APA PRESIDENTIAL ADDRESS
Protecting Victims of Violent Patients While
Protecting Confidentiality
Donald N. Bersoff
Drexel University
In this article I revisit Tarasoff v. Regents of the University
of California (1976), a case decided by the California
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Supreme Court in 1976 and followed in whole or in part in
This document is copyrighted by the American Psychological Association or one of its allied publishers.
approximately 33 U.S. jurisdictions. In Tarasoff, perhaps
the most notorious case in mental health law, the court held
that psychotherapists have a duty to protect intended vic-
tims of their patients’ conduct if that conduct presents a
serious danger of violence. The most prominent means of
protection is to warn potential victims and law enforcement
personnel. I argue that Tarasoff is bad law, bad social
science, and bad social policy. Because of therapists’ eth-
ical obligation to disclose the limits of confidentiality at the
beginning of treatment, informing patients of the thera-
pist’s duty to disclose threats may inhibit patients’ expres-
sions of violent urges. Alternatively, as in Tarasoff, once a
threat of violence is disclosed, the patient may never return
to therapy, making it impossible to work through the threat
of violence. I argue that rather than mandate the breaching
of confidentiality, it is sounder public policy to give ther-
apists the discretion to disclose so that they may take other
measures to protect potential victims. Finally, I briefly
touch on the impact of cultural disparity between therapist
and patient in these cases.
Keywords: confidentiality, ethics, violence prediction
W
ith the proliferation of mass shootings in the
recent past, the ongoing push for legislation
to control access to firearms, and the alleged
connection between mental illness and violence, it may be
particularly timely to revisit Tarasoff v. Regents of the
University of California (1976). Tarasoff is perhaps the
most notorious case in mental health law. For those who
may not have had the opportunity to read the case or hear
about it, the California Supreme Court in 1976 held that
psychotherapists have a duty to protect intended victims of
their patients’ conduct if that conduct “presents a serious
danger of violence to another” (p. 340). The court indicated
that the duty could be carried out in a number of ways but
suggested that the therapist could “warn the intended vic-
tim or others likely to apprise the victim of the danger, to
notify the police, or to take whatever other steps are rea-
sonably necessary under the circumstances” (p. 340). As a
result of a statute passed by the California legislature
subsequent to the decision, the duty to protect is more
circumscribed. According to the statute, “the duty shall be
discharged by the psychotherapist making reasonable ef-
July–August 2014 ● American Psychologist
© 2014 American Psychological Association 0003-066X/14/$12.00
Vol. 69, No. 5, 461– 467 http://dx.doi.org/10.1037/a0037198
forts to communicate the threat to the victim or victims and
to a law enforcement agency” (California Civil Code Sec-
tion 43.92). I argue in this article that Tarasoff is bad law,
bad social science, and bad social policy (Bersoff, 2002;
Slobogin, Rai, & Reisner, 2009). I also explore alternatives
to breaching confidentiality in order to warn, and I briefly
touch on the variable of culture in this particular case. To
begin, it may be helpful to flesh out the facts of Tarasoff,
as they are rather sparsely stated in the actual case (see
Everstine et al., 1980; Herbert, 2002). As you study the
following description of the case, I invite you, the reader, to
see if you can discern what I believe to be the most
important fact in the case.
Prosenjit Poddar was born and raised in the untouch-
able caste in rural India. In the fall of 1968 he came to the
University of California as a graduate student. The follow-
ing year, he met Tatiana Tarasoff at a dancing class. They
saw each other weekly for several months. On New Year’s
Eve, she kissed him. Poddar interpreted the kiss as signi-
fying a serious relationship. That was not her intent. When
she learned about Poddar’s belief, she disabused him of
that fact, saying she was seeing other men. Nevertheless, in
March 1969 he asked her to marry him, an offer which she
rejected. These rejections led Poddar to become quite un-
happy. He began to secretly audiotape their conversations,
trying to understand why she did not love him. He would
stay in his room in an apartment he shared with a room-
mate, obsessively listening to the tapes. He began telling
the roommate and other people about his thoughts of killing
her. His mental condition continued to deteriorate. So in the
summer of 1969 while Ms. Tarasoff was in Brazil for two
months on a vacation, at the suggestion of his roommate, he
voluntarily sought outpatient treatment at the university’s
counseling center. His initial interview was with a psychi-
atrist to whom he related his thoughts of killing Ms. Tara-
soff, although he did not specifically name her. The psy-
chiatrist prescribed antipsychotic and sleep medication and
Editor’s note. Donald N. Bersoff was president of the American Psy-
chological Association (APA) in 2013. This article is based on his pres-
idential address, delivered in Honolulu, Hawai’i, at APA’s 121st Annual
Convention on August 4, 2013.
