An Analysis of Institutional Review Board Policies for Enrollment of Adults with Impaired or Uncertain Decision-Making Capacity

Abstract

Background

Unwarranted exclusion of people with uncertain or impaired decision-making capacity from participation in research violates principles of justice and fairness and adversely impacts the health and welfare of these populations.

Methods

We conducted a cross-sectional study of institutional review board (IRB) policies for investigators and IRB members at 94 top-funded U.S. research institutions to better understand the guidance they provide to investigators who work with populations that have a wide range in decisional capacity. We collected data from publicly available websites and used deductive and inductive methods to develop our coding framework.

Results

We found that 41.5% of institutions had policies that require exclusion of people with uncertain or impaired decision-making capacity unless inclusion is scientifically justified. Only 5.3% had policies that require inclusion of these populations unless exclusion is scientifically justified. Eligibility criteria depended upon the risks of research in 54.3% of policies. Guidance on obtaining consent or assent was provided in 77.7% of policies and 44.7% provided guidance on assessing decision-making capacity. 30.9% of policies required that the IRB include a member who is knowledgeable of the needs and concerns of people with uncertain or impaired decision-making capacity when it reviews research pertaining to that population.

Conclusion

Some IRB policies at U.S. research institutions may be unfairly excluding people with uncertain or impaired decision-making from research participation. Institutions should review their IRB policies to ensure that these policies protect adults with uncertain or impaired decision-making capacity from harm but also do not exclude them from research unfairly.

1. Introduction

Decision-making capacity (DMC) is the ability to make a reasonable choice based on one’s values and the available information. To make a reasonable choice, one must understand the information and the consequences of different choices and be able to express a preference (Buchanan and Brock 1990). DMC is a functional ability that comes in degrees and can vary depending on the time, place, and context of the decision (Buchanan and Brock 1990). For example, a person may be capable of making a medical decision when they are alert and awake in the morning but less capable of making that same decision (uncertain or impaired DMC) when they have taken a medication that makes them groggy or confused. Various factors can compromise DMC, including medications, cognitive or developmental disabilities, psychiatric disorders, dementia, physical or emotional trauma, or stress (Chen et al. 2002; Barstow et al. 2018). DMC is different from competence, which refers to the ability to make a legally valid decision. In most countries, all adults are presumed to be legally competent unless they have been assessed to lack DMC and been adjudicated incompetent by a court (Berg et al. 2001). In situations where an adult lacks sufficient DMC to make a medical decision, a legally authorized representative (LAR), such as a health care agent, guardian, or close family member, may be able to make the decision for them (Berg et al. 2001).

Decision-making capacity is an essential element of informed consent in medicine (Barstow et al. 2018). Enrolling adults with uncertain or impaired DMC in research can create ethical dilemmas in biomedical research because there is a potential conflict between protecting vulnerable people from harm or exploitation and ensuring that they have access to research that may benefit them individually or as a group (Mastroianni and Kahn 2001; Friesen et al. 2023). Historically, some of the worst abuses of human research participants have involved vulnerable groups, such as prisoners, children, and socioeconomically disadvantaged people (Resnik 2025). Members of the National Commission for the Protection of Human Subjects of Biomedical or Behavioral Research (1979) were keenly aware of this legacy when they stated in their Belmont Report that researchers must protect vulnerable populations from “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 (National Commission 1979: 7).”

In 1981, the U.S. Department of Health and Human Services (HHS) incorporated ethical principles articulated in the Belmont Report into its regulations for protecting human participants in research, known as the Common Rule¹ (45 CFR 46), including the obligation to provide special protections for vulnerable groups. To approve a study that includes people who are “likely to be vulnerable to coercion or undue influence, such as children, prisoners, pregnant women², mentally disabled persons, or economically or educationally disadvantaged persons,” the IRB must determine that “additional safeguards have been included in the study to protect the rights and welfare of these subjects (45 CFR 46.111b; 1981 version).”

In 1991, HHS added special protections for pregnant women, fetuses, and neonates (45 CFR 46, subpart B); prisoners (subpart C); and children (subpart D), but not for adults with uncertain or impaired DMC. Although the National Commission (1978) published a report recommending that the Common Rule include regulations for people who are “mentally infirm,” HHS did not adopt these recommendations (Levine 1996). During the 1980s and 1990s, several bioethics commissions have also recommended that federal agencies should develop regulations for adults with uncertain or impaired DMC, but this advice has largely been ignored (Yanos et al. 2009). In the absence of regulations, ethical guidance for enrolling people with uncertain or impaired DMC in research has been provided by several organizations, including the National Bioethics Advisory Commission (NBAC 1998), the National Institutes of Health (NIH 2009), the Secretary’s Advisory Committee for Human Research Protections (SACHRP 2009), and the Council for the International Organizations of Medical Sciences (CIOMS 2016).

