I. INTRODUCTION
Advances abound in biotechnology and biomedicine, and cutting-edge clinical research is transforming medical practice by the minute.Footnote 1 While benefits are quite likely, biomedical research with people who suffer from mental illnessFootnote 2 must include enhanced protections to guard against harms to this potentially vulnerable population. Notions of harmful, abusive practices in psychiatric research and practice bring to mind the infamous eugenics movement of the early twentieth century;Footnote 3 unfortunately, however, abuse is not merely a thing of the distant past. In 2003, a young man with severe mental illness allegedly was coerced by his psychiatrist, against the wishes of his mother, to participate in “the CAFÉ study,” a clinical trial comparing atypical antipsychotics, including Seroquel, in first-episode psychosis; the study was conducted within the University of Minnesota Department of Psychiatry and sponsored by AstraZeneca, the manufacturer of Seroquel.Footnote 4 In 2004, while still enrolled in the study and taking Seroquel, the young man committed suicide by violent and gruesome means.Footnote 5 The State of Minnesota commissioned a 2015 report that showed evidence of coercive enrollment, inadequate supervision of both subjects and study personnel, conflicts of interest, and numerous misrepresentations and unethical cover-ups.Footnote 6 Though some bioethics and mental health policy professionals have signaled that human research protections for mentally ill people are insufficient,Footnote 7 any efforts to remedy deficiencies repeatedly miss the mark. The United States has never promulgated regulations or prescribed strict guidance on concrete approaches to addressing mental health vulnerabilities in clinical research; rather, vulnerability determinations have rested “on a common tradition of interpreting the regulations.”Footnote 8 This practice continues despite recent changes to the Common Rule.Footnote 9
Shortcomings in the regulatory management of biomedical research with people who suffer from mental illnesses are evident, especially considering the inclination of researchers, in practice, to narrowly focus on a participant’s consent-based decision-making capacity.Footnote 10 An exclusive focus on a mentally ill person’s decision-making capacity—as measured by standardized assessment instruments such as the MacCAT-CR)Footnote 11—when contemplating their vulnerability to coercion or undue influence ignores other sources of vulnerability for this population. These vulnerabilities include the lived experience of low self-esteem resulting from stigma,Footnote 12 sometimes being fearedFootnote 13 or discriminated against,Footnote 14 and social isolation.Footnote 15 Other contributions to vulnerability in this population include underemployment or unemployment,Footnote 16 housing insecurity,Footnote 17“limited availability of providers, the affordability of health insurance, and the cost of drugs.”Footnote 18 Additional sources of vulnerability may include the mental illness itself and psychotropic drugs used to treat the illness, both potentially altering the patient-participant’s authentic self through their effects on thought processes and choices.Footnote 19 Further, the power imbalance within the psychiatrist-researcher patient-participant relationship—sometimes referred to as “power-relational vulnerability” or deferential vulnerabilityFootnote 20—is another source of vulnerability that may go unrecognized.Footnote 21 Given the history of stigmatization and abuse against mentally ill people, both outsideFootnote 22 and within the research context,Footnote 23 regulations and guidance should impose heightened protections for this population to ensure proper enrollment and continued participation in biomedical research.
To say that all potential participants with mental illness are vulnerable to coercion or undue influence casts too wide a net.Footnote 24 Yet, an approach that views all vulnerabilities to coercion or undue influence emanating from mental illness as arising only in the context of decision-making capacity, at least as this term currently is assessed and applied in practice,Footnote 25 is far too narrow. A middle ground is needed to appropriately protect people with mental illness in biomedical research. This issue is complicated and will surely require complex solutions.
This Article does not set out to solve all the problems, but it is an effort to start a conversation, offer observations on the problematic application of the Revised Common Rule (“rCR”) in the context of mental illness, and make some initial recommendations. Further, this Article primarily addresses research with psychiatric patients who are not involuntarily hospitalized or experiencing active psychosis and whose decision-making capacities are not obviously compromised.Footnote 26 Part II of this Article provides a brief history of the Common Rule and relevant changes in the rCR, highlighting the lack of clarity on the meaning and parameters of “impaired decision-making capacity” and who is “vulnerable to coercion or undue influence” in the context of mental illness. Challenged here are assumptions that impaired decision-making capacity, as currently assessed and applied in practice, is and must be the rCR’s exclusive framework for assessing vulnerability to coercion or undue influence in potential participants suffering from mental illness. Part III discusses how the long-awaited changes to the Common Rule provide little clarity on who is vulnerable to coercion or undue influence outside four specifically named categories, what vulnerability of this type looks like, or how such vulnerability should be determined. Part IV maintains that absent firm leadership and guidance from the U.S. Department of Health and Human Services (“HHS”), expertise is needed regarding the unique circumstances of participants in psychiatric research. Part IV also contemplates sources of expertise, e.g., the psychiatrist who treats mental illness, the patient who suffers from and lives with mental illness, and the mentally ill person who already has served as a subject in biomedical mental health research. Part V suggests new language for binding guidance or further rule revisions. A final suggestion is to modify one scholar’s proposal to include research subjects in the research review process, proposing that if specific federal research regulations and guidance do not materialize, the next best alternative is to focus on who is at the institutional review board (“IRB”) table interpreting existing regulations.
