The analysis draws substantially from a research paper commissioned for this study, written by Drs. Farah Acher Kaiksow, M.D., M.P.P., and Jocelyn Carter, M.D., M.P.H. The full research paper can be found online at: nap.nationalacademies.org.
The processes and infrastructure of medical research have led to important advances in medical knowledge and therapies that have improved many lives; however, the existing system has also served to reduce participation by a diverse population. This chapter presents an overview of the range of factors, operating at multiple levels (participant and community characteristics, individual research studies, the institutions that conduct research, and the broader landscape agencies and policies that govern research), that serve as barriers to inclusion of underrepresented and excluded populations in clinical research.
Individual and community factors are often cited as reasons for lack of inclusion of underrepresented and excluded populations in clinical trials. The evidence suggests, however, that many of these concerns misrepresent barriers to participation in research or are surmountable with effort from research teams, funders, and policy makers. In addition to the barriers to inclusion that are often present in the life cycle of an individual study, a range of cultural, historical, and community-level factors influence feasibility and implementation of clinical research and directly influence study recruitment and retention.
Willingness to Participate
Overall, lack of willingness to participate is frequently given as the cause of poor representation of some populations in research. However, the evidence on this issue is clear: Asian, Black, Latinx Americans, and American Indian/Alaska Native individuals are no less likely than other groups, and in some cases are more likely, to participate in research if asked (Adeyemi et al., 2009; Arega et al., 2006; Bieniasz et al., 2003; Bishop et al., 2011; Byrd et al., 2011; Byrne et al., 2014; Ceballos et al., 2014; Evans et al., 2010; Gadegbeku et al., 2008; Garber et al., 2007; George et al., 2014; Guadagnolo et al., 2009; Hillyer et al., 2020; Kaplan et al., 2015; Langford et al., 2014; Manders et al., 2014; McElfish et al., 2018; Murphy and Thompson, 2009; Murphy et al., 2009; Priddy et al., 2006; Sanderson et al., 2013; Sprague et al., 2013; Thetford et al., 2021; Trant et al., 2020; Webb et al., 2010; Wendler et al., 2006).
A 2014 national review that included more than 4,500 Asian, Black, and Latinx Americans who were eligible for cancer trials found the same willingness to participate among all groups and equal enrollment rates (Langford et al., 2014). Study participation willingness was similar across racial/ethnic groups for studies focused on HIV, despite early narratives of stigma and discrimination related to the illness. Among Asian American, Black American, and white American college students in Atlanta, a 2006 study found no difference in willingness to participate in an HIV vaccine trial (Priddy et al., 2006). Results with older patients are equally convincing: among a population of 417 HIV-positive Black and Latinx people (60 percent male) in Chicago with an average age of 43, 95 percent would either agree to or consider participating in a study (Adeyemi et al., 2009). In this analysis, the strongest predictor of participation was simply being asked.
Rural populations are increasingly recognized as underserved, with underrepresented individuals from rural areas particularly at risk for poor health outcomes.1 Enrollment of rural populations into clinical research is especially challenging given structural barriers including access to health care and transportation issues. Yet people living in rural areas do not appear to be any less willing to participate, based on a large study of 5,256 people in Arkansas and a smaller study of 533 people in Alabama, Florida, Georgia, Louisiana, Mississippi, and Puerto Rico (McElfish et al., 2018; Thetford et al., 2021). Among the respondents in the Arkansas study, greater than 45 percent said they would participate in research if asked, with another 22 percent being undecided; only 32 percent said they would not participate (McElfish et al., 2018). The smaller multistate study further analyzed the data by ethnicity and rurality and found that among
1 The committee would like to note that it is unclear from the literature whether frontier populations are included in the research definition of “rural.” Frontier areas are sparsely populated rural areas that are isolated from population centers and services, and there is no universally accepted definition of rural that ensures frontier populations are included in this demographic (Coburn et al., 2017). When this report was written, the committee could not find any literature specific to frontier populations’ participation in clinical trials and clinical research.
Black and Latinx residents of both rural and urban areas, 75 percent were willing to participate in research, but greater than 90 percent had never been asked (Thetford et al., 2021).
Population-specific studies confirm what the more general studies cited above suggest: underrepresented populations are not necessarily underrepresented because they are unwilling to participate. Attitudes of 204 Black men about a variety of types of clinical research, including surveys, focus groups, clinical trials, and genetic studies, found that 74 percent endorsed a willingness to participate (Byrd et al., 2011). Regarding specific willingness to be randomized in a surgical versus nonoperative study of spinal disorders, Black Americans expressed equal willingness to be randomized as white Americans (Arega et al., 2006). The same results have been found in studies on Black individuals’ participation in HIV treatment trials, studies on aging, and recruitment for clinical trials on kidney disease (Evans et al., 2010; Gadegbeku et al., 2008; Garber et al., 2007). In the HIV treatment study, like the more general 2014 study mentioned above, the major barrier to participation of HIV-positive Black people was having never been asked (Garber et al., 2007).
Data suggest that Latinx populations may in fact be more likely to participate than other populations. A 2014 study of women in Texas reported that Latinx women were 44 percent more likely than non-Latinx women to participate in a gynecologic malignancy clinical trial (Manders et al., 2014). In New York City, Latinx patients were more than twice as likely to say they would join a cancer clinical trial compared with non-Latinx patients (47.7 percent and 20.8 percent, respectively) (Hillyer et al., 2020). A qualitative study of 59 Latinx men and women at the Texas-Mexico border demonstrated significant enthusiasm on the part of this group to get involved in research (Ceballos et al., 2014). “If I had the opportunity to participate in something like this, I’d love to,” said one respondent.
Although not as extensive, studies of American Indians echo those of Black and Latinx Americans. A study of American Indian college students found that, depending on the specifics of the trial, anywhere between 63 percent and 84 percent would probably or definitely agree to participate in a cancer clinical trial (Sprague et al., 2013). Only in cases where a significant amount of travel or risk of a confidentiality breach existed did willingness drop below 50 percent. In a separate study comparing American Indians with Asian, Black, Latinx, and white Americans, there was no difference between the groups in refusal to participate in a cancer clinical trial (Guadagnolo et al., 2009).
Although stated willingness is comparable across these underrepresented groups, it might differ from actual consent and participation rates; however, evidence suggests these are at least equal. A literature review, published in 2006, combined data from 20 studies that examined the consent rates of people of different races and ethnicities; 18 of these studies took place either entirely or mostly in the United States, while the remaining two studies took place in Europe, Australia, or New Zealand (Wendler et al., 2006). Combining data from these
studies to create a cohort of more than 70,000 individuals, this analysis found that Black and Latinx people had the same consent rates as white people. For clinical intervention studies, Latinx individuals actually had statistically significantly higher consent rates, 55.9 percent compared with 41.8 percent, respectively. A more recent study of 1,126 postpartum women in Philadelphia found that consenting women were actually more likely to come from underrepresented groups compared with those who did not consent (Webb et al., 2010).
The inability to channel the willingness of underrepresented individuals to participate in research has implications beyond lack of engagement in specific trials. In almost all papers on predictors of willingness to participate in research, prior exposure to or participation in research is associated with increased likelihood for participation and a more positive attitude toward research (Behringer-Massera et al., 2019; Byrne et al., 2014; Sprague et al., 2013; Webb et al., 2019). In a study of more than 7,800 people in Florida, the positive influence of prior exposure on future participation was higher for Black respondents than for white respondents (Webb et al., 2019). Unfortunately, misunderstanding or lack of knowledge about the willingness of underrepresented populations to enroll in clinical research has created a pattern by which failure of researchers and/or clinicians to ask these groups to participate contributes to their lack of enrollment, which further decreases their chances of future involvement, and thus the cycle continues.
