Legal and Ethics Considerations in Capacity Evaluation for Medical Aid in Dying

The Appropriate Evaluator

The initial logistical consideration in assessing the decisional capacity of patients seeking MAID is determining who will conduct the evaluation. All 11 U. S. jurisdictions that permit MAID assign this task to one or two physicians, although nine require a mental health professional to be involved under narrower circumstances (see Table 1). Assessing the frequency of such referrals is difficult because jurisdictions generally report the number of patients who require referral and ultimately die from MAID, but not those who are found to lack decisional capacity and die from other causes. For instance, although Weithorn reports that “5 percent of the patients who ultimately died from [MAID]” in Oregon “had been referred for mental health evaluation and found competent by the consulting mental health professional,” the lack of a known denominator makes these data challenging to interpret (Ref. 53, p 82). In any case, the absence of a legal mandate does not mean that individual hospitals cannot, or should not, require the involvement of psychiatrists, psychologists, or others with specialized training. For example, policy at the University of California San Francisco Medical Center (UCSFMC) requires that all patients seeking MAID be evaluated by a mental health provider even though state law does not require this additional safeguard.⁵⁴ In this regard, statutes serve merely as a floor, and health care institutions may be well advised to set more stringent requirements.

Even assigning the responsibility to a mental health professional or a psychiatrist does not fully resolve the question of what is the appropriate credentialing and training for the evaluator. In particular, when these cases arise in the hospital setting, the initial point of contact with behavioral health services is usually the consult-liaison (CL) psychiatry team. CL psychiatrists are generally well trained in decisional capacity assessment using the four skills model, but that does not mean they are necessarily the providers best suited for capacity assessment in MAID, especially if doing so were to require a different set of assessment criteria and tools. Alternatively, an outside psychiatrist with forensic training might be tasked with this distinctive form of assessment. Of note, that is the process that most hospitals use regarding capacity assessment for nonclinical matters, such as assessing a hospitalized patient’s testamentary capacity to execute a will. Such an approach would help ensure a truly independent evaluation, as the CL team may also be involved in providing treatment and emotional support to patients, which may be perceived as inconsistent with evaluating for MAID. In addition, separating evaluative and therapeutic roles facilitates open communication between patients and CL psychiatrists; patients might otherwise fear that complete honesty in pursuit of care (for example, regarding symptoms of cognitive decline) could impede their ability to pursue MAID. On the other hand, the CL team is more likely to be familiar with the patient’s personal narrative and the context for the patient’s decision and to have engaged in the sorts of conversations “about how that individual had come to choose MAID” that clinicians find most helpful in making such determinations (Ref. 48, p E360). A third option, entirely theoretical at present, might involve the use of a novel class of professionals (whether medical doctors (MDs), psychologists, or nurse practitioners (NPs)) who have undergone clinical training specifically in the field of MAID capacity assessment, a niche that will potentially expand in coming years as more jurisdictions consider legalization. Whichever approach is ultimately adopted, care should be taken in advance to ensure that the parties charged with such assessments feel both competent and empowered to conduct them.

Evaluator Liability and Consequences

A second logistical consideration is whether, and under what circumstances, evaluators may face either legal or professional consequences for conducting capacity evaluations for MAID. This is a subject that arises in clinical practice but has received short shrift in the existing medicolegal literature. Many jurisdictions that do permit MAID offer statutory protection to shield individuals participating in the practice from these consequences. Oregon’s Death with Dignity Act, for instance, states that “[n]o person shall be subject to civil or criminal liability or professional disciplinary action for participating in good faith” in MAID and also that “[n]o professional organization or association, or health care provider, may subject a person to censure, discipline, suspension, loss of license, loss of privileges, loss of membership or other penalty for participating or refusing to participate in good faith” in MAID.¹⁹ Whether the “loss of membership” provision is constitutional, in light of the U.S. Supreme Court’s decision in Boy Scouts of America v. Dale, remains unclear.⁵⁵ In that case, the Court faced the question of whether the Boy Scouts of America (BSA) might exclude a gay Eagle Scout and assistant Scoutmaster, James Dale, based on their sexual orientation.⁵⁵ In a five-to-four decision, the Court ruled that even a “public accommodation” like the BSA might exclude an individual from membership if “the presence of that person affects in a significant way the group’s ability to advocate public or private viewpoints,” regardless of any state law to the contrary (Ref. 55, p 648). Arguably, the same legal principle would allow a professional organization opposed to MAID to expel members with contrary viewpoints or engaged in MAID-related conduct, but Oregon’s statutory provision has not yet been challenged.

