Oh, the problems a footnote can cause—especially when it appears in a decision of the U.S. Supreme Court. In its landmark 1996 opinion in Jaffee v. Redmond, the Court recognized a psychotherapist-patient testimonial privilege for the federal courts: therapists can no longer be compelled to testify in federal proceedings regarding their patients' disclosures in treatment (1). Buried in the text, however, was a footnote that appeared to suggest that the privilege would not apply in cases where there was a "serious threat of harm." The controversy over the "dangerous patient exception" to the Jaffee privilege was born at that moment, and the federal courts have been trying to sort out the issue ever since. This column describes those efforts.

The decision in Jaffee was precipitated by a plaintiff in a civil case who tried to gain access to the records of the defendant's treating social worker (2). At that point, the status of a psychotherapist-patient privilege in the federal courts was unclear. However, when Congress adopted the Federal Rules of Evidence in 1975, it had given the courts the power to recognize testimonial privileges in light of the courts' "reason and experience" (3). Applying that authority, the Supreme Court produced a ringing endorsement of the value of psychotherapy and the importance of confidentiality.

"Effective psychotherapy," the Court said in its opinion, "…depends upon an atmosphere of confidence and trust in which the patient is willing to make a frank and complete disclosure of facts, emotions, memories, and fears." Consequently, the Court concluded, "The psychotherapist privilege serves the public interest by facilitating the provision of appropriate treatment for individuals suffering the effects of a mental or emotional problem. The mental health of our citizenry, no less than its physical health, is a public good of transcendent importance." The court ruled that so long as the communications in question met the requirements for the privilege, they would be protected from disclosure, no matter how important to the case the evidence might be.

Justice Stevens, who wrote the opinion for a 7—2 majority, thought that there was no reason to sketch the full dimensions of the privilege, leaving that for subsequent cases. However, he added in a footnote, "Although it would be premature to speculate about most future developments in the federal psychotherapist privilege, we do not doubt that there are situations in which the privilege must give way, for example, if a serious threat of harm to the patient or to others can be averted only by means of a disclosure by the therapist." The interpretation and weight to be given that footnote has been a subject of contention that has led to strikingly different opinions from the federal circuit courts.

One circuit court has held that the footnote means that the privilege is vitiated when the patient poses a serious threat that requires warning third parties (another circuit came to the same result, albeit on different grounds), but two sister circuits have strongly disagreed. The most powerful statement against a dangerous patient exception comes from the Ninth Circuit Court's opinion in U.S. v. Chase (4). Chase, who was seeing a psychiatrist for irritability, depression, and anger, voiced threats in therapy against a group of FBI agents, leading his psychiatrist to call the local police. He was arrested and charged under a federal statute, and testimony from his psychiatrist was admitted over his objections at trial. The Ninth Circuit, reviewing a decision of a three-judge panel of the court, held that the psychiatrist should not have been allowed to testify, although the error was harmless in this case and Chase's conviction was allowed to stand. In so ruling, the court found that the Jaffee footnote was referring to the duty of therapists to protect potential victims of their patients by issuing warnings—which was unaffected by the new privilege—rather than to limits on the protections afforded by privilege against compelled disclosure in court.

Having interpreted the footnote in Jaffee as not addressing an exception to the privilege, the court offered several arguments as to why a dangerous patient exception should not be adopted. To do so, said the court, would diminish the value of the psychotherapist-patient privileges that every state had created for its own state courts—almost always without a dangerous patient exception—because patients could never be certain that their therapists would not be compelled to disclose the contents of their sessions in federal court. In addition, from a policy perspective, the judges found that greater harm would be done by recognizing an exception and thus deterring potentially violent patients from seeking treatment than the good that would be gained by facilitating convictions in some cases. The Sixth Circuit had earlier reached the same conclusion on roughly similar grounds (5).

In contrast, the Tenth Circuit endorsed a dangerous patient exception, though it narrowly interpreted the circumstances under which it would apply. In U.S. v. Glass (6), the court was dealing with a man who had told his therapist, as reflected in the notes of the session, "he wanted to get in the history books like Hinkley [sic] and wanted to shoot Bill Clinton and Hilary [sic]." When the patient became noncompliant with his treatment plan, the facility notified the police, who in turn called the Secret Service. His therapist was allowed to testify. Glass was convicted, and he appealed. Looking to the precise wording of the Jaffee footnote, the Tenth Circuit ruled that the therapist's testimony would have to be excluded unless the terms of the footnote were met—that is, unless the trial court found that the threat he had uttered was serious and that disclosure of the threat at the time was necessary to avert harm to the President. If that were the case, the dangerous patient exception would apply and the therapist's testimony would be admissible.

