Introduction
Each year, approximately 500,000 people are diagnosed with dementia in the United States — a number projected to double by 2060.Reference Fang 1 Alzheimer’s disease is our seventh leading cause of death, 2 and the cost of dementia care in 2023 was estimated at $345 billion. 3 While diagnostic advances have been made for some forms of dementia, few therapeutic modalities are available, and there is a compelling need to develop new diagnostic and therapeutic modalities. 4 Dementia research presents unique ethical challenges, however, because cognitive deterioration impairs the capacity to consent to research participation, necessitating enrollment permission from someone other than the subject. Two approaches, sometimes used in conjunction, are available: obtaining permission from a surrogate and reliance upon prospective consent from the subject expressed in advance directives; neither is ethically adequate.Reference Harrison 5 Advance directives are often unavailable, and they rarely offer the specificity needed for ethically adequate guidance about research participation. Researchers most often rely on permission from surrogates, permission that is ethically questionable because the surrogate may lack knowledge of the study-specific values and preferences of the subject.
Supported decision making (SDM) is a process by which an adult person with cognitive impairment voluntarily enters into an agreement with another person(s), someone trusted by the impaired person, who will assist them with executing cognitive tasks that will allow them to continue to exercise self-determination.Reference Peterson, Karlawish and Largent 6 The impaired person (the principal) and the assisting person (the supporter) come to an agreement on the nature and extent of assistance the supporter will provide to the beneficiary. 7 SDM is recognized to some degree in thirty-nine states and the District of Columbia, 8 though none of the jurisdictions makes provisions for the use of SDM in clinical research. These variations between the provisions suggest gaps and inconsistencies that raise concerns for equity and justice. Initially advanced as an alternative to the ethically troubling aspects of guardianship, it has more recently gained attention as having value for persons with dynamic cognitive impairments. While the benefits for persons with dynamic cognitive impairment have been considered in the context of medical treatment, there has been little focus on its potential application to the enrollment of cognitively impaired subjects in clinical research. We argue that supported decision-making can be adapted for dementia patients who wish to participate in research but will lose the capacity to provide informed consent. Through early designation of a supporter — one willing to assist throughout the progression of the patient’s impairment — dementia patients can engage with that supporter (and eventual surrogate) on the very conditions that may lead to requests to enroll them in dementia research, ensuring that the surrogate will be well-informed about their values and preferences. This, in turn, will enable a more ethically defensible “substituted judgment”; that is, the surrogate will come nearest to making the decision the subject would want.
Our argument advances in three sections. We first examine the existing regulatory structure for surrogate consent to clinical research enrollment, demonstrating the inadequacies of uninformed surrogate consent. Our second section identifies weaknesses in reliance on advance directives, including those specifically designed for clinical research in dementia. Our third section offers an approach to the adoption of supported decision-making as a transitional approach to ethically optimal surrogate decisions, an approach we believe addresses many of the inadequacies of reliance on advance directives and uninformed surrogates.
Surrogate Decision-Making
In the United States, the legal framework for informed consent to research reflects the principles articulated in the Belmont Report 9 as codified in the Common Rule. 10 Regulations for the protection of human subjects advance the fundamental principle of respect for the values and preferences of the research subject, consistent with the ideal that no person should be used without their consent solely for the benefit of another. But many uncapacitated subjects in non-beneficial dementia research are being used solely for the benefit of another; therefore, the ethical importance of fully understanding a subject’s values and preferences is great. Despite the widely endorsed ethical importance of respecting the values and preferences of the research subject, the current regulatory structure tends to focus more on legal permission than on ethical adequacy. The placement of laws permitting surrogates to provide treatment consent reflects the original policy goals; they are often found in provisions exempting providers from liability for allegations of a lack of informed consent. 11 Laws framed to absolve a physician or other provider from medical malpractice liability (or battery) are not ideally suited to protect the interests of patients; indeed, one might argue that they have ethically conflicted goals.