Author’s note. Correspondence concerning this article should be ad-
dressed to Donald N. Bersoff, Earle Mack School of Law, Room 271,
Drexel University, 3320 Market Street, Philadelphia, PA 19104. E-mail:
dnb24@drexel.edu
461

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Donald N.
Bersoff
referred Poddar to a psychologist, Lawrence Moore, for
therapy, during which Poddar kept disclosing his intent to
kill the unnamed Ms. Tarasoff. Poddar kept his appoint-
ments for eight weeks, but in August 1969, Dr. Moore told
his patient that if he continued to talk about killing this
unidentified but readily identifiable woman he would have
him hospitalized. Faced with this ultimatum and the pos-
sible disclosure of his threats, Poddar never came back to
therapy.
Dr. Moore consulted with his psychiatrist supervisor,
who then wrote a letter to the campus police informing
them of their patient’s death threats. The campus police
found Poddar in his apartment, which he shared with his
new roommate who, ironically, was Ms. Tarasoff’s brother.
The campus police interviewed Poddar in front of the
brother. Poddar admitted he had a problematic relationship
with a woman but denied threatening to kill her. Poddar
reassured the police that he would stay away from the
woman, and they released him. The director of psychiatry
then asked the police to return Dr. Moore’s letter and
ordered that all copies of the letter and any therapy notes be
destroyed. Poddar was never hospitalized. Despite his
promise, Poddar continued to stalk Ms. Tarasoff. On Oc-
tober 27, 1969, Poddar went to Ms. Tarasoff’s home, but
she was not there. He returned later with a pellet gun and
a kitchen knife and found her alone. When she refused to
speak to him and ran from the house, Poddar caught up to
her and stabbed her to death. He then returned to her home
and called the police. Poddar was tried in criminal court.
He attempted to plead diminished capacity manslaughter,
but the trial judge did not give that instruction. He was tried
for first-degree murder but was convicted of second-degree
murder. He served five years in prison while a new lawyer
successfully appealed his conviction. In 1974, the state
462
supreme court vacated the conviction, but the state decided
not to retry Poddar. It was agreed that he would immedi-
ately return to India and never return to the United States.
He did so and later married a lawyer.
But it is not the criminal case that made this case
notorious. It is the civil case. Ms. Tarasoff’s parents
claimed that the defendant-therapists should have warned
their daughter about Poddar’s threats of violence. The
defendants responded that the plaintiffs failed to file a
cognizable cause of action; therapists, they said, have a
duty to their patients but not to private third parties. That
argument convinced the trial court, but the plaintiffs ap-
pealed and the California Supreme Court took the case to
review that dismissal. In a 1976 decision, which modified
one issued in 1974, the court held that the plaintiffs could
amend their complaint to state a cause of action against the
defendant mental health professionals for failing to protect
Ms. Tarasoff when their patient presented a serious danger
of violence to her.
By one recent count, 33 states have adopted the re-
quirements of Tarasoff in their own jurisdictions; 11 have
made the duty to warn discretionary (including the District
of Columbia), and 7 have not entertained the issue. To
illustrate the impact of Tarasoff, there are three of its
progeny I would like to describe.
First is Peck v. Counseling Service of Addison County
(1985). There, the Vermont Supreme Court extended the
duty to warn to a master’s-level counselor whose adoles-
cent patient threatened to burn down his father’s barn,
creating the possibility that the Tarasoff duty may extend to
protecting property, or at least preventing damage to prop-
erty that may also create a risk to people. The court asserted
that the father could have been in the barn.
The second case is Ewing v. Goldstein (2004). In that
case, the patient told his father during a family dinner that
he was very troubled that his former girlfriend was seeing
another man and that he was contemplating causing him
harm. Soon after, the father called his son’s therapist and
informed him of the threat. The therapist recommended
admission to a hospital. After one day in the hospital, the
son was discharged. On the following day the son killed the
boyfriend and then committed suicide. The victim’s parents
sued the therapist and the hospital in two separate suits.
