Commentary: Psychiatric Advance Directives at a Crossroads—When Can PADs be Overridden
http://www.100md.com
《美国精神病学与法律学报》
Dr. Appelbaum is Professor and Director, Division of Psychiatry, Law and Ethics, Department of Psychiatry, Columbia University College of Physicians and Surgeons, New York, NY.
Abstract
Current statutes enabling psychiatric advance directives (PADs) typically include provisions allowing override of patients’ choices by treatment staff. Lest the purpose of the PAD be vitiated by too broad an application of the override mechanism, its use should be carefully limited. In inpatient settings, voluntary patients should have the right to decline treatments in advance, although not an absolute right to demand treatments of their choosing. The situation of involuntary patients is more complex. Permitting PADs to trump commitment statutes would undercut the combined parens patriae/police power rationale for commitment, a path taken currently by no U.S. jurisdiction. Moreover, PADs should not be permitted to negate the usual mechanisms for involuntary treatment of committed patients; to do otherwise risks forcing facilities to confine indefinitely persons they cannot treat. Even in those circumstances, however, where PADs provide evidence of reasonable patient preferences (e.g., for one medication over another), the choices they embody should be respected.
Advance directives (ADs) seem to be an ideal mechanism for persons with psychiatric disorders to express their treatment preferences. In contrast to persons making end-of-life treatment decisions, the most common use for ADs, psychiatric patients generally have experienced both the disorder and its treatment previously and thus are in a strong position to make meaningful choices.1 Most psychiatric disorders are sufficiently episodic that patients who lose decisional capacity will regain it at some point—unlike patients with dementia, for example—at which time they can indicate their competent preferences with regard to future care.
Yet, as Swanson and colleagues2 point out, the special statutes being adopted in many states to authorize psychiatric advance directives (PADs) typically contain language that allows clinicians, usually psychiatrists, to override patients’ choices. They cite Pennsylvania’s recently adopted statute which, in addition to giving priority to the state’s civil commitment law, permits physicians who "cannot in good conscience comply" with a patient’s or surrogate’s instructions that are "contrary to accepted clinical practice and medical standards" to disregard those instructions (although efforts must be made to find a practitioner who is willing to comply) (Ref. 2, p 386). Read broadly, this provision would appear to vitiate the impact of a PAD statute in a large number of cases.
Language such as this, which allows clinician override of patients’ advance decisions, contradicts the intrinsic rationale of PADs: maximizing autonomous choice, even in circumstances of decisional incapacity. Perhaps it represents some ambivalence on the part of legislators about empowering persons with mental illnesses to make their own decisions. Yet available data suggest that patients’ decisions are typically quite reasonable,3 and it seems generally agreed that when patients and clinicians are stimulated to discuss patients’ preferences in advance—a process encouraged by the completion of a PAD—much good comes from the interaction.
Swanson and colleagues,2 however, make a strong case for sometimes overriding the decisions made by patients in their PADs or voiced by their surrogates (i.e., at least in cases of danger to the patient or to others and perhaps in other situations as well). So the challenge for policy-makers is to craft statutes that maximize the value of PADs, without counterproductive adherence to patients’ advance preferences when they may be detrimental for them and for the broader society around them. How would a reasonable PAD statute address this tension
PADs warrant a presumption that the choices they embody will be respected, unless a limited number of exceptions apply. Were any other presumption to be operative, there would be little point to enabling PADs in the first place. Advance decisions, though, come in two forms: requests for particular treatments and objections to other treatments; and they are applied to two patient populations: voluntary and involuntary. (For the purposes of this discussion, voluntary patients are those who are not subject to either inpatient or outpatient involuntary commitment proceedings at the time their PADs are invoked.) There is reason to suggest that PADs should operate in different ways for these different purposes and patient groups. To take the easiest case, just as competent voluntary patients have the right to decline any medical treatment—psychiatric or otherwise—so they should have the power to incorporate those wishes into a PAD and to have their objections respected in the future.
