Analysis and Proposed Actions Regarding the NBAC Report: Research Involving Persons with Mental Disorders That May Affect Decisionmaking Capacity. General Comments



    NBAC currently is conducting a review of the entire system of human subject protections. The WG recognizes that, upon completing this comprehensive review, NBAC may make additional recommendations affecting human research subjects with impaired decisionmaking capacity. The WG was mindful of that possibility in considering how HHS might address the recommendations contained in the current Report.

    Section I addresses a major overall concern regarding NBAC’s recommendations - the issue of who should be covered by any additional protections. Section II summarizes NBAC’s envisioned regulatory framework and describes the WG’s concerns about it and its implications.

    1. Scope of Applicability of NBAC’s Recommendations

      The WG found some ambiguity as to who is to be protected by the Report’s 21 recommendations. Certain NBAC recommendations refer to “persons [or subjects] with mental disorders that may affect decisionmaking capacity” (e.g., Recommendations 2 and 20), while others refer to “persons with mental disorders” (e.g., Recommendations 1, 3, and 18). Based upon an analysis of the entire Report, the WG interprets the intended scope of NBAC’s recommendations to be “persons with mental disorders that may affect decisionmaking capacity,” as listed in the title.


      The WG agrees with NBAC that research involving persons with mental disorders that may affect decisionmaking capacity merits additional scrutiny and perhaps additional protection. The WG further considered whether Department action should focus just on persons with decisional impairment resulting from a mental disorder or should extend the benefits of additional protections more broadly to all persons with impaired decisionmaking capacity. It concluded that a broader scope is warranted, based on two factors: the need for special protections in a broader population of persons with decisional impairment and the potential for stigmatization (perceived or real) if those special protections are limited to persons with mental disorders that may affect decisionmaking capacity.

      1. The scope of the Report’s recommendations seem appropriately applicable to all persons with decisional impairment, irrespective of diagnosis.

        NBAC focused on persons with mental disorders that may affect decisionmaking capacity [3] “in part because of this population’s difficult history of involvement in medical research.” (Report at p. ii). Instances of alleged abuse and public and scholarly criticism represented “a critical mass of concern” about the ethics of conducting research, particularly certain drug research, on mental illnesses and about “perceived gaps” in existing procedural protections. (Id. at p. 3). As NBAC noted, the Report continues a history of regulatory recommendations for additional protections in this area.

        In the chapter on “Scope,”‘ the Report states that the definition of mental disorders cannot be precisely pinned down, in part because of rapid changes in the field. The Report lists dementias, delirium, schizophrenia, depression -including bipolar disorder, mental retardation, and substance abuse disorders -as some, but not all, “of the disorders in which decisionmaking capacity may be affected.” (Report at p. 7). Nonetheless, other conditions associated with impaired decisionmaking were not included in the scope: “Other potentially vulnerable subjects whose decisionmaking capacity may be compromised by such factors as trauma (e.g., head injury) or physical illness (e.g., cancer or sepsis) will not be considered directly in this report.” (Report at p. 2, n.3). However, the Report does not exclude persons with these conditions from the asserted need for additional protections.

        The WG agrees with NBAC that current procedures for human subject protection should be improved to adequately protect research participants whose decisionmaking ability is impaired because of a mental disorder. The current HHS and Food and Drug Administration (FDA) regulations rely fundamentally on informed consent and independent review. However, the protection of informed consent is available only when subjects have the capacity to assess risks and benefits and to choose. Individuals who are decisionally impaired, including some with mental disorders, may not be fully capable of appreciating the range of risks and benefits of a particular protocol. An individual who lacks capacity may only be enrolled in research by a legally authorized representative (LAR); however, if the LAR has a conflict of interest or does not adequately protect the individual’s interests or choices regarding research, the individual can do nothing affirmative because of that lack of capacity. Because of the need to rely on surrogate decision-makers and susceptibility to undue influence and coercion, a subject who has impaired decisionmaking capacity’ due to a mental condition may not be adequately protected by current informed consent procedures.

        However, the Working Group notes that some physical disorders or conditions also may result in impaired capacity to make decisions and therefore to give voluntary informed consent to research. Either mental or physical conditions that impair decisionmaking capacity may make subjects vulnerable to coercion or undue influence and thus in need of protection. As NBAC observed, “many of the issues and concerns raised in this report (and, indeed, many of its recommended protections) could be applied to all persons with questionable or diminished capacity.” (Report at p. 5).

      2. Limiting the scope of protections may be perceived to be stigmatizing.

        Some commentators on the Report believe that focusing on “persons with mental disorders that may affect decisionmaking capacity” improperly stigmatizes a category of persons, most of whom are capable of making their own decisions. Some of the Report’s recommendations refer to “persons with mental disorders” without the additional modifier “that may affect decisionmaking capacity” - and this may contribute to the perception of a lack of sensitivity.