II. THE COMMON RULE THEN AND NOW: CHANGES PROVIDE LITTLE CLARITY
A. Efforts to Solidify Protections
Despite several attempts between the 1970s and 1990s to provide greater protections in biomedical research for people suffering from “mental disorders,”Footnote 27 the Common Rule’s codification in 1991Footnote 28 made no mention of people who suffer from mental illness and only mentioned the “mentally disabled.”Footnote 29 In 1996, the National Bioethics Advisory Commission (“NBAC”) began working to study the issue of “mental disorders” in the context of research participation and to formulate possible solutions to address voids in protections.Footnote 30 In December 1998, NBAC issued its report, Research Involving Persons with Mental Disorders That May Affect Decisionmaking Capacity.Footnote 31 NBAC limited its inquiry to decision-making capacity, ignoring other types of vulnerability arising in mental illnesses or “mental disorders.”Footnote 32 NBAC’s first recommendation related to IRBs and how, in order to close gaps and extend appropriate protections to people suffering from “mental disorders,” IRBs should be tailored to properly represent these potential participants’ interests.Footnote 33 Specifically, Recommendation I states:
All IRBs that regularly consider proposals involving persons with mental disorders should include at least two members who are familiar with the nature of these disorders and with the concerns of the population being studied. At least one of these IRB members should be a member of the population being studied, a family member of such a person, or a representative of an advocacy organization for this population.Footnote 34
Official guidance based on these 1998 recommendations would have been appropriate, but federal regulators issued no guidance specifically limiting risks to which mentally ill participants might be exposed in biomedical research.Footnote 35
Due to the heterogeneity of people who suffer from mental illnesses, and the variability of mental illnesses even within diagnostic categories, a detailed, one-size-fits-all rule may have been beyond realistic expectation.Footnote 36 But the fact that a task is challenging does not warrant avoiding or abandoning attempts to address it. Nearly thirty years after implementation of the original Common Rule, and over twenty years since the NBAC report, rather than mending inadequacies, the recent changes under the rCRFootnote 37 have made an already flawed regulatory framework even more out of sync with safeguarding people with mental illness who may need heightened protections in biomedical research. One major shortcoming of the rCR is that it gives IRBs more responsibility and discretion in deciding what it means to have “impaired decision-making capacity,” to whom that criteria should apply, and under what circumstances.Footnote 38 This approach allows for inconsistencies based on who is sitting at the IRB table on any given day, at any given institution. As discussed later, the IRB can play a significant role in addressing what is lacking in the rCR,Footnote 39 so increased IRB responsibility should not be criticized. Yet, IRB discretion in carrying out increased responsibility ad hoc should be curbed, rather than expanded, through binding guidance or modified regulation.
B. Absence of “Mental Illness” in Original and Revised Common Rule
Very early in the rulemaking process, drafts of what would become the original Common Rule referenced the “mentally infirm,”Footnote 40 though ultimately this term was deleted from the final rule. The term “mentally disabled” is the closest approximation to mental illness in the original Common Rule.Footnote 41 Who was “mentally disabled” or what constituted a mental disability was never properly defined under the original rule, and whether or not it was intended to include people with mental illness is unknown.Footnote 42 The Americans with Disabilities Act definition of “disability” includes a “mental impairment that substantially limits one or more major life activities of an individual,”Footnote 43 yet the term “mental disability” has many mainstream connotations, including congenital brain abnormalities and cognitive deficits from brain damage or neurodegenerative disease.Footnote 44 The mystery as to the meaning of “mentally disabled” in the Common Rule is moot, as this term was stricken in the recent revision, shown here:
45 C.F.R. § 46.107 IRB membership.
(a) … If an IRB regularly reviews research that involves a vulnerable category of subjects that is vulnerable to coercion or undue influence, such as children, prisoners, pregnant women, or handicapped individuals with impaired decision-making capacity, or economically or mentally disabled educationally disadvantaged persons, consideration shall be given to the inclusion of one or more individuals who are knowledgeable about and experienced in working with these categories of subjects.
…
(£e) An IRB may, in its discretion, invite individuals with competence in special areas to assist in the review of issues which that require expertise beyond or in addition to that available on the IRB. These individuals may not vote with the IRB.Footnote 45
Other notable changes in the rCR include the addition of those with “impaired decision-making capacity” as one example of individuals who may be vulnerable to coercion or undue influence, as well as instruction regarding possible IRB composition.Footnote 46 Mental illness is mentioned nowhere.Footnote 47
C. Impaired Decision-Making Capacity: An Exclusive Lens For Viewing Mental Illness?
The new but likewise vague concept “impaired decision-making capacity” now substitutes for “mentally disabled.”Footnote 48 The rCR groups individuals with impaired decision-making capacity with other categories of prospective subjects, “such as children, prisoners, … [and] economically or educationally disadvantaged” people, when describing the types of subjects meant to be included in those “vulnerable to coercion or undue influence.”Footnote 49
Before the Common Rule revisions, relevant literature and analyses of the propriety of enrolling mentally ill people in research rested primarily on concerns about decision-making capacity.Footnote 50 In both the original rule and the rCR, however, the use of “such as” in 45 C.F.R. §§ 46.107, 46.111 indicates that the list of vulnerable groups is neither exhaustive nor final.Footnote 51 The rCR allows for subjects who suffer from mental illness but who do possess traditional decisional capacity to not be considered vulnerable in the clinical research setting. But, intended or not, the rCR’s “such as” language also implies that factors besides impaired decision-making capacity may affect a person’s vulnerability to coercion or undue influence; that is, sometimes people who suffer from mental illness may fairly be considered a vulnerable, unnamed group under the rCR for reasons other than impaired decision-making capacity. Excluding bases outside traditional notions of decisional capacity, as it is assessed and applied in practice, in determining vulnerability to coercion or undue influence in this population makes simplistic and even harmful assumptions about the experience of mental illness.Footnote 52 A mentally ill person might possess traditional decision-making capacity based on a standard assessment and still be vulnerable to coercion or undue influence in biomedical research, and, thus, need additional, heightened protections.Footnote 53
D. Why “Impaired Decision-Making Capacity” is too Limited
Under the rCR, specific categories of subjects deemed to be per se vulnerable to coercion or undue influence—children, prisoners, and economically and educationally disadvantaged peopleFootnote 54—are readily discernable: a child is a person under age 18, a prisoner is an incarcerated person, and determining a person’s financial means and education level is a straightforward task. The same cannot be said of the category “individuals with impaired decision-making capacity.”Footnote 55 Categories of individuals vulnerable to coercion or undue influence can extend beyond those named in the rule; but, of those explicitly listed, “individuals with impaired decision-making capacity” is the one category with uncertain meaning. What about a person who has impaired decision-making capacity makes them vulnerable to coercion or undue influence? Where does mental illness fall within this category, if at all? Should “mental illness,” standing alone, qualify as a non-enumerated category left open under the “such as” language in 45 C.F.R. §§ 46.107, 46.111?Footnote 56 As stated above, if a person is mentally ill but has decision-making capacity, they may still be vulnerable to coercion or undue influence under the rCR.
