Virtual Mentor. October 2003, Volume 5, Number 10.
In Defense of Absolute Confidentiality
Defense of the position that patient-psychiatrist confidentiality should be absolute, otherwise patient trust is eroded is serious ways that affect the ability of all physicians to help their patients.
Kenneth Kipnis, PhD
An Examination of Confidentiality in Psychiatry after Tarasoff
In 1968 two students at the University of California at Berkley, Tatiana Tarasoff and Prosenjit Poddar, met and began dating. Poddar believed the relationship to be more serious than Tarasoff did and became preoccupied and withdrawn when she rejected him. In the summer of 1969 Tarasoff left the country to do field work. Poddar went to the university health service for treatment of his depression.
The psychiatrist at the health service prescribed Poddar a mild anti-psychotic and sent him to a psychologist for outpatient therapy. In these therapy sessions Poddar described fantasies of hurting an unnamed girl. The psychologist also found out from a third person that Poddar had been considering buying a gun and became concerned about Poddar's potential for violence. After consulting with his supervising psychiatrist and the psychiatrist who had initially evaluated Poddar, the psychologist called and wrote the campus police asking them to apprehend Poddar. When the campus police went to Poddar's apartment they found him to be, in their judgment, rational. The police warned Poddar to stay away from Tarasoff but did not take him into custody.
Two months later Poddar went to Tatiana Tarasoff's home. Tarasoff's mother told him Tatiana was not home and asked him to leave. Poddar returned later with a pellet gun and a butcher knife, found Tatiana home alone, and killed her.
Tatiana Tarasoff's parents filed a suit against the university campus police and the health services, arguing that Poddar should have been apprehended and their daughter should have been warned about his threats.
In its final ruling on the case in 1976 the California Supreme Court found that therapists have a duty to protect their patients' potential victims. Various state courts have struggled with how to define the duty a physician may owe patients' potential victims. In general the physician's duty to protect has extended to clearly foreseeable victims of clearly foreseeable threats .
In this issue devoted to ethical issues in the specialty of psychiatry we have solicited 2 commentaries that examine confidentiality in the practice of psychiatry after Tarasoff. Commentary 1.
by Kenneth Kipnis, PhD
The conventional wisdom on the ethics of medical confidentiality has been largely shaped by the Tarasoff case [1,2]. In 1969 Prosenjit Poddar, a student at Berkeley, told a university psychologist he would kill Tatiana Tarasoff who had spurned his affections. Reported to the campus police and held briefly, he was released and then did as he had said. The Tarasoff family sued the University of California for Tatiana's death and finally prevailed in its allegation that the university had failed in its duty to protect. Today it is hard to find discussions of the ethics of confidentiality that do not appeal to this parable and, occasionally, to the California Supreme Court moral: "The protective privilege ends where the public peril begins." Taking its cue from Tarasoff, the prevailing standard in medical ethics now holds that the obligation of confidentiality will give way when a doctor is aware that a patient will seriously injure some identified other person. This note is intended as both a challenge to this conventional wisdom and a preliminary defense of a medical confidentiality that does not contemplate such exceptions.
First one must highlight an error that infects much of the writing on this topic. The mistake is to move from a premise that some action is legally required to a conclusion that it is ethically required. The unhappy truth is that ethical obligations can conflict with legal ones. Journalists, for example, are sometimes ordered by the courts to reveal the identities of confidential sources. Although law demands disclosure, professional ethics requires silence. Reporters go to jail rather than betray sources.
Similarly in pediatrics, statutes may require doctors to report suspicions of abuse. Where there are protective agencies that are inept or overworked and foster care that is dangerous or unavailable, these reports are more likely to result in both the termination of therapy and further injury to the child than in either protection or proper care. To obey the law is most likely to abandon and even harm the child, both of which are ethically prohibited in medicine. To assume that legal obligations always trump or settle ethical ones is to blind oneself to the possibility of conflict. Professions have to face these dilemmas head-on instead of masking them with language that conflates legal standards and ethical ones. They must conceive professional ethics as largely separate from the law's mandate. And when law requires what professional responsibility prohibits (or prohibits what professional responsibility requires), professional organizations must press the public, legislatures, and courts to cease demanding that conscientious practitioners dishonor the duties of their craft.
