The Representation Gap: How Racially Homogeneous Clinical Trials Are Undermining Care for Millions of Americans

The United States is among the most racially and ethnically diverse nations in the world. Its clinical research enterprise, by most measures, is not. For decades, the trials that generate the evidence underpinning prescribing guidelines, surgical protocols, and preventive care recommendations have enrolled patient populations that bear little demographic resemblance to the country those recommendations are meant to serve. The consequences of this mismatch are not abstract. They manifest in treatment protocols that perform differently across populations, in drug dosing standards calibrated to physiologies that do not represent the majority of patients receiving them, and in a compounding erosion of trust between minority communities and the institutions asking them to participate in research.

A Documented Enrollment Disparity

The scale of underrepresentation in US clinical trials is well characterized in the literature, even if its causes remain contested. A 2021 analysis of trials supporting FDA drug approvals between 2015 and 2019 found that Black patients constituted approximately 8 percent of trial participants despite representing roughly 13 percent of the US population. Hispanic patients accounted for about 11 percent of enrollment against a population share exceeding 18 percent. Asian Americans, Native Hawaiians, and Pacific Islanders were similarly underrepresented, and American Indian and Alaska Native populations were so infrequently enrolled as to render their data essentially invisible in aggregate analyses.

Cardiovascular research illustrates the problem with particular clarity. Heart disease remains the leading cause of death in the United States and disproportionately burdens Black Americans, who experience higher rates of hypertension, heart failure, and cardiac mortality than any other demographic group. Yet landmark trials that established first-line treatment protocols for heart failure—studies that continue to inform current ACC/AHA guidelines—enrolled Black patients at rates well below their disease burden. When those protocols are applied uniformly, they are applied on the basis of evidence that was never collected from the populations at greatest risk.

When Homogeneous Evidence Meets a Heterogeneous Population

The clinical stakes of this imbalance become most visible when pharmacological responses diverge across populations in ways that homogeneous trial data could not have predicted. The history of BiDil, the first drug approved by the FDA specifically for use in a racial subgroup, illustrates the complexity. The medication—a combination of isosorbide dinitrate and hydralazine—demonstrated significant mortality benefits in a trial conducted exclusively in self-identified Black patients with heart failure, after earlier trials in mixed populations had failed to show a significant overall effect. The episode raised legitimate scientific questions about population-specific pharmacodynamics while simultaneously exposing how the absence of disaggregated data in earlier research had obscured a clinically meaningful signal.

Drug metabolism differences represent another domain where underrepresentation generates measurable risk. Polymorphisms in cytochrome P450 enzymes, which govern the metabolism of a wide range of commonly prescribed medications, vary in prevalence across racial and ethnic groups. CYP2C19 loss-of-function variants, which affect clopidogrel metabolism and therefore antiplatelet efficacy, are present in approximately 15 percent of white patients but in up to 30 percent of East Asian patients. When clopidogrel dosing standards are derived from predominantly white trial populations and applied uniformly, the therapeutic implications for patients with different metabolic profiles are not incidental—they are a predictable consequence of inadequate representation at the evidence-generation stage.

Oncology provides additional examples of consequential gaps. Certain EGFR mutations that predict responsiveness to targeted therapies in non-small cell lung cancer occur at higher frequencies in Asian patients than in white patients. Clinical trials that enrolled few Asian participants generated efficacy data that understated the potential benefit of these agents for a population that stood to gain most from them. Delayed access to effective targeted therapy in those patients is a direct downstream consequence of upstream enrollment decisions.

Barriers to Equitable Participation

Addressing the enrollment gap requires an honest accounting of why it exists. No single explanation suffices. Historical abuses—most infamously the Tuskegee Syphilis Study, in which Black men were deliberately denied effective treatment for syphilis over four decades—have left a legacy of institutional mistrust that continues to influence research participation decisions in Black communities. Surveys consistently find that distrust of medical institutions and concerns about exploitation rank among the most frequently cited reasons for non-participation in clinical research among Black and Hispanic respondents.

Structural barriers compound the trust deficit. Many phase II and III trials are conducted at large academic medical centers concentrated in metropolitan areas, with participation requirements that include frequent in-person visits, time away from work, and transportation to facilities that may be geographically or logistically inaccessible to lower-income patients. Since racial and ethnic minority populations in the United States are disproportionately represented in lower-income brackets, eligibility criteria and logistical demands that appear race-neutral can function as de facto exclusions.

Language access is a further constraint. Informed consent processes conducted only in English effectively exclude non-English-speaking Hispanic, Asian, and immigrant populations from participation, regardless of their formal eligibility. Research staff who share patients' linguistic and cultural backgrounds are associated with higher recruitment success in minority communities, yet such staff remain underrepresented in clinical research workforces.

Institutional and Policy Responses

The policy landscape is evolving, though implementation lags behind intent. The FDA's 2020 Action Plan for the Inclusion of Underrepresented Populations in Clinical Trials established expectations for sponsors to develop and submit diversity action plans for Phase III trials, and subsequent guidance has moved toward making those plans a formal regulatory requirement rather than a recommendation. The National Institutes of Health's longstanding policy requiring the inclusion of women and minorities in federally funded research, while foundational, has been criticized for lacking enforcement teeth and for setting enrollment targets without requiring analyses stratified by race and ethnicity.

Decentralized clinical trial designs represent one of the most structurally promising responses to participation barriers. By enabling remote consent, telemedicine-based assessments, and at-home sample collection, decentralized models reduce the geographic and logistical friction that has historically excluded patients without proximity to academic research centers. Early evidence suggests these designs meaningfully expand the demographic range of enrolled populations, though digital access disparities introduce their own equity considerations.

Community-based participatory research models offer a complementary path. By partnering with community organizations, federally qualified health centers, and trusted local institutions from the trial design stage rather than the recruitment stage, investigators can build the relationships that convert passive eligibility into active participation. Programs such as the NIH's All of Us Research Program, which has enrolled over 750,000 participants with a deliberate emphasis on historically underrepresented groups, demonstrate that inclusive enrollment is achievable at scale when institutional commitment and community trust are cultivated together.

Toward an Evidence Base That Reflects America

The argument for diversity in clinical trials is sometimes framed primarily as a matter of social equity—and it is that. But it is equally a matter of scientific validity. An evidence base generated from a narrow demographic slice of the population is not merely incomplete in a social sense; it is scientifically limited in ways that compromise the external validity of its conclusions. Treatment guidelines derived from such evidence carry embedded assumptions about biological and social homogeneity that do not hold in clinical practice.

For the United States to sustain a research enterprise genuinely capable of informing care across its full population, the field must treat demographic representativeness not as an aspirational footnote but as a core methodological standard—one that is designed into trials from inception, enforced through regulatory and funding mechanisms, and evaluated as rigorously as statistical power or endpoint validity. The patients waiting on the other side of that reform are not a subgroup. They are the majority.