 54, unpublished).” Quoting this language, the appellate court upheld lower court summary judgment for the defendant. On similar facts, a California court reached a different conclusion in Ewing v. Northridge Hospital Medical Center,18 as discussed earlier.
Richter is of further interest because it appears that the defendant-physician asserted the patient's privilege and refused to allow the plaintiff access to the record of the patient's treatment, which the court apparently allowed. If our reading is correct, the net effect of this (unpublished) decision is to make it impossible for plaintiffs to obtain the evidence necessary to create a cause of action for duty to protect!
Victim or Guardian of Victim Knew of Danger
In two cases, the court held that there is a Tarasoff duty, but it does not run to persons who are already aware of their danger.55,56 In Latimer v. Havenwyck Hospital,55 a man threatened his father and was hospitalized. After his discharge from the hospital, he shot his father. The hospital was found not to be liable, because the father had known of his son's threats to harm him before the hospitalization, and no new threats were issued during this hospitalization.
Treater Never Took Charge of the Patient
In two Virginia cases, the defendants were deemed not to have taken charge of the patient. In Nasser v. Parker,57 the Supreme Court of Virginia ruled that there is a duty to warn if a special relationship exists and if the psychiatrist and hospital have taken charge of the patient. Virginia is unique in deriving a duty to warn from taking charge, meaning the involuntary hospitalization of a patient. The usual doctor-patient relationship does not rise to the level of taking charge, so that in Nasser there was no special relationship and therefore no duty to warn.
In Sage v. United States,58 an active-duty medical officer shot random victims. The Federal Eastern District Court of Virginia's verdict was in favor of the United States because the government “had not taken charge of nor did it exercise control over the assailant” (Ref. 58, p 1), and the shooting was not foreseeable. Nasser and Sage occurred before Head v. Inova Health Services.30 In contrast, the Head court did not reference taking charge, but instead focused on the fact that the psychiatrists were not given a specific and immediate threat to harm, as necessitated by the Virginia statute.
Duty Discharged by Warning the Victim
The decision in Emerich v. Philadelphia Center for Human Development59 mirrors the statutory language in many states and essentially protects clinicians in Pennsylvania as if the legislature had passed a statute. The patient's boyfriend killed her after they separated. The victim planned to go to his apartment to retrieve her belongings. The patient had told the therapist on the day of the killing that he planned to kill his girlfriend. Later, the girlfriend called the therapist to ask whether she should go, and he strongly advised her not to. She went and was killed. The court held that the therapist had fulfilled his duty by urging her not to go.
The Emerich court said: [W]e find that, in Pennsylvania based upon the special relationship between a mental health professional and his patient, when the patient has communicated to the professional a specific and immediate threat of serious bodily injury against a specifically identified or readily identifiable third party and when the professional determines, or should determine under the standards of the mental health profession, that his patient presents a serious danger of violence to the third party, then the professional bears a duty to exercise reasonable care to protect by warning the third party against such danger [Ref. 59, p 1043].
In this case, the court held that the therapist had fulfilled this duty by advising the victim not to go, and, the court added: “We leave to another day the related issue of whether some broader duty to protect should be recognized in this Commonwealth” (Ref. 59, footnote 5).
Statutory Immunity
In four cases,60–63 the court ruled in favor of the defendant, not on the basis of immunity under the Tarasoff statute but on immunity under a different statute (Table 3).
States With Affirmative Duty Jurisdictions: Verdict or Summary Judgment for the Defendant Because of Statutory Immunity
In each of these cases, the defendants were protected by immunity because they were agents of state and county entities. This immunity protected workers in a community residence for ex-convicts60 and clinicians employed in a county mental health facility.61
In Tobis v. State,62 the facts involved a man who killed both his first and second wives. After the first killing, he was found not guilty by reason of insanity. Several years later, after his release from commitment, he remarried and then he shot and killed his second wife. Her survivors sued. The Tobis court referenced the state's Tarasoff statute (RCW 71.05), but held that it did not apply to procedures affecting the criminally committed and dismissed the case based on judicial immunity (2004). Defendants in Michael E. L. v. County of San Diego,63 Higgins v. Salt Lake County,61 and Ley v. State of California60 were all protected by state statutes granting immunity.