However, requests of competent voluntary patients for particular treatments are not automatically honored, and the same should be true of the requests in PADs. Treatment availability, medical appropriateness, and (unfortunately, when the first two criteria have been met) financing to cover the costs of care all factor into determinations of whether patients will receive a treatment they request. It is difficult to argue that currently incompetent patients should be able to claim, through their PADs, treatment that they could not demand directly if competent. The reference to "accepted clinical practice and medical standards" in the Pennsylvania statute addresses one of the legitimate bases on which a voluntary patient’s request might be denied. In addition, a model statute might explicitly recognize the issues of availability and reimbursement for the costs of care.
This is all well and good as far as voluntary patients are concerned, but in the psychiatric context the special situation of involuntary patients must be considered. Every state permits the civil commitment of persons with mental illnesses who represent a danger to other people or to themselves (including in the latter category, persons who cannot meet their basic needs), based on a mix of parens patriae and police power considerations. The former is premised on the belief that severe mental illnesses compromise decisionmaking and thus legitimate the state’s assumption of decisional power for persons who are likely to suffer or cause harm if left to their own devices; the latter rests on society’s very real interests in preventing physical harm to its members.
Ought PADs to determine whether committable patients are hospitalized and/or treated against their will It seems clear that the presence of a PAD that declines treatment in no way invalidates the police power rationale for commitment—namely, that harm is likely to patients or to others. But if the parens patriae concern that drives commitment is based on the belief that patients must be taken care of when they are unable to make good decisions for themselves, does the availability of a prior competent choice vitiate the state’s interest in making decisions for them Were this argument to be accepted, PADs would trump commitments based on parens patriae grounds. No matter how impaired or likely to come to harm, a patient with a valid PAD refusing hospitalization would walk free.
What are we to make of the fact that no state with a PAD statute allows this to occur Perhaps it speaks to the essential inextricability of parens patriae and police power rationales in the current statutory framework. That is, so long as commitment is predicated on dangerousness to others or self (based on the state’s police powers)—but is also limited to persons with mental illnesses (because of parens patriae concerns)—both rationales apply to every case. Mitigation of the strength of the parens patriae justification, therefore, may not negate the residual grounds for commitment, regardless of the patient’s advance preferences. In addition, it may be that the state’s parens patriae interests are invoked by the conjunction of distorted decisionmaking due to mental illness and the consequent risk of serious harm. That the patient has previously requested that she be allowed to endanger her well-being while under the influence of a mental disorder may simply not be material to the state’s interests in preventing harm. Whether on one or both of these grounds—and both seem independently persuasive—patients who qualify for involuntary commitment should not be able to trump that process by means of an advance directive.
Of course, even if PADs cannot control whether patients are hospitalized involuntarily, perhaps they should determine whether treatment can be administered over patients’ objections. This was the outcome of Hargrave v. Vermont, which challenged the state’s power to treat an involuntary patient whose PAD indicated her preference to avoid all antipsychotic medications.4 Although decided on the grounds that Vermont’s differential treatment of PADs (which could be overridden for involuntary patients) and nonpsychiatric ADs (which could not) violated the Americans with Disabilities Act, the decision seems to reflect the belief that allowing advance refusal in these circumstances is desirable policy.5
In many respects, the dilemma with which the Hargrave court grappled recapitulates an earlier battle over the right of involuntary but competent patients to refuse treatment.6 Psychiatrists argued that it made little sense to commit someone against his or her will—especially if a parens patriae rationale was operative—without administering the treatment that would facilitate a rapid return to society. Swanson et al.2 point to other interests that support this approach, including the societal burden of prolonged hospitalization in the absence of treatment. However, this perspective did not prevail in many jurisdictions when the right to refuse treatment was litigated; in those states, committed patients who retain decisional capacity are able to decline any non-emergent care. In so far as the application of PADs to committed patients threatens to reopen this issue, a similar outcome might be anticipated in those jurisdictions.
But whether the law should evolve in this direction is a different matter. Advance directives, as noted, developed in the context of decisions to terminate end-of-life treatment. Here, society has a strong interest in letting people decline continued care, since treatment usually leads only to short-term prolongation of life (if that), often at the cost of considerable suffering and at great financial cost. The situations in which PADs come into play are quite different. Patients who complete PADs usually have treatable conditions, with the prospect of substantial symptomatic improvement and often a marked increase in functioning. Moreover, allowing such patients to decline treatment through PADs imposes a burden on society of indefinite care for persons with serious mental disorders. Sound policy dictates avoiding these consequences by refusing to privilege PADs over the usual mechanisms for involuntary treatment of committed patients.