Any conversation about the low participation rates of underrepresented individuals in medical research must include the issue of distrust and/or mistrust of the health-care system. Whether caused by distrust (an individual’s sense that their trust has been violated by a specific act, person, or institution) or mistrust (a less specific but no less legitimate feeling that a person or institution may not be acting in an individual’s best interest) (Griffith et al., 2021), the legacy of both historical and contemporary abuses in medical research is an important factor driving the lack of engagement of underrepresented populations with both health care and research. This holds true across a range of underrepresented groups, including Asian American, Black American, Latinx American, and Mexican American (Adeyemi et al., 2009; Behringer-Massera et al., 2019; Bonevski et al., 2014; Braunstein et al., 2008; Buchbinder et al., 2004; Bussey-Jones et al., 2010; Byrd et al., 2011; Corbie-Smith et al., 2002; George et al., 2014; Hardie et al., 2011; Haynes-Maslow et al., 2014; Hoyo et al., 2003; Hughes et al., 2017; James et al., 2017; Lor and Bowers, 2018; Moreno-John et al., 2004; Murphy and Thompson, 2009; Murphy et al., 2009; Newman et al., 2006; Occa et al., 2018; Scharff et al., 2010; Smirnoff et al., 2018).
In qualitative studies with Black Americans, those who decline to participate or express lower willingness to participate frequently mention the offenses com-
mitted by the Tuskegee Syphilis Study as well as more recent personal stories of distrust as reasons for their declinations (Alsan and Eichmeyer, 2021; Behringer-Massera et al., 2019; Buchbinder et al., 2004; Byrd et al., 2011; Corbie-Smith et al., 2002; Scharff et al., 2010). The authors of a survey study about differences in willingness to participate in a cardiovascular drug trial suggest that in addition to not being asked, this type of distrust/mistrust can explain much of the participation gap between Black and white Americans (Braunstein et al., 2008). In a clinical trial exploring barriers and motivators to participation in clinical trials among 67 Black Americans, focus group themes included the perception that research would benefit white participants or the research institution more so than any underrepresented individuals enrolling in the study (BeLue et al., 2006). A study of 17 Black women at high risk for HIV found that, despite expressing favorable attitudes toward medical research in general, distrust was a commonly cited reason for not participating (Voytek et al., 2011). Similar studies exist regarding Black individuals’ participation in blood/tissue donation for genetic studies (Bussey-Jones et al., 2010), psychiatry research (Murphy et al., 2009), and cancer research (Haynes-Maslow et al., 2014). These studies propose historical abuses as a major source of distrust among Black Americans and further assert that this distrust is a large factor in their unwillingness to enroll in medical research.
This issue of trust is of course not limited to Black Americans, and reasons for the distrust vary depending on the group or individual. In interviews with an older population of Hmong individuals, specific concerns arose about possible researcher misuse of information that might lead to loss of financial support from governmental agencies (Lor and Bowers, 2018). The Havasupai Tribe case regarding the misuse of genetic samples and lack of complete informed consent reinforced existing distrust of medical researchers and discouraged tribe members from participating in further genetic research (Garrison, 2013). In a study of 50 Filipino and Native Hawaiian/Pacific Islander people, major focus group themes included negative feelings about the purpose and intent of the research (Gollin et al., 2005). Research into Latinx Americans and Mexican Americans as well as Asian Americans of Filipino descent suggests that at least some of their distrust is rooted in fear for their own or a family member’s immigration status (George et al., 2014; Hardie et al., 2011; Maxwell et al., 2005; Occa et al., 2018). Concerns about health insurance coverage have also been reported. A study of 88 Black Americans’ attitudes toward genetic research identified fear of the loss of health insurance coverage because of targeted discrimination as a barrier to participation (Sadler et al., 2010). Populations who participate in illegal or culturally stigmatized behaviors, including intravenous drug users, people with substance use disorders, LGBTQIA+ individuals, and people who are HIV positive, also may not trust that their personal information will be kept private by research teams (Bonevski et al., 2014; Voytek et al., 2011).
Although the committee agrees that distrust and mistrust are certainly factors that influence the participation of historically underrepresented groups in clini-
cal research, some studies have found that the distrust/mistrust is not associated with willingness to participate in medical research (Alhajji et al., 2020; Ford et al., 2008; Garber et al., 2007; Katz et al., 2008; Webb et al., 2019; Westergaard et al., 2014). A study of 5,139 Black individuals and 2,670 white individuals in Florida found that while Black respondents had mildly lower levels of trust in both researchers and research studies than white people, level of trust did not predict intent to participate for either group (Webb et al., 2019). Although mistrust is likely a factor, the studies that show it is not associated with willingness to participate in research may point to this not being an insurmountable problem and perhaps not the most important barrier. For example, a systematic review of 40 years of research on barriers to enrollment in cancer studies found that, although mistrust was the most commonly cited individual-level reason for not participating in research, the most common barriers overall were related to being offered the opportunity to participate.
Fundamentally, we may never know exactly how much historical and current discrimination and abuses influence underrepresented individuals’ participation in clinical research. The research done in this area may be limited by participants’ unwillingness to openly discuss trust issues with research teams that represent the very entities the participants distrust. Additionally, people with the highest levels of mistrust are unlikely to participate or to be represented in any research. Persistent and systemic efforts to delegitimize, underemphasize, or ignore the link between historical and contemporary occurrences of scientific misconduct/abuse and the mistrust of underrepresented populations toward research will certainly only continue to worsen current disparities in participation. Moreover, an inability or unwillingness of the research community to acknowledge and make efforts to address the roots of distrust/mistrust in underrepresented communities would stymie any movement toward increasing the trustworthiness of researchers in the view of underrepresented populations.
Social and Economic Factors
Although an individual’s socioeconomic status is the result of a multitude of factors both within (individual level) and outside (structural level) their control, socioeconomic issues are discussed here at the individual level in an attempt to describe how these issues drive individual decision-making.
American women and underrepresented individuals make less money and are more likely to live below the federal poverty line compared with white men (DOL, 2020; KFF, 2019). Reduced economic resources can make elective participation in research a challenge. Jobs with fewer options for earned time, sick days, vacation days, and remote work may make participation in research impossible. Individuals with lower incomes are also frequently responsible for caring for children, elderly family members, and sometimes both at the same time, while also working outside the home (Indorewalla et al., 2021). Under these circumstances,
even those individuals who do have the time to participate may not see the value in altering their regular routines, as this may pose much higher opportunity costs, including the loss of potential wages, than for those with more resources (Brown et al., 2000; Olin et al., 2002; Quiñones et al., 2020). One study on HIV/AIDS research in Black men with a history of drug use found that, despite their willingness, eligible participants were often not able to participate due to competing priorities related to work and family (Slomka et al., 2008). In focus groups and in-depth interviews with Asian American women assessing perceived barriers to participation in cervical cancer prevention research, reasons for nonparticipation included lack of time and inconvenience (Giarelli et al., 2011). Time conflicts and childcare responsibilities also emerged as barriers to participation in research for a study examining the perspectives of Black- and Latinx-immigrant participants (Calderon et al., 2006). A systematic review of barriers to study retention found that the most commonly reported barrier was competing priorities related to participants’ socially disadvantaged status (Bonevski et al., 2014).
Perhaps most important are issues of opportunity costs, which include the loss of any potential gains that participants might be able to make if they choose to participate in research rather than the other potential activities. Whether it be a one-time 10-minute survey or a years-long clinical trial, study participation requires time away from work, family, and other commitments. Given this, household financial position plays an outsized role in who gets included in clinical research. Worldwide, nearly 50 percent of the people who participate in clinical trials are considered “high income,” despite representing only 16 percent of the total population; conversely, the “lower middle class” makes up 38 percent of the population and 13.5 percent of the people who participate in clinical trials (Gilmore-Bykovskyi et al., 2021). A prospective study of cancer trials within the United States confirms that this global pattern holds true in the United States, even after accounting for factors such as age, race, and education (Unger et al., 2016).
Several studies have examined the importance of educational background or highest level of grade completion in research participation, many of which support the notion that educational status is more relevant than income level for the participation of Black Americans in research (Alhajji et al., 2020; Byrd et al., 2011). In a study of perceptions influencing research enrollment among low-income Black, Latinx, and white residents of New York City, respondents who had less than a high school education were more likely to have increased feelings of exploitation associated with research participation (Smirnoff et al., 2018). However, other studies have found no specific association with participants’ highest level of education and willingness to participate in research (Kaplan et al., 2015).