Six Cases from 10 Jurisdictions (9 States and the District of Columbia) with Permissive Statutes
In Green v. Ross64 the Florida Court of Appeals dismissed the case, holding “we are unable to conclude that the permissive language of section 491.0147 created an affirmative duty to warn so as to support a cause of action for a failure to warn” (Ref. 64, p 542). In Boynton v. Burglass65 the Florida court rejected Tarasoff, stating that the science is not sufficient to predict dangerousness. The court further stated that the Florida statute does “not require a psychiatrist to warn but is couched in permissive terms, and merely provides that a psychiatrist may disclose patient communications” (Ref. 65, p 451). In Santa Cruz v. Northwest Dade Community Health Center,66 the Florida court ruled that there is no special relationship between the mental health center and victims of a shooting.
An Illinois Appellate Court dismissed a nurse's complaint against a psychiatrist after a patient beat her. In Charleston v. Larson,67 the court held that a special relationship did not exist between psychiatrist and nurse sufficient to create a duty for a psychiatrist to protect her.
Praesel v. Johnson68 involved an ambulatory family practice patient with a known seizure disorder who seized while driving. She killed the driver of another car. Survivors sued the family practitioner, a neurologist, and a neurosurgeon, all of whom had seen the patient. The Texas court noted that prior Texas decisions held that there is no duty when the victim is not clearly identified, and they went on to say, “This court has not yet determined whether mental health practitioners have a duty to identifiable victims of a patient. We do not decide the question today” (Ref. 68, p 397). The court also noted that a physician in Texas has no affirmative duty to report to the Department of Motor Vehicles that a patient has a seizure disorder. Reporting is discretionary.
Weigold v. Patel69 is another driving case. The patient was a nurse who was on medication that interfered with her concentration. She knew she was not supposed to drive, but chose to drive anyway and caused an accidental death. The plaintiff sued the nurse's psychiatrist and her psychologist. The court held that it was her independent action in choosing to drive that was the proximate cause of the accident and death and found for the defendant on that basis. The court added in dicta that there is no duty to the general public.
Four Cases from 13 Nonstatutory States
No Tarasoff Duty in This Jurisdiction
In Williamson v. Liptzin,70 a floridly psychotic patient shot eight people, and one of the victims sued the psychiatrist. The North Carolina Court of Appeals ruled that the psychiatrist was not the proximate cause of the plaintiff's injuries, because the therapy had ended eight months earlier and there was no evidence that the patient was dangerous at the time of termination. Two years after Williamson, Gregory v. Kilbride71 concerned a psychiatrist who allegedly failed to commit a patient. The patient subsequently murdered his spouse and himself. As in Williamson, but more explicitly stated, the court held that “North Carolina does not recognize a psychiatrist's duty to warn third parties” (Ref. 71, p 114).
No Duty Under the Facts in the Case
In Hesler v. Osawatomie State Hospital,72 an involuntarily committed state hospital patient was given a pass. He caused a motor vehicle accident that killed the plaintiff's decedent. The Kansas Supreme Court found that a duty to control existed, but that involuntary commitment does not create a psychiatrist's duty to the general public. Although brought under §1983, this case adds to the case law holding that mental health professionals do not owe a duty to the public at large.
In Weitz v. Lovelace Health System, Inc.,73 a man killed his wife, his daughter, and himself. Survivors sued the outpatient therapists and the health care system that employed them. The federal court held first that New Mexico law would not consider that the outpatient relationship implies sufficient control to trigger a duty to protect. The court added as its own opinion that, “imposing a duty to control in the outpatient context would require providers to exercise a degree of care and oversight that would be practically unworkable” (Ref. 73, p 1181). Further, the court held that there was no duty to warn, because the victim already knew that she was in peril.