Numerous sources of vulnerability to coercion or undue influence for those suffering from mental illness are less apparent than common understandings of impaired decision-making capacity, yet no less important.Footnote 57 Living with mental illness in a society that misunderstands, fears, and stigmatizes mental illness gives rise to all sorts of vulnerabilities to coercion or undue influence.Footnote 58 Some of the symptoms experienced by people suffering from mental illness include despair, guilt,Footnote 59“ambivalence, diminished motivation, disengagement, or impulsivity,”Footnote 60 somatic disturbances, and suicidality.Footnote 61 Major depression causes “diminished ability to think or concentrate, or indecisiveness.”Footnote 62 In 1997, Professor Carl Elliott aptly wrote: “When a person is caught in the grip of depression, his values, beliefs, desires, and dispositions are dramatically different from when he is healthy. In some cases, they are so different that we might ask whether his decisions are truly his.”Footnote 63 In fact, because “symptoms can cloud a person’s judgment,” psychiatrists often “advise patients not to make major life choices while experiencing a depressive episode.”Footnote 64 Further, psychotropic drugs used to treat mental illness may affect thinking, behavior, values, judgment, and risk-taking assessments—in essence, these drugs may alter one’s personality, for better or worse—thereby distorting the potential research subject’s true self.Footnote 65 Furthermore, mental illness can lead to extreme isolation and loneliness,Footnote 66 and participation in biomedical research might provide an inappropriate or false sense of belonging and fellowship in a time of psychiatric crisis or ongoing distress.Footnote 67 Ultimately, “even the most well-informed and decisionally ‘fit’ individual cannot realize the spirit of informed consent if his or her ‘choice’ is inauthentic, symptom driven, compromised, or coerced.”Footnote 68
Many scholars concur that people suffering from mental illness or other conditions that impair decision-making capacity should not be enrolled in biomedical research without “sound scientific justification” for their inclusion.Footnote 69 Others believe that “[r]esearch decisional capacity” should not be determined by an assigned diagnosis or group membership, “but by characteristics unique to that person.”Footnote 70 Another body of literature questions whether people suffering from mental illness need any protection against their own potential decision-making deficiencies.Footnote 71 In all the above examples, the myopic emphasis on potential participants lacking or possessing decision-making capacity—and the narrow definition of decision-making capacity and common beliefs about its assessment in people with mental illness—ignores all other sources of vulnerability to coercion or undue influence in this population. Traditional decision-making ability may appear intact, but may be affected by other vulnerabilities to coercion or undue influence, including the mental illness, the influences of psychotropic drugs, and imbalances in power dynamics experienced between the patient-participant and psychiatrist-researcher.Footnote 72 These factors may go completely unrecognized in the widely accepted assessment of decision-making capacity in research, the MacCAT-CR.Footnote 73
The rCR leaves open how people with mental illness should be viewed in the context of vulnerability to coercion or undue influence.Footnote 74 To cabin the mentally ill subject’s vulnerability within traditional notions of decision-making capacity is short-sighted, though it is the trend in both the literature and in practice.Footnote 75 The rule’s indecisiveness, however, provides an opportunity to expand protections and finally close the gaps originally identified by NBAC.Footnote 76 Suggested solutions on how to accomplish this are set forth later in this Article.
III. WHAT IS VULNERABILITY TO COERCION OR UNDUE INFLUENCE, AND WHO IS VULNERABLE IN THIS CONTEXT?
A. Common Definitions
“Coercion” and “undue influence” find their origins in the law.Footnote 77 The concept of “vulnerability,” however, is rooted much more in ethics and social sciences than in law.Footnote 78 The Common Rule and the rCR do not define any of these terms,Footnote 79 but reaching a clear understanding of their meanings and implications in biomedical research is essential. Yet, many IRBs appear to use the terms “coercion” and “undue influence” interchangeably, without differentiation.Footnote 80 It has been argued that monitoring traditional notions of overt “coercion” and “undue influence” in research is insufficient, because other factors—social, cultural, economic, and political realities—may make research participation less voluntary and more a necessity.Footnote 81
1. Coercion and undue influence in relation to mental illness
Though “coercion” and “undue influence” are undefined in the Common Rule and rCR, on its website, HHS discusses the need for investigators to avoid engaging in coercion and undue influence.Footnote 82 Coercion is described as follows:
Coercion occurs when an overt or implicit threat of harm is intentionally presented by one person to another in order to obtain compliance. For example, an investigator might tell a prospective subject that he or she will lose access to needed health services if he or she does not participate in the research.Footnote 83
Professor Jill Fisher asserts that limiting the concept of coercion to intentional acts fails to address the societal factors that force individuals to make choices they might otherwise not, absent their adverse circumstances.Footnote 84 The concept is described as “structural coercion” that has the potential to undermine the propriety of enrolling certain subjects in research.Footnote 85 Presented in the context of a person becoming a research participant out of desperation for a cure, for the purpose of obtaining money, or accessing general health care, the idea of “structural coercion” easily relates to psychiatric patients in research. Structural coercion takes into account that “broader power dynamics” outside the researcher-subject relationship “might infringe on the voluntariness of consent”Footnote 86 and prospective research participants might actually “turn to research in order to mitigate the threat of structural violence.”Footnote 87 Again, Fisher presents structural coercion in the context of monetary incentives for participation, or participation to obtain basic health care or a miraculous cure, exploring the space where societal inequalities may limit opportunities and choices for potential research participants.Footnote 88 In this setting, there is something tangible, e.g., money, health care or experimental “treatment,” that the participant lacks and desperately needs.Footnote 89 Structural coercion aptly applies to people who live with mental illness, a population that experiences housing insecurity, underemployment and unemployment, and insufficient treatment options.Footnote 90 In the case of the psychiatric patient-participant, beyond a need for sources of income, mental health care and treatments, reasons for participating in research might include a sense of belonging to something, a yearning for human interaction, and even feelings of friendship (e.g., with study staff)Footnote 91 after years of social isolation.