This is not to say, however, that PADs should simply be disregarded once a patient is found to meet involuntary treatment criteria. When PADs express preferences among equally reasonable treatment choices (e.g., selecting a particular antipsychotic medication), these preferences should be respected. If a patient’s PAD rejects a specific treatment (e.g., electroconvulsive therapy; ECT), that treatment should be used only if no other treatment of likely effectiveness is available (e.g., except in emergencies, ECT could not be used unless antidepressants have been ineffective or there is strong reason to believe that they will be). Finally, although the Hargrave court gave short shrift to Vermont’s PAD statute, the law contained a provision that is well worth considering: patients who refused usual treatments (e.g., medications) were not treated for a period of time before involuntary treatment was initiated to see if they might improve sufficiently that it would be unnecessary to override their choices. This provision seems reasonably protective of patients’ wishes without imposing an excessive burden on either the patients or the treatment system.
If a balanced approach is taken to the use of PADs, their use is likely to grow, as is clinicians’ respect for patients’ choices. Should PADs be used as another weapon in the battle against civil commitment, however, support by clinicians for their completion and implementation, which is essential to widespread use, will wither. PADs now stand at a crossroads. Whether they proceed along a productive path depends heavily on the responses by legislators and judges to the issues raised by Swanson et al.2—to whom we should be grateful for underscoring the salience of these questions for the mental health policy agenda.
References
Appelbaum PS: Michigan’s sensible "living will" (Letter). N Engl J Med 301:788, 1979
Swanson JW, McCrary SV, Swartz MS, et al: Superseding psychiatric advance directives: ethical and legal considerations. J Am Acad Psychiatry Law 2006;34:385–94
Srebnik DS, Rutherford LT, Peto T, et al: The content and clinical utility of psychiatric advance directives. Psychiatr Serv 56:592–8, 2005
Hargrave v. Vermont, 340 F.3d 27 (2nd Cir. 2003)
Appelbaum PS: Psychiatric advance directives and the treatment of committed patients. Psychiatr Serv 55:751–2, 2004
Appelbaum PS: Almost a Revolution: Mental Health Law and the Limits of Change. New York: Oxford University Press, 1994(Paul S. Appelbaum, MD)
Abstract
Current statutes enabling psychiatric advance directives (PADs) typically include provisions allowing override of patients’ choices by treatment staff. Lest the purpose of the PAD be vitiated by too broad an application of the override mechanism, its use should be carefully limited. In inpatient settings, voluntary patients should have the right to decline treatments in advance, although not an absolute right to demand treatments of their choosing. The situation of involuntary patients is more complex. Permitting PADs to trump commitment statutes would undercut the combined parens patriae/police power rationale for commitment, a path taken currently by no U.S. jurisdiction. Moreover, PADs should not be permitted to negate the usual mechanisms for involuntary treatment of committed patients; to do otherwise risks forcing facilities to confine indefinitely persons they cannot treat. Even in those circumstances, however, where PADs provide evidence of reasonable patient preferences (e.g., for one medication over another), the choices they embody should be respected.
Advance directives (ADs) seem to be an ideal mechanism for persons with psychiatric disorders to express their treatment preferences. In contrast to persons making end-of-life treatment decisions, the most common use for ADs, psychiatric patients generally have experienced both the disorder and its treatment previously and thus are in a strong position to make meaningful choices.1 Most psychiatric disorders are sufficiently episodic that patients who lose decisional capacity will regain it at some point—unlike patients with dementia, for example—at which time they can indicate their competent preferences with regard to future care.
Yet, as Swanson and colleagues2 point out, the special statutes being adopted in many states to authorize psychiatric advance directives (PADs) typically contain language that allows clinicians, usually psychiatrists, to override patients’ choices. They cite Pennsylvania’s recently adopted statute which, in addition to giving priority to the state’s civil commitment law, permits physicians who "cannot in good conscience comply" with a patient’s or surrogate’s instructions that are "contrary to accepted clinical practice and medical standards" to disregard those instructions (although efforts must be made to find a practitioner who is willing to comply) (Ref. 2, p 386). Read broadly, this provision would appear to vitiate the impact of a PAD statute in a large number of cases.