In addition to Oregon, other jurisdictions offer more limited protections. For instance, Vermont’s statute prevents “civil or criminal liability or professional disciplinary action,” but not loss of membership in professional organizations.²⁰ Montana, which legalized MAID via judicial decision, does not protect physicians from professional consequences at all, although that state also does not yet require formal capacity assessment either.

In jurisdictions where MAID is legal, establishing protection from professional consequences, such as expulsion from medical organizations, may prove essential to availability. Otherwise, providers may be reluctant to provide capacity evaluations for MAID for fear of running afoul of organizations to which they belong. For instance, prior to November 2023, the American Medical Association’s Code of Ethics objected to MAID on the grounds that the practice, which it termed “physician-assisted suicide,” was “fundamentally incompatible with the physician’s role as healer.”⁵⁶ But, as early as 2019, the American Medical Association (AMA) also issued guidance that providers might provide MAID “according to the dictates of their conscience without violating their professional obligations.”⁵⁷ More recently, in 2023, the AMA agreed to review both its use of the term physician-assisted suicide (PAS) and its position on MAID.⁵⁸ At present, the AMA officially objects to the practice. Many state medical associations stand opposed as well. The Montana Medical Association’s position statement on MAID, for example, states that the organization “does not condone the deliberate act of precipitating the death of a patient,” which may deter providers from involving themselves in any aspect of the practice, including capacity assessment.⁵⁹ Physicians and psychiatrists engaging in capacity evaluations for MAID in states where the practice is permitted should know the extent to which they are protected from legal and professional consequences. If the goal is to ensure the availability of MAID, expansive “good faith” protections, such as Oregon’s, should be adopted. None of these protections, of course, can protect physicians from potential reputational damage or the possibility that other providers will decline to provide referrals to colleagues they perceive to be pro-MAID.

Far more legally complicated are capacity assessments for MAID for patients in jurisdictions in which the practice is not legal who are seeking to travel to other jurisdictions, whether domestic or abroad, where MAID is permitted to end their lives. In fact, many states have statutes that specifically criminalize facilitating suicide (see Table 2).^60,–,97 In New York State, for example, “a person is guilty” of a class E felony “when he intentionally causes or aids another person to attempt suicide” and “a person is guilty” of a class C felony, “when he intentionally causes or aids another person to commit suicide.”⁹⁸ Conducting a capacity assessment for MAID for a patient who seeks to travel elsewhere to end his own life legally may constitute a form of manslaughter in New York State and prove similarly criminal in many other states. In a number of jurisdictions, the law on this subject remains uncertain, depending on such factors as whether capacity assessment for referral is considered a “physical act” related to the patient’s subsequent death. These vague statutes create the hypothetical potential for severe penalties, even if they do not appear to be widely enforced against capacity evaluators. Moreover, because these capacity assessments rarely if ever come to the attention of law enforcement authorities, if enforcement would occur in a high-profile case remains unknown.

In practice, CL psychiatrists are frequently asked to conduct such assessments in states where MAID remains illegal, as patients may be reluctant to travel abroad for the procedure without prior documentation ensuring eligibility. Some psychiatrists, at least, apparently do perform such evaluations. In doing so, these providers may open themselves up to civil, as well as criminal, liability. In addition, the possibility exists that private employers, including medical schools, may be able to rely on violations of the AMA Code of Ethics as a basis for terminating faculty. This possibility arose in the matter of Bandy Lee, a psychiatrist fired by Yale for allegedly violating the “Goldwater rule” of the American Psychiatric Association (APA) that essentially prohibits diagnosing public figures whom one has not personally examined.^99,–,101 Although the APA does not enforce its “Principles of Medical Ethics with Annotations Especially Applicable to Psychiatry” on nonmembers, that does not necessarily prevent other institutions from courts to academic employers from relying on these principles as professional standards.⁹⁹ The potential impact of the codes of voluntary professional associations, some of which proscribe participation in MAID, has significant practical implications. State civil and criminal law in nearly all jurisdictions in which MAID is not yet legal remains unresolved on these complex subjects. Providers operating in Veterans Affairs facilities and other federally run institutions may face additional restrictions and consequences. Physicians and psychiatrists engaging in capacity evaluations for MAID in jurisdictions where the practice is not authorized for patients planning to obtain MAID elsewhere may be entirely unaware of the legal and professional risks involved in conducting such assessments.

