Introduction
Every jurisdiction in the United States permits capable persons to appoint third-party agents, often known as healthcare proxies or durable powers of attorney for healthcare, to make medical decisions on their behalf in the event of incapacitation.Footnote 1 A federal statute, the Patient Self-Determination Act (1991), and various state laws encourage individuals to do so.Footnote 2 In the absence of a proxy appointed through an advance directive, the vast majority of states have also established a mechanism by which family members or close friends—usually known as surrogates—can render decisions for incapacitated patients.Footnote 3 The purpose of such statutes is to advance patient autonomy.Footnote 4 While binding one’s future, incompetent self through a so-called “Ulysses contract” remains controversial among ethicists, the approach has largely been embraced by the medical and legal communities.Footnote 5 , Footnote 6 , Footnote 7 What remains less clear—and only addressed by statute in a minority of states—is what level of understanding is required to appoint a healthcare agent. The literature on the subject remains limited.Footnote 8 A consensus has emerged that evaluating capacity to appoint a proxy requires a different standard of ability from capacity to render specific healthcare decisions.Footnote 9 , Footnote 10 Far less clear is what that standard should be. Moye et al.Footnote 11 sum up the challenge in this way: “Any standard for capacity to appoint a healthcare [agent] must…take into account the relative risks and burdens of a stringent, high standard on one hand and a lax standard on the other.” Both extremes threaten to undermine the patient’s ability to effectuate her wishes, either by appointing the wrong decision-maker or failing to appoint the correct one. Too high a bar can also impede care if authorization for treatment is delayed.
The assumption of the few commentaries on the subject, as well as the existing statutes, is that the same standard should be used in all cases. According to this approach, while the “understanding and choices” of the specific patient may vary, as may “the context of the risks and benefits of the medical and interpersonal factors,” the general expectations regarding capacity to appoint a proxy do not.Footnote 12 Minimal attention is directed at the level of scrutiny that examiners should employ when investigating capacity for this purpose, and no consideration at all is afforded the patient’s previous stance toward advance directives themselves—although logically this history might have some meaningful bearing upon her current choices. After briefly reviewing the state of law and scholarship in this underdeveloped area, this essay proposes a novel strategy for determining whether a patient possesses the capacity to appoint a healthcare agent. This methodology, drawing upon the “sliding scale” model frequently employed in evaluating decisional capacity, recognizes four distinct sets of possible circumstances related to the relationship between the patient and the healthcare directive itself: (1) those in which the identity of the would-be proxy and the default surrogate are identical; (2) those in which no previous proxy was appointed, but the would-be proxy and the default surrogate differ; (3) those in which the would-be proxy will replace a previously appointed agent, but the change can be attributed to an intervening event; and (4) those in which the would-be proxy will replace a previously appointed agent, but the change cannot be attributed to an intervening event. Conceptualizing potential cases in this manner and classifying them into these discrete categories should help guide evaluators regarding how high a bar for certainty to establish and how much scrutiny to impose when making such determinations. As cases advance from category one to category four, the dangers of a lax standard increase, because there is an elevated risk that the appointment of an agent will undermine rather than effectuate the patient’s autonomous wishes. While this paper is focused on practices in the United States, the issues discussed and model proposed should prove relevant in any setting where clinicians are asked to assess a patient’s capacity to appoint a healthcare agent.