Undecided Cases
Affirmative Duty Statute States
In the first of 10 affirmative duty cases, Barbarin v. Dudley,74 a home-duty nurse sued a physician after a patient injured her. On the day of the injury, the physician warned the nurse that the patient had violently attacked her mother four years earlier, but that the patient had never communicated a specific threat during his subsequent course of treatment. The Fourth Circuit Court of Appeals held that the physician had fulfilled the statutory duty by warning the nurse about this patient before her injury. However, the same appeals court reversed and remanded the case to the trial court, saying that the physician may still have a common law duty to protect the nurse.
In Clay v. Telecare,75 the defendant had repeatedly failed to hospitalize a patient with schizophrenia who had a known history of violence and who had repeatedly threatened the victim. The court said the facts created triable questions of whether a duty to warn and a duty to hospitalize were present.
In Perreira v. State,76 a psychotic patient was discharged from a state hospital, and four months later he shot and killed a police officer. The surviving spouse sued the psychiatrist, the mental health facility, and the state. The killing occurred before Colorado passed its Tarasoff statute, and the Colorado Supreme Court held that the psychiatrist had a legal duty by common law to exercise due care in determining whether a patient presented a risk of serious bodily harm if released from involuntary commitment.
In Evans v. Morehead Clinic,77 the events occurred in a statutory state, but before the statute was enacted. Seven months after a discharge against medical advice from a state hospital, a patient shot a man whom he had previously accused of having an affair with his 80-year-old wife. The court held that, on the facts in this case, a duty to protect might exist. The court reversed summary judgment for the defendants and remanded the case for trial.
Ewing v. Northridge Hospital Medical Center,18 which falls in the affirmative duty category, is discussed extensively earlier in the article.
In New Jersey, the Tarasoff statute mandates a duty to warn and protect in cases of suicide as well as in cases of harm to third parties. N.J.S.A. 2A:62A-16b2. Marshall v. Klebanov78 was brought under a duty-to-protect theory after a female patient of a psychiatrist committed suicide two days before her scheduled appointment almost one month after her initial intake. Dr. Klebanov had originally scheduled the patient to return one week after the first appointment, but Ms. Marshall did not appear for reasons that remained disputed. The plaintiff contended that Marshall was turned away by the office receptionist because she could not make payment, but Klebanov claimed that he had never refused treatment of a patient due to lack of payment. Instead, Klebanov insisted that Marshall refused the visit because she felt well enough to wait for insurance certification several weeks later. In fact, Klebanov called her that day to inquire regarding her absence and rescheduled the appointment approximately three weeks later. The court held that a “duty to warn and protect … can also be incurred when the circumstances are such that a reasonable professional … would believe the patient intended to carry out an act of imminent, serious physical violence against a readily identifiable individual or against himself.” (Ref. 78, p 376, italics added). The court said that the fact of the long interval between appointments created a triable issue, and remanded the case for trial. However, the court said the psychiatrist did not incur a duty to protect under the statute, but that he could still be liable under common law for abandoning a depressed patient.
After a hung jury, the Nebraska appellate court ruled on a Tarasoff case of first impression in Nebraska (Munstermann v. Alegent Health-Immanuel Medical Center).79 The court said that the facts created a triable issue. Five days after the patient was discharged from hospital, he killed his girlfriend. His comments while in the hospital about his intentions toward his girlfriend were in dispute. The court at first noted that the Nebraska Tarasoff statute does not apply to psychiatrists! Instead, it applies to “any person licensed or certified as a mental health practitioner.”80 The court noted that the language of the statute specifically excluded practitioners of medicine or psychology, but went on to say in dicta that psychiatrists had a common law duty to protect. The court went further and held that the intent of the Nebraska legislature in enacting a Tarasoff statute was clearly to hold psychiatrists to the same duty as mental health practitioners. Thus, the Tarasoff statute in Nebraska applies to psychiatrists just as it does to all other mental health practitioners, and on that basis, the court remanded the case for retrial.
