Patient Diversity in Oncology Trials: Network Benchmarking Results
Published November 2025 — Network-wide enrollment data indicates that sites implementing community outreach programs and multilingual consent processes saw a 27% increase in enrollment of historically underrepresented populations in Phase II-III oncology studies.
The Diversity Imperative in Oncology Research
The underrepresentation of racial and ethnic minorities, older adults, and rural populations in oncology clinical trials is one of the most well-documented and persistent failures of the clinical research enterprise. Despite constituting approximately 40% of the U.S. population and bearing a disproportionate share of cancer incidence and mortality, Black, Hispanic, Asian, and Native American patients represent only 14-18% of oncology trial participants across industry and academic studies.
This underrepresentation has direct consequences for the generalizability of trial results, the detection of differential treatment responses across populations, and the equitable distribution of therapeutic innovation. Regulatory agencies — including the FDA through its 2024 Diversity Action Plan guidance and the EMA through its reflection paper on diversity in clinical trials — have made clear that diversity in clinical trials is no longer aspirational but operationally expected.
This research brief presents benchmarking results from 28 Phase II and Phase III oncology trials conducted across the Clinitiative network between 2022 and 2025, examining the specific site-level interventions associated with improved enrollment of historically underrepresented populations. The analysis identifies actionable strategies that produced measurable results — not aspirational frameworks, but operational practices with documented outcomes.
Key Findings
The analysis encompasses 28 oncology trials across 156 sites, with enrollment data for 6,840 patients for whom race, ethnicity, age, and geographic classification data were available.
Sites that implemented structured community outreach programs and multilingual consent processes achieved a 27% increase in enrollment of historically underrepresented populations compared to sites without these interventions, after controlling for catchment area demographics.
Sites that employed community health workers (CHWs) as patient navigators enrolled underrepresented patients at 3.2 times the rate of sites that relied solely on traditional physician referral pathways, making CHW partnerships the single most impactful diversity intervention identified in the dataset.
Sites offering consent documents and consent discussions in patients' preferred languages saw a 41% improvement in consent-to-enrollment conversion rates among non-English-speaking patients compared to sites that offered English-only consent with interpreter services.
Of the 28 studies in the dataset, 89% had FDA Diversity Action Plans either submitted or in preparation. Sites contributing to the diversity benchmarking effort described in this brief met or exceeded their DAP enrollment targets in 72% of cases.
Methodology
The benchmarking analysis employed a comparative effectiveness design, evaluating the relationship between specific site-level diversity interventions and enrollment outcomes across 156 sites participating in 28 Phase II-III oncology trials.
Site Intervention Classification
Each of the 156 sites was surveyed to document the diversity-related interventions in place during their participation in the trial. Interventions were categorized into six domains: community outreach programs (formal partnerships with community organizations, churches, advocacy groups), multilingual consent processes (translated documents and bilingual staff vs. interpreter-only services), community health worker or patient navigator programs, transportation and logistical support services, culturally adapted patient education materials, and extended or flexible clinic hours to accommodate working patients. Sites were classified as high-intervention (4+ domains active), moderate-intervention (2-3 domains), or low-intervention (0-1 domains).
Demographic Baseline and Catchment Adjustment
To avoid conflating site location demographics with intervention effectiveness, all enrollment diversity metrics were adjusted for catchment area demographic composition using census tract data. A site in a majority-minority metropolitan area would naturally enroll more diverse participants regardless of intervention; the analysis therefore measured enrollment diversity relative to the eligible population within each site’s catchment area. The primary outcome metric — the Diversity Enrollment Ratio (DER) — was calculated as the proportion of underrepresented populations enrolled divided by the proportion of underrepresented populations in the catchment area eligible population. A DER of 1.0 indicates proportional representation; values below 1.0 indicate underrepresentation.
