Superseding Psychiatric Advance Directives: Ethical and Legal Considerations

The Legal Backdrop for Overriding PADs

Current state laws that authorize PADs give doctors wide discretion to ignore them. Specifically, in cases in which the patient's advance choice of treatment (or choice to forgo treatment) conflicts with the doctor's view of the standard of care, PAD laws do not require that doctors follow the patient's wishes; more to the point, most of these laws provide broad legal immunity to doctors who, in good faith and consistent with clinical standards, decline to follow, in whole or in part, a patient's advance treatment instructions as documented in a PAD. Clinicians are obligated to follow whatever portions of the PAD they can, even if they override some particular instructions. However, that clinicians are granted discretion to decide which PAD instructions are to be followed and which are not weakens the instrument to some degree. Weakening PADs even further, civil commitment law trumps a PAD in every U.S. jurisdiction.

The override features typical of PADs are clearly illustrated in the language of one of the newest PAD statutes to be enacted in the United States: Pennsylvania's mental health advance directive and powers of attorney law (Act 194), which became effective on January 30, 2005. Pennsylvania's PAD statute is modeled after, and incorporates, many of the features of other states’ statutes, most of which were enacted in the 1990s. The Pennsylvania law allows competent individuals to specify their wishes directly regarding mental health treatment before becoming incapacitated, and to appoint an agent to carry out their wishes during future periods of incapacity. The law also reinforces the requirements of the federal Patient Self-Determination Act of 1990 by obligating facilities and providers of mental health treatment to ask patients whether they have a PAD when they are admitted for treatment and to inform all patients about the availability of PADs as part of discharge planning. Treatment facilities are also required to place a copy of the mental health advance directive or power of attorney in the patient's mental health record.

However, Pennsylvania's Act 194 also contains three specific sections devoted to ensuring that physicians can override these directives with few (if any) consequences. First, the section on construction of the law explicitly states that the PAD statute shall not be construed to affect “the ability to admit a person to a mental health facility under the voluntary and involuntary commitment provisions of the Mental Health Procedures Act.”⁶ Second, the section on compliance contains a clause allowing a physician, who “cannot in good conscience comply” with the instructions of an agent appointed under a PAD because the instructions are “contrary to accepted clinical practice and medical standards,” to refuse to comply so long as he or she makes “every reasonable effort” to assist in the transfer of the patient to another provider who is willing to comply. If reasonable efforts to transfer the patient fail, the patient may be discharged, at least under the statute.⁷ Third, the section on liability states that a physician who acts in good faith may not be subject to criminal or civil liability or disciplined for unprofessional conduct as a result of refusing to comply with a PAD, the provisions of which the provider deems violate “accepted clinical standards or medical standards of care.”⁸

Thus, according to Pennsylvania's (quite typical) PAD statute, as long as mental health care providers act in good faith on the basis of their perception of the standard of care, and make reasonable efforts to transfer the patient, the law provides virtually complete protection for the provider from the consequences of overriding a PAD. In effect, whereas PADs are supposed to promote psychiatric patients’ autonomy and self-determination, these instruments may actually reinforce doctors’ professional autonomy.

The Problem with Advance Directives

Psychiatrists’ responses to PADs can be understood in the larger context of the development of medical advance directives, which have faced similar challenges and problems over the years. Like PADs, medical advance directives were seen as having great potential to help guide difficult medical decisions for incompetent persons, and were expected to provide the means for people to fulfill their wishes regarding their own health care after they could no longer speak for themselves. In practice, however, such directives—particularly “living wills” containing written treatment instructions—often failed to live up to expectations.