Similar to the New Jersey statute, the Nebraska statute also includes language about duty to protect against harm to self. The duty to protect applies “when the patient has communicated a serious threat of physical violence against himself, herself or a reasonably identifiable victim or victims.”80
In Tamsen v. Weber,81 a patient who was involuntarily committed to an Arizona state hospital as a danger to himself was given unsupervised grounds privileges. The patient left the grounds and severely injured a random person. The court did not reference the Arizona statute, but reversed the summary judgment for the defendant and remanded the case for trial because, “A physician owes a duty to protect others by controlling an involuntarily committed patient with known or reasonably discernable dangerous propensities who is in the physician's care and custody.” (Ref. 81, p 24) The court found that the evidence in the case supported a reasonable inference that the hospital physician knew or should have known that the patient was dangerous.
Little v. All Phoenix South Community Mental Health Center82 involved an ambulatory patient who stabbed his wife, but she survived, and she sued the mental health centers, alleging that they owed a duty under Arizona's Tarasoff statute. The court agreed with the defendants that the wife's claim did not meet statutory criteria because there was not an explicit threat toward an identifiable victim. However, the court reversed and remanded the summary judgment for the defendants. The court held that the Tarasoff statute was unconstitutional because it “abrogated a general negligence cause of action for persons who were not ′clearly identified’ but who were foreseeable victims” (Ref. 82, p 1368).
In Shively v. Ken Crest Center for Exceptional Persons,83 a resident of a facility for the mentally impaired sexually assaulted a boy in the same apartment complex. The court did not reference Delaware's Tarasoff statute, but concluded that the facility “owed a duty to the minor plaintiff to take reasonable measures to protect him from its resident and warn him of the resident's reasonably foreseeable dangerous propensities. Public policy dictated that result” (Ref. 83, p 1). A motion for summary judgment for the defendant was denied.
States With Permissive Statutes
There were seven undecided cases in states with permissive statutes. In two cases brought under the Tarasoff theory, the appellate court specifically rejected the theory, but held that another duty existed and that the duty created a triable issue.84,85 In Bryson v. Banner Health System,84 an Alaska court said that a substance abuse facility owed a duty to protect its patients because it undertook to treat them. When one patient raped another, the duty to the victim created a triable issue. In a Florida case, O'Keefe v. Orea,85 a 17-year-old killed his father four days after the boy was discharged from the hospital. The psychiatrist had treated the son's parents in connection with their difficulty in controlling their son. The court held that the psychiatrist had two duties: a fiduciary duty to the parents of his minor patient, and a duty to the parents as his patients. The appellate court remanded the case for trial under a theory of medical negligence.
In Rivera v. New York City Health and Hospitals Corporation,86 a patient pushed a bystander in front of a subway train, and the survivors sued his health care providers and a shelter for failure to protect. The court found a possible duty to the public: A duty to the general public arising from the treatment of an outpatient turns on the facts. No bright line exists: in limited circumstances a mental health provider may be liable for failing to control or commit a voluntary patient who later harms a member of the public [Ref. 86, p 419].
In Bragg v. Valdez,87 a patient was discharged from an inpatient unit because he lacked insurance. He later assaulted his mother. The trial court dismissed the suit. The California Court of Appeal reversed and remanded. In this case, there was no communicated threat, but the court held that the Tarasoff statute applied only to medical decisions regarding clinical judgments about danger, not to decisions based on money alone. The court held that the psychiatrist owed a general duty under the commitment statute to anyone assaulted, because the discharge was not based on a medical judgment but was solely for monetary reasons.
In Logan v. Smith,88 an outpatient had repeatedly expressed anger toward his supervisors, but never toward the fellow workers whom he subsequently killed. In dismissing the defendant's motion for summary judgment, the court said the workers were part of a group of foreseeable victims.