Background
Both statute and case law are surprisingly sparse regarding the question of capacity to appoint a healthcare agent. This cannot be attributed solely to the relatively recent emergence of advance directives in medicine, because even many statutes that specifically consider the role of proxies do not address the issue. For instance, the 1993 Uniform Health-Care Decisions Act, which offers a clear definition of the capacity to render healthcare decisions, is entirely silent on the question of the capacity standard for appointing an agent to do so.Footnote 13 Although in theory the range of standards is potentially very broad, statute, case law, and commentary have all largely confined themselves to three particular approaches. The first approach views the appointment of a healthcare agent through the lens of contract law and contemplates the advance directive as analogous to a contractual arrangement between private parties. Most jurisdictions require, at a minimum, some combination of voluntariness, absence of duress, and sound mind to effectuate a contract. Several states, including Michigan, Massachusetts, and California, follow this barebones approach by statute regarding healthcare proxies. Michigan law requires that a proxy appointment must be “executed voluntarily” and that the authorizing party must be of “sound mind and under no duress, fraud, or undue influence.”Footnote 14 California states the premise even more directly: “A natural person having the capacity to contract may execute a power of attorney,” which appears to incorporate healthcare as well as financial decisions.Footnote 15 Massachusetts requires an appointment with “sound mind and under no constraint or undue influence.”Footnote 16 Tennessee has adopted this approach via case law.Footnote 17 (Complicating matters, “sound mind” is generally not defined in state statutes, but left to the courts.Footnote 18) Other states incorporate a more developed approach to the contract-based theory of agent appointment that includes “the ability to understand the nature of the agreement, the effect of the contract (i.e., what obligations are imposed), and the subject matter of the contract.”Footnote 19 For instance, Connecticut law spells out that to appoint an agent one must be “of sound mind and able to understand the nature and consequences of health care decisions at the time this document was signed.”Footnote 20
A second approach that has been raised in the literature, but not adopted by any United States jurisdiction, would follow the standards for testamentary capacity.Footnote 21 These standards are usually defined as some variation of the four-prong Greenwood-Baker test, after the remote English cases of Greenwood v. Greenwood (1790) and Harwood v. Baker (1840) in which they were initially established.Footnote 22 These include recognizing the nature of the act of creating a will, the extent of one’s wealth, the identity of one’s natural beneficiaries (“the natural objects of one’s bounty”), and the actual recipients under the document.Footnote 23 By analogy, executing a healthcare proxy might require a knowledge of the “natural” surrogates or default decision-makers under state law; no statutes actually do so, although some require that the selection be “appropriate”—but that does not inherently require an individual be aware of more “natural” or “appropriate” alternatives.
A third approach, adopted by statute in both Vermont and Utah, enumerates specific standards unique to the appointment of a healthcare proxy. In Vermont, to appoint an agent one must have an “ability to make and communicate a decision” regarding the choice of agent and possess “a basic understanding of what it means to have another individual make health care decisions for oneself and of who would be an appropriate individual to make those decisions….”Footnote 24 In Utah, factors to be considered in the capacity assessment are “whether the adult has expressed over time an intent to appoint the same person as agent”; “whether the choice of agent is consistent with past relationships and patterns of behavior between the adult and the prospective agent, or, if inconsistent, whether there is a reasonable justification for the change;” and “whether the adult’s expression of the intent to appoint the agent occurs at times when, or in settings where, the adult has the greatest ability to make and communicate decisions.”Footnote 25 Of note, Utah’s statute emphasizes that “an adult who is found to lack health care decision making capacity…may retain the capacity to appoint an agent.”Footnote 26 In fact, no jurisdiction applies the same standard for appointment of a healthcare agent as it does to making medical decisions for oneself, while every jurisdiction that addresses the issue suggests that the standard ought to be lower. At the extreme, New York State appears to set the bar at the level of legal incompetence. New York law states, “every adult shall be presumed competent to appoint a health care agent unless such person has been adjudged incompetent or otherwise adjudged not competent to appoint a health care agent, or unless a committee or guardian of the person has been appointed for the adult pursuant to…mental hygiene law or…the surrogate’s court procedure.”Footnote 27 New York’s rule appears to suggest that even a psychiatric determination of incapacity, without a formal judicial ruling on legal competence, is not enough to prevent the appointment of an agent—although whether a court will uphold an appointment by a clearly incapacitated person remains uncertain.
The first serious academic considerations of the issue did not arise until the twenty-first century, and they initially appeared in the context of authorization for research. Scott Kim and Paul Appelbaum observed as recently as 2006 that “there is virtually no literature on this topic, either theoretical or empirical. Even the most authoritative and comprehensive compendium on the evaluation of competencies does not discuss” capacity to appoint a proxy.Footnote 28 In the research setting, Kim and Appelbaum proposed three guiding principles that might determine whether individuals could consent to research on their behalf: (1) “the ability to understand—after a full explanation—the key issues regarding the nature of research” and “apply these facts to themselves,” (2) the ability “to indicate whether or not they would appoint a research proxy and to name a potential proxy,” and (3) to “explain why that person was chosen.”Footnote 29 Kim and Appelbaum refer to these principles as appreciation, choice, and reasoning, respectively.Footnote 30 This rubric, not far removed from the “four skills” approach that Appelbaum and Thomas Grisso had previously advanced for clinical decision-making, sets a relatively high bar for appointing an agent.Footnote 31 They also distinguish this form of capacity from capacity to create an “advance instructional research directive” or “living will for research participation,” which they acknowledge would require an even higher bar.Footnote 32 While their proposed standard might prove persuasive in a research setting, a lower one makes more sense in the clinical arena, where the patient stands to benefit directly from the choice of the optimal agent. Arguments for a lower bar in the clinical setting include the possibility that a default surrogate is not prescribed by law in the jurisdiction or not readily available.Footnote 33 The case for a lower bar is especially compelling if the “document can be revoked at any time,” and if, “when a patient disagrees with the agent, the disagreement invalidates the authority of the agent.”Footnote 34
In 2013, Moye et al. became the first to propose a capacity tool for proxy appointment specifically designed for the clinical setting. They argued for a two-prong test that required “understanding and choice” and elaborated that these could be demonstrated by displaying “capacity to understand the meaning” of giving “authority to another to make healthcare decisions” through the proxy in light of current or future “diminished capacity to consent to treatment,” as well as capacity to “determine” and “express a consistent choice” regarding an appropriate surrogate.”Footnote 35 More recently, Mark Navin et al. have proposed a less demanding test that requires merely that the patient “[e]xpress a consistent choice about a preferred [agent]” and “[d]emonstrate a basic understanding of what [an agent] does.”Footnote 36 Both approaches place considerable authority in the hands of physicians “to confirm, or refute, a patient’s capacity to execute an advance directive.”Footnote 37 Although proposed as distinct from evaluations of medical decision-making capacity, each of these approaches shows clear influence of the dominant models in that area.
