Education + Advocacy = Change

Click a topic below for an index of articles:




Financial or Socio-Economic Issues


Health Insurance



Institutional Issues

International Reports

Legal Concerns

Math Models or Methods to Predict Trends

Medical Issues

Our Sponsors

Occupational Concerns

Our Board

Religion and infectious diseases

State Governments

Stigma or Discrimination Issues


If you would like to submit an article to this website, email us at for a review of this paper

any words all words
Results per page:

“The only thing necessary for these diseases to the triumph is for good people and governments to do nothing.”



The Duty to Warn and Protect -

Impact on Practice

The authors discuss the concepts of the duty to protect and the associated threat to confidentiality and their impact on practice for Canadian psychiatrists. They review these concepts and provide a synthesis of legal cases impacting psychiatric practice. and conclude that the onus is on the psychiatrist to make him or herself aware of the current state of the legal obligation with respect to duty to protect. The evolving concept of duty to protect has and will continue to have significant impact on the practice of psychiatry.

The practice of medicine is increasingly subject to external review and legislation. Both have a significant impact on psychiatry in terms of risk assessment and the duty to warn or protect third parties (1–4). Tarasoff v. the Regents of the University of California (1974), the seminal American case, not only brought the issue to the forefront but also began the ongoing process of attempting to clarify the clinician’s obligation to protect society from potentially dangerous patients (5).

Since Tarasoff, there have been numerous related cases in the US and several pertinent cases in Canada (3,4,6,7–9). In response to the growing acceptance of a duty to warn and protect, many of the U.S. states have enacted legislation to define that duty or responsibility in law (10). At the same time, professional organizations have attempted to define the responsibilities of their membership in this regard, as the obligations may be unclear, and the act of warning or protecting may breach the time-honoured principle of confidentiality (11–13).

The duty to warn and protect is a relatively new concept and is a departure from traditional psychiatric practice. This so-called duty is to a large degree defined by legislation and case law.

Because this is an ongoing process, the exact nature of this duty remains difficult to define at any one time. What has begun to happen, however, is the growth of a body of literature on the “Duty to Warn and Protect”; a body of literature that becomes obsolete as new legislation and case law emerges. We describe the evolving concept of duty to protect and the often competing obligation of maintaining patient–physician confidentiality.

This paper provides an analysis of recent pertinent Canadian case law. The overall impact on the practice of psychiatry is discussed.

Duties to Warn and Protect
The issue of duties to protect or warn third parties at risk of violence from patients was first highlighted in the now well-known Tarasoff case (5,14). Unfortunately, in the subsequent zeal to discuss and describe and promote the outcome of this particular case, the original case, Tarasoff I, with its Duty to Warn, is often the only one quoted (5,15). In actual fact, it is Tarasoff II that uniquely distinguishes the ideas of the Duty to Warn and the Duty to Protect (14).

In Tarasoff, Mr Prosengiit Poddar told his treating psychologist that he planned to kill his former girlfriend, Ms Tatiania Tarasoff. In response, the clinician provided both oral and written warnings to the campus police (a fact not known by many clinicians), who interviewed Mr Poddar and subsequently released him. Mr Poddar later insinuated himself with Ms Tarasoff’s family and then killed her. Her parents initiated a lawsuit.

The case was eventually heard in the California Supreme Court, which, in Tarasoff I 1974, reversed the Appellate decision that there was no cause of action. The Supreme Court held that “the doctor bears a duty to use reasonable care to give threatened persons such warnings as are essential to avert foreseeable danger arising from the patient’s condition.” It also noted that “protective privilege ends when the public peril begins”— a now well-known maxim (5,7).

As a result of the defendant’s petition for a rehearing (with the American Psychiatric Association providing one of many amici briefs), the trial court was instructed to hear the lawsuit against the police and the various employees of the University of California. Because of the subsequent response, the California Supreme Court, in an unusual move, reheard the same case in 1976, later known as Tarasoff II. Tarasoff II held that “when a therapist determines, or pursuant to the standards of his profession should determine, that his patient presents a serious danger of violence to another, he incurs an obligation to use reasonable care to protect the intended victim against such danger. The discharge of this duty may require the therapist to take one or more various steps depending upon the nature of the case. Thus, it may call for him to warn the intended victim or if it is likely, to apprise the victim of a danger, to notify the police, or take whatever steps are reasonably necessary under the circumstances” (10). Generally, under common law, a person has neither a duty to control the conduct of another nor a duty to warn those at risk. Tarasoff and its progeny made an exception when the defendant had a special relationship with the person at risk of inflicting violence or the victim. In these situations, affirmative duties were supposed to benefit third parties (16).

