Advance directives have been one of the more promising innovations in recent years to give patients a greater voice in their psychiatric treatment (1). Completed when patients are competent, advance directives allow patients to appoint proxy decision makers and to make choices about particular treatments, all to take effect should patients later become incompetent to make decisions for themselves. Advance directives have been hailed as a way of encouraging patients and treaters to discuss future contingencies and to negotiate mutually acceptable approaches to care (2,3). All states have statutes that govern the use of advance directives, which can be applied to general medical and psychiatric care, and many states now have special provisions for advance directives for psychiatric care per se.

However, mental health professionals have always been concerned that advance directives could also be used in a less collaborative way. One of the earliest proponents of advance directives, Thomas Szasz—a fierce critic of psychiatric diagnosis and treatment—suggested that people with mental disorders use advance directives to preclude future treatment, especially treatment with medications (4). As Szasz saw it, if advance directives represented the unalterable choices of competent patients, there would be no way to override the preferences embodied in the directives. This suggestion raised the prospect of a class of patients who would be permanently untreatable, even if they later became psychotic and were hospitalized involuntarily. Now, in the wake of a decision by the U.S. Court of Appeals for the Second Circuit, that prospect seems closer to materializing.

The case, Hargrave v. Vermont, grew out of a complaint filed in 1999 on behalf of Nancy Hargrave, a woman with a history of paranoid schizophrenia and multiple admissions to the Vermont State Hospital (5). Hargrave had completed an advance directive—known in Vermont as a "durable power of attorney for health care," or DPOA—in which she designated a substitute decision maker in case she lost competence and in which she refused "any and all anti-psychotic, neuroleptic, psychotropic, or psychoactive medications." The major national law firm that represented Hargrave immediately filed suit to block the state of Vermont from overriding her advance directive should she ever again be involuntarily committed and obtained certification to represent the entire class of patients in similar situations.

Hargrave's target was Act 114, a 1998 Vermont statute that attempted to address the dilemma inherent in psychiatric advance directives. Although advance directives were intended to facilitate patients' participation in treatment decisions, they have, as noted, the potential to prevent all treatment, even of patients who are ill enough to qualify for civil commitment under the prevailing dangerousness standards. To mitigate this prospect, the Vermont legislature allowed hospital (or prison) staff to petition a court for permission to treat an incompetent involuntarily committed patient, notwithstanding an advance directive to the contrary. Before the court could authorize nonconsensual administration of medication, it had to allow the terms of the patient's advance directive to be implemented for 45 days. So a patient like Hargrave, who had declined all medications, would be permitted to go unmedicated for a 45-day period, after which the court could supercede the patient's refusal of treatment.

The core of Hargrave's challenge to the statute was based on Title II of the Americans With Disabilities Act (ADA), which requires that "no qualified individual with a disability shall, by reason of such disability, be excluded from participation in or be denied the benefits of the services, programs, or activities of a public entity, or be subjected to discrimination by any such entity" (6). Hargrave claimed that she and other members of her class were being discriminated against on the basis of mental disorder, given that only committed persons with mental illness could have their advance directives overridden under Act 114. And the public "services, programs, or activities" from which she was being excluded was the state's durable power of attorney for health care itself.

In response, the state of Vermont offered three arguments. First, because Hargrave had been involuntarily committed, Vermont claimed that she qualified under an exclusion to the ADA for persons who pose a "direct threat." Next, the state contended that the plaintiff was not being discriminated against on the basis of disability, because anyone who completed an advance directive was susceptible to having his or her choices superceded (the state has an alternative override mechanism that involves judicial appointment of a guardian), and in any event, it was the status of being civilly committed, not being mentally ill, that was the point of distinction here. Finally, Vermont looked to a federal regulatory provision that allows a public entity to continue existing practices, despite an ADA challenge, if the change being called for would "fundamentally alter the nature of the service, program, or activity" (7).

The Second Circuit, like the U.S. District Court that had originally heard the case, failed to find any of these contentions persuasive. With regard to the claim that Hargrave and other involuntarily committed patients constitute a direct threat, the three-judge panel noted that not all committed patients would be a threat to others, as required under the ADA, because many were hospitalized for danger to self. Even persons who were found to be dangerous to others at the time of commitment, the court held, could not be presumed still to be dangerous when override of their advance directives was sought. The court was similarly unpersuaded that some condition other than mental illness was the basis for the differential treatment, given that Act 114 applied only to persons with mental illness. And allowing advance directives to stand as written, the court decided, even when patients were committed, does not fundamentally alter the advance directive statute (although it might affect the provision of psychiatric treatment to involuntary patients), which the court held was the proper point of reference. Hence the court concluded that Act 114 violated the ADA and enjoined its enforcement.