In the 1990s, clinical investigators, ethicists, and patient advocacy groups began to critique the protectionist stance toward vulnerable populations found in the Belmont Report and Helsinki Declaration (World Medical Association 2013)³ by arguing that this philosophy undermines the goals of respecting individual autonomy and potentially benefitting vulnerable populations by offering the fair opportunity to participate in research (Mastroianni and Kahn 2001; Dresser 2001; Lyerly et al. 2008; Ross 2008; Rios 2016; DeCormier Plosky et al. 2022; Ouelette 2024). Today, the protection vs. inclusion dilemma is one of the central ethical issues in research with vulnerable populations (Resnik 2025; Friesen et al. 2023).

Recently, HHS took steps to promote the inclusion of people with disabilities in research. In May 2024, HHS finalized new regulations that implement Section 504 of the Rehabilitation Act. The Rehabilitation Act, which was enacted in 1973, is considered to be the first civil rights law that protects people with disabilities. The Rehabilitation Act governs recipients of federal funds and includes rules designed to prevent discrimination against disabled people in health care. The revisions to the Rehabilitation Act clarify the non-discrimination responsibilities of federal funding recipients and include changes in language designed to make the law consistent with the American with Disabilities Act. One of the important changes in the Rehabilitation Act is that it now applies to discrimination in health care. According to HHS, Section 504 of the Rehabilitation Act applies to clinical research, which implies that “study eligibility criteria [must] be written in a way that does not unnecessarily screen out people with disabilities whose research participation would not alter the intended purpose of the program of clinical research being undertaken (HHS 2024: 40089).”

Many of the scenarios described by this new HHS rulemaking as potential disability discrimination in clinical research are not uncommon in existing research practice. For instance, HHS provides as an example of potentially impermissible behavior the exclusion of “patients with cognitive disabilities from participating in a research study regarding cancer treatment based on a belief that they would not be able to provide informed consent (HHS 2024: 40089).” HHS similarly indicates that overly broad exclusion criteria, such as excluding all patients with visual impairment from a clinical trial of a diabetes intervention when the only population that is medically indicated for such an exclusion are those patients with diabetic retinopathy, may also run afoul of disability non-discrimination law. Institutions whose policies unnecessarily exclude patients with DMC may be in violation of federal civil rights law.

Although Section 504 now prohibits institutions from unnecessarily excluding people from research who have cognitive, intellectual, or other disabilities, which may be short-term, recurring, or long-term, several studies conducted before 504 was enacted found that people with uncertain or impaired DMC have been routinely excluded from clinical trials. Feldman et al. (2014), for example, examined 300 randomly selected clinical trials published in the six highest impact factor medical journals between 2007 and 2011 and found that only 2% of these studies included people with cognitive disabilities and that over 90% were designed in a manner that would automatically exclude people with these disabilities. Cammani et al. (2021) examined 2809 studies registered in ClinicalTrials.gov between June 2010 and June 2020 and found that 17.4% excluded individuals with cognitive impairment, 21.9% and 17.7% excluded individuals who could not give informed consent. McDonald et al. (2022) examined 248 Phase 2, 3, or 4 NIH-funded clinical trials listed in ClinicalTrials.gov between February 2018 and February 2021 and found that 74.6% of the studies had eligibility criteria that directly or indirectly exclude adults with cognitive disabilities. DeCormier-Plosky et al. (2022) examined 97 clinical trials registered in ClinicalTrials.gov and found that 68% excluded people with psychiatric conditions and 42% excluded people with cognitive disabilities.