Prior to Tarasoff, there was a clinical duty to patients, but subsequent to Tarasoff, the significant duty was now through patients. However, it is in Tarasoff that some of the critical issues around the Duty to Warn and Protect emerge. It is a common error to assume that Tarasoff I, that is the Duty to Warn, is the core concept. In fact, Tarasoff II in 1976 superceded Tarasoff I and articulated rather a Duty to Protect. It is from the distinction between the Duty to Warn and the Duty to Protect that some of the potential errors affecting practice arise (15,17,18). Further, the absence of discussion in Tarasoff of alternative solutions—such as hospitalization of potential offenders—that afford protection to society without breaching physician–patient confidentiality has impacted both subsequent practice and legislation (19).


While the purpose of informing or warning is to protect, in some situations these actions may be insufficient or even increase risk to the victim (17,20). The Courts shifted the duty of clinicians to warn third parties into areas where they never previously exercised control (3,21). This shift has been especially pronounced in the US, where the definition of third parties has been expanded to include unspecified persons at unforeseen times and places (4,9,16,21,22). One early, contentious question was whether physicians should warn third parties of the risk of contracting AIDS from their patients (23,24).

Even as psychiatrists began to become familiar with the Tarasoff ruling, its impact on the practice of psychiatry has come under increasing criticism (10,17–19,25,26). However, the courts have all but ignored challenges based on professional, ethical, and clinical grounds. US legislatures have been more responsive to clinicians’ concerns, attempting to clarify under which circumstances physicians had a duty to inform or protect (10). The courts have also made attempts at clarification (1,3,11).

Physicians face a conflict between their obligation of confidentiality toward their patients and their duty to protect others from potentially dangerous patients (18,27). Confidentiality can be defined as the ethical, professional, and legal obligation of a physician not to disclose what is communicated to him or her within the physician–patient relationship.

Confidentiality and trust between physician and patient have been considered essential to successful treatment (3,12). This ethical proscription against divulgence of patient confidences protects the legal right to privacy (10). The Hippocratic Oath states: “Whatsoever things I see or hear concerning the life of man, in any attendance on the sick or even apart therefrom, which ought not to be voiced about I will keep silent thereon, counting such things to be as sacred secrets”(28). Physicians have a fiduciary duty to their patients; they can and should protect their patients’ privacy and refuse inappropriate access to their files. In Canada, confidentiality of patient information is protected by statute except in certain specified situations.

It is the assurance of confidentiality that has allowed many patients to seek help who otherwise may not have done so. There is an acknowledgement, however, that confidentiality is not absolute. Examples of well-recognized exceptions to confidentiality are those related to child abuse, infectious diseases, and the various mental health acts (19,25). Other exceptions are reflected in the Federal Aeronautics Act and various Highway Traffic/Motor Vehicle Administration Acts.

On the other hand, there are professional bodies that have called for absolute confidentiality and have requested that psychiatrist–patient communications be given unqualified privilege (29). Physicians may still remain liable to legal action in the event they divulge the contents of their communication with patients, despite the existence of judicial developments or statutes that require them to do so in specified situations (and which then may provide a basis for a defence). “Duty to protect” situations are a case in point (19).

The right of a patient to bring a lawsuit for damages has been based on the implied contract to keep information confidential, recovery for invasion of privacy, recovery for tortuous breach of duty of confidentiality, and recovery based directly on statues prohibiting disclosure (licensing and privilege statutes). The various defences have raised the issue of consent, overriding interest of the public, the patients’ interest (Tarasoff), and reporting statutes (Tarasoff) (30).

The Impact on Practice: Balancing Confidentiality and the Duty to Protect

If one accepts that despite any moral, ethical, or professional objections, there exists a legal duty to protect, how can one proceed in a way that minimizes risk to all parties, including the clinician? As a starting point, we look at how the courts have viewed the obligation to third parties. Although some of the guidelines and statutes established to define responsibilities have considered “informing” or “warning” as sufficient, the major concern of the courts has ultimately been the protection of third parties (11).