The therapist asserted that the patient never made a threat
to him that would trigger a required warning or other
protective measures. Nevertheless, the California interme-
diate court held, “When the communication of the serious
threat of physical violence is received by the therapist from
a patient’s immediate family and is shared for the purpose
of facilitating and furthering the patient’s treatment, the
fact that the family member is not technically a ‘patient’ is
not crucial to the [relevant] statute’s purpose” (Ewing v.
Goldstein, 2004, p. 817). The holding begs the question of
how broadly “family member” should be defined. In fact, in
a footnote, the court opined, “We are not faced with and do
not address the situation in which a third party who is not
a member of the patient’s immediate family, but who may
be involved in his therapy in some manner (e.g., an inti-
July–August 2014 ● American Psychologist
 mate or close friend), conveys the information of the pa-
tient’s potential dangerousness to the therapist” (p. 819).
The third and final case also broadens the potential for
liability. It involves another disgruntled boyfriend, which
seems to be the paradigm factual setting for most of these
cases. In Hedlund v. Superior Court of Orange County
(1983), a therapy patient told the mental health profession-
als involved that he intended to physically harm his ex-
girlfriend. She was also a patient and the mother of a
6-year-old child. The therapists failed to warn the mother
of her danger, and she was seriously wounded by a shotgun
blast inflicted by the boyfriend while she was riding in a car
with her son. When she was shot, she threw herself over the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
boy, saving his life and preventing him from suffering a
This document is copyrighted by the American Psychological Association or one of its allied publishers.
serious injury. As a result of the attack, however, he be-
came psychologically traumatized. The California Supreme
Court held not only that the mother could sue for profes-
sional negligence for the therapists’ failure to warn her of
their dangerous patient but also that a foreseeable, innocent
bystander who witnesses the violence inflicted by a patient
may sue the patient’s therapist for negligent infliction of
emotional distress. Think of the possible exposure if the
potential victim is a schoolteacher, scout leader, or camp
counselor!
One of the major problems with Tarasoff and similar
cases is that the impetus to protect the intended victim is
the communication of a serious threat of violence to an-
other. If we parse “serious threat of violence,” there are two
determinations to be made in assessing the risk of violence
and whether to breach confidentiality. The first determina-
tion pertains to the seriousness of the threat. I interpret that
to mean how probable the threat is. No court that I know of
has defined how sure the therapist must be to protect the
potential victim from an actual threat. A recent statute
passed by the New York legislature requires disclosure if
the patient is “likely to engage in conduct that would result
in serious harm to . . . others.” As I interpret that likelihood
standard, it would require only slightly more than a 50%
possibility that the patient is dangerous. This in itself will
lead to a great many false positives and unwarranted
breaches of confidentiality. The second determination re-
quires the therapist to warn if the patient is going to be
violent. But what constitutes violence is not clear. If the
patient said, “I just took the lug nuts off my wife’s car and
in five minutes she’s going to drive down the interstate—
and he shows you the lug nuts—it is clear that this threat of
homicide obviously constitutes a serious threat of violence.
What about the following situations, however?
● The patient says he is going to go home tonight and
stab his wife in the arm. Asked if he is going to kill
her, he responds, “No, I just want her to bleed a bit.”
● The patient says he is going to go home tonight and
punch his wife in the jaw.
● The patient says he is going to go home tonight and
just cut off the tip of her pinky (on the nondominant
hand).
● The patient says he is going to go home tonight and
slap his wife in the face.
July–August 2014 ● American Psychologist
● The patient says he is going to go home tonight and
scream at his wife until she cries.
When I present these hypothetical situations to stu-
dents or practitioners, I get varied responses, illustrating a
definitional problem in fulfilling the mandate of Tarasoff.
There is one case that may help define the issue, however.