As mentioned previously, the population of adults with uncertain or impaired DMC is highly diverse and includes not only people with cognitive disabilities but also people who may have difficulty making decisions due to various psychiatric or medical conditions. While Section 504 makes it clear that recipients of federal funds are prohibited from unnecessarily excluding people with disabilities from research, unwarranted exclusion of any adult with impaired or uncertain DMC violates principles of justice and fairness and may adversely impact the health and welfare of the population of which that person is a member (Cammani et al. 2021; St. John et al. 2022). Therefore, it is incumbent on institutions to develop IRB policies that protect adults who have uncertain or impaired DMC from harm or exploitation but also to respect their rights to participate in research. Since investigators who are planning to conduct research with human participants must obtain approval from an institutional review board (IRB or research ethics board, REB), it is important to understand whether IRB policies are contributing to exclusionary research practices. The aim of our study was to analyze IRB policies at top-funded U.S. research institutions to better understand the guidance they provide to investigators concerning the inclusion of adults with uncertain or impaired DMC in research. To our knowledge, no previously published studies have collected data pertaining to these topics.

2. Method

This study was part of a larger study on IRB policies on vulnerable populations from the top 105 U.S. academic institutions ranked by research and development (R & D) expenditures for 2022, the most recent year available for which data were available (National Science Foundation 2024; Jonathan et al. Forthcoming). In this larger study, two researchers independently accessed and coded IRB policies for investigators and IRB members available from institutional websites. The policies were located on webpages devoted to IRB policies and standard operating procedures (in general) or policies and procedures focusing on vulnerable (or special) populations or informed consent. 94 institutions (87.6%) from this larger study had policies pertaining to the enrollment of people with impaired or uncertain DMC in research. We used deductive and inductive methods to develop our coding framework. The deductive part involved developing a set of questions to code these 94 policies based on a review of the NIH’s (2009) “points to consider” document, CIOMS, NBAC, and SACHRP guidelines, and an initial review of IRB policy documents. The inductive part involved refining questions based on discussions among members of the research team and initial review of policies. The coding questions were as follows:

1. Does the policy have language that a) requires exclusion of people uncertain or impaired DMC unless inclusion is scientifically justified (i.e., it is exclusionary); b) requires inclusion of people with uncertain or impaired DMC unless exclusion is scientifically justified (i.e., it is inclusionary); c) leaves exclusion or inclusion decisions to the discretion of the investigator (i.e., it is discretionary); or d) none of these options?

2. Does the policy have language that requires that inclusion/exclusion depend on the risk of the research to the potential participant?

3. Does the policy have language that provides guidance on assessing DMC?

4. Does the policy have language that provides guidance on using a LAR, including but not limited to, guidance on the conditions that trigger the use of an LAR, how to select an LAR, and standards for LAR decision-making?⁴

5. Does the policy have language that provides guidance on assessing assent or consent, including, but not limited to, guidance that defines assent, guidance that defines consent, guidance on the recommended age for obtaining assent, guidance for using a DMC assessment tool, and guidance for asking for a DMC consult?

6. Does the policy have language that requires that the IRB includes a member who is knowledgeable of the needs and concerns of people with uncertain or impaired DMC when it reviews research on these populations?

7. Does the policy provide guidance on supported decision-making?⁵

Two members of the research team independently reviewed and coded the policies. We calculated percent agreement as well as Cohen’s Kappa between the two coders for each of the seven questions. Cohen’s Kappa was calculated prior to resolving disagreements between coders. Cohen’s Kappa ranged from 0.159 to 0.482. We are not reporting data for the question about guided decision-making because the percent agreement was very poor and we have low confidence in the reliability of this result.⁶ Although agreement between the two coders was initially low, we took steps to enhance the reliability of the final coding, including holding conferences between the two coders to discuss and resolve their disagreements and using a third coder when necessary. Only the final, resolved coding is reported in this paper. After coding, we calculated the percentage of IRBs with policies addressing the questions of interest and identified example policy text.

3. Results

41.5% (n = 39) of institutions had policies with language that requires exclusion of people with uncertain or impaired DMC unless inclusion is scientifically justified; 5.3% (n = 5) had policies with language that requires inclusion of people with uncertain or impaired DMC unless exclusion is justified; 3.3% (n = 3) had policies with language that leaves exclusion or inclusion decisions to the discretion of the investigator; and 50% (n = 47) had policies with language that did not fit into any of these categories. Most policies (54.3%) had language making inclusion or exclusion dependent on the risks of research. For example, the University of Michigan’s (2023) policy has distinct requirements for including people with impaired decision-making in minimal risk research, more than minimal risk research with the prospect of direct benefit for the participants, and more than minimal risk research without the prospect of direct benefit. Other results were that 77.7% of policies had language that provided guidance on consent or assent and 44.7% had language that provided guidance on assessing decision-making capacity. 30.9% of policies had language requiring that the IRB include a member who is knowledgeable of the needs and concerns of people with uncertain or impaired DMC when it reviews research pertaining to that population. See Table 1 for complete results and Table 2 for examples of policy language.