Under the Common Rule, the surrogate is referred to as the Legally Authorized Representative (LAR). 12 For research proposals initiated after 2019, if a state has no law specific to surrogates for clinical research, LAR means “ … an individual recognized by institutional policy as acceptable for providing consent in the nonresearch context … to the subject’s participation in the procedure[s] involved in the research.” 13 There is no ethical dimension to the definition: no requirement, for example, that the surrogate have any knowledge of the values and preferences of the potential subject, no requirement that the surrogate be willing and available to remain involved after signing the consent form, and no requirement that the surrogate have capacity adequate to the enrollment decision. Few states have laws specific to surrogate consent for research participation;Reference Harrison and Trombetta 14 institutional policies often defer to state law for treatment decisions and, where the preferences of the subject are not known, the best interest of the subject. 15 This latter approach is particularly problematic, however; as we argue below, the idea of the “best interest” of an individual in non-beneficial clinical research is largely illusory, absent a specifically articulated desire of the subject. Where known, those desires deserve great respect, but the majority of potential study subjects have not articulated any desires. In their absence, the ethical foundation for surrogate permission is also illusory.
Despite the Common Rule requirement for additional protections for vulnerable subjects, Institutional Review Board (IRB) policies do not consistently provide adequate protection for subjects vulnerable due to cognitive impairments.Reference Harrison 16 IRB policies often defer to surrogate hierarchies for treatment decisions, 17 but they need not do so if they determine there is no applicable state law. Moreover, while the individual who is legally prioritized to make treatment decisions under state law may or may not be the person who is the LAR under institutional policy, there is no assurance that the person who meets institutional policy requirements will be the person most knowledgeable about the values and preferences of the patient. Indeed, it is uncertain the surrogate will have any knowledge of the patient’s values and preferences with respect to becoming a subject in a research study of the type proposed; we refer to such surrogates as uninforned surrogates. Under this framework, a subject lacking decision-making capacity for research enrollment has no assurance they will not be enrolled in research inconsistent with their values and preferences.
Long-standing reliance on surrogates for decision-making in treatment may have influenced research policies, but the ethical differences in treatment and research are highly significant and warrant distinct consideration. The distinction between treatment and research is so fundamental as to preclude the unquestioned adoption of treatment surrogates as an ethically acceptable approach in research. For cognitively impaired patients, recommendations for diagnostic and therapeutic interventions have as their sole purpose and justification the health care needs of the patient. The surrogate, often referred to as the “proxy” or the “next of kin” (colloquialisms that in themselves evidence views of who that person is or what their role is), is needed to provide legally effective permission such that treatment may proceed without claims of lack of consent. We tolerate the ethical inadequacies of treatment permission from a surrogate because the interventions for which permission is sought are evidence-based recommendations for the patient’s condition and are advanced by providers who are presumed to be exclusively devoted to the interest of the patient.
Unlike treatment, clinical research is not intended to confer any benefit on the research subject; it is by definition intended to lead to generalizable knowledge. The possibility that some subjects may benefit does not alter that fundamental distinction, especially with dementia research; most patients with advanced dementia will not survive long enough to benefit from research gains. The dementia patient will likely be serving as a test subject for modalities that, if ever approved, will not be available for years and will benefit, if anyone, unknown persons into the future. The legal device of obtaining surrogate permission for treatment, however ethically acceptable it may be, has no equivalent ethical defense in the research context.