The criteria proposed by both Moye et al. and Navin et al. have substantial advantages over a model that relies upon either legal methods based on contract theory or that defaults to the standards used to render other medical decisions. Unfortunately, neither specifically incorporates the patient’s overarching attitude toward advance directives, her particular engagement with them in the past, or how the advance directive is likely to alter the choice of agent. Yet if the goal is to determine how best to effectuate the patient’s authentic and autonomous wishes, then whether she has previously appointed a proxy, why she has or has not, and whether or not doing so will alter the identity of the agent or the scope of authority are arguably highly relevant to the evaluation process. Rather than one of many factors to be considered, this essay proposes that classification according to these criteria serves as the basis for determining what levels of certainty and scrutiny should be required when utilizing models of assessment, whether Moye et al. or Navin et al. or another.
A Category-Based Model
The principle behind the “sliding scale” model for clinical capacity assessment is that “the more dangerous the medical decision, the more stringent the standards of competency.”Footnote 38 The notion that the level of risk involved in a medical decision could be sorted into discrete categories was initially proposed by Loren Roth in 1977, and the concept of the scale more fully developed by James F. Drane, and by Allan Buchanan and Dan Brock.Footnote 39 , Footnote 40 , Footnote 41 The sliding scale can be used in clinical capacity assessment in conjunction with the four skills model—or, in theory, another model—to determine the level of scrutiny required regarding the patient’s decision. Logically, the same approach can be used regarding determination of capacity to appoint a proxy, but so far this method has not been proposed in the literature. What follows are four discrete categories that evaluators might find helpful in classifying such cases before proceeding to assessment. Excluded from this model are situations in which the patient arrives at the hospital with a previously executed proxy and the question arises as to whether the patient had capacity to appoint the proxy at that earlier date. While these cases do arise, they raise a different (albeit related) set of questions, and an argument can be made for honoring advance directives in such circumstances in the absence of clear evidence of incapacity—and imposing a much higher bar for invalidating them. That is a subject beyond the scope of this paper. Instead, the four categories below apply to level of certainty of accuracy and scrutiny of investigation at the time of evaluation to execute a proxy.
Proxy and Surrogate Convergence
In some situations, the proxy proposed by the patient and the default decision-maker in the absence of a proxy will be the same person. Since state surrogacy laws are designed to ensure that someone who knows the patient and her wishes well is the surrogate in the absence of a proxy, when the proxy and surrogate are identical, the likelihood is high that the proposed proxy will be a person well-known to the patient and more cognizant than most others of the patient’s wishes. In addition, no other party is being supplanted and, at least in regard to advance directives, no evidence indicates that the patient specifically did not want this person to serve as their decision-maker. One might wonder whether in this circumstance it even matters whether the third-party acts as proxy or as surrogate, if she is the same individual operating under the same standard (e.g., best interest, substituted judgment). Yet the difference is more than titular in a number of jurisdictions where the powers of appointed proxies are broader than those of default surrogates. For instance, in New York State, proxies possess more expansive powers regarding end-of-life decisions and the authorization of certain clinical procedures.Footnote 42 , Footnote 43 Allowing the third-party decision-maker to act as a proxy expands modestly the scope of her power but might prove important in speeding care and avoiding the delay of a court hearing. Such circumstances pose the least likelihood that honoring the proposed proxy appointment will undermine the patient’s authentic wishes, so the least scrutiny is indicated for such evaluations and the lowest bar for upholding the appointment.