Hargrave, then, stands for the proposition that the state, having established a statutory basis for medical advance directives, cannot exclude involuntarily committed psychiatric patients from its coverage. Although the Second Circuit's opinion applies directly only to Vermont and New York, it is an influential court, and its opinion may well be echoed in other circuits around the country. Advance directives may now constitute an ironclad bulwark against future involuntary treatment with medication—except in emergencies—even for incompetent, committed patients and even when the alternative is long-term institutional care.

In many respects, Hargrave represents a continuation of the battle over the right of psychiatric patients to refuse treatment that began in the 1970s. Indeed, the list of amici who filed briefs in support of Hargrave reflected the coalitions that were formed to push for a right to refuse treatment 30 years ago. But that battle ended ambiguously. Although some states were compelled by the courts to permit even committed patients to refuse medication unless they were found incompetent by a judge, other states still allow the treating physician—sometimes after a second opinion has been obtained—or a panel of clinicians to override refusal on clinical grounds (8). Even in states that require findings of incompetence and substituted judgment as to whether the patient, if competent, would have accepted the treatment, the vast majority (typically more than 90 percent) of cases that are adjudicated end with the court authorizing involuntary treatment with medication. The sense of many experienced observers is that when patients are psychotic and treatment seems clearly indicated, the courts find a way to justify administration of medication, sometimes despite the legal criteria (8).

If adopted more widely, however, Hargrave would appear to provide a tool whereby patients who are determined to avoid treatment with medications would be able (except in emergencies) to completely preclude such treatment. A reviewing court would be bound to honor the terms of the now-incompetent patient's advance directive and order that treatment be withheld. Judges or quasi-judicial decision makers would no longer have the discretion to apply "common-sense" criteria—for example, that patients with flagrant psychosis should be treated if possible—to mandate medication. Today, few severely ill committed patients avoid treatment with medications, regardless of the legal standard in their jurisdiction. Hargrave could change that. If large numbers of patients were to complete advance directives such as Nancy Hargrave's, declining all medication, hospitals might well begin to fill with patients whom they could neither treat nor discharge.

Are there legal mechanisms that could avoid this outcome without running afoul of the ADA? In the Hargrave case, the court itself noted that nothing in this decision precludes statutory revisions that do not single out persons who are disabled because of mental illness—for example, revisions that increase the competency threshold for executing a DPOA or that allow the override of the DPOA of any incompetent person whenever compliance with the DPOA would substantially burden the interests of the state. However, it is doubtful that raising the competence threshold would have much impact, and the court's suggestion regarding "interests of the state" that might warrant overriding any person's advance directive is, frankly, enigmatic.

But perhaps a clever legislator can find an opening here to blunt the impact of the decision. And there is no guarantee that other circuits, or even ultimately the U.S. Supreme Court, would necessarily agree with the Second Circuit's analysis. Of course, were the level of concern sufficient, it would always be possible for Congress to amend the ADA to exclude the class of persons at issue. Congress, though, is typically reluctant to tinker with major legislation, and the disability rights community would likely oppose firmly any amendment of the ADA.

Because the ultimate scope and impact of Hargrave may not be known until a decade from now, it is worthwhile to consider the possible effect of the decision on the use of advance directives for psychiatric treatment. Current research suggests that most patients who complete advance directives do not use these directives to decline all treatment with medication but rather to indicate preferences among alternative treatments or to inform future treaters of particular concerns—for example, the care of their pets while they are hospitalized. Although Hargrave may stoke some enthusiasm for advance directives among patients who are opposed to receiving any medication, it remains to be seen how common the phenomenon will become. Studies now under way will tell us more about the utility of advance directives in psychiatry—for example, whether, given the current state of the mental health system, advance directives actually have an impact on subsequent care (9). At a minimum, however, it seems likely that Hargrave, as it becomes more widely known, will chill enthusiasm for psychiatric advance directives among many clinicians. Because clinicians' suggestions that patients consider completing advance directives probably play an important role in encouraging the completion of such directives (10), Hargrave's legacy may be to inhibit the use of this once-promising tool.

Dr. Appelbaum, who is editor of this column, is A. F. Zeleznik distinguished professor and chair in the department of psychiatry at the University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655 (e-mail, appelbap@ummhc.org).