Historical Context

The passage of Oregon’s Death with Dignity Act in 1994 led to a dialogue in the medicolegal literature between those authorities who believed that impartial capacity assessment for MAID was possible and those who remained dubious. In the former camp, the team of psychologist James Werth at the University of Akron, G. Andrew H. Benjamin at the University of Washington, and Tony Farrenkopf of Portland, Oregon offered extensive guidelines for how to conduct capacity assessments for MAID consistent with the Oregon statute.¹⁰² These included use of the MacArthur Competence Assessment Tool for Clinical Research (MacCAT-CR), initially designed by Appelbaum and Grisso, although Werth and colleagues acknowledged that the data for this measure largely stemmed from research settings and that the designers had pointedly avoided addressing its applicability to decisional capacity related to choosing death.¹⁰³ Of note, the MacCAT-CR has never “been validated specifically for MAID,” (Ref. 49, p E359) and in practice, Canadian family physician Ellen Weibe and colleagues report that “experienced MAID assessors do not routinely require formal capacity assessments or tools to assess capacity in patients requesting MAID” (Ref. 49, p E362). In response, law professors Susan Martyn and Henry Bourguignon described capacity assessment in MAID as a “Trojan Horse” in that “the final determination of capacity” in MAID cases “will usually be a product of the hidden values and subjective judgment of the physician” (Ref. 52, p 395). Martyn and Bourguignon examined Dutch MAID practices and concluded that “Dutch doctors routinely report the degree of suffering or the perceived quality of life of a patient as the primary factor in determining when a request is granted” and that “in practice, physicians ignore capacity and voluntariness, occasionally listen to relatives or other third parties rather than the patient, and consistently perform euthanasia or PAS when they, the physicians, deem the patient’s quality of life miserable” (Ref. 52, p 398). They expressed concern that shifting standards of capacity based on the evaluating physician’s perception of the patient’s quality of life risked confusing PAS with euthanasia and turning MAID into “the physician’s prescription, not the voluntary and competent choice of the patient” (Ref. 52, p 399).

The contours of this debate, although of considerable importance in the ethics and policy questions regarding whether MAID should be legal, are not particularly relevant to clinicians once MAID has in fact been authorized by law. Arguably, many of Martyn and Bourguignon’s concerns may be accurate but may still not prove dispositive for legislatures and voters with regard to MAID’s legal status. Other considerations related to patient autonomy or the mitigation of human suffering may simply override these concerns. Unfortunately, the existential debate over MAID has transpired at the expense of a more nuanced discussion of three important concerns related to operationalizing capacity evaluation for MAID, which remain unsettled and insufficiently examined: the depth of understanding required to satisfy capacity for MAID, the levels of scrutiny and certainty required for a determination of capacity for MAID, and the conceptual challenge of “impairment” in MAID cases.

Level of Understanding

The level of understanding in clinical capacity evaluation refers to the evaluator’s assessment of the degree to which the patient fully understands the situation, its consequences, and the potential risks and benefits of any additional interventions (as well as choosing against any interventions at all). Even with the most sophisticated tools, such assessment is inherently subjective. Understanding differs from the objective factor of how much information must be provided by physicians to patients prior to rendering a clinical decision, a matter addressed in most American MAID statutes, often with considerable precision. Although capacity generally requires a threshold determination of whether a patient meets a particular standard, understanding occurs along a continuum.¹⁰⁴ After all, both the average college physics student and Albert Einstein may understand the theory of relativity, but their levels of understanding clearly differ. Establishing an operational model of capacity assessment requires determining where to locate the threshold for capacity on this continuum.