        Approximately one in five Americans experiences a mental disorder during the course of a year,[4] yet the preponderance are capable of making major life decisions. The WG agrees -- and NBAC’s Report takes pains to emphasize -- that a person is not decisionally incapacitated to consent to research merely because that person has a mental disorder or mental illness.[5] However, the Report’s focus was not expressly extended beyond persons with mental disorders and that in itself may appear stigmatizing or could lead to inadvertent stigmatization in practice.

      3. HHS should treat NBAC’s recommendations as potentially applicable to all research subjects whose decisionmaking may be impaired.

        NBAC “encourages” use of its Report as “guidance for conducting research on other persons [i.e., in addition to those with mental illness] whose decisionmaking capacity may be impaired.” (Report at p. 5). The WG accepts NBAC’s suggestion and concludes that the scope of additional protections should encompass all persons with decisional impairment, irrespective of diagnosis.[6] In particular, the WG believes that the Department should extend the safety net of protections to all individuals with a disorder that alters mentation and therefore may adversely affect decisionmaking capacity. However, the WG emphasizes that an individual with a condition that may impair decisionmaking capacity is not necessarily impaired.

        In view of the difficulty of determining who is vulnerable as a direct consequence of a condition that impairs decisional capacity, there is no expeditious way to identify, a priori, individuals with impaired decisional capacity; yet human subject protections rely primarily on prospective review by investigators and IRBs. Thus, investigators and IRBs should be alerted to those situations that warrant closer scrutiny. To be effective, additional guidance must clearly identify those who are or may be eligible for the additional protection; the IRB should be able to tell when it is reviewing a protocol that might encompass that vulnerable population.

        Accordingly, 45 CFR §46.401 (a) applies “to all research involving children as subjects.” Similarly, §46.301 applies “to all biomedical and behavioral research ... involving prisoners as subjects.” (Emphasis added). In both casts the regulations define who is included in the protected class; although the IRB may need to refer to local law (e.g., the age of majority in the jurisdiction) or ascertain some facts (e.g., whether the proposed subjects are incarcerated) in determining whether the additional protections apply.

        NBAC was concerned with a gap in existing regulatory protection for both (1) populations that are vulnerable because of decisional impairment and (2) subjects in certain classes of research that might lead to transient or other decisional impairment. The WG agrees that, in both instances, subjects deserve additional protection. At the same time, the WG recognizes that, in contrast to children and prisoners, it is very difficult to define who is included in the population.

        When the concern is with a particular type of research, the guidance must clearly identify for the IRB what that research is. Given the widespread incidence of mental disorders in the population, many protocols in different fields will -involve subjects with conditions that may affect decisionmaking, even those where the protocol has nothing to do with the study of those conditions and the investigators intend to recruit only adults with decisional capacity. Thus, for example; a protocol studying cholesterol levels or eyewitness testimony -- and including only incidentally a few subjects who experience mild depression or who are recovering alcoholics -- does not necessarily warrant the same degree of scrutiny as would a protocol to study a potential treatment for advanced Alzheimer’s disease. [7]

        With these factors in mind, the WG concludes that, in assessing NBAC’s recommendations, the Department should consider their applicability to research directed toward [8] persons whose mental or physical conditions are associated with decisional impairment. A condition is associated with decisional impairment when there is a reasonable likelihood that (1) the persons sought to be enrolled in the protocol have experienced or may experience decisional impairment as a result of their condition or disorder, or (2) the subjects may become decisionally impaired during the course of the study.

        Because some children are also persons whose condition or disorder is associated with decisional impairment, Department action in this area should clearly indicate that it does not supersede the protections of subpart D of 45 CFR Part 46, which regulates research involving children. Further, FDA has a regulation providing an exception to informed consent for subjects in need of emergency medical intervention who cannot give informed consent because of their life-threatening medical condition and do not have an LAR available. See 21 CFR §50.24. HHS announced a comparable waiver of the applicability of the 45 CFR Part 46 requirement for obtaining informed consent for this limited class of research (61 FR 51531, 10/2/96). Thus, any Department action should also clearly indicate that it does not supersede either these protections or the Secretarial waiver.

    2. Regulatory Framework Envisioned by NBAC

      Many of NBAC’s recommendations address the protections contained in existing HHS and FDA regulations. These include recommendations about subject selection, IRB review of research designs, evaluation of risks and benefits, informed consent, increased research and training, and increased support for IRBs. The WG agrees that these are features that may be valuable in protecting subjects whose decisionmaking capacity is impaired. The WG also agrees that increased attention to, and additional guidance in, these areas may be beneficial.