HHS describes “undue influence” as arising in the context of inappropriate rewards, but also as present in other instances in more subtle form.Footnote 92 For example, HHS notes that where undue influence depends on a specific individual’s situation, “it is often difficult for IRBs to draw a bright line delimiting undue influence” and that the IRB must “use its discretion in determining which circumstances give rise to undue influence.”Footnote 93 This task becomes difficult because IRBs in most circumstances do not evaluate a particular participant’s vulnerability to undue influence, and influence or “justifiable persuasion” is not per se unacceptable in research enrollment.Footnote 94
The power imbalance between the patient-participant and psychiatrist-researcher may unduly influence a participant’s choices.Footnote 95 This possibility is not surprising, as common sense suggests that the psychiatrist holds sway over the patient’s decisions about psychiatric interventions.Footnote 96 For example, a participant in a qualitative study eliciting opinions from psychiatric patients about deep brain stimulation (“DBS”) testified: “I have a psychiatrist who I have been with for a long time who has been prescribing medications for me and treating me and, now if he were to advocate for [DBS], and he had good reasons, that would be more persuasive.”Footnote 97 This type of trust and deference can be expected to exist in most psychiatric clinician-patient and researcher-subject relationships, implicating concerns about voluntarism in consent.Footnote 98 Moreover, some scholars suggest an additional vulnerability: the distrust those suffering from mental illness often have in themselves.Footnote 99 True autonomy is hindered “when agents ‘doubt their own worth and ability to make appropriate choices.’”Footnote 100 Without considering this additional aspect of vulnerability—a lack of self-worth and self-trust—studies on how to counteract vulnerabilities of mentally ill participants will be incomplete and inherently flawed.Footnote 101
Other less recognized sources of undue influence include the influence of psychotropic medications and changes in personality and preferences due to mental illness.Footnote 102 While mental illness or psychotropic medication may not change participants’ decision-making ability, they may change the decision they make.Footnote 103 This scenario continues to be ignored under federal regulations designed to protect vulnerable people in human subjects research.
2. Vulnerability in relation to mental illness
The Common Rule and rCR point to “a vulnerable category” or “a category of subjects that is vulnerable to coercion or undue influence,” respectively, in three contexts: (1) IRB membership,Footnote 104(2) equitable subject selection,Footnote 105 and (3) the potential need for additional safeguards in research involving vulnerable subjects.Footnote 106 Overall, the Common Rule did little to define vulnerability; instead, it “convey[ed] a consent-based conceptualization of vulnerability that lack[ed] internal clarity regarding its scope.”Footnote 107 This uncertain scope of vulnerability and the consent-based focus on decision-making capacity continue under the rCR.Footnote 108
For years scholars have attempted to forge meaning where little or no guidance exists.Footnote 109 The result is a disorganized, minimally useful literature focused on vulnerability.Footnote 110 The dictionary defines a “vulnerable” person as one capable of being physically or emotionally wounded, or open to attack or damage.Footnote 111 Worldwide, policies and guidelines on clinical research ethics rarely define “vulnerability,” but they generally “convey implicitly that vulnerability is fundamentally an inability to provide free and informed consent.”Footnote 112 Some assert that “mental illness creates an inherent vulnerability,”Footnote 113 while others maintain that such a notion risks overinclusion and furtherance of stigmatization and discrimination.Footnote 114
B. A Closer Look at the Origins of “Vulnerability”
Authored by the National Commission, the Belmont Report was the first official influential writing to identify vulnerability as a factor requiring significant attention in clinical research ethics.Footnote 115 The report states that one necessary consideration in assessing justifiability of research with humans is:
(iv) When vulnerable populations are involved in research, the appropriateness of involving them should itself be demonstrated. A number of variables go into such judgments, including the nature and degree of risk, the condition of the particular population involved, and the nature and level of the anticipated benefits.Footnote 116
Under “Selection of Subjects,” the report states that research targeting particular groups gives rise to injustice.Footnote 117 These groups include:
racial minorities, the economically disadvantaged, the very sick, and the institutionalized … Given their dependent status and their frequently compromised capacity for free consent, they should be protected against the danger of being involved in research solely for administrative convenience, or because they are easy to manipulate as a result of their illness or socioeconomic condition.Footnote 118
The National Commission recommended requiring additional protections in human subjects research involving pregnant women, fetuses, neonates, prisoners, children, and the mentally infirm.Footnote 119 While every other category was listed explicitly as a vulnerable population in the final Common Rule, the rule did not prescribe heightened protections for institutionalized “mentally infirm” subjects due to “concern that [such protections] might block valuable and appropriate research.”Footnote 120 In its final 1991 iteration, the Common Rule mentioned subjects “‘likely to be vulnerable to coercion or undue influence,’ including mentally disabled persons,”Footnote 121 but, as noted above, did not define mental disability.Footnote 122
Since publication of the Belmont Report, legal and nonlegal policies and guidelines relating to clinical research ethics have expanded worldwide; along with this expansion, the meanings and conceptual scopes of vulnerability and vulnerable populations have branched in nuanced directions.Footnote 123 Scholars have put forth various interpretations. For example, one analysis suggests that vulnerability is an “additional susceptibility to research-related harm or risk, or that additional safeguards are needed to protect the potential participant from research-related risk or harm.”Footnote 124 Another approach imagines layers of vulnerability “which arise from interactions between an individual’s characteristics and their environment, and which interact with one another to create an inextricably context-dependent vulnerability.”Footnote 125
One of the more influential interpretations of vulnerability in human research divides vulnerability into six subcategories: “cognitive, juridic, deferential, medical, allocational, and infrastructural.”Footnote 126 Categories most relevant to people suffering from mental illness are cognitive vulnerability and deferential vulnerability.Footnote 127 Cognitive vulnerability that impairs decision-making capacity may be either permanent, as with Alzheimer’s disease or brain injury, or temporary, e.g., an episode of major depressive disorder.Footnote 128 Deferential vulnerability highlights that an individual could possess cognitive decision-making capacity but still be influenced by others’ opinions “such that they are not able to decide based on their own preferences and values.”Footnote 129 A power imbalance between the researcher and a subject suffering from mental illness could serve as fertile ground for deferential vulnerability.Footnote 130 In fact, scholarly literature acknowledges “that vulnerability is a relational feature, borne of power asymmetries between participants and research staff, investigators and institutions.”Footnote 131