Table 1:

IRB policies concerning inclusion of adults with uncertain or impaired DMC in research

	n = 94
The policy requires exclusion of people with uncertain or impaired DMC unless inclusion is scientifically justified (E); requires inclusion unless exclusion is justified (I); leaves exclusion or inclusion decisions to the discretion of the investigator (D); or none of these options (N).	E = 41.5% (39) I = 5.3% (5) D = 3.2% (3) N= 50% (47)
The policy requires that inclusion is based on the risk of the research to the participant	54.3% (51)
The policy provides guidance on assessing decision-making capacity	44.7% (42)
The policy provides guidance on using a legally authorized representative for decision-making	74.5% (70)
The policy provides guidance on supported decision-making	3.2% (3)
The policy provides guidance on consent or assent	77.7% (73)
The policy requires that the IRB include a member who is knowledgeable of the needs and concerns of people with impaired DMC	30.9% (29)

Table 2:

Examples of IRB policies on including adults with uncertain or impaired DMC in research

Exclusionary language	Investigators who plan to conduct studies involving subjects with cognitive impairment will need to provide a scientific and/or scholarly justification of the inclusion of this population and assure that appropriate additional safeguards are in place to protect this population from coercion or undue influence (University of California at Los Angeles 2021). Research involving subjects without the ability to provide consent or with impaired decision-making capacity should only be conducted when the aims of the research cannot reasonably be achieved without their participation. Participation of such subjects in research cannot be justified solely on their availability or the convenience for the investigator (University of North Carolina at Chapel Hill 2024).
Inclusionary language	The IRB should be particularly cognizant of the special problems of research that involves the category of subjects who are vulnerable to coercion or undue influence, such as Children, Prisoners, Individuals with impaired decision-making capacity, Economically or educationally disadvantaged individuals…Depending on the research, exclusion of any of the above populations might be construed as unfair and attempts should be made to include these populations, with appropriate protections, if they are applicable to the research question (University of Vermont 2024). In general, all adults regardless of their diagnosis or condition, are presumed competent to consent to participate in research unless there is evidence to the contrary. When investigators propose to include individuals with questionable capacity, you must provide a plan for assessing the participants’ decision-making capacity (Northeastern University 2024).
Discretionary language	Individuals with reduced decision-making capacity require special protections. Investigators may decide to exclude or include them from their research. This section is intended to provide researchers with guidance to cover both situations (Massachusetts Institute of Technology 2024). The Principal Investigator is responsible for making a determination whether persons with decisional impairment are or will be included in the research, to describe the process for assessing consent capacity, to obtain IRB approval for inclusion of the population prior to enrollment, and to ensure that persons with decisional impairment are included in research with an adequate understanding of the study, when appropriate (University of Rochester 2019).
Risk-based inclusion language	Adult Individuals who lack consent capacity may be enrolled in research where the IRB finds: That the proposed research presents at least one of the following: No greater than minimal risk to subjects Greater probability of direct benefit to subjects than harm to subjects Greater than minimal risk and no prospect of direct benefit to individual subjects, but is likely to yield generalizable knowledge about subjects’ disorder or condition that is of vital importance for the understanding or amelioration of the disorder or condition (Indiana University 2024). An IRB may approve participation of adults with cognitive impairment or diminished decision-making capacity only under the following circumstances: Research Involving No More Than Minimal Risk [circumstances specified]… Research Involving Greater Than Minimal Risk but Presenting the Prospect of Direct Benefit to the Individual Participants [circumstances specified]… Research Presenting as Greater Than Minimal Risk Without the Prospect of Direct Benefit to the Individual Participants [circumstances specified] (University of Michigan 2023).

4. Discussion

Of the 94 institutions in this sample, 41.5% have policies with language that requires exclusion of people with uncertain or impaired DMC from research participation unless inclusion is scientifically justified. These policies reflect the protectionist approach to vulnerable populations found in the Belmont Report and the Helsinki Declaration. Further, a majority (54.3%) of institutions have policies with language that base inclusion on the degree of risk of the research, which is similar to the approach found in Subparts B, C, and D of the Common Rule. However, risk-based inclusion/exclusion is simplistic and misguided because, as noted above, adults with uncertain or impaired DMC are a highly diverse group, with different abilities, impairments, and medical conditions (Rosenstein and Miller 2008; Dickert and Largent 2024). Policies for adults with uncertain or impaired DMC should be responsive to the diversity inherent in this population (SACHRP 2009). A small percentage of institutions (3.2%) have language that leaves the inclusion or exclusion decision to the discretion of the investigator; however, discretionary inclusion/exclusion is problematic because it may lead to implicit bias and discrimination (Saunders and Midgette 2023).