Beginning with California in 1976, all states enacted advance directive statutes of some sort, including either living wills (containing instructions about particular treatments and medical conditions) or durable powers of attorney (appointing a surrogate decisionmaker) or both.⁹ The federal Patient Self-Determination Act of 1990 (PSDA) was later enacted to promote the use of written advance directives after the U.S. Supreme Court, in Cruzan,¹⁰ permitted states to apply a strict evidentiary standard of “clear and convincing evidence” to oral statements by patients who had not completed written directives. The PSDA requires health care facilities receiving federal funds to inform patients of their rights under state law to prepare an advance directive, to inquire and document whether patients have executed a directive, to ensure compliance with state laws by respecting advance directives, and to educate health care providers regarding these legal instruments.¹¹ Despite the PSDA, research suggests that the prevalence of written medical advance directives in the general public remains no higher than 25 percent (possibly much lower in many locations) and did not substantially increase after passage of the federal law.^12–16

A number of studies have suggested reasons why written medical advance directives have had limited success and have not been widely adopted. First, even when patients have executed written advance directives, physicians often are not aware of them. Perhaps more important, when physicians are aware of advance directives, these documents often have limited or no effect on clinical decisions.^14,17–20

That clinicians are granted (by law or in practical reality) considerable leeway to override medical advance directives at their discretion has contributed to the belief among many observers that advance directives (especially living wills) “have no teeth”—and that they may never be very effective vehicles for enacting patients’ true preferences. There are limited empirical data regarding the practice of overriding medical advance directives, but clinical experience and anecdotal reports suggest that overrides occur with some frequency.^21–25

Overriding patients’ advance directives raises several ethics concerns. It would seem that physicians have an ethics (if not always a legal) obligation to try to honor advance directives, in that doing so conveys respect for patients as persons and enhances their well-being—core values at the foundation of patient self-determination.²⁶ In contrast, when a physician deliberately overrides a patient's competently executed advance directive, the physician may violate the principles of informed consent to treatment, including the patient's right to refuse treatment, and, in effect, usurp the power that the law has invested in advance directives, to ensure that persons receive treatment that they would choose and do not receive treatment that they would refuse, when incapacitated.

Medical Advance Directives and PADs

Notwithstanding the aforementioned problems, there are legitimate reasons why physicians might override some advance directives in good faith. Brock²¹ has outlined three general types of scenarios in which health care providers might consider overriding a medical advance directive: (1) when there are good reasons to doubt that the advance directive accurately reflects what the patient would have wanted; (2) when the moral authority of the advance directive is questionable due to conflict with important current interests of the patient and/or changes in the patient's personal identity; and (3) when the interests of persons other than the patient warrant overriding the directive. In the following discussion, we review possible implications of these scenarios for PADs and the question of whether PADs in practice will be (or perhaps should be) overridden in certain situations. We then offer some analysis of the possible future of legal cases where issues of overriding PADs and fiscal concerns may collide.

Considering Brock's first scenario from a clinical point of view,²¹ uncertainty about what the patient actually wants might offer a reasonable justification for overriding a PAD. Severe and persistent mental illness can impair persons’ baseline ability to make and communicate reliable decisions about their own health care. Hence, at least some persons with psychiatric disabilities may never be the best judges of what is in their own best interest, or even the best voice of their own authentic preferences. Carl Elliott has phrased the problem this way: People do not always mean what they say; they do not always say what they want; and they do not always want what they say they want. That much is, if not exactly clear, at least uncontroversial. What is controversial is, recognizing this, how to proceed. How are we to interpret statements made by a patient who is now, by virtue of his medical condition, unable to interpret them for us? [Ref. 27, p 61].

However, from the perspective of patients with a long history of psychiatric treatment, PADs may actually convey treatment preferences much more accurately than medical advance directives or living wills do, to the extent that such preferences are shaped by previous personal encounters with the health care interventions in question. Precisely through their accumulated personal experience with the negative and positive aspects of treatment, psychiatric patients may gain a more authentic appreciation of the personal value of avoiding or receiving particular types of treatment in the future. In contrast, medical patients who execute living wills may lack this relevant experience and appreciation of future treatment contingencies. That patients complete advance directives having no familiarity with the actual situations or decisions they will face in the future grants physicians, under some circumstances, an ethical warrant to override such directives as invalid expressions of a person's true preference.