Proxy and Surrogate Divergence, No Prior Proxy
Likely, the most frequent of the four scenarios discussed are cases in which the patient has not previously appointed any proxy and wishes to appoint someone other than the default decision-maker. Two variations of this situation may arise. In the first, the patient wishes to appoint a proxy who is a different individual from the default surrogate under the jurisdiction’s surrogacy law. In the second, the patient wishes to appoint a proxy in a jurisdiction that does not have a surrogacy law at all.Footnote 44 A higher level of scrutiny and level of certainty is likely indicated here than in the first scenario, because the identity of the decision-maker and not just the scope of her powers is being altered. However, while another party or entity is being supplanted as the decision-maker, it is far more likely that the patient did not previously contemplate with sufficient seriousness who she preferred as an agent rather than that she intentionally chose to inform herself of the state surrogacy law and then relied upon its default provisions, rather than using a proxy. Not only are the scope of powers of proxies usually broader, and the certainty of their authority more secure, but most individuals knowledgeable enough to be familiar with surrogacy law are also likely wise enough not to rely upon it. (Needless to say, most physicians would also urge a patient to compete a proxy form, rather than relying on the default provisions of a jurisdiction’s surrogacy law.) As a result, in such cases, the evidence generally does not suggest that the patient did not previously want the proposed proxy to serve as their decision-maker so much as it suggests that they did not consider the issue. The patient may now have multiple reasons for appointing a proxy as opposed to relying upon a default surrogate or the courts—ranging from convenience, if the proposed proxy is present and the default surrogate is not, to a belief that the proposed proxy is more knowledgeable regarding her wishes and/or more likely to effectuate them. The certainty and scrutiny here should be higher than in the first scenario, but not so high as to place an excessive burden upon the patient.
Proxy Replacing Prior Agent, Intervening Event
A third situation arises when the patient wishes to appoint a proxy whose identity differs from a previously appointed proxy. A substantially higher level of scrutiny and the bar for certainty may be indicated here, because in this scenario clear evidence exists that the patient has previously given consideration to whom she preferred as her agent—and had chosen a person other than the proposed proxy. In such cases, it may prove helpful to investigate the reason for the patient’s changed preference. More deference should be afforded to the patient’s changed preference when an objective change has occurred in his circumstances. At the extreme, if the prior proxy has died or become incapacitated, the evaluator may consider this scenario as not calling for a significantly different approach than that used when no proxy has previously been appointed. More often, the change will be a divorce or estrangement—or, conversely, a marriage or the establishment of a new relationship. In fact, some states automatically invalidate a healthcare proxy upon divorce, but others do not.Footnote 45 As the significance of the alteration of circumstances becomes greater and its explanation for the change of preference more plausible, the argument for an increased level of scrutiny and higher bar for certainty prove less compelling.
Proxy Replacing Prior Agent, No Intervening Event
The fourth scenario that may arise is the case in which the patient has previously appointed an agent and now seeks to appoint a different agent without a significant alteration of circumstances. In this situation, clear evidence exists that the patient previously did not want the proposed proxy as her decision-maker, because she had identified another individual whom she preferred to fill that role. As a result, such requests should be evaluated with the most scrutiny and a high bar for certainty imposed. (Whether or not more or less regard should be afforded a proposed proxy who was previously named as an alternate or backup proxy is a complex question—on the one hand, that individual was previously contemplated and pointedly rejected; on the other, that individual was considered an appropriate agent by the patient, even if not the most desirable agent.) Overriding a proxy raises far more risk of undermining a patient’s authentic wishes than appointing one initially, and a sliding scale approach would argue for higher level of confidence under such circumstances.
Conclusions
Only one-third of American adults have completed an advance directive.Footnote 46 A far greater percentage—one estimate runs as high as 92%—discussed their end-of-life goals with their loved ones.Footnote 47 This disconnect suggests that there are many patients who will recognize the benefits of an agent or proxy once they are already ill and possibly when their cognition is limited. While both clinicians and legal authorities recognize that the level of capacity required to appoint a proxy can be lower than for clinical decision-making itself, they have not yet harnessed the sliding scale approach used in clinical capacity assessment for proxy appointment assessment. Yet all patients seeking to appoint a proxy are not similarly situated. While laws in some jurisdictions do establish a standard, none tell evaluators how closely to look or how certain they must be of their assessment. The hope is that by sorting particular cases into the categories discussed above prior to conducting an evaluation, and imposing a level of scrutiny (and, where legal, a standard of certainty) prior to assessment, the evaluator can help effectuate the patient’s wishes more than an assessment in the absence of such prior categorization.