Health literacy of patients and potential participants has been cited as a contributor to low participation in research, and low health literacy and numeracy skills are independently associated with less interest in research participation (Kripalani et al., 2019; Protheroe et al., 2009). However, as described earlier,
although there is varied understanding of medical and scientific topics among individuals, it is possible to engage across a spectrum of participants, if appropriate efforts are made. A study of Asian, Black, Latinx, and white men with prostate cancer in California found no difference between people with low health literacy compared to those with medium or high levels of health literacy regarding their willingness to participate in clinical trials (Kaplan et al., 2015). The same study used a questionnaire to assess general knowledge of clinical trials and, again, found no difference in willingness to participate based on the respondent’s understanding of research.
Challenges related to the frequent residential moves and lack of landline telephone access are also often cited as a primary reason for low enrollment and low retention of underrepresented populations (Bonevski et al., 2014; Otado et al., 2015). Reliable telephone access is a significant barrier for those living at or near the poverty line and has been associated with limited insurance coverage, healthcare access, and health behaviors (Bonevski et al., 2014). Perceptions of neighborhood safety have also been reported as reasons for reduced research participation of underrepresented individuals (Ceballos et al., 2014; Ejiogu et al., 2011).
The factors and problems that lead to the limited enrollment of underrepresented and excluded populations in clinical trials and research begin with and follow the life cycle of a project. While a substantial body of literature describes individual and community characteristics (e.g., childcare needs or limited public transportation) that may prohibit research enrollment, these issues remain unaddressed well after the study is designed, funded, and under way. Understanding and resolving underrepresentation in research requires careful examination of the research process itself. At the level of an individual research study, there are problems and factors that prevent the inclusion of underrepresented and excluded populations in clinical research at almost every stage in the process, including
- the development of research questions;
- the composition, training, and attitudes of the research team;
- research site selection;
- participant selection, including sampling and recruitment methods and inclusion and exclusion criteria;
- study protocols, including informed consent processes and remuneration; and
- development of multilingual recruitment and consent documents.
The authors of a systematic review of 40 years of cancer treatment or prevention trials summed up the issue well, writing that “because opportunity barriers largely reflect protocol design as well as the process of study implementation,
investigators play a major role in determining the extent to which trials are accessible to underrepresented groups” (Ford et al., 2008). In the sections below, the committee describes the many ways these problems and factors manifest throughout the course of the life cycle of an individual research study.
Research Questions: Drivers to Motivate Inclusion
Research questions are often driven by funding priorities and scientists’ interests and expertise, which constrains the range of questions that are asked and answered. Laypersons (patients, community members) are rarely a part of the process of developing and refining a research question, even when they are representative of the population the research team proposes to engage or help. Engaging patients and community members can take a variety of forms, from advisory boards to pilot testers (screening forms, scientific measures, intervention components) to true collaboration on design, implementation, and analysis (with shared funding and ownership of data).
Patient and Community Engagement
Engaging patients, community members, or other stakeholders in research has been identified as a useful strategy for enhancing participation of underrepresented groups in the research process and, ultimately, reducing health inequalities and improving population health outcomes (Nguyen et al., 2021). For example, activism by the HIV/AIDS community led to the first federally funded community advisory boards (CABs) and galvanized the siloed research establishment (Karris et al., 2020). CAB recommendations led to trials of combination therapies instead of one or two drugs at a time, the creation of a participant’s bill of rights, a robust informed consent process, and early vaccine trials (Strauss et al., 2001). The active involvement of CABs helped establish national research priorities, including emphasizing the needs of underserved groups such as women, and was considered critical to the overall quality of AIDS/HIV research (Karris et al., 2021; Strauss et al., 2001).
Patient engagement in research refers to patients or caregivers serving as partners or leaders in the research process, resulting in study decision-making that incorporates the experiences, expertise, and values of these stakeholders (Harrington et al., 2020). Better understanding of patient- and community-level concerns about research and their needs for participation in clinical trials can lead to more effective outreach tailored to specific individuals and populations and improved patient experience in clinical trials through less arduous screening, more responsiveness to inquiries from potential participants, and more attention to participants’ needs (Forsythe et al., 2019; Smith et al., 2015). Systematic reviews indicate that patient engagement in research enhances study enrollment rates and increases participant retention (Crocker et al., 2018; Domecq et
al., 2014). These patient engagement relationships may be very individualized and personal between a patient and clinician or care setting, unlike community engagement, which is likely built between an academic team and/or clinician that works with more than one individual (Kimminau et al., 2018). Community engagement in research involves inclusive participation of people affiliated by geography, sociodemographic characteristics, or shared interests (Wallerstein et al., 2018). Community-based participatory research, a form of community-engaged scholarship that emphasizes rigorous partnered processes and focuses on community priorities, has been associated with significantly higher recruitment and retention of minority participants in research (Las Nueces et al., 2012; Yancey et al., 2006) and better behavioral and clinical outcomes (O’Mara-Eves et al., 2013). In one systematic review of clinical research studies, patient and community involvement in designing recruitment and retention strategies, developing patient-facing information, helping to identify potential participants, or providing feedback on poor recruitment rates was associated with higher odds of a patient enrolling in a clinical trial (Crocker et al., 2018). An exploratory finding in these analyses was that the effect size was significantly higher when there was substantial involvement of patients or caregivers with lived experience of the condition being studied.
Among the commonly cited barriers to conducting patient- and community-engaged research are defining the community or patient partners for collaborations, capturing and addressing diverse viewpoints and perspectives, time and budget restrictions, and lack of researcher training in patient and community engagement strategies (Domecq et al., 2014; Levitan et al., 2018; Nguyen et al., 2021). Patient- and community-engaged research have been found to be feasible in many settings with careful planning, adequate training, and appropriate funding for the collaboration (Crocker et al., 2018; Domecq et al., 2014). Cost concerns may be mitigated by consideration of the financial benefits of these engagement approaches. An analysis of the financial value of patient engagement found that engagement activities (such as patient advisory panels or patient reviews of the protocol) can reduce the need for protocol amendments and their associated delays and costs, increase enrollment, and reduce study dropouts (Levitan et al., 2018). For a generic oncology new molecular entity, the study estimates that patient engagement activities that avoid one protocol amendment and improve enrollment, adherence, and retention is the equivalent of accelerating pre-phase 2 product launch by more than 2.5 years (and by 1.5 years for pre-phase 3).
It has been more than 50 years since the participatory research paradigm gained traction in the social and health sciences (Wallerstein and Duran, 2006). In the intervening years, most academic institutions, health organizations, and funders have recognized the need for and required, at least to some extent, community engagement in research. However, the pressures of academia and scientific research often preclude meaningful engagement of communities, which can be a slow and challenging process. Among these pressures are discipline-specific
and institutional definitions of rigor and productivity, funders’ focal interests, timelines, and financial investments in research projects, and the researcher’s own scientific and professional interests (Cornwall and Jewkes, 1995). For example, time for building partnerships for a single study is typically not possible given the constrains in budgets and the limited timelines of grants. This makes it challenging to do partnership building, data collection, implementation of the study, analysis, and community dissemination under one contract grant. Nevertheless, many institutions and departments have CABs, community representatives on the institutional review board, and/or institutes that focus on community-academic partnerships. Federal funding agencies, like the National Institutes of Health (NIH), have similar mechanisms for public engagement in the research enterprise (Agnew, 1998). The proliferation of these features of the academic landscape is an acknowledgment of the importance of community engagement, but true engagement and empowerment requires an approach focused on co-learning and generating knowledge, rather than perfunctory stops along the research trajectory.