For example, an experienced anesthesiologist might argue that a critically ill patient's advance instruction “never to put me on a ventilator if I am in a coma” does not reflect what the patient would actually want in a real situation confronted years later—particularly if the patient had never experienced being critically ill at the time the advance directive was executed, and if there is a reasonable probability of recovery. In contrast, a psychiatrist would be hard pressed to argue that a patient with schizophrenia who has been repeatedly hospitalized and treated with haloperidol lacks the personal experience to inform, in a PAD, refusal of this drug or refusal of hospitalization. Also, the range of treatment options and choices may be much more constrained in mental health care, compared with end-of-life medical care, making it easier to predict the contingencies one might actually face during a future mental health crisis. In short, highly specific PADs may be less susceptible to override than medical advance directives.

Of course, PAD instructions may also raise such doubts, in some situations, about whether a particular advance instruction accurately reflects what the person would have wanted. Consider, for example, a patient who refuses electroconvulsive therapy (ECT) in a PAD, but has never experienced ECT. Let us suppose the patient later experiences severe, life-threatening depression that is refractory to antidepressant medications. It is not difficult to imagine a physician second-guessing such a PAD by asking: “Would this person really have wanted to forego the best—and perhaps only—remaining effective treatment option?”

Brock's second criterion contains two components and addresses the complex issue of what is to be done, first, when the person has “suffered such profound cognitive changes that there are doubts about whether personal identity is maintained between the person who executed the advance directive and the present patient” (Ref. 21, p 55), and second, when the directive seems to be in conflict with important present interests of the person. On this score, PADs may be susceptible to override.

For some clinicians, the simple fact that PADs are typically completed by patients who have chronic disorders that impair thinking, judgment, insight, and basic perception of reality may call into question the validity of the advance instructional document. Cognitive impairment associated with major psychopathology may be long-lasting, and may even become a permanent feature of an altered personality. Such cases might cast doubt on the assumption of continuity of “identity” and agency between the patient's self as currently presented and the “prior” self as represented by the PAD document. For similar reasons, some ethicists have opposed PADs altogether as instruments of “self paternalism” or “Ulysses contracts” that inappropriately favor the documented preferences of a prior self over those of the present person, arguably a valid self and worthy of respect, even in a psychotic state.^28,29 Concerns about patient safety—suicidality, in particular—may also call into question the moral authority of a treatment-refusal PAD, due to potential conflict between the patient's autonomy interests and the patient's present safety and survival interests.

In the third scenario for overriding medical advance directives, the interests of others justify not honoring the directive. The example Brock²¹ offers for this situation is when the interests of a patient near death have substantially diminished and interests of the family are the driving force in medical decisionmaking—a phenomenon colloquially known as “treating the family.” The “other interests” may also involve the interests of physicians or society in general, rather than family members. For example, physicians may assert a prerogative to override a medical advance directive in cases where the doctor believes further life-sustaining treatment will be physiologically futile, yet the patient, via an advance directive, or the patient's family directly, is demanding that aggressive treatment be continued. One could argue that forcing a physician to prescribe aggressive treatment under such circumstances would violate his or her professional integrity and, more broadly, the integrity of the medical profession itself.³⁰ An additional argument along these lines might be made on the basis of distributive justice: that it is unfair to expend scarce medical resources on futile care for one patient merely because that patient requests such care in advance, when other patients who might benefit are going without needed treatment due to lack of resources.

Physicians’ legal defensiveness is another powerful factor representing “other party interests” that may influence PAD override decisions. In earlier studies,^31,32 we defined “legal defensiveness” broadly to refer to clinicians’ general level of concern about the implications of both civil and criminal law regarding their treatment decisions for seriously ill patients. Specifically, legal defensiveness in this context refers to the aggregate of clinicians’ attitudes and practices arising from, or attributed to, the perceived threat of legal sanction in response to their decisions to intervene (or not to intervene) in particular ways for incapacitated psychiatric patients.