In Ewing v. Goldstein (2004), a California intermediate
court indicated that the duty to warn is triggered not only
when the patient threatens to take a life but also when the
patient poses a risk of grave bodily injury to another. That
phrase, it said, is akin to mayhem or serious bodily injury
as defined by the California criminal code. Mayhem is
defined as unlawfully depriving a human being of a mem-
ber of his or her body, or disabling, disfiguring, or render-
ing it useless, or cutting or disabling the tongue, or putting
out an eye, or slitting the nose, ear, or lip. Serious bodily
injury is defined as a serious impairment of physical con-
dition, including loss of consciousness, concussion, bone
fracture, protracted loss or impairment of function of any
bodily member or organ, a wound requiring extensive
suturing, and serious disfigurement (see Ewing v. Gold-
stein, 2004, p. 821). In that light, the situation in which the
patient actually threatens to stab his wife in the arm would
qualify because it would require suturing. Cutting off the
tip of her pinky would also qualify because it would de-
prive her of a member of her body. Less clear is the punch
in the jaw. If it is probable that she would suffer a broken
jaw, a concussion, or would lose consciousness, it would
qualify for a warning; if none of these possibilities occur, it
is a closer question whether a warning would be required.
Apparently, a slap in the face does not qualify as grave
bodily injury. And, finally, the screaming, although possi-
bly causing emotional trauma, does not result in a physical
injury and is thus outside the purview of Tarasoff.
I mentioned at the outset that Tarasoff and its progeny
are grounded in bad law, bad social science, and bad social
policy. I would now like to make those arguments. First, as
to bad law, in Tarasoff, the California Supreme Court
began its legal analysis by referring to the general principle
that people owe no duty to control the conduct of others,
but it “carved out an exception to this rule in cases in which
the defendant stands in some special relationship to either
the person whose conduct needs to be controlled or in a
relationship to the foreseeable victim of that conduct”
(Tarasoff v. Regents of the University of California, 1976,
p. 343). In support of that proposition it referred to Sections
315–320 of the Restatement (Second) of Torts (1965). The
Restatement is a synthesis of basic principles of tort law
drafted by law professors, practitioners, and judges. The
court relied primarily on Section 315, which states:
There is no duty so to control the conduct of a third person as to
prevent him from causing physical harm to another unless
(a) a special relation exists between the actor [therapist] and the
third person [patient] which imposes a duty upon the actor to
control the third person’s conduct, or
(b) a special relation exists between the actor and the other
[potential victim] which gives the other a right to protection.
463
 There is no doubt that a special relation exists between the
therapist and the patient as stated in exception (a). But does
that relationship impose a duty to control the patient, the
third person? The commentary to Section 315 indicates that
the answer will be found in Sections 316 –319. Section 316
refers to the duty of a parent to control his or her child;
Section 317 refers to the duty of an employer to control the
conduct of his or her employee; and Section 318 refers to
the duty of a property owner to control the conduct of
someone who uses that property. None of these examples
are relevant to the Tarasoff duty. That leaves Section 319,
which states:
One [therapist/institution] who takes charge of the third person
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
[patient] whom he knows or should know to be likely to cause
This document is copyrighted by the American Psychological Association or one of its allied publishers.
bodily harm to others [potential victims] if not controlled is under
a duty to exercise reasonable care to control the third person to
prevent him from doing such harm.
On its face, that principle could arguably pertain to the
Tarasoff paradigm. But the commentary states that this rule
applies to two situations. The first deals with the duty to
prevent a patient with a contagious disease from being
released from a hospital and is thus irrelevant. The second
illustrates the rule with the following example: “A operates
a private sanitarium for the insane. Through the negligence
of the guards employed by A, B, a homicidal maniac, is
permitted to escape. B attacks and causes harm to C. A is
subject to liability to C.” Putting aside the archaic lan-
guage, this example clearly does not apply to the Tarasoff
case. Poddar was in outpatient therapy, not under the con-
trol of a hospital. And, although the case did not indicate
any diagnosis for Poddar, he did not appear to be seriously
mentally ill, that is, psychotic or bipolar, and thus was not
eligible for civil commitment. The other legal justification
for imposing the duty to warn that the court in Tarasoff
cited is a case called Merchants National Bank & Trust Co.
of Fargo v. United States (1967). That case is inapposite as
well because it also involved an inpatient. There, the estate
of a deceased wife sued for wrongful death, claiming that
the hospital negligently released her violent husband from
a VA hospital and made no attempt to monitor him, their
patient. In sum, then, in my view the court in Tarasoff
offered no valid legal support for imposing the duty to warn
on therapists who are treating nonhospitalizable outpa-
tients.