Research shows that health outcomes are improved when a patient and physician are of the same race. Alsan et al. (2019) found that Black doctors could reduce the Black-white male gap in cardiovascular deaths by 19 percent. Additional research has shown that hiring diverse staff and providing proper training for clinical staff are important facilitators for improved recruitment and retention of diverse clinical trials participants (Butler et al., 2013; Quinn et al., 2012). This is also one of the main facilitators to successful recruitment and retention of underrepresented populations in clinical research from the interviews discussed in Chapter 5. However, there is little research on, and therefore little evidence to suggest, that concordance between participants and the clinical workforce would increase participation in clinical research. Some studies have shown that diverse staff do not play a key role in participation in clinical research. For example, in one study of adults living with HIV, only 12 percent of respondents felt that having a research staff of the same race was important (Adeyemi et al., 2009). In another study of Black Americans who either elected or declined to participate in a study on kidney disease found that neither the gender nor the ethnicity of the recruiter had any influence on likelihood of enrollment (Gadegbeku et al., 2008). Similarly, a study promoting group management of heart failure among Black individuals found that most participants did not request a Black group leader (Rucker-Whitaker et al., 2006). However, some Black participants asked that the people helping to manage their diets provide culturally relevant suggestions. This may suggest that it is most important for staff to be able to give advice and relate to populations represented in clinical research, rather than being from the same racial or ethnic group as them. For example, respondents to a survey of Hmong--
speaking people said that speaking the same language was less important for participation than having a trusting relationship with researchers who were known and had created relationships within their communities (Lor and Bowers, 2018). This was also a finding from the interviews on facilitators to successful recruitment discussed in Chapter 5, which found that cultural and linguistic congruence with the target population was not enough and that gaining engagement and community buy-in for the study goals and desired outcomes was equally important.
Engaging with participants and building relationships requires genuine respect for individuals and their communities. The clinical psychologist Carl Rogers advocated for having an attitude that is “non-evaluative, nonjudgmental, without criticism, ridicule, depreciation, or reservations” for the patient (Patterson, 1985). This does not mean that physicians should change their values, but should not impose their own values and demand change from participants simply because they are the medical expert. In an editorial, Frosch and Tai-Seale (2014) suggest that “instead of lecturing (whether mentally or verbally) non-adherent patients, physicians can humbly inquire and ask the patient to reveal the reasons behind their behaviors, from which the physician can learn the barriers and identify potential levers for change.”
Despite a demonstrated willingness to participate in research, underrepresented populations are often not asked by researchers to participate in clinical studies (Adeyemi et al., 2009; Byrd et al., 2011; George et al., 2014; Katz et al., 2006; Murphy and Thompson, 2009; Webb et al., 2019). A contributing factor appears to be attitudes of the research staff and health-care providers who are responsible for recruitment. There is evidence that, while acknowledging the importance of diversity in an abstract way, many principal investigators may not see diversity as an important factor in their own work. A 2020 study of 313 researchers at a large research university found that while 87 percent of respondents believed that diversity was very or extremely important, only 38 percent reported that it was a priority in their own research programs (Passmore et al., 2020).
Principal investigators and study staff also bring their own biases to the research enterprise. Their perceptions about a potential participant’s reliability, health literacy, language skills, and social support, among other factors, all play into whether the potential participant will be offered information on enrollment (Joseph and Dohan, 2009). In one study, 92 percent of HIV/AIDS researchers felt that individuals with substance use disorders would need more support during trial participation than so-called traditional participants; 50–60 percent of these researchers believed that Black and Latinx individuals, as well as women, would also need additional support (King et al., 2007). In the same study, these researchers also had biases about their perceptions of the willingness of different groups to enroll in studies: 77 percent felt that white men were generally highly
interested compared with 33 percent for white women, 20 percent for Black men, 16 percent for Black women, 13 percent for Latinx men, and 11 percent for Latinx women—these numbers stand in contrast to the results of studies on willingness to participate, which show a high degree of willingness to participate in research among women and underrepresented minority populations (Adeyemi et al., 2009; Byrd et al., 2011; George et al., 2014; Katz et al., 2006; Murphy et al., 2009; Webb et al., 2019).
Investigator and staff biases may influence the amount of time and effort they expend recruiting participants from underrepresented populations. In one study, oncologists used far fewer words and spent significantly less time with Black patients than with white patients, in both the clinical care visit and discussion of clinical trial enrollment (Eggly et al., 2015). Additionally, discussion of clinical trials was less robust for Black patients, with more emphasis on voluntary participation and less focus on the purposes and risks of participation. A different survey of Black cancer patients found that only one-third of eligible patients reported being given written information on possible clinical trials (Brown et al., 2013). Among persons living with HIV/AIDS in the United States, Latinx respondents were less likely to know about research opportunities compared with both white and Black respondents, and Latinx and Black patients were less likely to be notified about possible enrollment by any member of any clinical or research team (Castillo-Mancilla et al., 2014).
Several studies have determined that the distance to health care and clinical research from a patient’s home, or home community, is also a factor that prevents participation. Most clinical research takes place at or near large academic centers that are less frequently used by some underrepresented populations compared with community health settings. The greater the distance between home communities and where patients are required to present for initial involvement, study visits, or exit interviews, the less likely they are to participate (Coakley et al., 2012; Sprague et al., 2013; Unger et al., 2016). Given the issue of distance, challenges with transportation have also been identified among the most common reasons for not participating in research studies (Brown et al., 2000). This relationship has been specifically established for Native Hawaiian and Pacific Islander populations, where individuals may be more likely to be living in remote areas and in under-resourced settings away from where research usually takes place (Giuliano et al., 2000). Conversely, a qualitative study assessing the effectiveness of offering transportation via a research van that would pick up participants in their home communities and then drive them to the research study site found that participants were highly satisfied with the convenience that transportation offered (Alcaraz et al., 2011). Research activities that do not offer transportation thus do so at the risk of excluding those without access.
Sampling and Recruitment Methods
Another factor that prevents recruitment of racial and ethnic minority population groups in research is the way that recruitment is typically performed. Sampling methods may decrease the chances of diverse enrollment. Often, random sampling methods simply do not result in large enough study populations to capture the needed diversity. Random sampling can miss people who may want to remain hidden for a myriad of reasons (e.g., fear of discrimination, prosecution) such as LGBTQIA+ individuals or people with substance use disorders (Bonevski et al., 2014). Different recruitment methods have been shown to work for different populations. Mass media, including television, radio, and newspaper ads, may work well for one group, while word-of-mouth is much more suitable for another (Bistricky et al., 2010; Coronado et al., 2012; King et al., 2011).
Inclusion and Exclusion Criteria
Another element of the existing research structure that serves as a factor or problem preventing participation for diverse research populations is the development and application of inclusion and exclusion criteria that restrict or undermine the inclusion of underrepresented and excluded populations. Eligibility criteria must be carefully designed and intentionally applied to address the question being evaluated and achieve accurate and meaningful results, yet these restrictions often lead to the unintentional and systematic exclusion of certain groups (Langford et al., 2014; McKee et al., 2013; Quiñones et al., 2020). For instance, asthma researchers trying to assess differences in bronchodilator response found that potential participants from underrepresented groups were more likely to have inadequate responses to the methacholine challenge, one of the inclusion criteria; however, the methacholine challenge cut-point may have lacked sensitivity for underrepresented populations given previously reported differences in methacholine responsiveness among different racial/ethnic groups (Hardie et al., 2010). Similarly, a lack of preexisting or baseline data may result in unintentional exclusion of underrepresented participants. Initial chart review to determine eligibility for a study on COPD (chronic obstructive pulmonary disease), for example, unintentionally missed patients without baseline spirometry data, despite the designers’ intentions to minimize exclusion criteria and maximize enrollment of underrepresented populations (Huang et al., 2019).
Lack of access to adequate health care is more common among underrepresented populations and can lead to delayed diagnoses and a more advanced form of disease, which can make individuals ineligible for study enrollment (Giuliano et al., 1998; Ward et al., 2004). Review of cancer trial recruitment among a medically underserved population that included American Indians found that restrictive inclusion criteria was one of the most common reasons for lack of enrollment
(Guadagnolo et al., 2009). Among the 88 potential American Indian participants, advanced stage/poor performance was the most commonly cited reason for non-enrollment (27 percent).
An analysis of the exclusion criteria for a study on smoking cessation found that Black and Latinx patients were more frequently excluded than white patients (Hooper et al., 2019). Additionally, in this analysis, white patients were usually excluded for a single reason, such as serious mental illness, difficulty with attendance, or medical conditions, whereas Black patients were more than twice as likely to be excluded for three or more reasons, such as smoking status, barriers to attendance, lack of motivation, or other health contraindications. Another report, also on eligibility for a smoking cessation study, found that despite being nearly twice as likely as white contacts to complete initial telephone screening, Black contacts were less likely to be eligible for enrollment (King et al., 2011). This difference persisted even when controlling for demographic factors such as education, gender, and income level. These analyses illustrate how the structure of current inclusion and exclusion criteria, intentionally or not, reduces opportunities for underrepresented individuals to participate in research.
Researchers are often not trained or skilled in explaining research methodologies or the potential positive impacts of research outcomes in ways that actively engage ethnically underrepresented populations (Bonevski et al., 2014; Hughes et al., 2017). Studies examining publicity and advertising for recruitment into research studies have identified a general failure to message the positive implications of research outcomes. Yet, there is evidence that this problem can be solved or mitigated. In a qualitative study exploring reasons for consent or refusal to participate in a comparative effectiveness study, researchers found that further explaining how a comparative effectiveness study works—for example, emphasizing that it does not test new medications—increased respondent’s positive views of the study (Behringer-Massera et al., 2019). A group of researchers in Baltimore, Maryland, were able to successfully recruit a diverse cohort of more than 3,700 participants into a 20-year longitudinal study on aging in part by focusing on the direct benefits to the enrollees of their participation (Ejiogu et al., 2011). Similar studies have shown increased interest in research when people believe the research might provide personal, familial, or societal benefits (Boise et al., 2017; Gadegbeku et al., 2008).
Finally, the length and complexity of the research process, especially the consent process and consent forms, has also been reported as a factor that prevents enrollment for underrepresented populations (Durant et al., 2014; Nipp, Hong and Packett, 2019; Hamel et al., 2016; Langford et al., 2014). In one study of people
living with HIV that included predominantly Black and Latinx individuals, 19 percent cited the consent form being too hard to understand as a reason why they did not participate (Adeyemi et al., 2009). However, this barrier is not unique to underrepresented populations, as overly complicated consent forms are a barrier for all groups to participate in research (Kass et al., 2011; Sauceda et al., 2021).
The long time frame of most research projects may also reduce willingness to participate or remain enrolled in a clinical trial. A model created using data from potential cancer research participants found that the longer the time between a potential participant’s consent to first contact by a study team member predicted probability of attrition; this effect was higher among racially underrepresented people compared with white individuals (Azfar-e-Alam et al., 2008). Current consent processes and consent forms are linguistically and culturally inappropriate for many underrepresented groups (Sauceda et al., 2021).
Additionally, researchers need to provide more appropriate recruitment materials, tailored to the language and literacy needs of potential research participants. The lack of suitable study materials in their respective languages has been shown to reduce participation of Asian, Creole, Hmong, Latinx, and American Indians, as well as Native Hawaiian/Pacific Islanders (Byrne et al., 2014; Calderon et al., 2006; Giarelli et al., 2011; Giuliano et al., 2000; Huang et al., 2013; Lawrence, 2000; Lor and Bowers, 2018; Nguyen et al., 2005; Occa et al., 2018; Tu et al., 2005). Even when language-specific materials are available, the quality and integrity of those materials may not be high. For many languages, verbatim translations are unlikely to capture the true meaning of the materials without incorporating commonly used idioms and culturally appropriate phrasing. This may especially be true for Spanish-speaking groups, as there are significant differences among the languages spoken in different Spanish-speaking countries (Occa et al., 2018). Translations that do not reflect the appropriate dialect or accepted verbal usage patterns can further discourage targeted populations from enrolling in a study. Fundamentally, it is critical that linguistic and literacy needs of diverse research participants are met.
All of the above-mentioned challenges—study design, outreach methods, choice of incentives, and research processes—are exacerbated by time and financial restrictions placed on researchers. Prioritizing speed, combined with a historically uninformed approach to minority recruitment, has led to a system in which research trials do not adequately prioritize enrollment of underrepresented populations.
Health-Care Access and Strong Primary Care
Closely related to socioeconomic status is access to health care. Lack of or limited health-care access is a root cause of inequitable health care throughout the United States. In a recent study of individuals just before and after age 65 (age of Medicare eligibility), Wallace et al. (2021) found that those eligible for
Medicare showed a marked reduction in racial and ethnic disparities of insurance coverage, access to care, and self-reported health. Besides obvious health-care consequences, this inequity also has implications for research. Patients who are not actively engaged with the health-care system will have limited opportunity for enrollment in studies.
Strong, trusting relationships with primary care providers (PCPs) have been noted to have significant impacts on research engagement (Adeyemi et al., 2009; Buchbinder et al., 2004; Friedman et al., 2020; Gadegbeku et al., 2008; Trantham et al., 2015). One study performed in five geographically diverse health-care centers (New York City; Baltimore, Maryland; Birmingham, Alabama; Iowa City, Iowa; and Boston, Massachusetts) found that the positive endorsement of a PCP led to increased likelihood of participation, while a negative attitude almost always led to a refusal to enroll (Buchbinder et al., 2004). In another study, simply having a PCP was the strongest predictor of clinical trial follow-up among a population of predominantly ethnically underrepresented individuals; socioeconomic status was not significantly associated with follow-up (Friedman et al., 2020). A North Carolina study on the involvement of Black male cancer survivors in research found that these patients and their families expressed significant trust in their physicians and would be open to enrollment in a research study if their physician suggested it (Trantham et al., 2015). Conversely, patients who are reluctant to visit their PCPs are more likely to be nonparticipants in medical research (Gadegbeku et al., 2008). This pattern also holds for other members of the health-care team, such as nurses, with patients reporting that they would not participate in a trial if their nurse does not recommend it (Adeyemi et al., 2009).
As described earlier in this chapter, several studies have also determined that the distance to health care and clinical research from patient home or home communities is also a problem, since most clinical research takes place at or near large academic centers that are less frequently used by underrepresented populations compared with community health settings. The greater the distance between home communities and where patients are required to present for initial involvement, study visits, or exit interviews, the less likely they are to participate (Coakley et al., 2012; Sprague et al., 2013; Unger et al., 2016).
LANDSCAPE FOR RESEARCH—COMMUNITY AND POLICY FACTORS THAT INFLUENCE THE REPRESENTATIVENESS OF CLINICAL TRIALS AND RESEARCH
Diversity interdigitates with each stage of the clinical trial and clinical research process. In the ideation stage, some questions might not be asked if there is not diversity among principal investigators and faculty driving the research questions, as described above. In addition, diversity of studies will also be affected by where the site is chosen for recruitment, and how it occurs.
The larger research enterprise and environment required to support diverse research studies present additional factors and problems that prevent the inclusion of a diverse research population.
Research Infrastructure to Facilitate Diversity in Clinical Trials and Clinical Research
Academic Medical Centers
As of 2019, academic medical centers comprise the nation’s 154 accredited medical schools and more than 400 major teaching hospitals and health systems. These institutions conduct 55 percent of the extramural medical research supported by the NIH and operate 98 percent of the nation’s 41 comprehensive cancer centers (Fisher, 2019). As such, these centers have substantial influence on the clinical trial enterprise.
Nevertheless, the traditional academic medical center structure creates substantial barriers to adequately consider diversity, equity, and inclusion in clinical trials and research. Sustainably and meaningfully engaging underrepresented and underserved populations often does not align with the traditional paradigm of promotion and tenure. Traditional academic centers mostly value teaching, research, and service, and a researcher’s success is mostly judged by their productivity in publishing research and obtaining grant funding. Applying principles of community-based participatory research (see the Patient and Community Engagement section, above) and recruiting diverse population groups into clinical trials and research is time-consuming and requires investments to build and sustain trust; these investments are often only minimally considered in promotion and tenure decisions. This scenario often creates little incentive for early-stage investigators to invest time and resources to build community relationships. Moreover, academic institutions often provide inadequate institutional resources for researchers to engage communities, especially beyond the lifespan of a single research project.
Additionally, recruitment and retention of diverse faculty and staff are challenges in many academic medical centers. Research shows that women, particularly those from racial and ethnic minority groups, as well as men from racial and ethnic minority groups, are underrepresented in medical faculty. In an analysis conducted by the Association of American Medical Colleges, only 3.6 percent of medical school faculty are Black, 3.2 percent are Hispanic, and only 0.2 percent are American Indian or Alaska Native (AAMC, 2020). Further, although overall, women are about at parity with men at the medical school level, women leave the profession as they move throughout the career pipeline. For example, in 2019, women composed 48 percent of medical school graduates but only 41 percent of the full-time faculty. Further, women made up only 37 percent of associate professors and 25 percent of full professors in academic medicine. The percent-
age of women chairs and deans is even lower, at 18 percent for both positions (AAMC, 2020). In addition, racial and ethnic subgroups of women face a double bind in medicine and are even more underrepresented at higher academic ranks. Although racial and ethnic subgroups of women represent 18 percent of the U.S. population, only 3.2 percent of full professors in medicine are women from racial and ethnic subgroups (Carapinha et al., 2017; NASEM, 2020a).
Although academic medical centers have a long-standing history of community service, many underrepresented and underserved communities lack trust in these institutions. Academic researchers have been referred to as “in-and-out” researchers or “parachute” researchers (Stefanoudis et al., 2021), where one takes and does not give back to the community that has enabled the research success. Wilkins and Alberti (2019) argue that to address health inequities and truly engage with communities to address their research needs, academic health centers will need “commitments from institutional leaders, infrastructure to support engagement, and changes in policies to fuel innovative partnerships, facilitate community partner integration, and reward community-engaged scholarship.” There is a growing need for academic medical centers to shift from community service to an “enterprise-wide approach” to community engagement in order to advance their missions of clinical care, education, and research. This shift in approach is critical to understanding, examining, and addressing the social determinants of health and structural barriers relevant to underrepresented and excluded communities.
Engagement opportunities across academic health centers and their benefits are described in Table 4-1.
Community Health Centers
Other types of health services organizations, including community health centers and rural health centers typically provide care to diverse population groups and represent an untapped resource for clinical trial and research recruitment. Federally qualified health centers (FQHCs) are grantees of the Health Resources and Services Administration, under Section 330 of the U.S. Public Health Service Act (P.L. 78-410), and include migrant health centers, health care for the homeless health centers, and public housing primary care centers. FQHC Look-Alikes meet all the requirements of health centers and reap most of the benefits of health center status, but do not receive a federal grant. Rural health centers can be public, nonprofit, or for-profit health-care facilities; however, they must be in rural, underserved areas.
These health centers provide care to a diverse population and have even shown reduced mortality in treatments compared with hospitals (Wennburg et al., 1998). For FQHCs and FQHC Look-Alikes, in particular, over two-thirds (68 percent) of patients who seek care have patient incomes at or below the poverty level, 22 percent are uninsured, and 47 percent are covered by Medicaid. More
|Mission||Community-Engagement Opportunity||Benefit to Community||Benefit to Academic Research Organizations|
|Research||Scientists, regardless of discipline, develop research questions in collaboration with community.a||Aligns research resources with local needs; increases connection to STEM mentors and training; develops community capacity to use research, seek grants, and increase community-based organization’s sustainability; and ensures data can be used to support local advocacy efforts.||Increases relevance of research and likelihood that findings will be broadly implemented; increases recruitment and retention in clinical studies; enhances scientists’ competitiveness by strengthening external validity; increases internal validity by adding community perspective to construct definitions and measurement tools or strategies; produces stories useful for marketing and advocacy; and develops trainees’ skills in communication, collaboration, and engagement.|
|Researchers work with community members to improve the relevance and conduct of studies, as well as the dissemination of findings and discoveries.|
|Research centers invite community members to serve on search committees and interview faculty applicants, and incorporate those perspectives into hiring decisions.||All of the above, plus provides the community the opportunity to exercise agency and influence decisions and increases opportunities for mutually beneficial projects.|
|Education||Educators integrate the community and community health needs assessments when developing interprofessional learning opportunities. Community-based learning is evaluated in terms of outputs and outcomes relevant for learners, community members, and the research organization itself.b||Ensures learner service aligns with community needs in respectful and valued ways; evaluation allows improvement to community-based organization’s program and exposure to evaluation science, which is important for the partner agency’s own improvement efforts; and learners’ passion and commitment present a different side of the health-care system.||Develops interprofessional competencies; develops trainees’ communication, collaboration, and engagement skills; exposes learners directly to local sociocultural contributors to health; and produces stories useful for marketing and advocacy purposes.|
|Mission||Community-Engagement Opportunity||Benefit to Community||Benefit to Academic Research Organizations|
|Learners across health professions directly contribute to local community health needs assessments processes as data collectors or analysts, or by presenting results to community groups.||Increases exposure and connection to learners, increases awareness of local health-improvement activities, and presents more opportunities to codesign community health needs assessments-related health interventions.||Provides additional labor for teaching hospitals’ community-related administrative functions; provides research practicums focused on survey design, focus group development and execution, data analysis, data reporting, program development, etc.; offers educators new opportunities to teach about social determinants of health, population heath, public health, etc.; and provides graduate medical education involvement and contributes to instruction on health and health-care disparities.|
|Program directors routinely model the stratification of their patient and participant data by sociodemographic characteristics to identify health-care inequities. Trainees partner with community members, patients, and faculty to develop interventions.||Results in improvements to work flows more likely to benefit patients’ and community members’ health outcomes.||Contributes to instruction on health and health-care disparities; targeted disparity-focused quality improvement efforts can have an effect on overall measured quality; when implemented in an accountable care organization or similar setting, can result in increased shared savings; advances scholarly output; and increases trainees’ patient and community-engagement skills.|
|Mission||Community-Engagement Opportunity||Benefit to Community||Benefit to Academic Research Organizations|
|Clinical care||Clinical teams use data across multiple levels—clinical, sociodemographic, and neighborhood—to tailor care plans in ways that are responsive to the health and the environmental or social profiles of their patients.||Improves health outcomes, enhances knowledge of and access to community assets, and increases demand or support for local community-based organizations’ programs.||Improves quality of care, particularly on measures related to readmissions, cost, and resource use; enhances physician and provider wellness through increased ability to manage patients’ social factors; increases efficiency and effectiveness of hospital community health or prevention efforts by enhancing alignment or reducing redundancy with local initiatives; and advances scholarly output.|
|Clinicians and care teams, through their electronic health records, have robust linkages to hospitals’ community health-improvement efforts and make appropriate and timely referrals to community assets that can provide social support and resources for patients and their families.|
|Care team members spend time at community-based referral partners meeting staff, engaging patients, and learning about local social service processes to improve their community knowledge and profile and to increase their ability to make appropriate, knowledgeable referrals.|
SOURCE: Table adapted from Wilkins and Alberti, 2019.
than 28 million patients received care at 1 of 1,375 FQHC delivery sites in 2020 (an additional 679,000 were served at 87 FQHC Look-Alikes in 2020), with FQHCs and FQHC Look-Alikes treating nearly one in seven uninsured people in the United States (HRSA, 2020a, 2020b). Greater than 62 percent of FQHC and FQHC Look-Alike patients are members of racial/ethnic minority populations, including 37 percent who are Latino and 26 percent who are African American. More than one-quarter (28 percent) of patients are best served in a language other than English (HRSA, 2022). Similarly, rural health centers serve more than 8 million people across 4,400 delivery sites in 45 states (NARHC, 2022). As such, community health centers are an ideal setting to recruit diverse participants into clinical trials and research. This research activity, however, will require infrastructure and support for the community health centers.
Nevertheless, the barriers to clinical trials and research recruitment at community health centers are multifactorial. Health-care providers who work in community settings outside of academic centers may have limited knowledge about available research opportunities. This may be particularly true in rural communities (Paskett et al., 2002). The same is even true for physicians near academic medical centers: in a survey of more than 100 physicians in New Jersey, lack of awareness of cancer research opportunities was reported by 95 percent of PCPs, 84 percent of non-oncology specialists, and even 50 percent of oncologists (Hudson et al., 2005).
While electronic health record (EHR)–based prescreening of patients for trial eligibility has been a common practice in many health systems (Canavan et al., 2006; Sullivan, 2004; Wilcox et al., 2009), it poses challenges for many health systems, especially those lacking sufficient EHR infrastructure. Some clinics may be unable to successfully query the EHR using study inclusion and exclusion criteria. Often EHRs have heterogeneous data structures that can make it difficult to consistently apply study inclusion and exclusion criteria across sites (for multisite studies) (Hersh et al., 2013; O’Brien et al., 2021). Health centers that lack onsite specialty care services, may inadequately track the delivery of clinical services completed offsite. Many of these problems are more pronounced at community health centers, which may have limited data infrastructure and fewer staff trained to carry out research functions. Initiatives such as the Community Health Applied Research Network are working to address some of these infrastructure challenges and may serve as a model to expand and grow capacity for FQHCs (see Box 4-1).