In one of these studies, Swanson and McCrary³¹ examined the effects of physicians’ legally defensive attitudes on their responses to hypothetical end-of-life treatment scenarios in a survey of 301 physicians practicing in academic medical centers in Texas. We found that physicians with attitudes of extreme legal defensiveness were more likely to define what constituted futile versus beneficial treatment at an arbitrary threshold which, in effect, maximized the physician's latitude and prerogative to override patients’ (ostensibly reasonable) preferences for end-of-life treatment abatement. These findings suggest that some physicians (though a minority) tend to assume an adversarial position in their consideration of treatment decisions for critically ill patients—an attitude that anticipates, and thus may actually create, conflict with these patients or their surrogates.

Considering PADs along these lines, a legally defensive psychiatrist might be expected to override a patient's advance refusal of treatment primarily out of concern for being held legally liable for any adverse consequences that might follow from honoring the patient's request—such as the possibility that the patient might engage in violent behavior if left untreated.

More broadly, the “other party interests” scenario may arise for PADs when physicians feel professionally or socially obligated to provide treatment that the patient has refused in advance. For example, the potential conflict between the interests of patients and those of physicians and others was highlighted in a recent decision of the U.S. Court of Appeals for the 2nd Circuit (Hargrave v. Vermont),³³ which struck down a state law that allowed mental health professionals to override a person's advance refusal of psychotropic medications through a general health care proxy.^34–36 Specifically, the court ruled that the Vermont override law, which applied only to persons with psychiatric disorders, was discriminatory on the basis of disability and thus violated the Americans with Disabilities Act, Title 3. Opponents of the court's decision in Hargrave have argued, among other things, that priority should have been given to the larger interests of society in this matter—for example, the interests of taxpayers who might ultimately have to pay for longer hospital stays for psychiatric patients with PADs that refuse antipsychotic medications.

Brock's third type of override scenario²¹ has a special application in the case of patients with PADs who also meet criteria for involuntary civil commitment. As we have already suggested, in every U.S. jurisdiction with a PAD statute, PADs may be overridden by a civil commitment order for patients whose condition qualifies for such intervention either (1) under the doctrine of parens patriae or (2) through exercise of the state's “police powers” to protect public health and public safety (e.g., in the case of a patient with a violent history who is considered to pose a danger to others).³⁷

There is some irony in the fact that involuntary commitment legally trumps a PAD, because PADs have been promoted to stakeholders explicitly as a means to avoid or decrease the incidence of involuntary treatment. The potential for conflict between the interests of patients and those of others, including the safety interests of the general public, highlights a key weakness inherent in PADs. There is no clearly identified legal mechanism, other than involuntary commitment, to enforce a patient's own advance wishes for treatment if the patient later resists such treatment when incapacitated. Likewise, there is no legal mechanism that will enforce a patient's right to refuse all intervention at a time when the patient poses an imminent danger to self or others.

Brock's third criterion may also apply in cases of iatrogenic illness (i.e., when the patient's condition has been caused, or exacerbated, by the physician himself or herself). Research suggests, for example, that physicians are significantly more likely to override a patient's do-not-resuscitate (DNR) order when the cardiopulmonary arrest is due to a complication of treatment, especially when the complication arises from a physician's error.^38,39 Similarly, surgeons and anesthesiologists often demand that patients’ DNR orders be suspended in perioperative settings before they will agree to perform or facilitate palliative surgical procedures for terminally ill patients.^40–42 The reasoning provided in both of these examples includes: (1) fear of peer and professional condemnation for a patient death (especially where hospital policy views all deaths in surgical contexts as “unexpected”); (2) personal feelings of failure when physicians are more directly responsible for a patient's death; and (3) fear of litigation for negligence or even homicide.^41,42