If a clinician becomes concerned about a patient’s risk of harm to third parties, there are various algorithms to assist in making good decisions to protect them (2,31). Appelbaum proposed a three-part model for the duty to protect: assessment, selection of course of action, and implementation (17). Truscott presented guidelines for Canadian psychiatrists after Wenden v. Trikha (7,11). Most algorithms include as part of the first question a determination of whether the patient poses a danger to others (Figure 1). A risk-assessment instrument should be considered, along with formal or informal consultation. If the patient is not deemed a risk, assessments and interventions are continued as appropriate for the patient’s clinical condition. If there is a danger to others and the danger is due to a serious mental illness, the appropriate response is to consider or proceed with hospitalization. If the patient is already hospitalized, steps should be taken to increase security to prevent unauthorized leave from the hospital. If the risk is not due to mental illness, the question then becomes whether the danger is imminent or at least “causes a sense of urgency.” If the danger is imminent, then hospitalization may still be an option to rule out an underlying mental illness. The next step is to determine whether the potential victim(s) are readily identifiable or belong to a reasonably identifiable group. Various options still remain. One can consider notifying the identified victims, law enforcement authorities, or any other people who might be in a position to prevent the follow-through of violence. Family interventions can be useful as potential victims are often family members (1). Good documentation of each step is essential.

It is sometimes feasible to use the relevant provincial mental health act to protect an intended victim and at the same time limit confidentiality breaches to those associated with the act (32). For example, all provinces have mental health acts that allow for the detention of individuals. At the point of decertification or discharge, the issue of the duty to protect would need to be revisited. It is important to note that, if a determination is made to breach confidentiality, obtaining consent from the patient obviates many of the purported difficulties. Any breach of confidentiality should occur only when all other efforts to work with the patient have failed. Open negotiations with the patient in a climate of therapeutic beneficence often result in a solution satisfactory to all parties. This point cannot be emphasized too strongly.

Recognizing the dilemma facing their members, various professional bodies have attempted to protect and guide them by crafting guidelines (in the hope that they would become statute). James Beck, in a still-relevant article published in 1987 (9), suggested a model statute that should read:
No monetary liability, no cause of action may arise against a Licensed Mental Health service provider for failing to predict, warn or take precautions to provide protection from a patients violent behaviour if two circumstances exist: the provider has made reasonable and professional efforts to assess the likelihood of such violent behaviour, and those efforts have failed to yield a threat of physical violence against a clearly identified or reasonably identifiable victim or victims.

Any licensed Mental Health service provider who determines that his or her patient poses a threat of violence to another shall discharge the legal duty to protect the other by any professional, reasonable course of action. Such actions shall include but are not limited to communicating the threat to victim or victims: notifying the appropriate law enforcement agency; arranging for the patient’s hospitalization voluntarily or petitioning for involuntary hospitalization.

Ontario’s Medical Expert Panel on the Duty to Inform felt that the duty existed when the patient revealed an intention of committing serious harm and that the threat was more likely than not to be carried out. The victim could be a person or group, and the police, and possibly the victim, should be notified. In less clear circumstances a risk assessment should be done, possibly by a colleague (13). Following the publication of the consensus statement of Ontario’s Medical Expert Panel on Duty to Inform (13), the College of Physicians and Surgeons of Ontario adopted this statement, thereby affirming it as the standard of practice in Ontario.

The profession has been criticized for appearing to embrace the duty to protect. Criticisms include the argument that the prediction of dangerousness places too stringent a responsibility on psychiatrists (10,26). Despite data to suggest that the predictive ability of psychiatrists has improved, evidence for the specificity and utility of predictions is not overwhelming (33–37), and many potential victims are left vulnerable. The knowledge of the duty to warn has prompted some clinicians to define these responsibilities to the patient at the outset of therapy. Not only can this lead to a situation where potentially treatable conditions are left unaddressed due to the silence of the patients about their symptoms, but the warning about the limitations of confidentiality with new patients can subtly or not so subtly influence which issues are dealt with and may cause clinicians to skew their practice toward seeing “safer” patients. A cynical view may be that clinicians have done this to “game” the onerous burden of the duty to warn. How then do we get sex offenders and other individuals who have violent impulses to come for treatment and to be open about their strivings and symptoms?

Tarasoff and subsequent developments have reinforced our role as agents for social control (26). This mantle was a difficult one to wear for many clinicians, perhaps even more so with the addition of the duty to protect and the widening “zone of danger.” The new standard for practice shifted the emphasis of psychiatry from healing patients to protecting society. The fact that the contentious parts of the duty to protect cases stem from those patients who are not acutely ill illustrates the dramatic shift in the role of the psychiatrist. Actions leading to a reduction in serious violent behaviours generally achieve widespread support and thus may be politically difficult to challenge on ethical grounds. Nonetheless, in this new role, physicians use their powers to control non-illness related behaviour (38).



The Impact on the Profession: Pertinent Canadian Cases
There have been Canadian cases which have provided some guidance since Tarasoff . In the case of Wenden v. Trikha (7), a voluntary patient left the hospital without the knowledge of hospital staff and collided with the plaintiff’s vehicle. Ms Wenden, who was severely injured in the motor vehicle accident, commenced a tort action against the hospital and the psychiatrist, claiming they owed her a duty of care and were in breach of that duty.

The trial court held that hospitals and doctors owe duty of care to the patients and staff in the hospital. Although the hospital and psychiatrist were not found liable, a duty might be owed to others as well if there is a relationship of sufficient proximity between the hospital or doctor and the injured party. Physicians’ liability hinges on the particular risk posed by a patient, the predictability of future dangerous behaviour, and the ability to identify the person or group exposed to the risk. Tarasoff was considered and found to be too narrow. The trial judge spoke of a duty to protect: “. . . it is only fair and reasonable that both a hospital and a psychiatrist who becomes aware that a patient presents a serious danger to the well-being of a third party or parties owe a duty to take reasonable care to protect such a person or persons if the requisite proximity of a relationship exists between them.”

In an analogous case (Pittman Estate v. Bain), Dr Bain was found negligent of his duty to protect the wife of his patient who, he learned, had contracted HIV (39). The implication here is in the duty owed to a third party, and the duty may include informing the potential victim.

The highest court in the land, the Supreme Court of Canada, recently attempted to enunciate clearly the legal position in a case referred to as Smith v. Jones (40). Mr Jones was charged with aggravated sexual assault of a prostitute. Dr Smith, a psychiatrist, was retained by defence counsel to assess the accused for purposes of preparing a defence to assist in sentencing. During the course of the assessment, Mr Jones described his fantasies and plans to kidnap, rape, and kill prostitutes. Mr Jones pleaded guilty, and the matter was put over for sentencing. Dr Smith informed the defence counsel that, in his opinion, the accused represented a threat to society. Defence counsel, in a subsequent communication, requested that Dr Smith maintain confidentiality under the solicitor–client privilege previously agreed upon. Dr Smith, however, commenced an action in civil court for a declaration entitling him to disclose the information.

The British Columbia Supreme Court ruled that Dr Smith is required to disclose. On appeal, the British Columbia Court of Appeal said Dr Smith is permitted to disclose. On further appeal to Supreme Court of Canada, the majority ruled that danger to public can overrule solicitor–client privilege—the highest privilege recognized by the courts. The Supreme Court did not make clear whether disclosure was mandatory or discretionary.

The court specifically noted that solicitor–client privilege is the highest privilege recognized, and if a public safety exception applies to this, it will apply to all types of privilege and confidentiality. The court noted that it was inappropriate for it to consider the exact steps an expert might take to prevent harm to the public. It went on to say that it might be appropriate to notify the potential victim, the police, or a crown prosecutor, depending on the specific circumstances. Tarasoff was considered but not applied in this case.

This is the clearest statement of a duty to warn by the highest court in the land. This case laid to rest any doubt that public safety outweighs the interest of doctor–patient confidentiality.

Bearing this in mind, the physician should take note of the algorithm described above (Figure 1). In practice, clinicians will have to err on the side of attempting to treat patients, hospitalize them, or as a last resort, inform any potential victims when confronted with a patient with major mental illness who represents a risk to others. Mossman, in a thoughtful and provocative article, points out that despite the increasing accuracy of risk-prediction instruments and indeed guided clinical judgment, in practice clinicians will be increasingly forced to err on the side of safety (41). This, of course, is somewhat of a paradox in these days of shrinking inpatient facilities and a trend toward community treatment. As noted above, we would stress good practice, clear documentation, and consultation.