In Schlegel v. New Milford Hospital,89 the court found a possible duty to control when an outpatient killed his mother less than a day after being discharged from a hospital emergency room. The patient had a long history of violence and substance abuse, had behaved bizarrely toward his mother, had very low serum potassium, and was actively psychotic in the emergency room. He was given one dose of chlorpromazine and discharged to his mother's care without any further evaluation and without a treatment plan.
In Mason v. IHS Cedars Treatment Center of Desoto Texas,90 three recently discharged inpatients were in a single-car accident that killed one and seriously injured the other two. The suit claimed negligent release, arguing inter alia that the prior relationship of the patients in the hospital raised questions about the appropriateness of discharging them at the same time. The court reversed a trial court verdict of summary judgment for one psychiatrist, and remanded, not on a Tarasoff basis but rather on a theory that discharge of the patients may have been the proximate cause of the injury and that the discharge itself may have fallen below the standard of care.
Six Cases Decided for the Plaintiff
States With Affirmative Duty Statutes
Negligence Determined on the Facts
Turner v. Jordan91 involved a manic inpatient with a history of serious violence including past violence on the unit. Staff were frightened of him. After the patient, who was in good control and not suicidal, refused medication, Dr. Jordan, the attending psychiatrist, wrote a note stating that the treatment plan was to encourage the patient to sign out against medical advice. Later that day, the patient attacked and severely injured a nurse. She sued Jordan, and she won. The appellate court upheld the trial court verdict for the nurse. In Walker's Adm. v. Simmons,92 a psychiatric patient killed her five-year-old cousin after her parents signed her out of a psychiatric treatment center for a weekend visit. The family alleged that this patient had communicated threats to kill family members, peers, and staff on numerous occasions before the weekend visit. The court found sufficient cause of action against the psychiatrist and treatment center under Virginia's mandatory Tarasoff statute, because a threat had been communicated.
Statute Misinterpreted by Trial Court
In a Colorado case, Halverson v. Pikes Peak Family Counseling and Mental Health Center,93 an inpatient who had a history of violent behavior assaulted the plaintiff. She asserted that the defendants failed to protect her even though they “had notice” (Ref. 93, p 234) of this patient's dangerousness and his prior aggressive behavior toward her. The trial court issued a summary judgment for the defendant. The appellate court reversed and remanded, holding that the district court had erred when it held that under the Colorado Tarasoff statute, there was no duty to protect unless the victim herself had told the defendants that her assailant had made a serious threat. The plaintiff later prevailed in trial, and the defendant was denied certiorari by the Supreme Court.
Not a Tarasoff Case on the Facts, but Negligent Treatment
In Sheron v. Lutheran Medical Center,94 the spouse of an overdose victim brought suit against a psychiatrist and the medical center. The Colorado court of appeals refused to extend the Tarasoff statute to suicide victims, but did affirm the plaintiff's judgment based on negligence, because the psychiatric specialist did not recommend hospitalization during an evaluation the day before the victim's death when she admitted to an overdose with pills.
Permissive Statute States
In Garamella v. New York Medical College,95 the third author was the plaintiff's expert at trial. The psychiatrist defendant was both the resident's psychoanalyst and a supervising psychiatrist in his training program. The patient was a pedophile who believed that child-adult sex was good and right and told the analyst about his plans to enter a child psychiatry fellowship. The analyst scrupulously maintained his patient's confidentiality. The patient ultimately raped a boy whom he treated while moonlighting as an emergency room doctor. The boy's parents sued and the trial court decided for the plaintiff. Subsequent to the publicity attendant on the lawsuit, other victims came forward and the resident was convicted and sentenced to prison for his crimes.