Regarding PADs, analogous situations could arise if clinicians are faced with PAD refusals of treatment for patients whose mental health crises are ostensibly “caused,” or at least made worse, by previous episodes of inadequate treatment or preventable adverse side effects of treatment. For example, consider a PAD refusal of all antipsychotic medications (similar to the actual request of Nancy Hargrave, the plaintiff in Hargrave v. Vermont) by a currently psychotic patient who has been treated in the past only with old-line neuroleptics that have caused unpleasant extrapyramidal symptoms and tardive dyskinesia. A psychiatrist might consider overriding such a patient's PAD and prescribing a newer (and more expensive) pharmacotherapy regimen (e.g., olanzapine) which is known to be more tolerable and perhaps more effective in relieving psychotic symptoms. The psychiatrist's ethical and practical reasoning here might be: (1) mental health professionals—or, broadly, the mental health care system—are partly responsible for the patient's past adverse experiences with treatment; (2) these past experiences have contributed both to the present crisis and to the patient's distorted preference to forego all medication, a preference that may be considered misinformed because it includes refusal even of medications that the patient has never tried and which the psychiatrist believes would benefit the patient; and (3) therefore, under these circumstances, it would be irresponsible and perhaps negligent for the psychiatrist to honor the patient's PAD refusal. A similar situation may occur in the case of a PAD refusal of hospitalization, when the psychiatrist believes that the alternatives available for intervention and treatment in the community are poor, or that lack of access to adequate community-based care has contributed to the patient's current relapse.

Legal Challenges to Overriding PADs

All of these reasons for overriding PADs assume what the state statutes make explicit—that it is legally permissible for doctors to do so. But the question arises, how firm is the legal footing on which the override features of the current PAD statutes rest? The case of Hargrave v. Vermont remains today the latest word from the federal courts on the overriding of PADs. Yet, the fiscal and social concerns of some critics of this decision are not easily dismissed. The practical implication of the court's decision was to impose an obligation on the state of Vermont to provide inpatient custodial care to acutely ill patients like Nancy Hargrave on an indefinite basis, allowing their symptoms to remain untreated if they have documented in advance a (presumably competent) wish to forego medication.

Vermont did not seek a writ of certiorari from the U.S. Supreme Court to review the Second Circuit Court's decision in Hargrave. Presumably, state officials decided against appealing this decision because the number of patients affected would likely be very small, and the state was willing to assume financial responsibility in the event that extended inpatient care would be needed in such cases. However, other states may be less willing to face the prospect of financing essentially custodial care for severely ill but otherwise treatable psychiatric patients—particularly in the absence of a determination of dangerousness.

No other case like Hargrave appears to have been adjudicated in any other jurisdiction to date. However, with more and more states adopting PAD statutes, and new research showing a large latent demand for PADs among consumers in public mental health systems throughout the United States,⁵ it is perhaps not unlikely that another case will arise to challenge the states’ authority to override advance refusals of treatment. It is difficult to predict what the outcome would be. On the one hand, if a similar case were to be litigated in one of the 22 states with PAD statutes that currently allow physicians to override PADs (essentially at their discretion, on the basis of their perception of the standard of care), the precedent of Hargrave and the federal ADA might make it quite difficult for states to defend these special override provisions, in that they apply only to psychiatric patients’ directives. On the other hand, critics of the Hargrave decisions (and fiscal conservatives generally) might take a different tack. It could be argued that even if Nancy Hargrave had the right, while competent, to refuse treatment with medication during a future period of incapacity, she was not entitled to have the state pay for extended inpatient care so that she could remain indefinitely untreated with medication. It might follow from such an argument that funding for psychiatric hospital care could be denied to any patient who refuses medication in advance, provided the patient was not dangerous.

Some of the states’ PAD statutes already include language suggesting that PAD instructions do not entitle patients to receive services that are not “feasible” or would not otherwise be paid for (i.e., in the absence of a PAD). For example, a PAD request to be admitted to a private drug abuse rehabilitation program carries no obligation for the provider to offer this to the patient on any different terms than would have obtained without the PAD. One could imagine broadening these statutory “feasibility” provisions to include denying payment for any hospitalization without medication, on the grounds that such a hospitalization would be unfeasible, and would be longer and more expensive than would have been the case without the PAD refusal of medication. The potential shift of the burden of care for treatment-resistant psychiatric patients from public resources to family, private, or other resources might also be raised as an ethics and public policy issue.