Cancer Type Distribution
The 28 oncology trials covered eight cancer types: non-small cell lung cancer (6 studies), breast cancer (5 studies), colorectal cancer (4 studies), prostate cancer (3 studies), hematological malignancies (3 studies), head and neck cancer (3 studies), melanoma (2 studies), and pancreatic cancer (2 studies). This distribution allowed sub-analysis by cancer type, which is important because the magnitude of racial disparities in incidence and mortality varies substantially across cancer types and influences the baseline diversity opportunity at each site.
Community Outreach and Engagement Strategies
The analysis identified four community outreach models with documented enrollment diversity impact. Sites implementing these models showed consistent improvements across cancer types and geographic regions.
Community Health Worker Partnerships
The single most impactful intervention was the integration of community health workers (CHWs) into the trial recruitment workflow. CHWs serve as trusted intermediaries between clinical research sites and communities that have historically low participation in clinical trials. In the 24 sites that employed CHWs, the Diversity Enrollment Ratio increased from 0.52 to 0.81 — a 56% improvement toward proportional representation. CHWs were most effective when they were members of the target communities themselves, had existing relationships with community organizations, and were integrated into the site research team rather than operating as external contractors. The average cost of a CHW program was $48,000 per site per year, which was offset by the enrollment efficiency gains (0.8 additional patients per month) and the reduced need for enrollment extension periods.
Faith-Based and Community Organization Partnerships
Sites that established formal partnerships with churches, mosques, community centers, and advocacy organizations in underserved neighborhoods saw a 34% increase in referrals from underrepresented populations. The key success factor was not simply distributing information through these organizations but rather engaging organizational leaders as research ambassadors who could address community concerns about clinical trial participation — including historical mistrust, fear of being treated as research subjects rather than patients, and practical concerns about cost and time commitment. Sites that hosted educational events at community organizations (rather than at the clinical site) generated 2.1x more referrals than sites that relied on brochure distribution alone.
Culturally Adapted Patient Education
Standard clinical trial education materials are typically written at a college reading level and assume familiarity with medical research concepts. Sites that developed culturally adapted materials — written at or below an 8th-grade reading level, available in multiple languages, and incorporating imagery and narratives relevant to the target communities — saw a 22% increase in consent rates among underrepresented patients. Importantly, culturally adapted materials were not simply translated versions of standard materials; they were developed in collaboration with community advisory boards to address the specific concerns, questions, and decision-making frameworks of each target population.
Transportation and Logistical Support
Transportation barriers disproportionately affect low-income and rural patients, creating a structural obstacle to trial participation that no amount of outreach can overcome without direct intervention. Sites that offered transportation support — including ride-sharing vouchers, shuttle services, or mileage reimbursement — saw a 19% improvement in enrollment of patients living more than 25 miles from the site. Four sites in the dataset partnered with local non-profit transportation services, reducing per-patient transportation costs from $45 per visit (ride-sharing vouchers) to $18 per visit while providing more reliable and patient-friendly service.
Multilingual Consent Processes and Language Access
Language access emerged as a critical determinant of enrollment diversity, particularly in regions with large Spanish-speaking, Mandarin-speaking, and Vietnamese-speaking populations. The analysis compared three models of language access during the consent process: interpreter-only services (telephone or in-person interpreter for consent discussion, with English-language consent forms), translated documents with interpreter support (IRB-approved translated consent forms supplemented by interpreter services), and full bilingual consent (translated documents, bilingual research staff conducting the consent discussion, and bilingual study materials throughout the trial).
Sites offering full bilingual consent achieved a consent-to-enrollment conversion rate of 78% among non-English-speaking patients, compared to 55% for translated documents with interpreter support and 41% for interpreter-only services. The gap is attributable to the relationship-building advantage of bilingual staff: patients consistently reported higher trust levels when the person explaining the study spoke their language natively and could address nuanced questions about risks, benefits, and alternatives without the mediation of a third-party interpreter.