States With No Tarasoff Statute
In Long v. Broadlawns Medical Center,96 the hospital promised a patient's family that they would warn them before discharging the patient. They did not warn at discharge, and the patient killed a family member. The family sued. The Iowa court stated explicitly that it was not imposing a Tarasoff duty to warn on psychotherapists, but said that a common law duty to warn existed in this case because the hospital had promised to warn and then failed to do so. The court noted that this common law duty survived the family's knowledge of the patient's violent propensities. Although this is a plaintiff verdict, it does not find the defendant negligent for breach of a Tarasoff duty. The court explicitly states that it is not imposing such a duty.
Discussion
The primary finding of this review is that defendants are now rarely held to be negligent on grounds of failing to warn or protect. In reviewing 21 years of legal history, we found just four cases in which psychotherapists were found liable for breach of a Tarasoff duty. In two other plaintiff verdicts, the court explicitly rejected Tarasoff as a basis for the verdict, finding instead a common law basis for the breach of duty. This is a welcome change in how the law views the legal duty of practicing clinicians. In the early days after Tarasoff, courts found negligence in cases with fact patterns similar to those that now are being decided for defendants (for example, Refs. 3–7). This evolution is the result of legislative actions, increased judicial sympathy toward the clinician confronted with threats of violence, and social climate change.
The verdicts favoring the defendants are noteworthy for the bases on which courts returned them. In North Carolina, the courts have said that there is no Tarasoff duty,70,71 and in Virginia, the Tarasoff duty may be most applicable to involuntary inpatient settings.57,58 In statutory states that either affirmatively mandate or permit clinicians to warn, the courts consistently found that there was no duty under the statute, because the patient had not communicated a threat to the therapist (for example, Refs. 33, 68). Seven of these are cases that, before the statute, might well have led to verdicts for the plaintiffs.26–32 In each there was some history of violence, prior threats, and at least a presumption of motive. In an additional five cases, the facts were borderline (i.e., similar fact patterns had led to verdicts for the plaintiff in at least some early cases), the patient knew the victim, and there was some basis to infer motivation, but no prior history of violence and no specific threat communicated to the clinician.33–37
Perhaps equally important, whether in a statutory or nonstatutory state, the courts almost always found that defendants owed no duty to the public at large. These cases involved motor vehicle deaths, patients injuring patients or hospital employees, patients who were violent on a pass from the hospital, and random encounters between strangers (for example, Refs. 38, 69, 72). This holding is in marked contrast to earlier plaintiff verdicts.4,5 Only once, in Richter,54 did a court find that a duty might extend, under some circumstances, to the public at large.
Further limiting the duty, courts held in four cases that the duty did not extend to persons who were aware of their danger, irrespective of whether the therapist had warned the victim.55,56,69,73 One involved a husband who killed his wife and daughter, and one involved a man who killed his father. In one of these cases, the court also noted that there is no duty to control outpatients. These verdicts contrast with the much earlier verdict in Jablonski,3 in which the court found a psychiatrist liable even though the victim had ample knowledge of her peril.
In two cases, the court held that the therapist could not be held liable for violence that occurred too long after termination: two months36 and eight months70 after therapy ended. Again, these cases contrast with the earlier case in which the psychiatrist was found liable for violence occurring to a random victim of a motor vehicle accident five and one-half months after the patient had left the hospital.7
It is also striking that in four cases, the courts went beyond the facts to make a comment on the limits of a possible duty. Although given an opportunity by the facts of the case before them, three courts specifically declined to rule on whether there should be a Tarasoff duty in their jurisdiction59,68,73 and the fourth70 came close to suggesting that the statute created a barrier to a finding of negligence, even when the defendant breached the standard of care.