As for bad social science, the Tarasoff defendants
argued that imposing the duty to protect “is unworkable
because therapists cannot accurately predict whether or not
a patient will resort to violence” (Tarasoff v. Regents of the
University of California, 1976, p. 344). What was true in
1976 is largely still true. The court acknowledged that
therapists would have difficulty in forecasting whether their
patients would present a serious danger of violence. They
expected that the predicting therapist would only need to
exercise a reasonable degree of skill, knowledge, and care
ordinarily possessed by similar professionals under similar
circumstances— basically the malpractice standard. Unfor-
tunately, the standard in the profession is that it is ex-
tremely difficult to accurately predict which outpatients
who express violent thoughts or fantasies will actually
engage in violence. As the concurring judge asked, “Per-
haps the psychiatrist is an expert at deciding whether a
464
person is mentally ill, but is he an expert at predicting
which of the persons so diagnosed are dangerous?” (Tara-
soff v. Regents of the University of California, 1976, p.
354). The sophisticated research done by the MacArthur
Project’s Research Network on Mental Health and the Law
found that the only population for which accurate assess-
ments of risk can be made are acutely psychotic, usually
paranoid, active substance abusers who recently committed
an act of violence (see Monahan et al., 2001). That descrip-
tion fits neither Poddar nor most of the outpatients thera-
pists treat.
As for bad social policy, I asked the reader earlier to
think about what the most important fact in the Tarasoff
case is. To my mind, the crucial fact was not an uttered
threat of future violence but the fact that once Poddar knew
that his confidences would be disclosed to others or that he
would be involuntarily committed, he never returned to
therapy. One wonders if Ms. Tarasoff would be alive today
if the psychologist-therapist and his psychiatrist-supervisor
had not been so quick to call the police but rather had
worked with their patient for the two months between the
threat and the killing. The dissent, with whom I agree,
argued that the holding in Tarasoff would deter those
requiring treatment from seeking therapy because they
would have no assurance of confidentiality:
Given the importance of confidentiality to the practice of psychi-
atry [sic], it becomes clear the duty to warn imposed by the
majority will cripple the use and effectiveness of psychiatry.
Many people, potentially violent—yet susceptible to treatment—
will be deterred from seeking it; those seeking it will be inhibited
from making revelations necessary to effective treatment; and
forcing the psychiatrist to violate the patient’s trust will destroy
the interpersonal relationship by which treatment is effected.
(Tarasoff v. Regents of the University of California, 1976, p. 360)
In sum, the dissent asserted that “the majority’s duty to
warn will not only impair treatment of many who would
never become violent but worse, will result in a net in-
crease in violence” (Tarasoff v. Regents of the University of
California, 1976, p. 361). That, of course, is an empirical
question that cannot be ethically studied. Fairness dictates
that I note that other commentators have suggested that a
good client– clinician relationship is based on trust rather
than on confidentiality. Thus, they argue that if trust is
maintained, and the therapeutic alliance is strong, the ther-
apeutic relationship may endure even after a mandated
breach in confidentiality (see Truscott, Evans, & Mansell,
1995). Nevertheless, it takes many sessions to build a
relationship of trust and to develop a strong therapeutic
alliance. It does not help the clinician faced with the
utterance of violent threats early in treatment.
In any event, clinicians in 33 jurisdictions are com-
pelled to follow Tarasoff and in seven others to decide
whether to take action. Therefore, the clinician’s obligation
to disclose threats to potential victims imposes two corol-
lary duties, one ethical and the other clinical. The first duty
is to engage in informed consent. All professional associ-
ations that have as their members mental health practitio-
ners require in their codes of ethics that they inform po-
tential clients of certain essential pieces of information at
the onset of treatment. For example, the American Psycho-
July–August 2014 ● American Psychologist
 logical Association’s (2010) Ethical Principles of Psychol-
ogists and Code of Conduct states: “When obtaining in-
formed consent to therapy . . . psychologists inform clients/
patients as early as is feasible in the therapeutic relationship
about the nature and anticipated course of therapy, fees,
involvement of third parties, and limits of confidentiality”
(Standard 10.01, italics added). This ethical mandate, one
can hypothesize, will have the unintended consequence of
causing just what the dissent in Tarasoff argued would
happen—it will inhibit the patient from revealing violent
fantasies, urges, and threats for fear that the therapist will
breach confidentiality and warn a putative potential victim
and inform the police.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
The second duty is to perform a competent assessment
This document is copyrighted by the American Psychological Association or one of its allied publishers.