      Some of NBAC’s recommendations call for HHS and other agencies to require IRBs to implement procedures that are available, but not explicitly required, under current regulations when decisional capacity is an issue. For example, NBAC would encourage, and in some cases require, greater participation on the IRB by individuals familiar with the nature of the disorders and the concerns of the population being studied. Further, for greater than minimal risk research, NBAC recommends that IRBs generally require a formal capacity assessment These are important protections that IRBs already have aiithority to invoke at their discretion, and the WG agrees that IRBs should do so whenever they determine such actions to be necessary and appropriate for the rights and welfare of subjects.

      Certain of NBAC’s recommendations work in concert to call for a new regulatory framework. This recommended framework would require IRBs to classify all proposed research involving persons with mental disorders that may affect decisionmaking capacity as either (I) minimal risk, (2) greater than minimal risk but offering the prospect of direct medical benefit to subjects, or (3) greater than minimal risk with no prospect of direct medical benefit to the subject. (See Recommendations I0­I2). The current regulations do not require a classification of risk.

      In accord with the current regulations, the NBAC recommendations would permit any adult who can provide informed consent to choose to participate in IRB-approved research involving any degree of risk. If the person has impaired decisionmaking capacity, the NBAC recommendations would also permit enrollment in the research when the IRB waives the informed consent requirement in accordance with the current regulations. Even without a waiver of informed consent, a person who lacks capacity could be enrolled in minimal risk research (Recommendation 10) by an LAR, provided the LAR meets certain conditions (See Recommendation I4). If the research offers the prospect of “direct medical benefit,” then regardless of risk, a person who lacks capacity may still be enrolled in the research by an LAR, subject to the same conditions. (Recommendation 11).

      However, if the proposed research does not offer the prospect of direct medical benefit and involves any degree of risk greater than minimal risk, NBAC’s envisioned regulatory framework would impose a limit that currently does not exist. (Recommendation 12). The subject could not be enrolled in the research by an LAR, except under one of two circumstances. In addition to IRB approval, either (I) the subject must have executed an advance directive to participate in that type of research, called a “Prospective Authorization” under Recommendation I3 or (2) the research must be reviewed by a “Special Standing Panel” (SSP) convened by the Secretary and that panel must conclude that the protocol “offers the possibility of substantial benefit to the population under study, that its risks to subjects are reasonable in relation to this possible benefit, and that it could not be conducted without the proposed population,” and the panel must be “satisfied that all appropriate safeguards are incorporated.” (Recommendation 2). If the SSP subsequently were to promulgate guidelines covering certain instances of that type of protocol, then only IRB approval and surrogate consent would be required unless the guidelines require more.


      1. The envisioned regulatory framework could result in an unnecessarily lengthy review process.

        Under NBAC’s envisioned regulatory framework, IRBs generally would not be able to approve research without direct medical benefit that poses any risk greater than minimal and involves a person unable to give informed consent, unless that person has executed a Prospective Authorization specifically authorizing this type of research and his or her LAR consents. Prospective Authorizations will not provide a significant alternative to SSP review. For some conditions affecting cognition, the individuals do not have, and will probably never have, the decisional capacity required to execute a Prospective Authorization. Even among those able to execute a Prospective Authorization, very few are likely to do so.

        Thus, unless the research offers the prospect of direct medical benefit and regardless of whether it offers the prospect of direct non-medical benefits, the SSP, at least until it has promulgated guidance for IRBs, would be required to review all research involving even a slight increase above minimal risk. Not only physical, but psychological, social, and other non-medical risks must be included in determining the protocol’s risk level. Moreover, SSP review would be required for all research supported by HHS where the risk cannot be reduced to minimal risk - i.e., not only for clinical research but also for research in the areas of epidemiology, social and behavioral sciences, program evaluation, health services, and health promotion. This could slow significantly the progress of important research protocols that pose reasonable - albeit greater than minimal - risk.

        By contrast, the introduction of an intermediate risk category, as in the HHS children’s regulations, Subpart D of 45 CFR 46, along with the ability to consider direct non-medical benefits, would allow IRBs to determine when the risks posed by those protocols are justifiable and what the appropriate’ safeguards are, without being required to defer to a national review body. If such an alternative to NBAC’s recommendations were adopted, the feasibility of the SSP concept would be materially greater.

      2. The envisioned regulatory framework would alter IRB authority in ways that could produce anomalous results.

        The Report envisions a framework of protections akin to that of the regulations governing research with children but with some key differences. The following comparison presents two examples.

        Section 46.406 of the children’s regulations permits the child’s parents or guardians to enroll the child in some no-direct-benefit research that “represents a minor increase over minimal risk.” For example, if an IRB were to determine that a protocol with no direct benefit posed a minor increase over minimal risk but was “likely to yield generalizable knowledge . . . of vital importance” to understanding or ameliorating autism, a 17-year-old individual with autism could be enrolled with parental consent.