Although the four skills model is widely used when evaluating capacity for most clinical decisions, some jurisdictions have chosen a different standard for certain forms of decision-making. For instance, Utah and Vermont establish distinct standards for capacity to appoint health care proxies, an approach followed in practice by many physicians elsewhere as well.^19,94,105 In addition, capacity to consent to voluntary admission to a psychiatric hospital requires a different standard in some jurisdictions, an approach endorsed by a task force of the APA in 1993.¹⁰⁶ In practice, most physicians use a lower standard when a patient seeks to appoint a close relative as a temporary decision-maker, such as the emergency room patient who states, “Doctor, I don’t understand my medical condition, but I’ve been married for many years, and I trust my spouse to make all of my decisions right now.” In short, the principle that certain clinical decisions are best served by a different approach to capacity assessment has been widely accepted in medical practice for many years. A compelling case exists for including MAID in this category of interventions requiring a distinct approach. The four skills model is designed to assess capacity for treatment and to consider “the values that patients place upon the risk and benefit of each treatment in question” (Ref. 40, p 1636). These risks are generally measured in terms of “probabilities” of mortality and morbidity.⁴⁰ Yet, in the case of MAID, mortality is not the risk to be avoided but rather the benefit being sought. Although one might ultimately decide that the same level of understanding is necessary for MAID as for other medical procedures, this conclusion is not at all intuitive. Because MAID is at least arguably different in kind from most other medical interventions, one might reasonably argue that novel approaches are needed for capacity assessment in this area. The intent here is not to suggest that the MacArthur Competence Assessment Tool does not have important value in the research setting, or that it does not offer helpful guidance in developing tools for capacity assessment for MAID in the clinical setting, only that it ought to be thought of as a starting point and not the only possible approach.

One line of reasoning favors a higher level of understanding and appreciation to consent to MAID. Requiring an increased degree of understanding for high stakes decisions has strong backing in the literature, including in the early works of Drane and of Buchanan and Brock.^39,50 Werth and colleagues have noted that “many commentators have suggested that as the consequences of the capacity decision become more severe, the test should become more stringent” (Ref. 102, p 356). Arguably, because the result of MAID is death, the stakes are extremely high. (Of course, in jurisdictions that only permit MAID for terminally ill patients, the stakes might be considered lower for MAID than in cases of medical decision-making in which the patient has a meaningful choice between death and long-term survival, such as forgoing chronic dialysis.) Another line of reasoning, probably more persuasive, argues for a lower standard of understanding for patients seeking MAID. A full appreciation of the details and nuances of one’s medical condition, although often essential for comprehending the risks and benefits of various proposed interventions, seems superfluous when the question before patients is straightforward, stark, and binary. Indeed, rejecting a lower standard may reflect parentalistic tendencies that have since been rejected in other areas of decisional capacity assessment.¹⁰⁷ Although awareness of alternative options, such as palliative care, may certainly help inform the patient’s decision regarding MAID, and disclosure of such other measures is required by several state statutes, the ultimate matter to be decided by the patient is whether the patient wishes to die via MAID or not die via MAID. Appreciating the clinical risks of MAID may not have a meaningful impact on patient decision-making, as the risk is the benefit. That is not to say that patients should not be informed that the MAID process might in theory result in discomfort or pain or that self-administered medications might not function as attended, but these outcomes are extremely rare, and no reports appear available in the literature. Nonclinical risks, such as the impact of choosing MAID on interpersonal relationships and finances, may indeed exist and prove highly consequential, but these consequences are generally not considered in any decisional capacity evaluations.

The choice of MAID may be emotionally fraught but is generally easier to understand than most medical procedures, which do involve various potential risks, side effects, and outcomes of different likelihoods. In the absence of psychiatric factors that may diminish capacity, as discussed below, confirming that the patient recognizes that MAID will lead to a volitional death and seeks this outcome appears to be a sufficient standard of understanding for capacity. The alternative, erring excessively on the side of caution, risks letting “a fair number of patients suffer” (Ref. 108, p 590). Obviously, understanding is not the only factor to be considered; concerns for duress or psychiatric impairment may prove relevant, as discussed below, but none of these factors bear on the level of understanding required for MAID capacity itself. Although society may be wise to create safeguards against abuses of MAID, relying on artificially elevated capacity standards is neither the most effective or consistent mechanism for achieving these ends.