of the risk of violence. Almost all courts impose liability
only if the therapist is actually told by the patient that
violence will occur. The clinician, however, must discrim-
inate between violent fantasies and urges and genuine
threats. This requires knowledge of basic concepts of risk
assessment, such as base rates and protective versus risk
factors and the extant risk assessment research (Monahan,
1993). Most importantly, risk assessment involves gather-
ing risk-relevant data from multiple sources. As Monahan
(1993) noted, “Most of the Tarasoff-like cases on which I
have worked have faulted clinicians not for making an
inaccurate prediction but for failing to gather information
that would have made a reasonable effort at prediction
possible” (p. 243). He, and others (see Conroy & Murrie,
2007; Heilbrun, 2009), suggest that clinicians gather infor-
mation from records of past treatment, interview significant
others, review current records, and interpret the results of
psychological testing, particularly specialized risk assess-
ment measures and preferably an actuarial measure or a
structured professional judgment measure (Tolman &
Rotzien, 2007). Of course, this presumes that the clinician
has the time to perform this comprehensive assessment of
risk. A clinician in the lug nut situation does not have the
luxury of time.
My antagonism to Tarasoff does not mean that I
advocate letting potentially violent patients go unchecked.
Truly violent aggressors at some time lose their right to
absolute protection when they threaten to use deadly force.
Immunity statutes that protect clinicians from liability if
they try to warn potential victims and inform the authorities
simply make it too easy for us to betray our fidelity to our
patients and to become society’s police force. As one
psychology intern expressed it to me, “If faced with the
decision to protect confidentiality or warn who I think is a
potential victim of violence, I’m going to breach confiden-
tiality.” In contrast, I think we should be obliged to attempt
all other viable options before we abrogate the principle of
fidelity and unilaterally disclose confidential communica-
tions to private third parties.
I would agree with the Supreme Court’s view on this
issue in Jaffee v. Redmond (1996), the case that created a
psychotherapist–patient privilege in the federal rules of
evidence. The Court created the privilege because it be-
lieved that protecting the confidential communications be-
tween therapists and their patients served both private and
July–August 2014 ● American Psychologist
public interests. “Effective psychotherapy,” they said, “de-
pends upon an atmosphere of confidence and trust in which
the patient is willing to make a frank and complete disclo-
sure of facts, emotions, memories, and fears. . . . [T]he
mere possibility of disclosure may impede development of
the confidential relationship necessary for successful treat-
ment” (Jaffee v. Redmond, 1996, p. 10). Failing to enact the
privilege, the Court opined, would chill confidential com-
munications between therapists and their patients. Most
relevant to our discussion is the Court’s view as expressed
in a footnote: “[W]e do not doubt that there are situations
in which the [psychotherapist–patient] privilege must give
way . . . if a serious threat of harm to the patient or to others
can be averted only by means of disclosure by the thera-
pist” (Jaffee v. Redmond, 1996, p. 18). To me, the most
important word in that sentence is “only.” It implies that
breaching confidentiality should be a last resort.
So, in the last section of this presentation I want to
explore what the clinician can do to protect a potential
victim without breaching confidentiality. I would like to
frame the discussion by describing what the English phi-
losopher W. D. Ross (1930) called prima facie duties or
moral values. These duties or values are nonmaleficence,
beneficence, justice, fidelity, and autonomy. Perhaps the
bedrock ethical duty required of all health care profession-
als is nonmaleficence—above all, do no harm. At the least,
when therapy ends, our clients should not be worse off as
a result of our interventions than when they began therapy.
Beneficence is the most commonsense principle; indeed,
the concept of psychology, psychiatry, and social work as
“helping professions” is its embodiment. It refers to prac-
titioners’ responsibility to benefit those they treat. The
danger regarding this principle is that it can devolve into
paternalism; we act in what we think is the best interest of
our patients even if it is contrary to what the patients think
is in their best interest. The starkest example in mental
health is, of course, involuntary civil commitment. Justice
refers to the recognition that fairness entitles all persons
access to and benefit from the contributions of mental
health practitioners and to be protected from unjust prac-
tices. Fidelity refers to the obligations of faithfulness and
loyalty inherent in the client– clinician relationship. It is
this principle that is most endangered by Tarasoff and its
progeny. As the California Supreme Court alliteratively
stated, “The protective privilege ends where the public
peril begins” (Tarasoff v. Regents of the University of
California, 1976, p. 347). At some point, the court says, our
duty to the client is superseded by our duty to a private
third party. The final moral principle is autonomy or self-
determination, guaranteeing people the freedom to think,
choose, and act so long as their actions do not infringe
unduly on the rights of others. A humorous example of this
principle is embodied in a New Yorker cartoon of many
years ago in which a teenager confronting her parents says,
“I just want to ruin my life in my own way.”