Another conception of vulnerability asserts that all research subjects are vulnerable; this conception should be challenged, then dismissed.Footnote 132 The notion of universal vulnerability in clinical research—which can be described as a “categorical claim” of vulnerability—is “both clearly right and fundamentally problematic.”Footnote 133 While all human beings are susceptible to harm, ascribing vulnerability to all human subjects serves to “undermine rather than promote” protections for clinical research subjects.Footnote 134 Such a generalized definition supports the need for research regulation in general,Footnote 135 but if we insist on employing “special protections” for all subjects, then we are “providing special protections for no one.”Footnote 136 It is argued, however, that attempts to formulate specific lists of vulnerable populations can lead to stigmatization and the harmful suggestion that persons on these lists are “everywhere and always” incapable of taking care of themselves and their interests.Footnote 137 The Canadian Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (“TCPS2”) states that “individuals should not automatically be considered vulnerable simply because of assumptions made about the vulnerability of the group to which they belong.”Footnote 138
Lists and categories may strip individuals of autonomy not only in the research setting, but in other areas of life as well.Footnote 139 While categories may be problematic, so too is the current “IRB free-for-all” in interpreting U.S. federal regulationsFootnote 140 as they relate to people who suffer from mental illness.Footnote 141 Without lists, firm definitions, or guidance, a category of subjects may be deemed “vulnerable to coercion or undue influence” at one institution on a given day and “not vulnerable to coercion or undue influence” at another institution, or even the same institution, the very next. Erring on the side of more protections may be better than insufficient and inconsistent protections.Footnote 142 Requiring consideration of heightened protections should not stymie research with mentally ill participants, but should aim to add structure and consistency to the research endeavor so those who might need extra protections are afforded them.
To counter the risk for “over-inclusiveness” in describing people who require additional protections in research, some scholars rely on comparative and contextual analyses.Footnote 143 A comparative approach to defining vulnerability highlights that protections needed for a particular subclass may be more than are necessary for the class as a whole.Footnote 144 A contextual analysis additionally contemplates under what circumstances this particular subclass requires additional protections.Footnote 145 Another relevant consideration is “situational” vulnerability, which involves temporary vulnerability.Footnote 146
C. Voluntarism must be Considered in Assessing Mentally ill Participants’ Vulnerabilities to “Coercion or Undue Influence”
Bioethicists have conjured numerous interpretations and conceptions of what it means to be vulnerable, which makes it all the more remarkable that most interpretations of the Common Rule and rCR employ impaired decision-making capacityFootnote 147 as the sole criterion for determining mentally ill participants’ vulnerability to coercion or undue influence.Footnote 148 Again, while specifically naming persons who are vulnerable—prisoners, children, educationally or economically disadvantaged individuals, as well as individuals with impaired decision-making capacity—the introductory language “such as” clearly leaves the field open.Footnote 149 Impaired decision-making capacity is but one example of vulnerability to coercion or undue influence:Footnote 150 whether or not this was the intent of those who drafted the rCR, the words “such as” are akin to the legalese “including but not limited to.”
Asserted earlier in the Article, the need for additional protections for participants suffering from mental illness may not necessarily arise from an impairment in decision-making ability, but rather from a sometimes undetectable susceptibility to coercion or undue influence from a variety of sources.Footnote 151 For these individuals, the effects of mental illness, psychotropic medications, power dynamics, and “structural coercion”Footnote 152 may not alter their ability to make a decision, but can change the decision they make.Footnote 153 The rCR ignores this possibility, not necessarily as written but as interpreted and applied in practice.
This type of susceptibility—the risk of altered decision-making—is the very type of situation where the concept of “voluntarism” is implicated.Footnote 154 Professor Laura Weiss Roberts asserts that informed consent has three components: information, decisional capacity, and voluntarism.Footnote 155 Voluntarism is “the individual’s ability to act in accordance with one’s authentic sense of what is good, right, and best in light of one’s situation, values, and prior history,” and “involves the capacity to make [choices] freely and in the absence of coercion.”Footnote 156 Unfortunately, voluntarism is often overlooked, if considered at all.Footnote 157 Instances may arise where, for example, mental illness or psychotropic medications unduly influence thought processes, preferences, and judgments and may, in fact, alter the patient’s decision to participate in clinical research.Footnote 158 Under these circumstances, choices may not be truly voluntary because they do not reflect the participant’s true self.Footnote 159 Due to these concerns, the concept of voluntarism should be a key factor in determining a mentally ill participant’s vulnerability to coercion or undue influence.
IV. CONSULTING THE EXPERTS
One scholar suggests that appropriate protections for research participants need not focus on the “concept of vulnerability” or involve attempts to create lists or categories of vulnerable populations.Footnote 160 Instead, this scholar argues that experts must develop guidelines to determine group and individual vulnerabilities based on the contexts of particular studies and certain circumstances, and then formulate context-specific protections.Footnote 161 This approach relies on “expertise regarding the specific individuals and contexts in question.”Footnote 162 But, it raises obvious questions: How much representation of individual participants is feasible? What is meant by expertise? Who is an expert? Might expertise differ based on context? And finally, is the expert the psychiatrist or the psychiatric patient? To answer this last question, perhaps both individuals possess expertise that is complementary but not necessarily overlapping. Another source of expertise might be leaders at HHS, the National Institutes of Mental Health, and mental health advocacy groups.