Individuals who receive care at community health centers may frequently change their address or phone number or may be houseless, creating obstacles to being recruited for study participation. These individuals also may face competing demands, such as work or caregiving responsibilities, or may lack transportation to attend research-related appointments. Patients’ health insurance coverage, which may be inconsistent or variable, may present barriers to obtaining the needed clinical services required for study participation. A high participant no--
show rate and the need to translate study materials into multiple languages (and/or enlist interpreter services) may impose additional study costs.
Specialty associations, such as the Association of Black Cardiologists and the Association of Black Gastroenterologists and Hepatologists, among others, can serve as effective organizations to promote clinical trial recruitment in traditionally underrepresented population groups (Ofili et al., 2019).
Drug and Device Companies and Clinical Trial Recruitment Centers
As health care evolves toward precision medicine, it is essential that the biologic differences among populations—and how these differences affect pathology, response, tolerability, and outcome—are comprehensively investigated in the context of clinical trials. Pharmaceutical and device companies have an essential role in developing and implementing successful strategies, measurable
outcomes, and robust outreach plans to include diverse populations efficiently and effectively in clinical trials.
Overly restrictive study design, stringent eligibility criteria, and continuous activation of clinical trials in sites based on their academic prominence or speed of enrollment often has resulted in the exclusion of underserved patient populations (much to the detriment of inclusive research). This has contributed to the widening disparities between patients who are expected to benefit from the new research in day-to-day clinical practice. It is clear that eliminating the factors and problems that limit trial participation would improve the generalizability of results. Problems that prevent the inclusion of diverse populations in industry-funded clinical trials include patient out-of-pocket costs, which are often not covered in the informed consent process; industry pressures to gather data quickly; and the selection of easy-to-recruit samples being incentivized (Iltis, 2004). Payment structures often pay per participant, which further incentivizes institutions to focus recruitment on populations that are easiest to recruit. Although many of these problems are not unique to industry-sponsored trials and are present in federally funded research as well, most clinical trials are industry-funded and the business demands of industry make these problems particularly acute.
Institutional Review Boards2
All research that involves human subjects must be reviewed and approved by an institutional review board (IRB) (see the Common Rule, 45 CFR 46).3 IRBs are charged with protecting the rights and welfare of human subjects who participate in research. The evaluation of human subjects’ rights and welfare is guided, in part, by key ethical principles established in international and national guidelines such as the Declaration of Helsinki) and the Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979; WMA, 2008). The Belmont Report, issued in 1979, was commissioned by law in response to the abuse of human subjects in the U.S. Public Health Service Syphilis Study at Tuskegee (Brandt, 1978). The report explores the boundaries of medical research, the determination of risk versus benefit in research, the appropriate selection of human subjects for participation in research, and the fundamentals of informed consent. Importantly, it also
2 This section relates to IRBs that exist under Food and Drug Administration regulations. It is important to acknowledge that research done on tribal lands falls under the individual tribes’ IRBs, because tribes are sovereign nations. See Kuhn et al., 2020.
3 See also Consideration of the Principle of Justice under 45 CFR part 46, July 22, 2021, at https://www.hhs.gov/ohrp/sachrp-committee/recommendations/attachment-a-consideration-of-the-principleof-justice-45-cfr-46.html.
outlines key ethical principles to guide research with human subjects. These ethical principles include (1) respect for persons, which refers to the right to self-determination, or autonomous decision-making; (2) beneficence, which refers to the obligation to protect the well-being of human subjects; and (3) justice, which refers to the fair distribution of the benefits and burdens of research participation.
The ethical principles outlined in the Belmont Report are operationalized through the day-to-day work of IRBs, which operate under the guidelines and administration of the Health and Human Services (HHS) Office for Human Research Protections (OHRP). The Code of Federal Regulations (45 CFR 46) guides the structure and function of IRBs, and particularly, the IRB review process. According to the Code of Federal Regulations, IRBs must have at least five members, and those members must have sufficient knowledge or experience to evaluate research activities proposed by investigators affiliated with the institution. IRBs are charged with ensuring that risks to human subjects are minimized through the use of sound scientific processes and their review focuses on the following key elements of research proposals: risks and benefits to human subjects, safety, protections of privacy, equitable selection of human subjects, and informed consent, with particular attention to coercion and undue influence. The latter element of IRB review can present barriers to enrolling excluded and underrepresented populations.
The ethical principle of respect for persons is operationalized in the research consent process, which is meant to support participants’ right to autonomous decision-making, and protect participants with diminished capacity for self-determination. The Belmont Report describes capacity for self-determination as fluid—increasing with maturity, but potentially lost in some natural and social circumstances such as severe illness, cognitive disability, or restricted liberty. The Code of Federal Regulations adds more clarity to this idea by its identification of groups whose vulnerability demands increased protection beyond those afforded to all human subjects in research. Those groups include children, prisoners, persons with impaired decision-making capacity, and economically or educationally disadvantaged persons (45 CFR 46). The code describes these groups as vulnerable to coercion and undue influence in research participation, and therefore directs IRBs to pay particular attention to their consent to research. The code does not, however, define coercion or undue influence, thereby leaving the interpretation to IRBs, who have largely focused on the potential for compensation and incentives to be unduly influential or coercive (Largent and Lynch, 2017).
Most research with human subjects involves some form of compensation for participation. Forms and amounts of compensation—sometimes referred to as incentives—vary by study, with one-time surveys and interviews typically offering smaller incentives compared with lengthy clinical trials that involve medical interventions and frequent study visits. IRBs tend to lean toward viewing higher payments as coercive, and err on the side of keeping payments low (Largent and Lynch, 2017). However, coercion requires “the overt threat of harm”
to gain another person’s compliance (DHEW, 1979). Thus, some ethicists argue that research payments cannot be considered coercive, which is a perspective articulated by the director of HHS OHRP (Largent and Lynch, 2017; Meeker-O’Connell and Menikoff, 2021). Undue influence refers to an offer that encourages the potential recipient to do something that is unreasonably against their best interests or values (Emanuel, 2005). It does not refer to an offer that encourages the potential recipient to do something reasonable that they might not do in its absence. IRBs cannot approve studies that pose unreasonable risk to potential participants, which means that any approved study should be considered a reasonable undertaking for its target population, on the whole. Thus, it is difficult to argue incentives are a form of undue influence; yet, research indicates IRB members are concerned about coercion and undue influence when substantial payments are offered to research participants (Largent et al., 2012).
IRB members’ concern about coercion and undue influence in the form of incentives reflect their commitment to the canonical principle of respect for persons. However, limiting incentives may ultimately compromise other equally important principles, including beneficence and justice. Some research provides direct benefit to participants, thereby supporting their well-being. In the absence of sufficient payment or other supports (e.g., food, transportation, childcare), persons who might benefit from research participation are prohibited from doing so. The people most likely to bear an excess economic burden of research participation, especially in the absence of substantial support, are those who are in hourly jobs, or live far from academic research centers, or have dependents for whom they must provide care (Nipp et al., 2016). Excluded and underrepresented populations are more likely to be in these social circumstances. Thus, without adequate support, their ability to participate is restricted, they miss opportunities to enhance their well-being, and the distribution of research benefits and burdens is unjust. The underrepresentation of particular demographic groups also limits the opportunity to generate sufficient data on the safety and efficacy of new therapeutics for them; this may create injustice in delayed access to interventions, or in unforeseen differential outcomes (Hume et al., 2017; Knopf et al., 2020). Although not the focus of this report, it is also important to recognize the negative impact IRBs can have on the enrollment of adolescents in clinical trials and clinical research. Since many IRBs require guardian consent, this may disallow adolescent participation if a parent or guardian is not comfortable with or is distrustful of clinical trials and clinical research. Allowing adolescents to make independent decisions on whether they would like to enroll in a clinical trial or not may reduce barriers to enrollment and further understanding of health disparities (e.g., sexual health or substance use) in adolescent populations (Fisher and Mustanski, 2014; Fisher et al., 2021; Gilbert et al., 2015; Knopf et al., 2017).
