Highlights
In nine pivotal U.S. coronary stent premarket approval (PMA) trials spanning 2003 to 2018, racial and ethnic minority participants were substantially underrepresented, and women were also enrolled below their expected share of the intended-use population (IUP).
Site geography and surrounding community demographics were strongly linked to minority enrollment. In contrast, female enrollment showed much weaker site-level patterning, suggesting that improving sex representation will likely require strategies beyond site selection alone.
Modeling suggested that shifting enrollment toward sites that already enroll more minority participants could bring Black and Hispanic representation closer to parity without reducing non-Hispanic White representation, underscoring site selection as a practical lever for trial design.
Background
Coronary stents are class III cardiovascular devices, which means they undergo the most rigorous U.S. regulatory pathway before marketing. Premarket approval studies are intended to generate evidence in the population most likely to receive the device in practice. In principle, that should mean trial participants closely resemble the IUP. In reality, cardiovascular device studies have repeatedly struggled to enroll women and racial and ethnic minority groups in proportions that match disease burden and clinical use.
This mismatch matters for several reasons. First, if a pivotal trial underrepresents key patient groups, estimates of safety and effectiveness may be less reliable for those patients. Second, underrepresentation can weaken confidence in labeling and shared decision-making. Third, gaps in trial diversity can compound existing health inequities, especially in conditions such as coronary artery disease where outcomes already differ by sex, race, ethnicity, and access to care.
What has been less clear is whether the choice of research sites itself influences who gets enrolled. Site selection is one of the earliest decisions in device development, and it is often based on investigator experience, enrollment speed, device expertise, and operational feasibility. The study by Batchelor and colleagues directly addressed whether site characteristics predict the demographic makeup of coronary stent PMA enrollment and whether altering site selection could improve representativeness.
Study Design
This was a pooled analysis of 8,859 U.S. participants enrolled across 196 sites in nine pivotal coronary stent PMA studies conducted between 2003 and 2018. The investigators examined site-level characteristics including U.S. region, surrounding county demographics, teaching status, Veterans Administration affiliation, trial volume, female principal investigator (PI) involvement, and the number of acute hospital beds.
The main outcomes were minority enrollment and female enrollment. The authors used multivariable regression to identify predictors of enrollment patterns and then modeled participant-to-prevalence ratios (PPRs) under different site selection scenarios. PPR compares the proportion of a demographic group enrolled in a trial with that group’s prevalence in the relevant U.S. intended-use population. In general, a PPR below 0.80 suggests underrepresentation, while values near 1.0 indicate closer alignment with population prevalence.
This design is observational and model-based rather than interventional. It cannot prove that changing sites alone will guarantee better diversity, but it does provide a useful estimate of how much representation might improve if enrollment were deliberately redistributed.
Key Findings
Minority enrollment was markedly below expected population share
Across the pooled sample, minority participants accounted for 12% of enrollees, corresponding to a PPR of 0.48. That level of representation is well below the threshold generally considered acceptable for population alignment. In practical terms, minority participants were enrolled at less than half of the prevalence expected from the intended-use population.
Importantly, minority enrollment was not evenly distributed across sites. The variation was large enough that site-level features explained a meaningful proportion of the differences in who was enrolled. The multivariable model identified several predictors of minority enrollment, including U.S. region and the demographic and socioeconomic profile of the surrounding county. Sites in the West and South were associated with higher minority enrollment, and county minority population, population density, and per-capita income also helped explain enrollment patterns. Overall, the model accounted for about half of the variance in minority enrollment (R2 = 0.50; P < 0.001).
This is a notable finding because it suggests that minority representation is not purely a function of broad national trends or random chance. Instead, where a trial is conducted appears to matter substantially, presumably through a combination of local patient demographics, referral networks, investigator experience, and community engagement.
Women were also underrepresented, but site effects were weaker
Women made up 30% of participants, with a PPR of 0.77. That indicates modest underrepresentation rather than the more pronounced gap seen for racial and ethnic minorities. Still, in a disease area where women are often older, present with different symptom patterns, and may experience distinct barriers to referral and enrollment, a 30% enrollment share may not fully reflect real-world need.
Compared with minority enrollment, female enrollment showed less variation across sites and was much harder to predict using the measured site characteristics. The only significant site-level factor reported in the abstract was non-VA status, and the model explained only 9.5% of the variance (R2 = 0.095; P < 0.001). The investigators also noted that female PI involvement was very uncommon, occurring at fewer than 6% of sites, which limited the ability to assess whether investigator gender meaningfully influenced enrollment.
The contrast between minority and female enrollment is clinically important. It implies that the reasons women are underenrolled may be more diffuse and less dependent on geography than the reasons minority participants are underenrolled. Those barriers may include eligibility criteria, referral bias, time and travel burden, caregiving responsibilities, perception of risk, or lack of targeted outreach.