A. The Experienced Research Participant
One novel proposal suggests that experienced research participants have expertise, and directly engaging them in conversation will better inform research ethics and oversight.Footnote 163 This approach recommends appointment of experienced research subjects to research ethics and policy development committees and IRBs.Footnote 164 Alternatively, institutions could establish advisory groups of research participants with varied experiences who could be called upon for specific issues by their human subjects protection programs.Footnote 165 More generally, it is asserted that engaging research participants with first-hand experiences is necessary to develop research ethics and policy and to strengthen the factual and moral bases of guidelines and regulations.Footnote 166
Some challenges of this proposal are readily apparent. The recommendation of including research subjects as IRB members may be unrealistic for certain populations, such as those in studies on terminal diseases.Footnote 167 For these groups, an unfortunate but inevitable attrition rate would deplete the pool of individuals available to serve as IRB members. Participants in studies implicating stigmatized diseases—e.g., clinical trials for antidepressant drugs—could experience repercussions in their personal, social, and professional lives if it is somehow disclosed that they serve as a representative of those suffering from mental illness.Footnote 168 For someone carrying the burden of mental illness and trying to get well or maintain wellness, this may be too much to ask.
B. Representation Through Empirical Research
Another approach suggests accessing the expertise of veteran research participants less directly, relying on empirical research as the preferred method of including participants’ views.Footnote 169 Built into this empirical research, however, is a power imbalance between researchers and participants.Footnote 170 With empirical studies that “express” the affected population’s perspectives, the “research community remains in control of the research and continues to determine its appropriateness, and research participants remain outside of the room except for their indirect voices in the data.”Footnote 171 This point warrants repeating: ethics research on clinical research subjects is framed and interpreted by researchers.Footnote 172 And inherent in research design are the “‘value judgments’ that necessarily reflect the background beliefs, the vested interests and the ideology of the investigator.”Footnote 173 Nonetheless, a recent qualitative study, hereafter “the Lawrence study,” is a fine example of indirect representation through examining depressed patients’ attitudes about DBS, an experimental intervention for treatment-resistant depression.Footnote 174 Psychiatry has a tempered hopefulness about the effectiveness of DBS, but optimism is based on inconsistent results seen in case reports and small open-label trials with no control groups.Footnote 175 The limited data on efficacy raises “ethical questions about vulnerability” in potential DBS patients.Footnote 176 Recognizing concern within the research community about unrealistic hopes, feelings of desperation, or inability to adequately weigh risks and benefits, the Lawrence study interviewed treatment-resistant depressed patients to assess “interest level, hopes, fears, and decision-making processes.”Footnote 177 Study participants appeared to understand the risks of DBS, were aware of their own limitations in processing information, and expressed cautious approaches to considering engagement in clinical research trials.Footnote 178
While qualitative studies often allow research participants opportunities for elaboration and originality of thought, the researcher devises the initial prompts.Footnote 179 Criticisms of the Lawrence study point out that the authors summarily and prematurely dismissed relational vulnerabilities, both “power-relational vulnerability”Footnote 180 and “identity-relational vulnerability.”Footnote 181 It is also suggested that participants in the Lawrence study should have been asked additional questions to fully analyze relational vulnerabilities and avoid misleading conclusions.Footnote 182 Even so, first-person accounts of prospective research participants who suffer from mental illness, and evidence-based research focused on their experiences, undoubtedly are valuable sources of information for mental health research.Footnote 183
C. Expertise on Institutional Review Boards
As shown in Part II, the rCR still requires consideration of including on IRBs those with knowledge about and experience working with subjects from potentially vulnerable categories.Footnote 184 IRB community members, however, can be quite seasoned by life experience or education, without having worked with a particular population. The rule’s “working with” requirement is an unnecessary limitation. Further, while the rCR requires that every IRB have at least one community member,Footnote 185 providing the opportunity to recruit community members experienced with particular populations, little evidence demonstrates that IRB community members are assigned to reviews based on their specific knowledge or interests.Footnote 186
It is possible to recruit and assemble a pool of IRB community members who have acquired expertise or highly relevant knowledge through various avenues, supplementary to or by means other than employment.Footnote 187 The discretionary mandate—effectively no mandate at all—that “consideration shall be given to the inclusion of one or more individuals who are knowledgeable about and experienced in working with these categories of subjects” is a missed opportunity to actually mandate or strongly encourage inclusion of the rich knowledge base accessible through targeted IRB community member recruitment and assignment.Footnote 188
V. POSSIBLE SOLUTIONS
A. What HHS Could/Should Do
HHS must either further revise the rCR or issue binding guidance on biomedical research with people who suffer from mental illness. Three points should be kept in mind: (1) some who live with mental illness or conduct research with this community assert that “it is stigmatizing to equate a mental illness with incapacity,”Footnote 189 that “no one-to-one relationship between a psychiatric disorder and decisional incapacity” exists,Footnote 190 and that it is inappropriate to label all people with mental illness as vulnerable;Footnote 191(2) as interpreted and applied in practice, decisional capacity is the exclusive mechanism under the rCR for viewing the vulnerabilities to coercion and undue influence in people with mental illness;Footnote 192 and (3) individuals with mental illness can be deemed to have traditional decision-making capacity based on widely accepted assessments, but still be vulnerable to coercion or undue influence for other reasons.Footnote 193 HHS does not need to amend rCR language to designate all people who suffer from mental illness as vulnerable. This approach would likely be met with substantial pushback from researchers,Footnote 194 community advocacy groups, and perhaps many individuals who live with mental illness. But, the current use of “such as … individuals with impaired decision-making capacity” creates a catch-all category for vulnerabilities to coercion or undue influence arising from mental illness that is simply too narrow.
A rather straightforward proposal would be to amend the rCR by adding a reference to “subjects who are potentially vulnerable to coercion or undue influence, such as individuals who suffer from mental illness” in both “Criteria for IRB Approval of Research” at 45 C.F.R. §§ 46.111(a)(3), (b) and “IRB Membership” at 45 C.F.R. § 46.107. The use of “potentially vulnerable” would likely satisfy legitimate objections to labeling all people with mental illness as vulnerable to coercion or undue influence. Saying people with mental illness are potentially vulnerable to coercion or undue influence is less essentialist and judgmental than saying they are vulnerable, always.
Alternatively, HHS could bypass any need for additional formal rulemaking, instead issuing binding guidance to clearly address and fill the enduring gaps in protections for mentally ill people in biomedical research.Footnote 195 First, HHS should provide definitions or guidance on the rCR’s vague concepts, e.g., the parameters of being “vulnerable to coercion or undue influence.” HHS should also answer whether it intended for IRBs to exclude all factors beyond traditional notions of decision-making capacity when considering how to properly protect mentally ill participants against vulnerabilities to coercion or undue influence: if that was the department’s intent, this stance must be revisited. Without officially revising the regulation’s language, binding guidance could offer clarifications. Additionally, effective guidance related to operationalizing the charge to IRBs under 45 C.F.R. § 46.111(b), that would facilitate needed protections for mentally ill participants, might look something like the proposed guidance set forth in Figure 1:
Proposed Guidance

To consistently and comprehensively implement this proposed guidance, IRBs and investigators would benefit from consensus on assessment of voluntarism/autonomous consent, and standardized tools such as template language for protocols and consents.
To realize additional well-informed and meaningful policy changes, HHS should assemble and seek advice from a panel of lay experts—people who live with, or have a close connection to another who lives with, mental illness. HHS should ask what members of this population want researchers to consider. Involving those who have experienced or witnessed psychiatric crises would be especially useful, as these individuals have meaningful perspective on living through decompensation and coming out the other side, and the effects of various interventions and circumstances along the way. While mental health advocacy organizations could also be consulted, any ties to large pharmaceutical corporations should disqualify such groups from providing input, as the sponsorship and influence of these corporations may lead to bias and conflicts of interest.Footnote 197 No solution will be perfect: the goal should be to craft policy approaches that first listen to those who have personal experience with mental illnesses, and then use information gathered to better protect this community in biomedical research.Footnote 198
B. People with Mental-Illness Experience are Integral to Research Review Process
People who have experienced a mental illness crisis and now manage their illness reasonably well truly know the vulnerability both of being in crisis and carrying the sometimes constant worry of relapse.Footnote 199 Recovery is a balancing act, and who would know better the competing forces and challenges of slogging through life with mental illness than those who live that experience each day?Footnote 200 The idea of involving affected populations in research related to their communities is not new. In fact, academic researchers in the United Kingdom are required, and private industry researchers strongly encouraged, to involve mentally ill people in research affecting their community.Footnote 201 Such approaches for involvement from the very early stages of research—even research idea generation—are admirable, but requiring such involvement in the United States would entail a major restructuring of how we conduct publicly funded research. While worthy of serious contemplation in the United States, extensive revamping of research regulations is beyond the scope of this Article and not a feasible option for promptly addressing the issues presented here.Footnote 202
In the absence of HHS revising the rCR, clarifying definitions or providing guidance on enrollment of mentally ill people in research, or in addition to such actions, this Article proposes a modification to how IRBs are assembled in order to enhance representation of potentially vulnerable mentally ill participants. The proposal could be implemented as a policy measure not subject to public comment and the official rulemaking process. A template exists in the National Institutes of Health’s (“NIH”) restructuring of its former institute-specific IRB system into a centralized Intramural IRB.Footnote 203 In eliminating institute-specific protocol reviews, the NIH implemented a mechanism to ensure scientific expertise is represented at each IRB table: physicians and scientists from the various institutes may be called upon for more technical, disease-specific reviews.Footnote 204 These specialists are considered “subject matter experts,” and whenever necessary, subject matter experts will be appointed to IRB panels.Footnote 205 A similar arrangement could be implemented in recruiting IRB community members.
NBAC’s 1998 recommendations to include people suffering from mental disorders or their advocates as IRB members were in line with historical international initiatives aimed at improving “the quality of research review by carefully attending to the concerns of particular communities of individuals” that may be vulnerable.Footnote 206 NBAC’s recommendationsFootnote 207 reflected an awareness that family members of those who may not be available or able to serve on IRBs are in a unique position to share perspectives and represent certain otherwise voiceless segments of the population in research ethics determinations.Footnote 208 Likewise in 2008 to 2009, the Secretary’s Advisory Committee on Human Research Protections (“SACHRP”) approved similar recommendations by the Subcommittee on Inclusion of Individuals with Impaired Decision-making in Research (“SIIIDR”): “IRBs should give special consideration, as appropriate, to the involvement of the following types of individuals in the review process: i. Patients, former patients, patient advocates or family members or others who can represent the views and perspectives of the research participants.”Footnote 209 Interestingly, in the Lawrence study, when asked whom they would consult regarding decisions about undergoing a DBS procedure, participants ranked close family members, romantic partners, and friends higher than psychiatrists, doctors, therapists, or experts in DBS.Footnote 210
If former mental illness research participants are available and able, their voices at the IRB table surely would be invaluable. But, one need not have served as a research participant to have pertinent information on the unique concerns mentally ill research participants may have. A family member who has spent a dozen years securing medical treatment and medication for a patient suffering from bipolar disorder or schizophrenia will surely possess knowledge of participants’ specific needs. Further, a person who lives with mental illness themselves and is in remission—though never a research participant—would bring highly pertinent perspective to the research review process. Official guidance and clear definitions would be of substantial value; however, institutions and their IRBs could move immediately to work on and implement positive changes. Simply changing who is at the IRB table interpreting regulations and guidance is a feasible way to more appropriately represent mentally ill people in research.Footnote 211
As noted above, under its new IRB structure, the NIH matches physician and scientist subject matter experts to specific protocols in panel reviews.Footnote 212 Likewise, a similar “subject matter expert” or “subject matter area of interest” pairing of community members to protocols could be both appropriate and effective. Where a research institution maintains IRB community member pools, community members could be asked to specify areas of experience or interest, without being required to reveal their own protected health information or that of family members; these members could then be assigned to review protocols relating to their indicated experience and interest. Such targeted representation on clinical research protocol reviews would allow the institution to capitalize on the vast and varied knowledge bases of community members, increasing the likelihood that potentially vulnerable populations’ unique needs are properly articulated, understood, and addressed. While taking some effort to set up, this modification would satisfy concerns about community representation without placing extra burdens on former research participantsFootnote 213 who may not be well, able, or available to serve on IRBs.
