Three Lessons Learned About Power and Improving Recruitment of Underrepresented Populations in Clinical Trials

In the Veterans Health Administration, the rate of alcohol use disorder remains high among veterans who identify as American of African descent or Black (1, 2). Although effective therapeutics for alcohol use disorder exist, identification of novel, safe, and effective medications remains a top priority (3). However, most clinical trials foreground White participants, a bias contributing to disparities in access to empirically supported treatments (4). Clinical trials that equitably include African American or Black participants are needed to provide efficacy data and improve the quality of health care for underserved populations.

A U.S. Department of Veterans Affairs clinical trial, titled “Topiramate Treatment of Alcohol Use Disorder in African Americans” (the VA-TOP study), aimed to test the efficacy of the medication topiramate for the treatment of African American or Black veterans with alcohol use disorder (5). However, a challenge arose in meeting trial recruitment goals (target N=160 participants). Although successful in recruiting, obtaining consent from, and randomly assigning a number of African American or Black patients (N=35) to treatment, the study’s researchers were unable to enroll the number of patients necessary for statistical power calculations, and the study was discontinued. This experience is not uncommon. Well-publicized barriers to participation by underrepresented groups include participants’ discomfort with the research process, time and resource constraints, discriminatory research practices, and lack of diversity in the research workforce (4). To eliminate the proliferation of flawed and underpowered clinical trials, a nuanced understanding of challenges with recruitment is needed.

We suggest that power dynamics must be recognized and embraced so that underserved populations can participate equitably (6, 7). Traditionally, power in clinical research has been thought to be wielded exclusively by the investigator. However, although rarely acknowledged, participants also hold power (e.g., through their ability to decide under what conditions they will participate in research). When researchers ignore or are naïve about such power dynamics, imbalances in power result, creating recruitment challenges.

Late into recruitment for the VA-TOP study, it was discovered that 50% (N=505 of 1,010) of potential participants contacted by the study team declined to participate. This high rate of declined participation had not been anticipated and initially puzzled the study team. To better understand nonparticipation, the team collected data about reported concerns of the contacted veterans who had declined participation (N=505), which included the following: skepticism of medication research (N=141, 28%), beliefs that one’s drinking was not problematic or a preference to reduce it on one’s own (N=67, 13%), the time commitment required (N=62, 12%), and transportation difficulties (N=15, 3%). Once these problems were known, the research team approached the Veterans Community Advisory Board (VCAB) for input into how to address these issues. However, recruitment for the clinical trial had fallen too far behind its stated goals, leading to discontinuation of the study. Importantly, the VCAB’s input revealed the decision map that veteran participants follow and emphasized several important lessons. Therefore, we encourage investigators who engage with historically understudied populations to consciously reflect on power dynamics and their effects on recruitment. In the following, we offer recommendations for ensuring an equitable balance of power between investigators and participants in clinical trials.

Lessons Learned From the VA-TOP Study

Meaningfully Connect With the Population

We urge investigators to form meaningful relationships with the population of focus as early as possible and to consider representation of the chosen population on their research team from the outset. A team of individuals from different racial and ethnic backgrounds is beneficial regardless of the type of research undertaken but could be pivotal for research with underserved populations. More crucial for behavioral health research is the stigmatization around conditions such as alcohol use disorder, which can be ameliorated by having members of the population of interest on the study team (6). Although greater attention to solutions that ensure diversity in the research workforce is needed, efforts by investigators to include participants’ voices in research can still be meaningful (8). For example, the VCAB recommended simple fixes to ensure that the study population saw itself represented on the study team: that veterans be among those performing consent procedures and that an African American or Black representative from the VCAB observe consent procedures to offer specific feedback. When a research team neglects inclusion of members from the population of interest on the study team, power imbalances occur, such as unfounded assumptions by investigators regarding acceptability of research processes, patronizing and condescending conversations with participants, and research interpretation and outcomes that fail to mirror the realities of the patients involved and of the larger population of patients that clinical trials are designed to serve (7).