The standard for judgments by an LAR is equally flawed. IRB policies typically advocate the use of substituted judgment, but many institutions have policies that allow or even instruct LARs to use a “best interests” test if they do not know the values and preferences of the subject.Reference Wright 18 If the proposal is for non-beneficial research (as most dementia studies are) and involves greater than minimal risk (as most dementia studies do), it cannot be in the clinical interest of a dementia patient to become a study subject. The sole ethically sound basis for a determination that it would be in the best interest of the patient to become a subject is where there is a prior unequivocal expression by a still-capacitated person expressing altruistic values and preferences that are aligned with a particular study. That cannot be confirmed on any basis other than direct knowledge of the study-specific values and preferences of the person. As the only ethically defensible best interest test requires that knowledge, no uninformed surrogate can provide ethically valid enrollment permission. 19
Surrogates appointed by a person under a durable power of attorney for health care or similar statutorily authorized instrument are conceptually preferable to default surrogates, but those persons may also be uninformed. The facially appealing difference is that the patient-appointed surrogate is exactly that: it is the patient’s choice and, therefore, presumably someone the patient trusts to make decisions. That veneer provides a reassuring gloss, but even a superficial scratch may reveal a lack of knowledge of whether the patient-cum-subject has altruistic values and preferences that are aligned with a particular study. It may be seductive to defer to the appointed surrogate because it is legally valid to do so and because, presumably, patients would not choose a surrogate if they did not trust them to make decisions. Even so, we cannot assume that the appointment of a decision-maker, especially if only specified for treatment decisions, involved any discussions about participation in clinical research. As with a default surrogate, the appointed surrogate has legal authority to decide, but that alone does not confer ethical validity on the decision. Indeed, even in treatment consent, patient-appointed surrogates are not consistently more accurate than statutory default surrogates.Reference Shalowitz, Garret-Mayer and Wendler 20
One further point about surrogate accuracy warrants consideration. In the treatment context, surrogates, more often than not, make decisions that align with patient preferences, though not overwhelmingly so. Shalowitz et al. reviewed 16 studies examining patient/proxy correspondence in a variety of clinical settings and found that patient-designated proxies and next-of-kin surrogates correctly predicted patients’ end-of-life treatment preferences in 68% of cases. 21 Significantly for our argument in favor of SDM, studies have shown that surrogates were least accurate in making decisions for dementia patients, and where they were not aligned with patient preferences, they often subjected the patient to more than they would have wanted.Reference Wilkinson, Wenger and Shugarman 22 These findings are consistent with our concern that an uninformed surrogate is likely to err in the direction of enrolling subjects in studies that would not align with their preferences.
Given that there are significant ethical concerns with reliance on uninformed surrogates for research consent, we turn to the ethical implications of the other established alternative, reliance upon advance directives.
Advance Directives
The term “advance directive” may refer to specified instructions from a person about the care they do or do not wish to receive, to a document that directs the appointment of a surrogate (the durable power of attorney for health care), or a document combing both features. These forms of extended decision-making evolved from the need for binding constraints on what might be done at the end of life. The Patient Self Determination Act, passed by Congress in 1990, required hospitals and other health care facilities accepting federal funds to inquire whether patients have an advance directive and to inform them of their rights under relevant state law. 23 However, at that time, state law was highly variable. In 1993, the Uniform Law Commission approved the Uniform Health-Care Decisions Act (UHCDA), 24 some form of which was ultimately adopted by most states. The UHCDA underwent significant amendment in July 2023 to reflect three decades of experience with advance directives.Reference Kohn 25 The amendments expanded the definition of decision-making capacity to include decisions made with support, thus acknowledging a role for SDM. However, despite decades of experience, advance directives for treatment remain subject to criticism, with studies finding them ineffective. 26
A more recent development is the promotion of dementia-specific advance directives. Primarily designed for treatment decisions, they have faced criticism.Reference Dresser 27 Daniel Sulmasy concludes they are “misguided [and are] at best, a rough guide to the patient’s authentic values … ” and suggests that surrogates appointed under durable powers of attorney for health care are preferable to a mere advance directive.Reference Sulmasy 28 Sulmasy argues that an appointed surrogate “can narrate the story of the person’s life, providing insight into what the person with dementia has always valued most, sharing who and what the person has loved, what beliefs and commitments she has held, what values have governed her life, what she has hoped for and feared, both in the face of a debilitating disease and overall.” 29 That is true, of course, only where there has been an extended personal intimacy between the surrogate and the patient, the kind of intimacy that would arise from a SDM relationship. Where the surrogate has been a supporter, we agree that the surrogate “can help determine what the real interests of this unique patient actually are in these circumstances and help to make a judgment that is authentic to the person the patient is.” 30
An alternative with even greater specificity of purpose is the research advance directive; it allows capacitated persons to express altruistic wishes for participation in clinically non-beneficial research on the neurodegenerative disease they are facing. Research advance directives are not a recent idea; they have been proposed at least since the late 1980s.Reference Heinrichs 31 Berghmans, in his 1998 argument about the limited value of research advance directives, referenced nineteen papers, though only four originated from the 1980s; the remaining fifteen were pubished in the mid-1990s.Reference Berghmans 32 This early interest in research advance directives for cognitively impaired patients did not gain significant traction; in a pair of reports on advance directives in dementia issued by the Department of Health and Human Services in 2007, research applications were not even mentioned.Reference Stein 33 More recent scholarship has remained limited.