In an effort to do no harm, to act with fidelity toward
our patients while doing good and supporting self-determi-
nation, what are some actions we as clinicians can take
short of breaching confidentiality? Assuming that the ther-
465
 apist determines that the patient is a high risk for violence,
there are a number of possibilities, particularly if the threat
is not one like the lug nut situation. In general, it is
desirable to seek consultation from respected peers or su-
pervisors with the goal of developing viable options for
preventing violence. Consultation also has the advantage of
helping to protect the clinician from the allegation that the
risk management plan was below the standard of care and
thus negligent (see Monahan, 1993). One clear option is to
recommend hospitalization. Voluntary admission is highly
preferable to civil commitment, as the latter would neces-
sarily involve breaching confidentiality. If, as in the hypo-
theticals I posed earlier, the potential violence will be
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
carried out in several hours, the clinician could cancel later
This document is copyrighted by the American Psychological Association or one of its allied publishers.
appointments with other patients and intensify treatment.
“I’m going to work with you for the next several hours until
I know you will not act on your violent threat” is one
suggestion. If there is more time, the clinician can refer the
patient for appropriate medication. Finally, even in the lug
nut situation, where the threat is imminent, the clinician can
seek to act beneficently without breaching confidentiality.
The therapist can say, “Listen buddy, you have two minutes
to get on the phone and tell your wife not to drive the car.
I’ll be glad to talk to her after that if you wish. But if you
don’t call her right away, I’ll have to do it.” This message
promotes both fidelity to the client and self-determination.
The client is given the opportunity to take responsibility for
his actions before the clinician unilaterally is compelled to
do so.
There are a myriad of interventions that clinicians can
make instead of breaching confidentiality. But in states that
make disclosure mandatory, to avoid litigation clinicians
will take the safe route. There are times when the law must
exercise control over professional functioning. But it is my
belief that the law’s regulation in the area of violence
prediction and containment is misplaced. The 11 states that
grant discretion to mental health professionals to warn
allow them to act with more nuance and professionalism.
Permitting discretion aligns therapists’ duty to warn with
that of attorneys. Under Section 1.6 of the American Bar
Association’s (2012) Model Rules of Professional Conduct,
lawyers have the discretion to reveal confidences uttered by
their clients if they reasonably believe it necessary to
prevent death or substantial bodily harm. It is difficult for
me to ascertain why mental health professionals should be
treated differently by the law than are lawyers.
There is one final point. As I noted in the facts of
Tarasoff, Poddar came from the untouchable caste in India.
From an American perspective, his stalking of and obses-
sion with Ms. Tarasoff was pathological. But I wonder if
there would have been a different outcome and a different
tack taken if Poddar’s therapist also came from the Indian
culture. We are fast becoming a plurality nation. By 2040,
non-Hispanic Whites will be in the minority, but we are not
making significant enough efforts to admit and retain train-
ees from the Middle East, the Asian subcontinent, Africa,
and Central and South America. There will be a significant
disconnect in the near future between those we serve and
466
those doing the serving. The failure to correct this problem
may lead to future Tarasoff-like tragedies.