Although IRBs certainly have a role to play in increasing the representation of excluded and underrepresented populations in clinical trials and clinical
research, these bodies are focused on protecting individuals, not communities. Engaging with community advisory boards offers an opportunity for researchers to anticipate and address community concerns and to help communities understand the risk of the proposed research (Quinn, 2004; Strauss et al., 2001). CABs can also facilitate the involvement of community members on local IRBs, offering additional protections to community members and helping to alleviate issues of trust.
Research funders have several roles and responsibilities that can influence the diversity of clinical trials. Traditionally, funders’ roles include prioritizing research topics, approaches, and methods; receiving and evaluating grant applications; selecting suitable proposals for funding; and evaluating the output of the research (Brantnell et al., 2015; Kessler Foundation, 2011). In each of these stages, funders have opportunities to promote diversity, but they also face constraints that may limit the effectiveness of efforts to enhance representativeness.
Funding Priorities. Funders set and implement research funding agendas that can ultimately affect the clinical research that is conducted and the scope of these projects. Sharing this agenda through published criteria and grantee informational sessions provides opportunities for emphasizing the ethical, scientific, and clinical importance of diversity in clinical trials. Moreover, the research agenda is typically informed by few scientists from underrepresented groups. Within the National Cancer Institute’s (NCI) Intramural Research Program, for example, only 1 percent of senior investigators (those granted tenure by the deputy director for intramural research) are Black and 2 percent identify as Hispanic. There are no Black and Hispanic senior scientists and clinicians (managers of large institutes’ or centers’ research departments) at the NCI. Instead, three-quarters or more of the NCI’s senior scientists and investigators are white. Nearly two-thirds of R01s are awarded to white applicants, with Black scientists and Hispanic scientists making up only 1 percent and 5 percent, respectively, of awardees (Ong, 2021). Further, although the NIH provides diversity supplements to investigators to support a diverse and inclusive workforce, many are limited to 2 years for the training of junior investigators, which means very little time for establishing partnerships, recruitment, and retention for projects.
Funders can also prespecify diversity targets for the research studies. This approach has been successful in several large research studies, including the NIH Diabetes Prevention Program and Systolic Blood Pressure Intervention Trial, or SPRINT (Group, 2015; Knowler et al., 2002). SPRINT, for example, which examined blood pressure in 9,361 people, set specific recruitment targets and ensured that trial sites were diverse and could bring in diverse patients to achieve recruitment goals (Ambrosius et al., 2014; Greer, 2015).
Proposal Reviews. Through the review process, funding agencies can give
priority to projects that include sufficient numbers of underrepresented persons. The NIH has implemented initiatives designed to foster the inclusion of underrepresented groups in NIH-supported clinical research trials and to incorporate valid analyses by sex and gender (NIH, 2001b, 2017a). Federal funders, such as the NIH, National Science Foundation, and Patient-Centered Outcomes Research Institute, require enrollment tables that are incorporated as part of the review process. The weight that reviewers give data may vary, however, and can be applied inconsistently. Further, these enrollment tables are not part of the score-driving criteria, which limits the impact they have on funding decisions.
In addition to review criteria, those reviewing grant applications and making funding decisions influence the type of research that is carried out. Studies have shown that NIH study sections, which review and decide which clinical research grants get funded, are overwhelmingly white. According to one study, 2.4 percent of study section members in the period FY 2011–2015 were African American/Black compared with 77.8 percent who were white (Hoppe et al., 2019). The Center for Scientific Review at the NIH is tasked with improving disparities in peer review and has stated that “there must be diversity with respect to the geographic distribution, gender, race, and ethnicity of the membership of study sections” (NIH, 2020b). It is important to acknowledge the availability of these data that illustrate the lack of diversity among academic medical centers and study sections for publicly funded biomedical research. Moreover, data on investigators in industry and other private entities are not publicly available.
Funding for Recruitment and Retention. Recruitment and retention of diverse participants can be costly; one study estimates recruitment costs ranging from $129.15 to $336.48 per enrolled patient (Penberthy et al., 2012). Recruitment can require higher staffing levels, more frequent contacts, longer accrual periods, additional funding, and more flexible funding to enhance trial accessibility for low-income participants, those with caregiving responsibilities, workers without flexible hours, and individuals with other competing priorities. For example, transportation to trial sites is often a deterrent to participation of underrepresented populations, and need for childcare can limit participation of caregivers, who are more likely to be female. Collaboration with community organizations to colocate services in community venues, such as faith institutions, or provide mobile services may increase the reach and effectiveness of the clinical trial recruitment and retention efforts. Many funding agencies often underestimate the increased effort and financial resources needed to ensure diversity in research studies. Flexible funding that can be used to promote or augment these strategies can play a critical role in increasing trial diversity and warrants consideration by IRBs (see Institutional Review Boards section, above).
Post-award Reporting and Monitoring. Review and examination of participant accrual and review of adverse events is a routine role of funders as well as the Food and Drug Administration and individual data safety monitoring boards. In this phase, systematic, timely, and transparent collection and reporting of trial
diversity metrics is a requirement for intervening to modify trial protocols to promote more inclusive recruitment or avoid differential disenrollment (Artiga et al., 2021).
Evaluating the Output of the Research. In addition to clinical effectiveness outcomes, funders often examine the impact of their research using metrics such as publications and patents produced. Making the diversity of participants that are recruited and retained an explicit outcome to be evaluated and reported can be an important strategy for enhancing patient and community trust in the research process, increasing the applicability of the research findings to women, minority communities, and older adults, and influence payers’ and providers’ acceptance of the findings for groups who were not adequately represented in the research.
In many ways, medical journals serve as the gatekeepers to medical knowledge, holding the key to publishing studies that advance clinical practice and improve health. Thus, journals yield great power along with accountability for what is and is not published in their pages. In the past year, leading medical journals have acknowledged that “they must do better towards inclusion and antiracism in all journal related activities,” and many have issued initiatives and calls to action to increase diversity, equity, and inclusion. In October 2021, the New England Journal of Medicine announced it would begin requiring authors to submit a supplementary table describing the disease or health problem under study, its distribution in the population (e.g., by race, ethnicity, and sex), and representativeness of enrolled study participants (NEJM, 2021). This is an important step, but whether it becomes a significant factor in determining the acceptance of manuscript submissions remains unclear. Further, journals still have a long way to go, as reflected by representativeness of editors and by their rate of publications in these areas.
Regarding representation, among the 346 editors and editorial board members across JAMA and the JAMA Network journals, 71 percent are white, 19 percent Asian, 6 percent Black, and 4 percent Hispanic; 38 percent are women (Fontanarosa et al., 2021). In a review of 444 leading medical journals, women represented only 21 percent (94) of editors in chief (Pinho-Gomes et al., 2021), and this rate has changed little over the past decade (Jacobs et al., 2021). Of 215 leading surgery journals, only 7 percent of editors are women (Kibbe and Freischlag, 2020). For publication rates on issues of diversity or health disparities, a recent review indicated that the proportion of articles on these topics relative to all articles published was only 7 percent at the Journal of General Internal Medicine and was less than 2 percent at other leading general medicine journals (Jackson et al., 2021). Moreover, few papers in any journal addressed “racism” in their title, abstract, or key words (Jackson et al., 2021; Rhea et al., 2020).
While some journal editors have expressed skepticism about the power of
journals, and post hoc publication, to influence inclusion in clinical research, there is in fact evidence to suggest that journals can exert a large impact. The requirement of adherence to reporting standards, such as CONSORT (Consolidated Standards of Reporting Trials) (Moher et al., 2010) and other standards, and mandatory trial registration, such as ClinicalTrials.gov, have standardized and raised the quality of research design, reporting, monitoring, and transparency for clinical trials. There is no question that similar influence could be exerted with the requirement for mandatory standards and reporting on diverse inclusion in clinical studies. Moreover, journals can influence diversity and inclusion across multiple domains (Rivara et al., 2021). Ultimately, a unified and concerted effort by medical journals, such as through the International Collaboration on Standards and Policies through the Royal Society of Chemistry’s Joint Commitment for Action on Inclusion and Diversity in Publishing (RSC, n.d.), may pave the way for ongoing and long overdue change.
This page intentionally left blank.