Language such as this, which allows clinician override of patients’ advance decisions, contradicts the intrinsic rationale of PADs: maximizing autonomous choice, even in circumstances of decisional incapacity. Perhaps it represents some ambivalence on the part of legislators about empowering persons with mental illnesses to make their own decisions. Yet available data suggest that patients’ decisions are typically quite reasonable,3 and it seems generally agreed that when patients and clinicians are stimulated to discuss patients’ preferences in advance—a process encouraged by the completion of a PAD—much good comes from the interaction.
Swanson and colleagues,2 however, make a strong case for sometimes overriding the decisions made by patients in their PADs or voiced by their surrogates (i.e., at least in cases of danger to the patient or to others and perhaps in other situations as well). So the challenge for policy-makers is to craft statutes that maximize the value of PADs, without counterproductive adherence to patients’ advance preferences when they may be detrimental for them and for the broader society around them. How would a reasonable PAD statute address this tension
PADs warrant a presumption that the choices they embody will be respected, unless a limited number of exceptions apply. Were any other presumption to be operative, there would be little point to enabling PADs in the first place. Advance decisions, though, come in two forms: requests for particular treatments and objections to other treatments; and they are applied to two patient populations: voluntary and involuntary. (For the purposes of this discussion, voluntary patients are those who are not subject to either inpatient or outpatient involuntary commitment proceedings at the time their PADs are invoked.) There is reason to suggest that PADs should operate in different ways for these different purposes and patient groups. To take the easiest case, just as competent voluntary patients have the right to decline any medical treatment—psychiatric or otherwise—so they should have the power to incorporate those wishes into a PAD and to have their objections respected in the future.
However, requests of competent voluntary patients for particular treatments are not automatically honored, and the same should be true of the requests in PADs. Treatment availability, medical appropriateness, and (unfortunately, when the first two criteria have been met) financing to cover the costs of care all factor into determinations of whether patients will receive a treatment they request. It is difficult to argue that currently incompetent patients should be able to claim, through their PADs, treatment that they could not demand directly if competent. The reference to "accepted clinical practice and medical standards" in the Pennsylvania statute addresses one of the legitimate bases on which a voluntary patient’s request might be denied. In addition, a model statute might explicitly recognize the issues of availability and reimbursement for the costs of care.
This is all well and good as far as voluntary patients are concerned, but in the psychiatric context the special situation of involuntary patients must be considered. Every state permits the civil commitment of persons with mental illnesses who represent a danger to other people or to themselves (including in the latter category, persons who cannot meet their basic needs), based on a mix of parens patriae and police power considerations. The former is premised on the belief that severe mental illnesses compromise decisionmaking and thus legitimate the state’s assumption of decisional power for persons who are likely to suffer or cause harm if left to their own devices; the latter rests on society’s very real interests in preventing physical harm to its members.
Ought PADs to determine whether committable patients are hospitalized and/or treated against their will It seems clear that the presence of a PAD that declines treatment in no way invalidates the police power rationale for commitment—namely, that harm is likely to patients or to others. But if the parens patriae concern that drives commitment is based on the belief that patients must be taken care of when they are unable to make good decisions for themselves, does the availability of a prior competent choice vitiate the state’s interest in making decisions for them Were this argument to be accepted, PADs would trump commitments based on parens patriae grounds. No matter how impaired or likely to come to harm, a patient with a valid PAD refusing hospitalization would walk free.
What are we to make of the fact that no state with a PAD statute allows this to occur Perhaps it speaks to the essential inextricability of parens patriae and police power rationales in the current statutory framework. That is, so long as commitment is predicated on dangerousness to others or self (based on the state’s police powers)—but is also limited to persons with mental illnesses (because of parens patriae concerns)—both rationales apply to every case. Mitigation of the strength of the parens patriae justification, therefore, may not negate the residual grounds for commitment, regardless of the patient’s advance preferences. In addition, it may be that the state’s parens patriae interests are invoked by the conjunction of distorted decisionmaking due to mental illness and the consequent risk of serious harm. That the patient has previously requested that she be allowed to endanger her well-being while under the influence of a mental disorder may simply not be material to the state’s interests in preventing harm. Whether on one or both of these grounds—and both seem independently persuasive—patients who qualify for involuntary commitment should not be able to trump that process by means of an advance directive.