Berghmans notes several difficulties with research advance directives: (1) they are not likely to be completed very often; (2) for them to be completed while the person still has adequate capacity, there must be earlier detection of advancing dementia, and early screening and detection of an untreatable disease is itself not morally neutral; (3) for a person to give valid advance consent for a study involving greater than minimal risk, they must have a high level of decision-making capacity, one not necessarily extant for those with even mild cognitive impairment; (4) there will be difficulty in formulating specific values and preferences in terms that are neither too vague nor too restrictive; and (5) they are subject to the philosophical argument that the person who completed them is no longer the same person as the subsequently incapacitated person, therefore the research advance directive no longer has moral authority because it was executed by a person who no longer is. 34 We believe that each of these criticisms can be addressed in a way that advances our argument for decision support as the best preparation for ethically adequate surrogate decision-making.
The low rate of completion for advance directives is perhaps the concern least well-addressed by SDM, at least to the extent advance directives may not be completed due to reluctance to engage thoughtfully in advance with difficult decisions. This reluctance may be mirrored in efforts to reach SDM agreements. The solution may lie in leveraging an earlier dementia diagnosis to encourage clinicians to advocate for care planning discussions that include research participation. The question of adequate decision-making capacity for executing a directive is important, but Berghmans’ concern may be mitigated by more recent findings. Persons with dementia may retain decision-making abilities for years after a dementia diagnosis; a meta-analysis of studies of decision-making concludes that “a diagnosis of dementia does not imply incapacity.” 35 Even individuals in moderate to severe stages of dementia can “reliably report on their care values and preferences, well-being, and quality of life.”Reference Wilkins 36 This emerging evidence challenges the presumption that diminished cognitive function negates the validity of advance directives and, by extension, reinforces the potential role for SDM in bridging the gap between early-stage preferences and later-stage participation.
As with treatment advance directives, research advance directives suffer from disclosure limitations. One useful framework distinguishes “type” disclosures, in which the general type of research is disclosed in the document, and “token” disclosures, in which the potential subject is given information about the specific trial or trials.Reference Scholten 37 In research advance directives, token disclosure would be virtually impossible because it would be impossible to know in advance what specific trial or trials will be underway when the person loses capacity. Type disclosure, on the other hand, provides far less specificity than is required for obtaining informed consent from persons who retain capacity. 38 Heinrichs persuasively argues that this creates an ethically troubling double standard. 39 We agree that the inability to anticipate future research projects is a significant limitation of all advance directives.
Reliance on earlier diagnosis as a way to engage patients while they still have adequate capacity would be more susceptible to Berghmans’ concern about moral neutrality if diagnostic testing on asymptomatic patients were pursued solely for the purpose of facilitating SDM (an approach we do not endorse), but where earlier diagnosis is the result of a clinically indicated response to signs and symptoms, the practical and emotional advantages may be significant. Dementia is relational in that it profoundly affects caregivers and family members who must navigate its progressive demands. Therefore, the opportunity to prepare can provide meaningful benefits that mitigate the moral objection that early discussion subjects a person to emotional distress without offsetting benefit. Moreover, appropriately asking in advance about research preferences does not violate the principle of respect for persons or threaten the interests of vulnerable persons. Rather, it serves a morally positive function from the perspective of justice in affording persons who would like to participate in research a means to do so.