VI. CONCLUSION
As clinical research with people who suffer from mental illness moves forward in new and exciting directions, eager anticipation of resulting benefits must be balanced by mindfulness of this community’s history of stigmatization and abuse. Beyond common concerns about research with vulnerable populations, the prospects of using new technologies in research efforts aimed at curing psychiatric illness—e.g., gene editing—potentially will bring with them extreme risks and major ethics considerations; serious thought about special regulations for these new biotechnologies must begin in the near term.Footnote 214 Here and now, however, gaps in protections endure for mentally ill people in “ordinary” biomedical research, such as research using combinations of psychotropic medications that are not considered novel biotechnologies.
As recognized by many—including NBAC—during the past several decades, awareness and acknowledgment of the gaps in protections is requiredFootnote 215: an earnest effort to resolve still unanswered questions about how adequately to protect mentally ill people in biomedical research is needed to finally close that gap. Unyielding adherence to traditional notions of decision-making capacity is a doorstop to closing the gaps in protections because other potential vulnerabilities to coercion or undue influence for mentally ill participants will remain unaddressed and ignored.
Standing alone, the rCR is not problematic; rather, it is the misguided application of the rule in the context of mental illness. Perhaps those in positions to interpret the rule, e.g., busy professionals at research institutions, are taking the most direct path to save time and simplify approaches to protect against vulnerabilities to coercion or undue influence. It is doubtful these approaches stem from lack of care about people who live with mental illness, and more likely that decisions are influenced by lack of time. Choices on implementation of protective policy that lean toward expediency, however, may leave unfulfilled the duty to protect people with mental illness in biomedical research.
Forty years after the Belmont Report, even with the 2018 revisions in the rCR, “we remain without sufficiently thorough policy guidelines for working with persons [who suffer from] mental illness.”Footnote 216 Perhaps in the next round of rulemaking, if ever that happens, the rCR will be further revised to address these deficiencies. Some scholars have encouraged an alternative approach of more frequent modifications to the Common Rule, facilitated by mechanisms such as framework rulemaking and strategic consultation.Footnote 217 Rather immediately, new guidance and policies could interpret the regulations in ways allowing novel approaches to safeguard potentially vulnerable participants.Footnote 218
In any event, it is important to recognize that most of the literature on the acceptability of research with mentally ill people is produced by researchers who conduct research on psychiatric interventions, some of which is sponsored by major pharmaceutical corporations.Footnote 219 Further, strong support regarding the acceptability of such research comes from community advocacy groups that largely are sponsored by big pharmaceutical corporations.Footnote 220 Without imputing ill intent to these parties, their perspectives may be biased toward the argument that mentally ill people are unfairly and improperly excluded from research. Limiting considerations to traditional decision-making capacity may be appealing to drug manufacturers and the researchers they sponsor, but this approach is simply not enough. The challenge of finding the right balance is difficult, but that should not dissuade an attempt to do so.
While confusion persists about the meaning of vulnerability to coercion or undue influence in clinical research with mentally ill people, creativity is needed in formulating new approaches to facilitate important research on mental illness while protecting potentially vulnerable participants. No solution can or will be perfect. The most straightforward remedy would be for HHS to further revise the rCR or provide clarity on the parameters of “vulnerability due to coercion or undue influence” and what this concept might include in real-life circumstances. Additionally, HHS should clarify its stance on “impaired decision-making capacity” and its relation to mentally ill participants. Also needed is binding guidance, like that suggested in Part V, on how to expand our thinking about mentally ill research participants’ vulnerabilities to coercion or undue influence, beyond traditional consent-based decision-making capacity. HHS should speak to the practical application of these concepts so as to afford appropriate and consistent protections for mentally ill research subjects. In the absence of specific HHS leadership, or in addition to such guidance, changes in federal agency policies could require that any IRB reviewing protocols for research with this community must include affected people in remission or family members of severely mentally ill people.
Looking to community members who possess experience-based knowledge about mental illness and who, thus, may appropriately represent the interests of future participants would be a major step toward progress. These stakeholders should be invited to participate in the process of brainstorming, planning, and drafting binding guidance, beyond the traditional public comment for regulatory changes. Asking stakeholders—including community members in remission or family members of those struggling—to assist HHS in determining necessary changes, and bringing these same groups to the IRB table to review research, will enhance protections for potential research participants who are suffering from mental illness and unable themselves to pull up a chair.
The opinions here expressed are those of the Author and completely independent of the National Institutes of Health. The Author thanks Ellen Wright Clayton for supporting this scholarship and for comments on earlier drafts. Thanks, also, to Leslie Meltzer Henry for her guidance and encouragement, and both Richard Bonnie and Xuemei Ding for their hospitality at the University of Virginia and for facilitating access to university libraries. Finally, thanks to the anonymous peer reviewers for their observations, probing questions, and helpful suggestions, and AJLM editors Jessa Boubker and Sharon Jaquez for their dedication and meticulous care in preparing this Article for publication.