Of course, even if PADs cannot control whether patients are hospitalized involuntarily, perhaps they should determine whether treatment can be administered over patients’ objections. This was the outcome of Hargrave v. Vermont, which challenged the state’s power to treat an involuntary patient whose PAD indicated her preference to avoid all antipsychotic medications.4 Although decided on the grounds that Vermont’s differential treatment of PADs (which could be overridden for involuntary patients) and nonpsychiatric ADs (which could not) violated the Americans with Disabilities Act, the decision seems to reflect the belief that allowing advance refusal in these circumstances is desirable policy.5
In many respects, the dilemma with which the Hargrave court grappled recapitulates an earlier battle over the right of involuntary but competent patients to refuse treatment.6 Psychiatrists argued that it made little sense to commit someone against his or her will—especially if a parens patriae rationale was operative—without administering the treatment that would facilitate a rapid return to society. Swanson et al.2 point to other interests that support this approach, including the societal burden of prolonged hospitalization in the absence of treatment. However, this perspective did not prevail in many jurisdictions when the right to refuse treatment was litigated; in those states, committed patients who retain decisional capacity are able to decline any non-emergent care. In so far as the application of PADs to committed patients threatens to reopen this issue, a similar outcome might be anticipated in those jurisdictions.
But whether the law should evolve in this direction is a different matter. Advance directives, as noted, developed in the context of decisions to terminate end-of-life treatment. Here, society has a strong interest in letting people decline continued care, since treatment usually leads only to short-term prolongation of life (if that), often at the cost of considerable suffering and at great financial cost. The situations in which PADs come into play are quite different. Patients who complete PADs usually have treatable conditions, with the prospect of substantial symptomatic improvement and often a marked increase in functioning. Moreover, allowing such patients to decline treatment through PADs imposes a burden on society of indefinite care for persons with serious mental disorders. Sound policy dictates avoiding these consequences by refusing to privilege PADs over the usual mechanisms for involuntary treatment of committed patients.
This is not to say, however, that PADs should simply be disregarded once a patient is found to meet involuntary treatment criteria. When PADs express preferences among equally reasonable treatment choices (e.g., selecting a particular antipsychotic medication), these preferences should be respected. If a patient’s PAD rejects a specific treatment (e.g., electroconvulsive therapy; ECT), that treatment should be used only if no other treatment of likely effectiveness is available (e.g., except in emergencies, ECT could not be used unless antidepressants have been ineffective or there is strong reason to believe that they will be). Finally, although the Hargrave court gave short shrift to Vermont’s PAD statute, the law contained a provision that is well worth considering: patients who refused usual treatments (e.g., medications) were not treated for a period of time before involuntary treatment was initiated to see if they might improve sufficiently that it would be unnecessary to override their choices. This provision seems reasonably protective of patients’ wishes without imposing an excessive burden on either the patients or the treatment system.
If a balanced approach is taken to the use of PADs, their use is likely to grow, as is clinicians’ respect for patients’ choices. Should PADs be used as another weapon in the battle against civil commitment, however, support by clinicians for their completion and implementation, which is essential to widespread use, will wither. PADs now stand at a crossroads. Whether they proceed along a productive path depends heavily on the responses by legislators and judges to the issues raised by Swanson et al.2—to whom we should be grateful for underscoring the salience of these questions for the mental health policy agenda.
References
Appelbaum PS: Michigan’s sensible "living will" (Letter). N Engl J Med 301:788, 1979
Swanson JW, McCrary SV, Swartz MS, et al: Superseding psychiatric advance directives: ethical and legal considerations. J Am Acad Psychiatry Law 2006;34:385–94
Srebnik DS, Rutherford LT, Peto T, et al: The content and clinical utility of psychiatric advance directives. Psychiatr Serv 56:592–8, 2005
Hargrave v. Vermont, 340 F.3d 27 (2nd Cir. 2003)
Appelbaum PS: Psychiatric advance directives and the treatment of committed patients. Psychiatr Serv 55:751–2, 2004
Appelbaum PS: Almost a Revolution: Mental Health Law and the Limits of Change. New York: Oxford University Press, 1994(Paul S. Appelbaum, MD)