Among Berghmans’ objections, perhaps the most ethically challenging is the “someone else” problem, the argument that where significant cognitive impairment causes profound changes in personality, the impaired person is a new and different person. Two competing positions are often associated with Ronald Dworkin, on the one hand, and Derek Parfit and Rebecca Dresser on the other. Dworkin argues that becoming demented is exactly that, a process of becoming, and that the process reflects points along the continuum of a single life.Reference Dworkin 40 Dworkin further posits “critical interests,” interests having more fundamental and therefore greater moral weight than experiential interests acquired along this continuum; 41 it is these critical interests from which core values and preferences arise, and therefore the values and preferences we respect in making non-beneficial research participation decisions are tied to and align with these critical interests more than what the person experiences during the loss of capacity. Dworkin’s view is more widely endorsed.
Parfit argues that severe neurological deterioration disrupts the psychological continuity necessary to defend the single-life construct.Reference Parfit 42 Sometimes referred to as “the someone else problem”Reference DeGrazia 43 or “numerical personal identity” 44 (Parfit also uses the terms “series persons” and “successive selves” 45), the view is essentially that the person who loses capacity becomes a different person. Applied to advance directives, the argument is that where neurological changes have been so great as to break the psychological continuity of a person, an advance directive no longer has any moral validity because it was written by a different person. The premise that an advance directive executed by one person has no moral validity for another person is unassailable, but the second premise, that neurological damage may cause so much of a psychological break in the continuity of personhood as to render the current person someone morally different, is open to criticism.
Though Parfit was first in time, Dresser’s position is perhaps better known because of her direct response to Dworkin; describing Dworkin’s view as “elegant theory,” Dresser questions whether a person who has not yet undergone profound psychological transformation can know what they will want post-transformation.Reference Dresser 46 Her argument, like Dworkin’s, is focused largely on advance directives for dementia patients in end-of-life settings, but we acknowledge relevant parallels to the context of enrolling dementia patients in research.
Responding to Parfit and Dresser is beyond the scope of our argument, but even if they are correct that dementia can produce a fundamental rupture in personal identity, that rupture remains speculative and should not be used to justify overriding the clearly expressed wishes of a capacitated person. In the context of dementia research, we believe it goes too far to deprive a person of a clearly expressed altruistic interest in participating in studies of the disease that affects them so personally. Paul Menzel expresses it eloquently: “[W]e should never treat a person who has had a whole life in which she was often competent as if she were a never-competent person…. The patient before us has a past that even now is part of their identity. The then-self/now-self problem does not go away, but neither does it destroy a directive’s moral weight.”Reference Menzel 47 The presence of a trusted supporter participating in the person’s process of becoming serves as a crucial bridge between the individual’s past and present selves, thereby ensuring continuity of agency. By recognizing and integrating shifts in identity, the supporter extends the individual’s cognition onto the future.
Emily Walsh advocates for deference to the present preferences of dementia patients where they conflict with an earlier advance directive.Reference Walsh 48 Although Walsh does not mention research consent, her analysis seems aligned with our argument. Describing dementia as a “cognitive transformative experience … one which alters a person’s cognitive capacities in such a way that may change the way the person thinks about their preferences, values, and beliefs,” 49 Walsh posits the alterations as gradual rather than radical: “[D]ementia patients do not wake up a radically different kind of being. Instead, they have gradual preference revelations dependent on which stage of dementia they find themselves in.” 50 When the advance directive is coupled with an appointed surrogate who has served as a trusted decision supporter throughout the individual’s progression of cognitive decline, that supporter is uniquely positioned to monitor preference revelations that may shape or refine previously expressed values and preferences regarding research participation. This continuity ensures respect not only for the person they once were, but also for the person they have become. By protecting the person who underwent this transformative progression from a reductive “then versus now” characterization, the supporter-as-surrogate preserves both retained and emergent preferences; this approach preserves the individual’s identity over time and counters the perception that cognitive impairment has rendered them a different person.