Although the Tenth Circuit offered little analysis in support of the exception it recognized, simply relying on the wording of the footnote, a more recent case from the Fifth Circuit provided something of a supportive rationale. The court managed to resolve the case, U.S. v. Auster (7), without having to decide directly whether a dangerous patient exception exists. But the opinion suggested that when the patient was aware that a warning might be issued if a threat were made—and almost all states have recognized a duty to protect third parties (8)—the marginal benefit from protecting the patient's disclosures in court was small and was outweighed by the costs of making conviction more difficult. Although Auster's threats had been communicated previously to his potential targets, the court's rationale might apply equally to patients who had merely been told at the outset of therapy of the possibility of disclosures in the event that others were endangered. Both Sixth and Tenth Circuits had rejected such a rationale.

When divergence of this degree is manifest in the appellate courts, it is usually an indication that both sides of the argument have something to be said for them. Supporters of the dangerous patient exception have what would seem to be a strong legal argument. Although the Jaffee footnote was incidental to the decision (what lawyers refer to as "dicta"), and thus of uncertain precedential value, it appears to indicate that the inclinations of the U.S. Supreme Court are to create an exception to privilege in some cases. The arguments of opponents that the footnote was meant to endorse disclosures to law enforcement or potential victims, rather than testimony in court, or was limited to commitment hearings are weak. In its language, the footnote refers explicitly to the "privilege … giv[ing] way," which could only be applicable to the courtroom setting. Moreover, there is no federal commitment law; hence commitment is governed by state law and takes place in state courts, where federal evidentiary rules do not apply. Thus it is improbable at best that the Court was referring to that context (9).

In addition, supporters of the exception can make a good case for the importance of therapists' testimony in trials of dangerous persons. Although it has been argued that such testimony is unnecessary, because such persons can be civilly committed if not convicted, that is not always correct. Patients with personality disorders, in particular, may represent long-term threats to another person but not qualify for acute hospitalization. Even if dangerous patients are committed, they are likely to be released after a much shorter period than if they were incarcerated, and even if treatment is effective, the risk may well recur. So it is not hard to imagine situations in which a therapist's testimony in a criminal trial may be the most effective—and sometimes the only—way of protecting a potential victim.

On the other hand, courts that have refused to recognize a dangerous patient exception—although they may resort to verbal gymnastics to explain away the apparent meaning of the footnote—have powerful policy justifications to offer. As one commentator put the argument, "it is nonsensical to interpret a footnote in dictum to stand for the proposition that a 'dangerous patient exception' should be recognized when it goes against the very core rationale for the privilege" (10). Because all privileges involve tradeoffs that impede the fact-finding functions of the courts, the Supreme Court was well aware that some adjudications would be more difficult. However, it concluded that this loss was warranted by the gain to society's mental health from protecting the confidentiality of communications in therapy.

Even if the dangerous patient exception were narrowly construed, it would inhibit help seeking by the very people in whose treatment society has the strongest interest—those who have harmed or are likely to harm others. The claim that such people are already deterred from treatment by knowledge that the therapist may be required to disclose their threats to potential victims, though perhaps true in some cases, underestimates these patients' abilities to appreciate a difference between a disclosure aimed at preventing harm and one intended only to facilitate punishment. Moreover, a federal exception would have an impact on state privilege statutes that almost all avoid dangerous patient exceptions. Finally, opponents of the exception contend that it is unnecessary to protect public safety because the civil commitment process remains available.

Perhaps the most worrisome aspect of the dangerous patient exception is its vulnerability to being used expansively for purposes beyond those originally envisioned. California, which has had a statutory dangerous patient exception for many years, is the prime example of how the courts can exploit an exception once created. A series of California decisions has developed the principle that once a threat is uttered, the privilege is lost for all purposes. Hence, therapists' testimony regarding patient threats has been ruled admissible at death penalty hearings after guilt has been decided and to prove premeditation so as to facilitate conviction for first-degree murder, rather than a lesser level of homicide. Entire sessions have been stripped of the protections of privilege because threats were uttered therein, and clinicians have testified not just about threats per se but about the entire context of therapy that led them to view the risk as real (11,12). Indeed, in a number of cases patients have been charged with crimes exclusively on the basis of threats voiced to their therapists, though they had already been hospitalized and had never actually attempted to harm another person (13). None of these uses of the exception relate directly to its avowed purpose: to protect the potential victim from harm.

With the split among the federal circuits, it seems clear that the U.S. Supreme Court ultimately will need to speak to the question of whether a dangerous patient exception exists in the federal courts and, if so, what its dimensions may be. Construed in its narrowest form—as applying only when essential to protect potential victims (for example, when a threatening patient is not otherwise committable)—it may be justified by the interests of public safety. But unless that line can be held, it would be preferable not to limit the privilege with regard to dangerous patients at all, lest the exception overwhelm the general rule that patients' communications in therapy are deserving of protection.