Moreover, it is well established that investment in the research study by the specific population that is the focus of the research will build trust, offer credibility, and engender an equitable balance of power between the research team and participants (9). Research shows that being in visible partnership with champions from within the community is effective for recruitment (6, 10, 11). It is important to consider what resources will be needed to facilitate these partnerships (e.g., staff salaries and consultant fees) and to add costs of these resources to grant budgets. The expectations that will be placed on the selected partners for the study also need to be considered. Although formal responsibilities can be outlined, investigators should also consider implicit institutional biases that members of the community might encounter (7). It is key to reflect on these issues from the outset and to establish plans for how community members will be meaningful members of the research team—with shared power and agency. These practices promote equity and inclusion within a research ecosystem and make members of the community true research partners.

Ensure That Eligibility Criteria Are Inclusive

Because investigators designate study eligibility criteria, they are in a position of power as gatekeepers. Importantly, eligibility criteria are commonly modeled on previous clinical trials consisting of predominantly White samples. Consequently, individuals from racial and ethnic minority groups may be disproportionately excluded owing to a host of discriminatory factors based in histories of marginalization (12). We witnessed this phenomenon in the VA-TOP study: top reasons for excluding potential participants during chart review (N=2,669 reviews) were use of ineligible medications (N=347, 13%), a history of schizophrenia or bipolar spectrum disorder diagnoses in the medical record (N=239, 9%), and chronic health conditions (N=82, 3%).

When developing eligibility criteria for a condition such as alcohol use disorder, researchers should consider limiting medication exclusions to those solely based on safety for the specific study medication. Additionally, researchers should be aware of potentially stigmatizing diagnoses that may be historically disproportionately used among individuals from marginalized groups, such as people with bipolar disorder or schizophrenia (13, 14). Specifically, we caution against using a lifetime diagnosis in a medical record as the basis for exclusion. Instead, we suggest beginning with a lower screening threshold and adding clinical assessments that sensitively ask questions about comorbid conditions and diagnoses.

Seek Continuous Partnered Collaboration

Although it was not initially apparent why participant recruitment into the VA-TOP trial became challenging, problems were quickly identified by the VCAB, reinforcing the advantages of including the community early on. First, several changes to the consent form were suggested, including simplifying the language, emphasizing the benefits of the research, and providing clearer descriptions of the study medication and possible adverse effects. Another important suggestion was that participants have an opportunity to meet with the principal investigator during informed consent visits to ask questions about the study, facilitating trust. Additionally, the payment schedule was altered to better compensate participants for their time. Finally, recruitment materials were redesigned after the VCAB noted stock images that were stigmatizing and discriminatory.

Meaningful inclusion of the study population centers empowerment: power is reflected in decisions that count (7, 9). In clinical research, especially in studies focused on stigmatized behavioral health conditions such as alcohol use disorder, power is unequally distributed between a study team and study participants. Although consulting with the population about ways to share power will raise issues that may evoke discomfort, such consultation is invaluable (7). Without such collaboration, participants may behave passively, as observed in the VA-TOP study, exercising what power they do have through drifting away from the study rather than firmly expressing reasons for their unwillingness to participate.

We recommend that investigators seek collaboration (rather than mere feedback) continuously from study conceptualization to dissemination. Investigators should consider the institutional environment and who in the institution can offer insight, even if such contributors are external to the research program. Most importantly, researchers should empower collaborators with real decision-making capacities, ensuring not only that their perspectives and experiences are being sought but also meaningfully incorporated (7, 10, 15). Ongoing collaboration with the population of interest will ensure that study design and materials are culturally sensitive, increase the population’s investment in the research, and facilitate greater ecological validity and clinical applicability, preventing many of the issues experienced with the VA-TOP study. Further, collaboration equitably shares power within clinical research processes and outcomes.

Implications and Conclusions

The inclusion of marginalized groups in clinical trials is crucial to building culturally informed treatment practices. Investigators should create research environments that mirror the needs of populations underrepresented in research, rather than interpreting low recruitment as a pathology of the population itself. Even though investigators traditionally have been perceived to hold all power in clinical research, it remains important to acknowledge the power that participants hold and will exert when studies are not culturally sensitive. Ideally, future research will incorporate thoughtful choices that meaningfully empower the population researchers are seeking to recruit throughout all phases of the study. Such intentions will increase knowledge of effective medication treatments for patients from underserved populations, boost engagement among individuals from the population of interest, and reduce health disparities. We conclude with the words of a VA-TOP participant and contributing author (C.L.) about his experience in the study after the VCAB’s recommendations were implemented: