Meta-Consent in Research on Decisional Capacity: A “Catch-22”?

Abstract

Empirical studies of ethical issues, which have increased in number and scope in recent years, may themselves raise both practical and ethical issues. One example of such an issue is the question of who may be legitimately enrolled in studies of decision-making capacity; must all participants in studies of consent capacity have capacity to consent? This question may pose a “Catch-22”: For example, if some of the participants in a study of consent capacity are deemed by a particular standard to be incapable of consent. In weighing the risks and benefits of studies of consent capacity, how should reviewers consider the context of actual versus hypothetical trials for which the participant's consent is being sought? Here, we explore these “meta-consent” issues by describing the dimensions of the issue and potential solutions, centering around the concept of “active assent” (requiring expressed understanding of the purpose of the study and its voluntary nature, as well as expression of a choice to participate).

Introduction

The study of ethical issues in clinical research has undergone impressive developments in the use of empirical techniques.^1–12 Yet such empirical bioethics studies may themselves raise both practical and ethical dilemmas.^13–16 Practically speaking, finding clinical research studies on which to “piggyback” empirical ethics studies can be challenging.¹³ Ethical dimensions of ethics research have been less explored,^15,16 particularly in contrast to the substantial literature on ethical issues in clinical research,⁷ which has primarily focused on informed consent and decision-making capacity—although numerous other important ethical aspects of clinical research also deserve study.^17–23

In studies of decision-making capacity, a common yet elusive issue arises: simply put, who can be enrolled in such studies? Must all participants in studies of consent capacity have capacity to consent? Depending on how this question is answered, a “Catch-22”²⁴ may arise. For example, what if some of the participants in studies of capacity to consent are found (by a given standard) to lack capacity to consent to research? If the study is piggybacked on a clinical trial, should those participants be ineligible to enroll in the clinical trial? What are the consent investigator's obligations when studying capacity to consent to an actual trial? How should institutional review boards (IRBs) weigh the risks and benefits of studies of consent capacity, and how should they take into account the context of an actual (versus hypothetical) trial for which the participant's consent is being sought? In this article we explore these issues, laying out the dimensions of the “meta-consent” issue, and then discuss some potential solutions to problems or questions that empirical ethics investigators may encounter.

Dimensions of the Problem

As described in The Belmont Report, ethical research should be guided by the principles of respect for persons, beneficence, and justice.²⁵ Protections for vulnerable populations derive directly from these principles. At the same time, these principles are broadly articulated and do not provide specific guidance for empirical ethics investigators faced with the “meta-consent” dilemma. In fact, many studies of capacity to consent to research (or treatment) have focused on populations potentially at risk for impaired decisional capacity.^{2,3,10,26–33} In and of itself, this focus is desirable from a scientific standpoint, since broad enrollment in capacity studies has enabled a fuller characterization of the actual range and variability of decision-making phenomena, their correlates, and opportunities for the improvement of consent procedures.

We also lack consensus about what criteria or standards should be used to classify people as possessing or lacking capacity to consent to various forms of research.^11,34–38 Some errors in categorization are thus likely. The weight given to avoiding incorrectly categorizing people as “incompetent to decide” when they are in fact competent to make the decision at hand versus calling someone “competent to decide” when they are in fact incompetent will necessarily vary by the context, including the procedures, level of risk, and possible benefits of the proposed study.^37,39–42

Finally, we note that regarding consent for “parent studies,” concerns about decisional capacity among those with psychiatric conditions have received more discussion and attention than for most nonpsychiatric medical populations (as evidenced by the National Bioethics Advisory Commission's decision to focus their 1998 report⁴³ and recommendations regarding decisional capacity on those with mental disorders). Although issues of consent are relevant to those with schizophrenia and other serious mental illnesses, it does participants an injustice (and violates the autonomy component of respect for persons) to assume a person lacks decision-making capacity based solely on diagnosis. It is equally an injustice (and a violation of the second component of respect for persons, protection of those with diminished autonomy) to assume that those without psychiatric conditions invariably possess decisional capacity. Rather, these judgments must be made for individual protocols and participants, regardless of whether the protocol involves people with schizophrenia or those with other medical conditions that place some patients at risk for impaired capacity.

Possible Capacity Scenarios

For the purposes of this discussion, we refer to the process of consenting to the capacity- or consent-related study as the “capacity study consent” (CC) and to the process of consenting to the actual or hypothetical clinical protocol as “parent study consent” (PC). The purpose of the former study might be to evaluate determinants of PC capacity/incapacity, to evaluate methods of assessing capacity, and/or to find methods of enhancing consent procedures.