Tarasoff and the cases that adopt its holding present
the therapist with complex problems requiring nuanced
decisions. The therapist must tread a thin line between
protecting confidentiality and protecting the potential vic-
tim. An error on either side of this line can lead to liabil-
ity— either for malpractice in unreasonably breaching con-
fidentiality or for wrongful death for failure to warn in the
face of a real threat. But beyond the private interests at
stake, these cases raise a more fundamental issue, the
dilution of privacy. Like the ozone layer of the atmosphere,
there has been a gradual but constant erosion of this fun-
damental value, particularly after the destruction of the
World Trade Center. As the late columnist William Safire
stated, “After 9/11, the passion went out of advocacy of
privacy. The right to be let alone had to be balanced against
the right to stay alive” (Safire, 2004, p. A27). The late Max
Siegel, a past president of the American Psychological
Association, made the following comments after Tarasoff
was decided:
This was a day in court for the law and not for the mental health
professions. If the psychologist had accepted the view of absolute,
inviolate confidentiality, he might have been able to have kept
Poddar in treatment, saved the life of Tatiana Tarasoff, and
avoided what was to become the Tarasoff decision. (Siegel, 1979,
p. 253)
It is a bit too late to think the law will support the principle
of absolute inviolate confidentiality. But it is not too late to
contemplate the consequences of making mental health
professionals agents of the state, torn between their duty
and ethical obligations to clients and their responsibilities
to the larger society in which they live.
REFERENCES
American Bar Association. (2012). Model rules of professional conduct.
Chicago, IL.: Author.
American Psychological Association. (2010). Ethical principles of psy-
chologists and code of conduct (2002, Amended June 1, 2010). Re-
trieved from http://www.apa.org/ethics/code/index.aspx
Bersoff, D. N. (2002). Some contrarian concerns about law, psychology,
and public policy. Law and Human Behavior, 26, 565–574. doi:
10.1023/A:1020260123477
California Civil Code Section 43.92.
Conroy, M. A., & Murrie, D. C. (2007). Forensic assessment of violence
risk: A guide for risk assessment and risk management. Hoboken, NJ.:
Wiley. doi:10.1002/9781118269671
Everstine, L., Everstine, D. S., Heymann, G. M., True, R. H., Frey, D. H.,
Johnson, H. G., & Seiden, R. H. (1980). Privacy and confidentiality in
psychotherapy. American Psychologist, 35, 828 – 840. doi:10.1037/
0003-066X.35.9.828
Ewing v. Goldstein, 120 Cal. App. 4th 807 (2004).
Hedlund v. Superior Court of Orange County, 669 P.2d 41(1983).
Heilbrun, K. (2009). Evaluation for risk violence in adults. New York,
NY: Oxford University Press.
Herbert, P. B. (2002). The duty to warn: A reconsideration and critique.
Journal of the American Academy of Psychiatry and Law, 30, 417– 424.
Jaffee v. Redmond, 518 U.S. 1 (1996).
Merchants National Bank & Trust Co. of Fargo v. United States, 272 F.
Supp. 409 (D.N.D. 1967).
Monahan, J. (1993). Limiting therapist exposure to Tarasoff liability:
Guidelines for risk containment. American Psychologist, 48, 242–250.
doi:10.1037/0003-066X.48.3.242
July–August 2014 ● American Psychologist
 Monahan, J., Steadman, H. J., Silver, E., Appelbaum, P. S., Robbins,
P. C., Mulvey, E. P., . . . Banks, S. (2001). Rethinking risk assessment:
The MacArthur Study of Mental Disorder and Violence. New York,
NY: Oxford University Press.
Peck v. Counseling Service of Addison County, 499 A. 2d 422 (Vt., 1985).
Restatement (Second) of Torts, Sections 315–319 (1965).
Ross, W. D. (1930). The right and the good. Oxford, England: Clarendon.
Safire, W. (2004, March 10). Privacy in retreat. The New York Times, p. A27.
Siegel, M. (1979). Privacy, ethics, and confidentiality. Professional Psy-
chology, 10, 249 –258. doi:10.1037/0735-7028.10.2.249
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
July–August 2014 ● American Psychologist
Slobogin, C., Rai, A., & Reisner, R. (2009). Law and the mental health
system: Civil and criminal aspects (5th ed.). St. Paul, MN: Thomson/West.
Tarasoff v. Regents of the University of California, 551 P.2d 334 (1976).
Tolman, A. O., & Rotzien, A. L. (2007). Conducting risk evaluations for
future violence: Ethical practice is possible. Professional Psychology:
Research and Practice, 38, 71–79. doi:10.1037/0735-7028.38.1.71
Truscott, D., Evans, J., & Mansell, S. (1995). Outpatient psychotherapy
with dangerous clients: A model for clinical decision making. Profes-
sional Psychology: Research and Practice, 26, 484 – 490. doi:10.1037/
0735-7028.26.5.484
467