Seventeen remanded cases have yet to be decided, and we therefore might see a further change in how the courts decide these cases. However, the six plaintiff verdicts illustrate Tarasoff's inability to protect clinicians from clinical assessments that the courts found to be seriously lacking or missing altogether (Bragg87). Although clinicians would prefer concrete rules to govern their decisions regarding danger, we agree with the reasoning in Rivera86 that “no bright line exists” when assessing risk and making appropriate decisions that mirror the assessment. In contrast to Felthous,10 we do not view a threat communicated to a therapist as an aggressive act that warrants disclosure, and we do not favor this proposed further expansion of the duty when it would entail an unwarranted breach of confidentiality. Clinicians are appropriately duty bound to breach confidentiality when necessary to protect against a legitimate risk of danger.
Tarasoff statutes limit the duty to protect to cases in which therapists may reasonably be expected to have anticipated violence, but they do not completely insulate clinicians from a court's finding that the clinician or facility failed to exercise sound clinical judgment. On the facts as presented in Clay,75 the plaintiffs could reasonably argue that a man with schizophrenia and a known history of violence who had threatened the victim should have been hospitalized. Marshall78 represents facts that can legitimately be a source of disagreement as to whether the psychiatrist breached a duty to his patient. Arguably, a one-month interval before a follow-up appointment is too long for a severely depressed patient, but the facts suggest that the long interval was a function of difficulties raised by the plaintiff to being seen earlier. From Ewing,18 clinicians can surmise that decisions to warn must consider all of the data, including the statements from and the credibility of collateral sources, and we would not rely on Richter54 to absolve this responsibility. Further, in Logan,88 risk assessments must consider the whole situation and not hinge on a singular communicated threat.
In Garamella,95 not only was there clinical judgment that led the psychiatrist to do nothing in the face of a potential danger to children, but there was also a serious systems concern. At that time, in that residency, a psychoanalytic training experience was offered as part of the residency, in which the analysts were also the supervisors, clearly creating blurred boundaries with obvious potential conflicts. After this verdict, the residency program, and the psychoanalytic experience were separated.
The remaining reversed and remanded and plaintiff cases also highlight a particular setting in which clinicians should exercise heightened attention to risk of danger: the control or involuntary confinement of patients. When considering Tarasoff, courts are more prone to hold clinicians liable in settings where patients are under the control of a clinician or where the clinician denied inpatient commitment. Of the 23 remanded or plaintiff cases, the facts in 13 such cases59,77,79–85,87,90–93,96 involved inpatient and residential settings, two cases89,93 involved the emergency room evaluation of a psychiatric patient, and one case75 concerned the repeated denial of hospitalization for a patient with schizophrenia. Further as stated earlier, there have been two cases in Virginia that appear to derive a duty to warn exclusively from taking charge, meaning the involuntary hospitalization of a patient.57,58
Being responsible for a patient in a confined or residential setting carries a potential special duty to anyone with whom the patient may come in contact. In Turner,91 a nurse was attacked. In Halverson93 and Bryson,84 patients were injured on inpatient units. In Mason,90 a court questioned the judgment of discharging three patients on the same day when they later were in a car accident together. In Shively,83 the court found that a residential setting may carry a special duty to warn and protect any of the residents from a sexually deviant resident. Given that courts are especially open to finding a duty in these cases, clinicians should be particularly vigilant in performing and documenting risk assessments in settings with a heightened degree of control.
Courts varied in how they decided cases in which patients injured others on passes from hospitalization or involuntarily committed patients who escaped confinement and subsequently injured a person. Most courts have found for the defendant clinician under Tarasoff because the victim was unidentifiable.43–45,47 However, Tamsen82 was remanded because the court was concerned about whether a psychiatrist breached his duty of due care by granting unsupervised grounds privileges to a potentially dangerous inpatient. In Walker's Adm.,92 the psychiatrist was found liable because the threat was communicated before the harm occurred.
Emergency room psychiatric evaluations that reject involuntary confinement of a dangerous patient are not unequivocally protected by Tarasoff statutes. In Schlegel,89 a psychotic patient killed his mother 24 hours after he was discharged following an emergency room evaluation. Plaintiffs prevailed in Sheron94 after a clinician failed to hospitalize a patient who had overdosed.