There may be some practical effects of this case law upon the profession (40). Patients could be warned of the limits of confidentiality at the beginning of treatment and, as in any other dialogue, there should be constant reminders and discussions about this. This may be particularly relevant when dealing with, for instance, sex offenders or people with anger problems. A written contract with the patient, outlining the exceptions to confidentiality, may be helpful but will not replace a dialogue or explanation. This may prevent certain patients from revealing their true fantasies or intentions, thereby tying the hands of psychiatrists who may have been able to intervene in the various ways noted above. As the psychiatrist’s role becomes more widely known, this could indeed prevent people from even seeking psychiatric help.

To the forensic psychiatrist dealing with a defence counsel, the situation may be even more problematic. The defence counsel would have to feel comfortable that his client will not reveal a future threat to public safety to the psychiatrist before making a referral. This could have a significant effect on the practice of forensic psychiatry.
Certain programs, such as court diversion programs, may have to be reviewed, bearing in mind this duty to warn and protect. It is possible that the defence bar will reconsider acceptance of these programs, thereby depriving clients of a useful and expeditious conduit to treatment.

The concept of the need to take measures to protect third parties from violent acts of clinicians’ patients has become very complex in both the US and Canada (2). It is important to distinguish the clinician’s duty to protect as a “legal obligation to take reasonable measures to protect third parties from one’s patient’s foreseeable harmful conduct” (2) from the moral or ethical arguments to do or not to do this. The principle to confidentiality, which is in fact a fiduciary duty of the physician, for the most part stands in opposition of the legal obligation to protect third parties (19,25). It is important for physicians to understand these obligations and to take the time to understand the legislation, case law, statutes and guidelines impacting on these issues. Notwithstanding the moral, ethical or professional debate about the duty to protect, Canadian psychiatrists now have a clear articulation of this duty from the Supreme Court. A better understanding of the legal obligations and the concepts driving them may enable the physician or psychiatrist to discharge his or her duties in ways that minimize or prevent confidentiality breaches while still providing good clinical care. By stepping back from the legally derived imperatives, it is important to recognize that Tarasoff and the resultant legal developments have impacted the practice of psychiatry in a way that can fundamentally change the time-hallowed relationship between patient and physician.