Transition from Decision Supporter to Informed Surrogate
Having framed SDM as an ethically preferable alternative to the inadequacies of uninformed surrogates and (standing alone) advance directives for dementia research, we are in a position to articulate our argument for the adoption of SDM as transitional to surrogate decision-making. Essential to this approach is that the supporter be legally recognized as the prioritized surrogate. This recognition ensures that the decision supporter during partial cognitive impairment is legally prioritized as the person best qualified to serve as a surrogate once a full transfer of decisional authority is necessary. Although the adoption of SDM may not alter the point at which a person no longer has cognitive capacity adequate to make decisions about research enrollment, it may extend the time during which that person can form an understanding of the research proposal; whether that is a cognitive consideration or a more practical consideration of the availability of a supporter having the time needed to go over complex information repeatedly is an open question for further research. Accordingly, for consent purposes, the supporter should have legal priority over all other statutorily identified surrogates.
We suggest five reasons why a decision supporter chosen well in advance of loss of capacity is the ethically ideal surrogate in the research context. First, that person is appointed by the patient while they still have the capacity to do so, evidencing the patient’s trust in the appointed person. Second, during their involvement throughout the progressive cognitive deterioration, the supporter develops an intimate knowledge about the patient’s values, preferences and wishes. Third, in addition to having legal authority, the supporter will have greater credibility with future researchers. This is particularly true if the supporter is appointed early in the progression of the disease and has time to demonstrate their psychological intimacy and practical involvement with the patient. Fourth, unlike statutory default surrogates who may be suddenly assigned with little or no prior involvement, the supporter has already demonstrated a sustained commitment to the protection of the subject’s well-being during treatment. And fifth, though we concede that there are practical limitations, early identification of the supporter allows for targeted training in the ethical responsibilities and facilitates a gradual adapting to the surrogate role as the patient increasingly depends on their support.
Our literature review did not identify formal proposals for the supporter-to-surrogate transition we advocate, though Bierer and colleagues argue for the integration of supported decision-making into clinical research;Reference Bierer 51 we believe they are the first to suggest the value of SDM as a transition to surrogate decision-making in the context of research. Our findings are consistent with Wright’s broader observation that there is not yet good research on how the concept has been used in dementia. 52 While we believe we are the first to explicitly advance this transition for research surrogates, a more general SDM framework advanced by Peterson et al. includes three steps that we adapt to advance our conception: identifying the domains of life in which support will be needed or desired; identifying the kinds of support that will be desired; and establishing a formal decision-making agreement. 53 We build on this foundation by explicitly adding a fourth step, the early selection of an appropriate supporter.
Our first step begins with the early choice of a suitable supporter. Ideally, this first step is triggered by the occurrence of a clinical event — such as a diagnosis of mild cognitive impairment (MCI) or early onset dementia — that awakens awareness of the value of supported decision-making. A thoughtful discussion with a healthcare provider at this stage can serve as an ethical and practical trigger for introducing SDM. However, because many patients with MCI do not develop dementia and may never need support with decisions, research is needed to refine the diagnoses or stages in diagnosis that would be an ethically appropriate trigger for SDM conversations between patient and provider. Once a diagnosis of dementia is confirmed, however, we argue it is ethically appropriate for providers to engage patients in a discussion of SDM and encourage them to discuss SDM with family or other individuals who may serve as supporters and future surrogates. This first step is critical in the research context because, as we argue above, there is — other than previously expressed atruism — no ethically defensible “best interest” that justifies enrolling a vulnerable person in non-beneficial studies with greater than minimal risk. The supporter-as-surrogate for research must have intimate knowledge of the beneficiary’s values and preferences, maintain consistent involvement throughout the progression of cognitive decline, and be available for continued involvement on research-related decisions and participation. Also important are the often-neglected criteria of ensuring that the supporter-as-surrogate is someone who has and is expected to retain decision-making capacity adequate to the role 54 and candid consideration of whether a potential supporter may have foreseeable conflicts of interest with respect to research enrollment consent. 55
Once a supporter is chosen, our second step is to identify clinical research as among the domains of life in which the beneficiary seeks support. While we do not advocate for a separate SDM agreement for research purposes, we emphasize the need for a clear delineation of research enrollment within a broader SDM framework. At this stage, the beneficiary should have the opportunity to direct the supporter to decline research participation if that aligns with their preferences. This safeguard is the best way to avoid false-positive enrollment decisions by surrogates and ensures that any participation aligns with previously expressed wishes.
The third step in our model is the specification of the types of research protocols in which the beneficiary wishes to enroll. Here, a clear delineation must be made between treatment consent and research enrollment consent. Researchers and IRBs will have a greater level of confidence in the decision of a supporter-as-surrogate if the agreement clearly specifies the types of studies in which the subject wishes to participate. Given the inherent unpredictability of future research, token-level details will be difficult to predict with adequate specificity, necessitating a structured approach to type-level specification. The beneficiary may have an open-ended wish to be enrolled in any type of dementia study, but it is intuitive that such an unspecified open-ended grant would be less confidently viewed years later by family, researchers and IRBs. Little additional effort would be involved in specifying types to ensure future confidence in the ethical validity of the agreement. For example, a subject may wish to participate only in minimal-risk studies or may be willing to assume greater than minimal risk; they may be willing to enroll in drug studies, or perhaps only certain kinds of drug studies; they may want to be involved in imaging studies with or without contrast; they may wish to avoid studies that necessitate extensive travel or appointment commitments for supporters, or studies they know would cause emotional distress for their family or others important to their lives. This degree of type specificity strengthens the legitimacy of future surrogate decisions and adds to the confidence researchers and IRBs will have in the discretion exercised by the supporter-as-surrogate.
Our fourth step is the formal memorialization of the subject’s wishes in a legally recognized written agreement. Peterson et al. emphasize the importance of a formal agreement that “provides important moral grounding [in that] it creates expectations of duty [that if not met] could result in moral wrongs.” 56 They also underscore its value in providing clarity for third parties such as clinicians or attorneys so that they “understand the parameters of the supported decision-making agreement and place faith in the beneficiary’s decision.” 57 For our model, that value is extended to fostering confidence in the supporter-as-surrogate’s decisions. Legislative recognition of a formal agreement is also crucial; the prioritized authority of the supporter-as-surrogate must be clear. Further work is needed to clarify the optimal contours of these agreements, but we argue that, at minimum, they should include: a clear statement of the principal’s intent that the agreement remains valid after the principal loses decision-making capacity and that the supporter will become the surrogate; clear directives about the types of research in which the principal wishes (or does not wish) to participate; and a presumption against enrollment in any research that does not align with the stated values and preferences of the principal.
Lastly, we note that although the adoption of SDM may not alter the point at which a person no longer has cognitive capacity adequate to make decisions about research enrollment, it may extend the time during which that person can form an understanding of the research proposal and therefore meangfully participate in the enrollment decision. Whether that is a cognitive consideration or a more practical consideration of the availability of an informed supporter having the time needed to go over complex information repeatedly and facilitate the potential subject’s understanding is an open queston for further research.
Conclusion
Supported decision-making offers an ethically positive alternative to existing frameworks for obtaining permission to enroll cognitively impaired subjects in clinical research. For the subject who retains some decisional capacity, supported decision-making maximizes relational autonomy. For the subject who no longer has the capacity for the decision to be made, a surrogate who has been a decision supporter provides the most ethically reliable approach — respecting the subject’s values and preferences while addressing the ethical fiction that surrogate permission is the same as the subject’s informed consent. A structured framework for supported decision-making for research should be advanced and the legal priority of appointed decision supporters over other statutory default surrogates should be codified in all states.
It is important to clarify that our advocacy for early adoption of decision supporters is not driven by a desire to expand study cohorts. While broader enrollment is a potential outcome — one that would be welcomed if ethically appropriate — our primary focus is on protecting the interests of subjects vulnerable due to cognitive impairment. We are focused on protecting them from unwanted enrollment and from the denial of participation in studies consistent with their preferences. In this sense, our approach aligns with Eric Fromm’s distinction between “freedom from” and “freedom to.”Reference Bergsma and Thomasma 58 We also draw from the perspective of bioethicist Paul Ramsey, with whom we focus our attention always “upon the [person] who is the patient in all procedures and the partner in all investigations, and away from that celebrated ‘non-patient,’ the future of medical science.”Reference Ramsey 59 Vulnerable subjects deserve nothing less.
Disclosures
The authors have nothing to disclose.
Open access