In our opinion, the breach in Turner91 was egregious. “Against medical advice” means what it says. Encouraging your patient to sign out against medical advice is tantamount to ignoring the patient and failing to provide or attempt to provide appropriate hospital-based treatment. In Bragg,87 the needs of a patient were again ignored when the hospital treatment was terminated based on a lack of insurance. Long96 is an example of a common law duty. When the hospital personnel undertook to warn the family, they created a duty to do what they promised. After that promise, the family reasonably expected to be warned and reasonably concluded that they could rely on the hospital. Because their damage appeared to be a direct result of the hospital's failure, there was a clear legal basis, having nothing to do with Tarasoff, for the plaintiffs to recover.
The statutes appear generally to promote a useful social policy, limiting the duty to protect to cases in which victims are identified or reasonably identifiable. In our view, the law of duty to protect should be interpreted so that known pedophiles are restrained from free access to children. Society makes rules limiting access of convicted pedophiles to children, and our view is consistent with this broader social policy. Children, whether named or not, are clearly foreseeable victims of pedophiles. The defendant's verdict in Doe53 in a statutory state contrasts with the plaintiff verdict in Garamella95 in which the appellate court found that children were foreseeable victims.
Courts in statutory states have generally been consistent in their interpretation of the law, with a few exceptions. In Little,82 an Arizona court declared the state's Tarasoff statute unconstitutional. It is also notable that 7 of the 17 remanded cases come from states with permission-to-warn statutes. That there are only nine states and the District of Columbia with such permissive statutes could indicate that permission, rather than a strict mandate to warn, can leave the courts more room to interpret and perhaps to increase the liability for clinicians. However in Green64 and Boynton,65 the Florida court, with a permissive statute, granted summary judgment to the defendants because the statute did not create an affirmative duty to warn, essentially negating an inherent duty to warn.
Of the 70 cases, the only one that appeared troubling was Barbarin74 in which a non-mental health clinician warned a home nurse regarding a patient on the day of the attack about a violent act that had occurred four years ago. The patient had not communicated a threat since, but the Fourth Circuit Court of Appeals decided that the physician may still have had a common law duty to protect the nurse, even though he had fulfilled the statutory duty to warn. We do not know the outcome of this case, but it is remarkable that this was the only case in which prudent clinical judgment, including a warning to a victim about the potential violent proclivities of a patient, was not sufficient to allow a verdict for the defendant.
A weakness of this review is that we have no way of capturing most of the cases decided before trial or at the trial court level that are not appealed. Thus, it is possible that there are many more plaintiff settlements or verdicts than we have reported. If this were true, we would expect that insurance company records would reflect payouts for duty to protect cases. In fact, Tarasoff cases do not appear in the summaries that insurance companies report. The APA-endorsed insurance company does not even list Tarasoff cases as a category of cases for which they have defended psychiatrists, let alone for which they have paid out a claim. The same is true for the insurance company in Massachusetts that insures many psychiatrists in that state. On this basis, it seems doubtful that there are many plaintiff verdicts escaping notice. It is true that trial verdicts are not published legal cases, but if the plaintiff verdicts existed, insurance companies would report their experience with them. What we cannot capture, if they exist, are cases settled at or before trial that lead to defendant verdicts or to dropped cases.
Finally, it was a surprise to us, and we have not seen it referenced in other review articles, that the New Jersey and Nebraska Tarasoff statutes reference violence against self as well as against others. We regard this as of academic, not practical, interest; we doubt that a suit after a suicide brought under a Tarasoff theory is more likely to succeed than one brought under the more usual theory of negligence.
In summary, we conclude that there is little basis for clinicians to fear being sued successfully for a bad outcome if the clinical practice has been reasonable. Generally, statutes appear to have played a constructive role in the development of the law of duty to protect.
Footnotes
-
Disclosures of financial or other conflicts of interest: None.
- American Academy of Psychiatry and the Law