1. McNiel DE, Binder MD, Fulton FM. Management of threats of violence under California’s Duty-to-Protect statute. Am J Psychiatry 1998;155:1097–1101.
2. Felthous, AR. The clinician’s duty to protect third parties. The Psychiatric Clinics of North America 1999;22:49–59.
3. Kaufman M. Post-Tarasoff legal developments and the mental health literature. Bull Menninger Clinic 1991;55:308–22.
4. Dickens BM. Legal issues in medical management of violent and threatening patients. Can J Psychiatry 1986;31:772–80.
5. Tarasoff v. Regents of the University of California, 188 Cal. Rptr. 129, 529 P2d 533, 1974.
6. Barefoot v. Estelle, 51 LW 5190,1983.
7. Wenden v. Trikha, Royal Alexandra Hospital and Yaltho, 14 CCLT (2d) 225 (Alta. C.A.), 1993.
8. Smith v. Jones, 169 D.L.R. (4th) 385 (S.C.C.), 1999.
9. Beck JC. The psychotherapist’s duty to protect third parties from harm. Mental and Physical Disability Law Reporter 1987; 11:141–7.
10. Appelbaum PS, Gutheil TG. In: Clinical Handbook of Psychiatry and the Law (2nd ed). Baltimore: Williams & Wilkins; 1991.
11. Truscott D, Crook KH. Tarasoff in the Canadian Context: Wenden and the Duty to Protect. Can J Psychiatry 1993;38:84–9.
12. American Psychiatric Association. “Principles of Medical Ethics with Annotation Especially Applicable to Psychiatry” (1993 ed), Section 4.
13. Ferris LE, Barkun H, Carlisle J, Hoffman B, Katz C, Silverman M. Defining the physician’s duty to warn: consensus statement of Ontario’s Medical Expert Panel on Duty to Inform. Can Med Assoc J 1998; 158:1473–9.
14. Tarasoff v. Regents of the University of California, 17 Cal.3d 425,551 P2d 334,1976.
15. Runck B. Survey Shows Therapists Misunderstand Tarasoff Rule. Hospital and Community Psychiatry May 1984;35(5): 429–30.
16. Bloom H, Bay M. A Practical Guide to Mental Health Capacity and Consent Law of Ontario. Toronto: Thomson Canada; 1996.
17. Appelbaum PS. Tarasoff and the clinician: problems in fulfilling the duty to protect. Am J Psychiatry 1985;142:425–29.
18. Roth LH, Meisel A. Dangerousness, confidentiality, and the duty to warn. Am J Psychiatry 1977;134:508–11.
19. Glancy GD, Regehr C, Bryant AG. Confidentiality in crisis: Part I—The duty to inform. Can J Psychiatry 1998;43:1001–5.
20. Huber MG, Balon R, Labbate LA, Brandt-Youtz S, Hammer JH, Mufti R. A survey of police officers’ experience with Tarasoff warnings in two states. Psychiatr Serv 2000;.51:807–9.
21. Mills MJ, Sullivan G, Eth S. Protecting third parties: a decade after Tarasoff. Am J Psychiatry 1987;144:68–74.
22. Goldman DL, Jacob T. Anatomy of a second generation Tarasoff case. Can. J. Psychiatry 1991;36:35–8.
23. Zonana H. Warning third parties at risk of AIDS: APA’s policy is a reasonable approach. Hospital and Community Psychiatry February 1989;40 (2):162–4.
24. Perry S. Warning third parties at risk of AIDS: APA’s policy is a barrier to treatment. Hosp Community Psychiatry February 1989;40(2):158–61.
25. Glancy GD, Regehr C, Bryant AG. Confidentiality in crisis: Part II—Confidentiality of treatment records. Can J Psychiatry 1998;43:1006–11.
26. Gurevitz H. Tarasoff: Protective privilege versus public peril. Am J Psychiatry 1977;134:289–92.
27. Leong GB, Eth S, Silva JA. The psychotherapist as witness for the prosecution: the criminalization of Tarasoff. Am J Psychiatry 1991;149:1011–5.
28. Edelstein, Ludwig. “The Hippocratic Oath: test, translation & interpretation. Bull Hist Med (Suppl 1); Chapter 16: 402–6. Baltimore: John Hopkins Press; 1943.
29. Shane F. Confidentiality and dangerousness in the doctor–patient relationship: the position of the Canadian Psychiatric Association. Can J Psychiatry 1988;30:293–6.
30. Macbeth JE, Wheeler AN, Sither JW, Onek JN. Legal and risk management issues in the practice of psychiatry. Washington: The Psychiatric Purchasing Group Inc (PPG); 1994.
31. Mills MJ. The so-called duty to warn: the psychotherapeutic duty to protect third parties from patients’ violent acts. Behav Sci Law 1984;2:237–57.
32. Sharpe G. The law and medicine in Canada. Toronto: Butterworths; 1987.
33. McNeil D, Binder R. Clinical assessment of the risk of violence among psychiatric inpatients. Am J Psychiatry 1991;148:1317–21.
34. Webster C, Douglas K, Eaves D, Hart S. The HCR-20: Assessing risk for violence, Version 2. Burnaby, BC, Canada: Mental Health, Law, and Policy Institute, Simon Fraser University; 1997.
35. Harris G, Rice M, Quinsey V. Violent recidivism of mentally disordered offenders: the development of a statistical prediction instrument. Criminal Justice and Behavior 1993;20:315–35.
36. Steadman JH, Silver E, Monahan J, and others: A classification tree approach to the development of actuarial violence risk assessment tools. Law Hum Behav 2000;24:83–100.
37. Monahan J, Steadman HJ, Appelbaum PS, and others. Developing a clinically useful actuarial tool for assessing violence risk. Br J Psychiatry 2000;176:312–9.
38. Steadman HJ. Dangerousness to risk assessment of community violence: taking stock at the turn of the century. Journal of the American Academy of Psychiatry and the Law 2000;28:265–71.
39. Pittman Estate v. Bain (1994) 112 DLR (4th) 258 Ontario Central Division.
40. O’Shaughnessy RJ. Smith v. Jones: Privileged communication versus the public interest. Canadian Psychiatric Association Bulletin 1999;31:80–1.
41. Mossman D. Commentary: Assessing the risk of violence – are “accurate” predictions useful? Journal of the American Academy of Psychiatry Law 2000;28:272–81.