Site selection could improve racial and ethnic representation without harming White enrollment
Perhaps the most actionable part of the analysis was the modeling of alternative site selection strategies. The authors estimated that if enrollment were shifted away from sites that historically enrolled fewer minority participants and toward sites that enrolled more minority participants, representation of Black and Hispanic participants could be brought to PPRs of at least 0.80. That would move these groups much closer to acceptable alignment with the intended-use population.
Equally important, this improvement did not appear to require a tradeoff in non-Hispanic White enrollment. In the model, non-Hispanic White representation remained at a PPR of 1.00. That finding challenges the common assumption that improving diversity necessarily reduces access or enrollment for the majority population. At least in this setting, the model suggests that better site selection could improve equity without creating a zero-sum outcome.
For trial planners, this is more than an academic point. Site selection is modifiable. If sponsors and investigators can identify centers that are more likely to recruit diverse patients, they can build representation into the trial from the outset rather than trying to repair it late in the enrollment process.
Expert Commentary
This study adds an important operational dimension to the long-running discussion about diversity in cardiovascular trials. Many efforts to improve representation focus on participant-level barriers, such as transportation, consent language, or mistrust. Those factors remain important, but the Batchelor analysis shows that the architecture of the trial itself also matters. In other words, diversity is not only about how patients are approached; it is also about where they are approached.
The finding that minority enrollment is strongly shaped by site characteristics is especially relevant to sponsor and contract research organization planning. Centers embedded in more diverse, denser, and economically distinct communities may have access to broader patient pools and more established pathways for reaching underrepresented groups. Selecting such sites early could improve the external validity of pivotal device data without materially changing the scientific question being tested.
At the same time, the modest explanatory power for female enrollment is a reminder that sex imbalance is harder to solve with geography alone. Women’s underrepresentation may reflect referral patterns, eligibility criteria, procedural preferences, or care-seeking behavior that extend beyond site catchment demographics. Meaningful improvement may require a broader toolkit: more inclusive eligibility criteria, sex-specific recruitment plans, flexible scheduling, community outreach, clinician education, and possibly more diverse investigator teams.
The low proportion of female PIs is another signal worth noting. Although the abstract does not establish a causal link, investigator diversity may influence trial culture, patient trust, or referral pathways. The limited number of female PIs in this analysis also highlights how difficult it remains to study this question rigorously. If sponsors want to know whether investigator gender affects enrollment, they will need trials that include enough women leaders to make the question answerable.
There are also methodological limitations to keep in mind. This was an observational, retrospective analysis of U.S. coronary stent PMA studies, so the results may not generalize to other device categories, non-U.S. systems, or later-stage postmarket studies. Site-level variables such as teaching status or acute hospital beds may be proxies for deeper structural factors that were not directly measured, including referral pipelines, community trust, linguistic access, and sponsor selection preferences. The PPR metric is useful, but it simplifies representativeness to one dimension and does not fully capture clinical heterogeneity within racial, ethnic, or sex groups.
Even with those caveats, the practical message is strong: if sponsors want pivotal coronary stent trials to mirror the U.S. population more closely, site choice should be treated as a diversity intervention, not merely a logistical detail.
Conclusion
Batchelor and colleagues show that U.S. coronary stent PMA studies still fall short of fully representing the intended-use population, especially for racial and ethnic minority groups. Minority enrollment varied substantially by site and was strongly associated with regional and county-level characteristics, making site selection a plausible lever for improving representation. Women were also underenrolled, but their participation was less predictably tied to site features, suggesting that different solutions will be needed.
For clinicians, regulators, and trial designers, the lesson is straightforward: diversity should be engineered into cardiovascular device development from the start. Better site selection may help close racial and ethnic gaps, but meaningful progress for women will likely require additional design, recruitment, and leadership changes.
Funding and clinicaltrials.gov
Funding details were not provided in the abstract supplied here. A single clinicaltrials.gov identifier was not listed, which is not unexpected for a pooled analysis of multiple pivotal PMA studies.
References
1. Batchelor WB, Califf R, Mehran R, Stone G, Blumer V, O’Connor C, Sharma G, Fiuzat M, Douglas P, Coylewright M, Yancy CW, Baron SJ, Kandzari DE, Abbott JD, Echols MR, Rymer JA, Krucoff MW, Spitzer E, Damluji AA. Aligning Coronary Stent Trial Enrollment With the U.S. Intended-Use Population: Implications of Site Selection. J Am Coll Cardiol. 2026;87(16):2163-2176. PMID: 42053200.
2. U.S. Food and Drug Administration. Diversity Plans to Improve Enrollment of Participants from Underrepresented Racial and Ethnic Populations in Clinical Studies: Guidance for Industry. Final guidance, 2024.
3. U.S. Food and Drug Administration. Enhancing the Diversity of Clinical Trial Populations—Eligibility Criteria, Enrollment Practices, and Trial Designs: Guidance for Industry. Final guidance, 2020.