Although laws cannot create ethical obligations by fiat, professions need to distinguish between the state's reasonable interests in the work of doctors—eg, preventing serious harm to children—and the specific legal mandates a state imposes—eg, requiring doctors to report a suspicion of child abuse to a state agency. Just as patients can make ill-considered demands that should not be satisfied so, too, can the state.
It is assumed in what follows that the state has a legitimate interest in preventing harm to people, and that doctors have an ethical obligation to further that public objective. The focus in this short essay is on the shape of that obligation, as it applies narrowly to adult patients, like Prosenjit Poddar, who present for treatment under their own steam. We set aside cases involving (1) children brought in by parents, (2) patients referred for independent medical evaluation, (3) patients in the custody of health care institutions, (4) health care that is the subject of litigation, (5) gunshot and knife wounds and the like, and (6) workers' compensation cases. A longer discussion could cover these.
Here we consider only the patient who somehow volunteers evidence that is ample to support a professional judgment that he or she is going to inflict death or serious injury on an identified other person. We suppose (1) that the evidence emerges under circumstances where it would ordinarily be covered by a professional obligation of confidentiality and (2) that a report would mobilize social mechanisms that will protect the person at risk. But for confidentiality, so the thinking goes, the injury would be prevented.
The most persuasive argument for breaching confidentiality may be as follows. The state's interest in preventing harm is weighty. Medicine has an obligation to protect the well-being of the community. Because the seriousness of the threatened injury outweighs the damage done to the patient by breaching confidentiality, the obligation of confidentiality must give way to the duty to prevent harm to others. Accordingly reporting is obligatory when it averts bad outcomes in this way. Of course clinicians should try to obtain waivers of confidentiality before reporting so avoiding the need to breach. But failure to obtain a waiver does not, on this argument, affect the overriding obligation to report.
As powerful as this justification is, there are problems with it.
First, if the profession accepts that its broad assurance of confidentiality must sometimes be breached, then any such assurances are fraudulent, and the profession should stop making them. If there are exceptions, clinicians have a duty to be forthcoming about what they are and how they work. Patients should know up front when they can trust doctors and when they can't.
Accordingly, the argument for breaching confidentiality has to be modified to support a qualified confidentiality rule, one that acknowledges a duty to report under defined circumstances. (In contrast, an unqualified rule contemplates no exceptions.) Instead of making promises, and then breaking them, doctors must qualify their promises so they won't have to break them. Commentators who have walked through the issues surrounding confidentiality have long understood the ethical necessity of "Miranda warnings" [3,4]. If doctors are ethically obligated to report, they need to say in advance what will be passed along, to whom, and what could happen then. They should never encourage or accept trust only to betray patients afterwards.
But now a second problem emerges. If prospective patients must know in advance that a doctor will report evidence that they will injure others, they will only be willing to disclose such evidence if they are willing to accept that others will know. If it is important to them that evidence not be reported, they will have a weighty reason not to disclose it to those who will do so. There are 2 groups of prospective patients: the first is willing to have reports made to others; the second is deterred from disclosure by fear of a report.
Consider the first group. Under a no-exceptions confidentiality rule, if the patient is willing to have reports made to others, a doctor should be able to obtain a waiver of confidentiality. Once that occurs, the ethical dilemma disappears and an unqualified confidentiality rule will work just as well as qualified confidentiality. The at-risk party will be protected just the same but with appropriate permission from the patient. In these cases there is no need to trim back the obligation of confidentiality since patients are willing to waive it.