Finally, as noted by Mirza and Appel, the capacity evaluation process itself is “far from benign” for some patients, “especially those from racial and ethnic minorities” (Ref. 37, p 36). Carpenter and Merz have argued that the burden may be exacerbated in patients seeking MAID who “are likely to be quite ill” and “may have limited tolerance for lengthy conversations” (Ref. 1, p 252). In light of these concerns, simplifying the process furthers the goal of improving patient welfare and more intrusive explorations of capacity should only occur when they are clinically necessary.

Levels of Scrutiny and Certainty

The levels of scrutiny and certainty required in capacity assessment are distinct both from the underlying question of understanding and from each other. Scrutiny refers to how carefully the evaluator endeavors to ensure that the patient meets the predetermined capacity standard. In clinical practice, high stakes decisions usually result in more careful scrutiny of a patient’s decisional capacity. Certainty, closely related but distinct, refers to how confident the evaluator is that the evidence obtained in the assessment establishes that the patient has met that standard. American law does not dictate a specific standard for certainty in capacity assessment, so considerable variation exists between evaluators.

The practice of applying increased levels of scrutiny and demanding levels of certainty in capacity evaluations for high stakes medical decisions has generally been applied to MAID. For example, John Peteet has observed that, in the context of MAID, “assessment of a patient for ‘capacity’ requires the psychiatrist to go beyond checking off the four criteria [of the ‘four skills’ model] to actively working to help expand the patient’s capacity for envisioning other options.”¹⁰⁹ Merely checking off that the four criteria have been formally met, which is all too common, is conceptually problematic. As Susan Stefan has cogently observed that whether or not “there should be heightened standards around a decision to die,” these should exist independent of “requiring a higher bar for competence” (Ref. 110, p 44). A case can certainly be made that physicians should require a patent to envision alternatives to MAID, but that bears on the understanding required, not on the scrutiny or certainty required. Unfortunately, evaluators often conflate these distinct elements of the process. A higher level of scrutiny and certainty may be deemed preferable in high stakes, irreversible decisions, such as choosing MAID, but it does not logically follow that increased requirements for scrutiny and certainty cannot exist simultaneously with a lower standard for understanding. In fact, such an approach may prove desirable. Although the underlying ethics concern is complex and controversial, a first step in establishing an effective capacity evaluation process for MAID is distinguishing these distinct elements and determining appropriate standards for each independently.

The Challenge of Impairment

The most fraught concern in operationalizing capacity assessment for MAID relates to the consideration to be afforded patients whose choices are influenced by psychiatric conditions. Most American statutes specifically require a behavioral health evaluation under such circumstances, presumably with the goal of determining whether capacity is impaired by the patient’s psychiatric state. Such a distinction may provide a false sense of assurance for legislatures but, barring rare cases of psychosis, likely obfuscates more than it clarifies. As Lois Weithorn has saliently noted, “persons in the later stages of terminal disease experience emotional suffering to a greater extent than do persons in the general population” and investigators “report challenges of distinguishing between the presence of a mental disorder and the psychological distress attendant to the grief, loss, and suffering that often accompanies the dying process” (Ref. 53, p 82). Depression in response to a terminal diagnosis may certainly be distressing and may merit treatment in situations where the patient consents. (Involuntary psychiatric treatment for patients with depression may also be indicated in some cases but is beyond the scope of this article.) Still, why such depression should invalidate a choice for MAID, if all other criteria are met, remains unclear to some.¹⁰⁷ For instance, patients with “poor medical prognoses and/or low quality of life (PMP/LQL)” may experience depression “yet voice a desire for death under circumstances in which they would prefer death even if not depressed” (Ref. 111, p 360). Although a values-based capacity assessment tool might prove more help in such cases than the four skills model, neither approach fully resolves the underlying phenomenological problem of determining whether the choice for MAID is true to the patient’s authentic self or is a product of pathology.⁴³ This question is of great significance in moral philosophy and ethics, but as the underlying distinction is theoretical rather than empirical, it is not one that any capacity assessment method or tool, no matter how precise, can ever resolve. Rather, to operationalize capacity assessment for MAID, the best approach is to accept that this distinction will remain unresolvable in most cases. In some cases, however, a clear determination between an authentic desire and one diminished by psychiatric pathology is possible. In such cases, having a clear policy is important. In cases where such a determination cannot be made, having a default policy may also prove highly useful. Yet, although establishing such a default standard is important, doing so should not be conflated with traditional capacity questions related to understanding and impairment.