Our findings may help explain why adults with uncertain or impaired DMC are often excluded from clinical trials, because many of the policies we reviewed permit or encourage exclusion. However, as mentioned earlier, policies that unnecessarily exclude people with cognitive, intellectual, or other disabilities may violate Section 504 of the Rehabilitation Act, and policies that exclude other populations with impaired or uncertain DMC may conflict with to the Belmont Report’s ethical principle of justice. Because unwarranted exclusion of people with uncertain or impaired DMC may be contrary to the principle of justice, stakeholders who are not legally required to comply with Section 504, such as private research sponsors and commercial IRBs, should be mindful of the ethical implications of the exclusion/inclusion policies and take appropriate steps to promote fair inclusion of people uncertain or impaired DMC in research.

We recommend, therefore, that institutions review their IRB policies to ensure safe and appropriate of enrollment of people with uncertain or impaired DMC in research. Policies should:

1. Require inclusion of adults with uncertain or impaired DMC unless exclusion is scientifically or ethically justified;

2. Address when and how adults with uncertain or impaired DMC should be included in research and which additional safeguards for their inclusion may be appropriate;

3. Provide guidance on assessing decision-making capacity, using supported decisionmaking [for people with DMC], using a legally authorized representative and obtaining assent [for people who lack DMC];⁷

4. Require reassessment of decisional capacity if the participant’s medical status changes, and

5. Require that the IRB includes a member who is knowledgeable of the needs and concerns of people with uncertain or impaired DMC when it reviews research on these populations.

Limitations

Our study has several limitations.

First, the institutions in our sample were from the U.S., so our findings are not generalizable to other countries. Second, our sample was limited to academic institutions, such as universities and medical schools, and did not include private IRBs. Third, we did not collect data on the institutional factors that have influenced policy development. We were only trying to describe policies not to explain how the originated.

Research agenda

Future research should address questions not answered or addressed by our study, such as the content of policies from countries outside the U.S. and private IRBs, how IRBs review studies that may include subjects with impaired decision-making, the specific guidance IRBs offer to investigators concerning DMC assessment, and the views of people with uncertain or compromised DMC and their family members concerning participation in research.

Educational implications

Education in human research ethics should include discussion of issues related to enrolling people with questionable or impaired DMC in research, such inclusion and exclusion criteria, informed consent, supported decision-making, and DMC assessment.

Best Practices

Academic institutions should consider developing policies for participation of adults with questionable or impaired decision-making in research.

Conclusion

IRB policies may be contributing to the inappropriate exclusion of adults with uncertain or impaired DMC from clinical research. Institutions should review their IRB policies to ensure that they protect adults with uncertain or impaired DMC from harm but do not exclude them unfairly.

Biographies

Emily Nguyen is an undergraduate student at the University of Virginia. She collected and coded data and edited/reviewed manuscript. She is interested in bioethics.

David Appiah is a graduate student and nurse at Sanford World Clinics-Adenta, Sanford Health, Ghana. He collected and coded data and edited/reviewed manuscript. He is interested in bioethics.

Ari Ne’eman is an Assistant Professor in the Department of Health Policy and Management at Harvard University T.H. Chan School of Public Health. He helped design the study and interpret the data and edited/reviewed manuscript. He is interested in bioethics.

Min Shi is a biostatistician at NIEHS/NIH. She analyzed the data and edited/reviewed manuscript. She is interested in biostatistics.

Barbara E. Bierer is Professor of Medicine (Pediatrics) at Brigham and Women’s Hospital and Harvard Medical School. She helped design the study and interpret the data and edited/reviewed manuscript. He is interested in bioethics.

Willyanne DeCormier Plosky is the Program Director of Multi-Regional Clinical Trials at Brigham and Women’s Hospital and Harvard Medical School. She helped design the study and interpret the data and edited/reviewed manuscript. He is interested in bioethics.

David B. Resnik is bioethicist at NIEHS/NIH. He conceived of the idea, designed the study and interpreted the data and wrote the first draft of the manuscript. He is interested in bioethics.