But now consider the second group: those who do not want credible threats reported. These prospective patients control the evidence doctors need to secure protection for parties at risk. If the patient cannot be drawn into a therapeutic alliance—a relationship of trust and confidence—then doctors (1) will not receive the evidence, and therefore (2) they will not be able to report it, and therefore (3) they will not be able to mobilize protection. Reporting rules do not protect at-risk parties in these cases. In contrast, a no-exceptions confidentiality rule has a better chance of getting the evidence on the table, at least to the extent that honest promises of confidentiality can make it so. To be sure, clinicians would have to set aside the 'Should I report?' conundrum and search for creative solutions instead. Perhaps patients can agree to protective measures that will only be implemented under conditions they accept. Perhaps there are pharmaceutical, counseling, or monitoring interventions that can help manage antisocial intent. Perhaps patients will give up weapons or consent to referral or commitment. Patients may be persuaded to comply in order to protect themselves rather than those at risk. To be sure, these strategies will not always work to prevent harm, but they will sometimes. The nub of the matter is that they can never work if they can't be implemented. And they can't be implemented if the fear of reporting deters patients from disclosure. In these cases there is no justification for trimming back the obligation of confidentiality since that reduces protection to parties at risk, increasing public peril.
The argument here is that, paradoxically, ethical and legal duties to report make it less likely that endangered parties will be protected. Depending on the prospective patient, these duties are either unnecessary (when waivers can be obtained) or counterproductive (when disclosure is deterred and nonreporting interventions are prevented). While doctors should accept an overriding obligation to prevent public peril, they do not honor that obligation by breaching or chipping away at confidentiality. The protective purpose to be furthered by reporting is defeated by the practice of reporting. The best public protection is achieved where doctors do their best work and, there, trust is probably the most important prerequisite. Physicians damage both their professional capabilities and their communities when they compromise their trustworthiness.
What may trouble doctors is a fear that they will learn about an endangered person and be barred by a no-exceptions confidentiality rule from doing anything. (Actually there is only one thing they cannot do: report. All other paths are open.) Even if a reporting rule keeps many prospective patients out of the office, or silences them if they are in the office, it protects doctors from the moral risk of having to allow injury to third parties when a simple report would prevent it. This distress is significant and has to be faced.
Here we must highlight a second error infecting much of the literature on this issue: the conflation of personal morality and professional ethics. Like law, personal morality can also conflict with professional responsibility. A surgeon who is a Jehovah's Witness may be morally prohibited from administering blood transfusions to patients needing them. A Roman Catholic doctor may be unable to suggest medically indicated reproduction-related services. And despite understandable moral misgivings, doctors everywhere must be prepared to administer treatments they know will end the lives of some patients. While personal morality should play a decisive role in career choice—one who is morally opposed to capital punishment should pass up work as an executioner—it shouldn't play a decisive role within medical ethics.
Many enter medicine believing that good citizens prevent serious injury to others, even if that means breaking promises. But the task of professional ethics in medicine is to set out principles that, if broadly followed, will allow the profession to discharge its collective responsibilities to patients and society. Confidentiality, I have argued, gets more patients into treatment, brings about better outcomes for more of them, and best prevents harm to third parties. Ethically, it is praiseworthy for honorable people to belong to a profession that, on balance, reduces the amount of harm done to others, even though those professionals must sometimes knowingly allow harm to occur. Although doctors may then feel guilty for knowingly allowing harm to occur, they are not guilty of anything. They are acting exactly as it is reasonable to want doctors to act.
It is hard enough to create therapeutic alliances that meet patients' needs. But if doctors take on the added duty to mobilize protective responses to their patients without waivers of confidentiality, their work may become impossible in too many important cases. And all of us will be the worse for that. The thinking that places the moral comfort of clinicians above the well-being of antisocial patients and their erstwhile victims is in conflict with the requirements of professional responsibility, properly understood. While it will be a challenge for many doctors to measure up to this standard, no one ever said it would be easy to be a good physician.
I am indebted to Leanne Logan, Roy Perrett, Ron Bontekoe, James Tiles, Faith Lagay, Vrinda Dalmiya, and Dudley M. Stewart, MD, for suggestions that have improved the quality of this paper.
Kenneth Kipnis, PhD, is a professor in the department of philosophy at the University of Hawaii at Manoa.
The viewpoints expressed on this site are those of the authors and do not necessarily reflect the views and policies of the AMA.
© 2003 American Medical Association. All Rights Reserved.