What would be the legal implications of such a policy? Perhaps there are some clues in the claims asserted by the parties in Hargrave—particularly certain constitutional arguments that were raised, but overshadowed by the ADA issue on which the courts actually decided the case. Both the district court and the Second Circuit in Hargrave based their decisions solely on statutory grounds (i.e., that the federal ADA superseded the state statute because of its facially discriminatory application). However, in the district court, both parties also made constitutional arguments, with Vermont claiming that to permit a durable power of attorney for health care to trump the state's authority to treat committed persons would be contrary to the state's police and parens patriae powers, while the plaintiff argued that Vermont's actions were a violation of her procedural and substantive due process rights under the Fourteenth Amendment of the U.S. Constitution. These constitutional arguments were not adjudicated by either the District Court⁴³ or the Second Circuit.³³

But suppose that the Vermont state legislature (or any state legislature, in response to a similar case) were to pass a statute denying funds for inpatient care of nondangerous persons who competently refuse medication in advance of a mental health crisis. On the one side, such a statute might be construed in some cases to conflict with the state's parens patriae obligation to care for persons who cannot care for themselves. On the other side, it might raise, in a new context, some of the central questions that have animated the right-to-die debate. Do individuals have a fundamental right to refuse treatment? If so, does such a right supersede a state's parens patriae authority and, perhaps, even obligate the state to fund its exercise with public resources? Further claims of discrimination might be raised, too, by comparison with the lot of terminally ill patients who refuse life-sustaining intervention and are not thereby denied publicly funded custodial hospital care.

Such a statute probably would face a constitutional challenge, which would ultimately have to be decided in the federal courts. The U.S. Supreme Court has consistently held, except in cases where rights previously determined to be fundamental (such as the right to relocate from state to state) are penalized by denial of state benefits, that persons have no constitutional right to receive public financing of social services and, hence, that state discretion in limiting any such benefit or service is very broad.^44–48 Because the U.S. Supreme Court has never specifically held that the right of competent persons to refuse treatment is fundamental (despite limited dictum to that effect in Cruzan¹⁰) and has instead allowed states to impose evidentiary restrictions and other obstacles to patient autonomy in such cases, it is not at all clear that the current Court would impose on such state statutes the judicial standard of strict scrutiny reserved for fundamental rights. The state's arguments in such a case could take the following forms: (1) even if Nancy Hargrave has a right to refuse medication, she has no corresponding right requiring the state to fund her indefinite care in the absence of her willingness to take medication (at least on a trial basis); and (2) the state has a legitimate interest, perhaps even a compelling one, to take steps to prevent unnecessary harm to mentally ill persons by providing them with treatment proven to ameliorate their symptoms. Such arguments have been persuasive in similar cases on the basis that decisions allocating state monies do not penalize the exercise of a right; they simply fail to fund its exercise. It seems likely that if a state passed a statute like that described above, many courts would not overrule it on constitutional grounds. If circuit courts of appeal did reach varying conclusions, it might persuade the U.S. Supreme Court to grant certiorari.

Further, consider social and humanitarian concerns that could be raised in cases like that of Nancy Hargrave. Even if the patient prefers to remain unmedicated, what are the implications of a state policy that confines such persons with mental illness, possibly adding to their suffering? Is such a position wise social policy? Should the state provide an incentive for patients to accept treatment by forcing them to rely on private health insurance in such cases (if they have it)? These and similar issues have been raised eloquently by Dr. Paul Applebaum, who notes, “If large numbers of patients were to complete advance directives such as Nancy Hargrave's, declining all medications, hospitals might well begin to fill with patients whom they could neither treat nor discharge” (Ref. 35, p 752).