        However, the envisioned regulatory framework would not permit even a legally appointed guardian to enroll an adult who lacks decisional capacity in any research with greater than minimal risk and no prospect of direct benefit - unless the potential subject has provided a Prospective Authorization or the research is approved by the standing panel that Recommendation 2 would create (or the IRB is authorized to act in accord with pertinent guidance developed by that panel). Thus, absent the exceptions noted, the same individual at 19 could not be enrolled in that protocol by his or her LAR(s), even by parents who have been formally appointed guardians by the court and would consent; whereas the IRB could approve research that would enrollment that individual at age 17. NBAC’s recommendations therefore would limit IRB authority more for the adult than for the child.

        A comparison of the safeguards for potentially beneficial research reveals another anomaly. Under §46.405 children may be enrolled in greater than minimal risk research if the research “holds out the prospect of a direct benefit” for the child. By contrast, under NBAC’s Recommendation 11, adults who lack decisionmaking capacity may only be enrolled in greater than minimal risk research that “offers the prospect of direct medical benefit,” which would exclude consideration of other possible benefits to subjects, including psychological, educational or social benefits. (Emphasis added). Thus, the envisioned regulatory framework would narrow the types of benefits that an IRB may consider in determining whether a greater than minimal risk protocol offers a direct benefit.

        There are of course important differences between children and adults with impaired decisionmaking capacity. In particular, children are legally presumed to lack decisional capacity, whether or not they do; whereas adults are presumed to possess decisional capacity unless proven otherwise. Also, the individuals who are legally authorized to represent children are more often readily identifiable than those for decisionally impaired adults.

        There are also key similarities, e.g., that the degree of actual capacity varies, that State law determines who may represent subjects’ interests, and that States may exercise their parens patriae power to limit the risk of harm that an LAR may permit. Thus, safeguards for the two groups need not be identical. However, in the opinion of the WG, these factual differences do not justify limiting IRB authority for adults with impaired decisionmaking capacity more than for children in similar situations. The WG is not aware of any evidence that indicates the categories of research risk set up by the children’s regulations, which have been in place for 17 years, are unsatisfactory. The WG is also concerned that the existence of two frameworks for approving protocols based on two different sets of risk level may lead to confusion and thus contribute to inadequate compliance by investigators and IRBs.

[3] “[T]he primary concern of this report is with the potential effect of neurologic or psychiatric conditions on the decisional capacity of potential research subjects.” Report at p. 20. “The specific concern of this report, however, is with persons whose decisional impairments may be related to the presence of what we currently understand to be a mental disorder.” Id. Part of NBAC’s limiting its focus in this way may result from its conclusion that additional steps needed to be taken to “both enhance existing protections and facilitate broad public support for continued research on mental disorders.” Harold T. Shapiro, Letter transmitting the Report (January 8, 1999) (emphasis added).

[4] Mental Health Report, Executive Summary, p. xii.

[5] For example, at pages 5-6, the Report states:

NBAC was mindful of the concerns that could arise from a focus on individuals who are members of a group (i.e., persons with certain mental disorders) rather than on persons who share a common functional characteristic (i.e., questionable decisionmaking capacity). This focus could raise the specter of equating mental disorders with incapacity and thus potentially stigmatize these individuals. NBAC shares this concern and recognizes that not all persons with mental disorders have impaired decisionmaking capacity.

To assume that a diagnosis of a mental disorder implies that a person is incapable of deciding whether to participate in a research protocol is prejudicial and incorrect.... Although persons with mental disorders are not necessarily decisionally impaired, much less decisionally incapable, any evidence that places a person’s decisionmaking ability into question should trigger a clinical assessment Any disorder that alters mentation may adversely affect decisionmaking ability ....

[6] Concerns have been raised about the need for input from communities that would be affected most by the expansion of the recommendations ‘scope. No particular action is needed to, address those since any proposed Departmental action should provide ample opportunity for comment by the public.

[7] Although “NBAC principally focused its attention on those who may be primarily considered for research protocols because it is their particular mental disorder that is being studied” (Report at p. 5), NBAC’s recommendations encompass more than such protocols. Even within the field of mental disorders, NBAC’s scope would include protocols about a condition that may cause impairment, even if it is improbable that the subjects will be decisionally impaired (because, for example, their condition is only mild, or at an early stage).

[8] See, e.g., Subpart B of the Regulations, which appears to be concerned with both research involving pregnant women as subjects and research “activities directed toward pregnant women.” § 46.201 (a)(2) and the heading of § 46.207.

View full report


"nbac.pdf" (pdf, 171.03Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®