There are 4 possible combinations of capacity (Table 1): competent CC and competent PC (Case A); competent CC and incompetent PC (Case B); incompetent CC and competent PC (Case C); and incompetent CC and incompetent PC (Case D).

Table 1.

Possible Combinations of Competence and Incompetence for Studies of Consent

	Parent Study Consent (PC)
Capacity Study Consent (CC)	Competent	Incompetent
Competent	A	B
Incompetent	C*	D

^*Case C is probably least likely, although it is at least conceivable. For instance, studies of capacity are somewhat abstract; thus a participant could fail to grasp the nature and purpose of a study of consent capacity while simultaneously adequately comprehending the more tangible nature of a clinical trial.

Some of these combinations are more likely than others and/or pose different levels of ethical dilemma. The first cell in Table 1 (A) is the cleanest: competent to consent to the consent study and competent to consent to the parent study. This combination poses no ethical dilemma as participants can be readily enrolled in both studies. Another common combination is competency to consent to the consent study but incompetency to consent to the parent study (cell B in Table 1). This combination arises frequently because capacity/consent studies are nearly always procedurally simple and carry minimal risk (generally involving little more than time and interviews), whereas the parent protocols may be substantially more complex in terms of procedures and/or the risk-benefit analysis. In theory, this combination poses no barrier to consent research; in practice, however, some investigators may view it as a disincentive for linking their study to a study of capacity to consent. That is, the following question may arise: given the current lack of a perfectly reliable and valid means of determining decisional capacity,³⁷ what effect will the presence of the capacity study have on enrollment for the parent study (assuming the capacity evaluations go beyond those routinely required for such a parent study)? If the extra knowledge of suspected incapacity to consent means one cannot permit the participant to consent to the parent study, then some clinical trials investigators are likely to resist linking their studies to such capacity studies.

The third and fourth scenarios (cells C and D in Table 1) involve cases where participants are incompetent to consent to the capacity/consent study, and either incompetent to consent to the parent study, too (a common scenario for those with the greatest cognitive impairment) or competent to consent to the parent study (this scenario probably being the least likely). In either of these scenarios, the immediate question is whether, or under what conditions, can consent capacity be studied among those who lack consent capacity? There are several possible answers to this question.

Possible Solutions

One solution to the problem would be not to enroll participants with questionable capacity to consent to capacity research, and then to extrapolate any findings from higher functioning individuals to the entire population. The problem with such extrapolation is that it assumes linear relationships among variables of interest that may not be applicable at all ability levels. For instance, in a study of enhanced consent strategies, it may be that techniques that improve understanding, appreciation, or reasoning among those with mild cognitive impairment differ from those most likely help those with more severe impairment.

Because capacity research itself finds justification in its value to the participants (and people like them) through the development of improved consent procedures and a better understanding of decisional abilities (and risk factors for incapacity) in various populations, it can also be argued that prohibiting the enrollment of people with suboptimal consent capacity in consent- and capacity-related studies represents a violation of justice. This justice concern is essentially what motivates our practices regarding research with other decisionally impaired populations, such as people with dementia or children—ie, it determines when such research is permitted and under what circumstances.

Although there is no full consensus, scientifically sound and otherwise ethical research with decisionally impaired but actively assenting individuals is most commonly viewed as acceptable when the protocol meets the minimal risk standard (cf 45 CFR 46.102(i)).⁴⁴ It is important to note that “minimal risk” does not mean zero-risk. Virtually no human activity would meet a zero-risk standard, but in most situations, the readily conceivable risks from capacity/consent research appear to fall well within the range of minimal risk. As defined by federal regulations, “Minimal risk means that the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests.”

Assuming a decisionally impaired individual actively assents to participate in a minimal-risk study, should there be an additional requirement to obtain consent from a family member or legally authorized representative? As noted in The Belmont Report, the principle of respect for persons has 2 components “first, that individuals should be treated as autonomous agents, and second, that persons with diminished autonomy are entitled to protection.”²⁵ The balance given to these 2 components seems to be a question on which reasonable people could disagree. For the enrollment of children, their assent is not sufficient to enroll them in minimal-risk research; investigators must also obtain legal consent from a parent or guardian. There is not a clearly parallel assumed authority in terms of decisionally impaired adults. Therefore, in the context of research that is scientifically valid, minimal risk, and otherwise ethically appropriate, we believe that there is no added “harm” against which individuals need be protected, so the right to be treated as an autonomous agent and the dignity thereby afforded take precedence. Yet, a subtle issue remains in the preceding assertion. If “active assent” is taken to mean more than simply the verbalization of “okay” or a parallel statement—as a conscious affirmation—then some minimal level of understanding would seem to be involved.

Tentatively, we propose that some expressed understanding of the purpose of the study (eg, that one is being asked to be in a study of the consent process) and the voluntary nature of participation, as well as expression of a choice to participate, should be required elements of active assent. Under this proposal, there would be very few people who would be incapable of enrolling in a consent study because the threshold would be very low (whether such agreement to participate is appropriately labeled “active assent” or consent with a low threshold is a matter of semantics). It would still enable the inclusion of people with a broad range of performance on measures of capacity for the parent study (whether real or hypothetical), including most of those who lack sufficient capacity to consent to a greater than minimal risk study. The exclusion of those few people who would be incapable to assent/consent for the capacity study itself is appropriate, as these are individuals where consent from a legally authorized representative would be necessary for the parent study in any case. Indeed, this is the option often adopted in dementia research;^45–47 yet, this would be a small minority of individuals with schizophrenia or other forms of serious mental illness.

In essence, these provisions would virtually eliminate most instances of impaired capacity to consent to capacity/consent research (ie, the second row of Table 1) without doing away entirely with the requirement for consent. Investigators must still contend with Case B in Table 1, where the participant is found incapable of consenting to the parent study. Presuming the parent study is not minimal risk, the second aspect of respect for persons (protection of those with diminished autonomy) is germane. In cases where the parent protocol represents a minor increase over minimal risk while offering the possibility of direct benefit, consideration may still need to be given to weighing autonomy considerations against the need for protection, thereby establishing a relatively low capacity threshold. In parent protocols that exceed that level of risk, however, there will be cases where the participant can enroll him- or herself in the consent study but not in the parent study. In those instances, we suggest considerations of ethical enrollment might follow requirements for child research (particularly 45 CFR 46.405, 46.406, and 46.407), which balance both risk and potential for direct benefit, along with requirements for active consent from the participant (where possible) and consent from a legally authorized representative. Many states currently lack clear provisions for proxy research consent, with a few exceptions (see recent California law on surrogate consent for research).⁴⁸

There is an additional and relatively unique issue that might be raised regarding “risk” in capacity/consent studies that are piggybacked on other studies. This is the risk, however unlikely, that a participant will be found to be incapable of consenting to the parent study and, in the absence of an appropriate proxy for legal consent, may lose the opportunity to participate in that parent study. This potential for lost opportunity would probably not meet the technical definition of “greater than minimal risk”; however, there are some contexts where patient-participants may be eager to enroll in clinical trials (eg, a poor prognosis condition for which there are no currently approved effective treatments). From such an individual's perspective, the possibility of being excluded from the parent study could be perceived as an important risk or harm that should be weighed when deciding whether to participate in a consent/capacity study. Therefore, depending on the nature of the parent protocol, that risk might be another element that consent/capacity investigators should be certain participants understand.

Some investigators who are involved in greater than minimal risk research might be hesitant to add additional burdens to research enrollment. However, in such instances we would argue that it is not only ethically mandated but also in investigators' own interests to ensure that those who are permitted to enroll are either fully capable of making that choice, or that appropriate methods of proxy consent have been instituted.

Conclusion

Studying decision-making capacity and other aspects of informed consent raises interesting dilemmas for the capacity investigator. These issues have been minimally addressed by the empirical ethics research community.¹⁵ As in all human research, the interests of participants must be given paramount precedence. Yet, as we have argued, functional prohibition of consent research with those at the lower ranges of functioning might also be a disservice to those individuals, particularly when the consent research is both valuable and minimal risk. We argue for the explicit adoption of the sliding scale concept of capacity; this would essentially enable almost all potential participants to enroll in the capacity research, given a low threshold for capacity needed to consent to minimal-risk research (which characterizes most research on consent capacity and/or the consent process).

Although empirical ethics investigators may have managed to find ways to conduct their studies and gain approval from their institutional review boards, in other instances, concerns (sometimes exaggerated) about capacity to consent may actually be hindering efforts to conduct important research by erecting unnecessary obstacles. Without data on this question, however, it is difficult to say, other than anecdotally, whether this is occurring. If true, this is troubling, because the focus of evidence-based ethics work is precisely to enhance the underlying ethical basis of research—a crucial task in psychiatric research and all research involving potentially vulnerable populations.^49,50