Also consider how such a case could present a direct conflict between a statute such as Pennsylvania's and the Hargrave decision, whereas under the Pennsylvania-type statute, a patient could be discharged if attempts to transfer were unsuccessful, but in a jurisdiction bound by Hargrave, the patient could not be discharged unless he or she recovered (presumably without medication) and no longer met involuntary commitment criteria. Because Nancy Hargrave chose to remain unmedicated, she probably had very distinct preferences (which could have been vitiated) about whether she would want to do so in an institutional setting or be allowed her freedom. In a vigorous exchange with Dr. Appelbaum, Michael Allen, a staff attorney at the Bazelon Center for Mental Health Law, argues that the “consumer's wishes” are dominant and that if psychiatrists were no longer allowed to override a PAD, “trust building, peer support, talk therapy, and other naturalistic supports” would be acceptable (and implicitly feasible) alternatives.³⁴ Dr. Appelbaum's response argues that Allen's solution ignores one inevitable fact—that in some cases of severe mental disorders, trust cannot be built, nor alliances established, despite the best efforts of providers.⁴⁹ The combined impact of these problems could present untenable conflicts for both physician and patient. Dr. Appelbaum further suggests that a major practical consequence of Hargrave is that psychiatrists will be reluctant to encourage (or perhaps even actively discourage) their patients to sign PADs. This result could constitute the worst of both worlds, where lawmakers believed they had given patients a legal way to implement their choices, but instead conflicts between statutes and case law combine with provider reluctance in a way that, paradoxically, disempowers patients without offering treatment to ameliorate their suffering. Such an outcome should be avoided if at all possible.

Conclusions

The new PAD statutes contain a large contradiction: among the provisions of these “let-the-patient-decide” laws are exceptions, which, in effect, render PADs as “let the doctor decide after all” (or, perhaps more accurately, “let the doctor decide whether the patient gets to decide”). PADs are qualified and nested in larger structures of law and policy that protect the interests of parties other than the patient, and which, in situations of conflict involving the treatment of incapacitated patients, tend to favor the clinician's professional judgment over the patient's manifest wishes to avoid standard treatment.

It is too soon to hazard any confident prediction about whether PADs, and the practice of overriding them, will become a prevalent or problematic feature of the mental health services landscape in the years to come. The legal prognosis for state laws that currently protect physicians’ prerogatives to override PADs is also uncertain. We can, however, speculate that the successful implementation of PADs will depend, to no small degree, on clinicians’ individual decisions to honor, and in some situations to set aside, patients’ wishes documented in PADs. Having reasonable safeguards for clinicians who decline PAD requests or refusals of treatment (i.e., when such requests clearly deviate from ethics- and evidence-based standards of care) may be not only prudent, but necessary for clinicians to support broad implementation of PADs and for patients in general—the vast majority of whom will never use PADs to refuse all treatment—to derive any benefit from them.^50,51

By their nature, PADs are complex vehicles of communication. They may be used to accomplish a variety of goals and may serve several functions: proscription as well as prescription of future treatment, engagement of a trusted third party as a surrogate decisionmaker, and the so-called Ulysses contract—a form of self protection against the potentially adverse consequences of one's own decisions during a future state of mind impaired by acute psychiatric illness. PADs almost unavoidably raise the specter of advance refusals of treatment during such states and, thus, the possibility of direct conflict with the professional standards and scruples of treating clinicians. That such clinicians are also, these days, acutely attuned to the nuances of legal risk and that they would generally rather be sued by disenfranchised psychiatric patients than by their family members (or, in the worst case, by the victims of the patients) brings to front and center the possibility that such directives will be overridden. And what then? Again, it is too soon to tell, but the fate of PADs, and the rights of clinicians to override them, will probably depend on the reasons for, as much as the results of, clinicians’ decisions to disregard patients’ advance wishes in individual cases. Some of the reasons that mental health professionals might override PADs may correspond to recognized and justifiable criteria for overriding medical advance directives in general. Other reasons and motives, however, may simply represent an inappropriate exercise of medical paternalism.⁵² Legal policy development, formulation, and application of clear guidelines for overriding PADs and educational outreach to clinicians regarding this and other issues surrounding PADs are all needed, but are hampered by the lack of empirical research on these topics.