From a regulatory perspective, all three models met IRB requirements for informed consent. However, the quality of understanding — as measured by consent comprehension assessments administered to a subset of patients — was significantly higher in the full bilingual consent group (92% comprehension score) compared to interpreter-only (74% comprehension score). This finding has implications beyond enrollment metrics: higher consent comprehension is associated with better protocol adherence, lower dropout rates, and reduced risk of consent-related regulatory findings.
The practical challenge of implementing multilingual consent is the cost and complexity of translating documents for each study. Sites that maintained a library of pre-translated common consent language sections — covering general clinical trial concepts, HIPAA provisions, and standard risk disclosures — reduced per-study translation costs by 60% and translation turnaround time from 21 days to 7 days, making multilingual consent operationally feasible even for sites with limited budgets.
Geographic Analysis of Diversity Performance
The geographic distribution of diversity-performing sites reveals important patterns for site selection strategy and network development.
Urban sites in the dataset (n=86) had the highest baseline diversity opportunity due to catchment area demographics, but also showed the widest performance variance. The top quartile of urban sites achieved Diversity Enrollment Ratios above 0.90 (near-proportional representation), while the bottom quartile fell below 0.35. The differentiating factor was not location but rather the intensity and authenticity of community engagement — urban sites in diverse neighborhoods that did not invest in outreach performed no better than suburban sites in homogeneous areas.
Suburban sites (n=48) showed moderate diversity performance overall but demonstrated the greatest improvement when interventions were implemented. The median DER for suburban sites increased from 0.38 to 0.62 with high-intervention programs — a 63% improvement — compared to a 42% improvement at urban sites. This suggests that suburban sites have untapped diversity potential that is accessible through targeted outreach to nearby underserved communities.
Rural sites (n=22) faced the greatest structural barriers to diversity enrollment, including smaller and more homogeneous catchment populations, fewer community organizations to partner with, and longer travel distances for patients. However, rural sites serving Native American, agricultural worker, or Appalachian communities achieved meaningful diversity improvements through culturally specific outreach. Three rural sites partnered with tribal health organizations and achieved DERs above 0.75 for Native American enrollment.
FDA Diversity Action Plan Alignment
The FDA’s 2024 guidance on Diversity Action Plans (DAPs) requires sponsors of Phase III trials and certain Phase II trials to submit plans describing their strategies for enrolling demographically representative study populations. Among the 28 studies in our dataset, 25 (89%) had DAPs either submitted or in active preparation at the time of analysis.
Of the 25 studies with DAPs, 18 (72%) met or exceeded their stated enrollment diversity targets. The 7 studies that fell short shared common characteristics: they set targets without corresponding site-level intervention budgets, they relied on existing site portfolios rather than activating new sites in high-diversity catchment areas, and they lacked mechanisms for mid-study diversity monitoring that would have enabled corrective action during the enrollment period.
The studies that successfully met their DAP targets shared three operational characteristics: they allocated dedicated budget for diversity interventions at the site level (median $35,000 per site for high-priority diversity sites), they included diversity enrollment metrics in regular enrollment tracking dashboards alongside overall enrollment numbers, and they designated a diversity enrollment lead at the sponsor level with authority to recommend site-level interventions or new site activations when diversity metrics fell below targets.
The alignment between our network benchmarking results and FDA expectations is strong: the interventions that produced the greatest diversity improvements in our dataset — community health worker partnerships, multilingual consent, community organization engagement, and transportation support — correspond directly to the types of strategies the FDA has indicated it expects to see in credible DAPs. Sponsors can use the quantitative benchmarks from this analysis to set realistic diversity targets and allocate intervention budgets with confidence that the investments will produce measurable results.
Conclusions
The benchmarking results from 28 oncology trials and 156 sites demonstrate that meaningful improvements in enrollment diversity are achievable through deliberate, funded, site-level interventions. The 27% overall increase in underrepresented enrollment at high-intervention sites is not a ceiling — it is a baseline that can be exceeded as sites and sponsors gain experience with diversity-focused operational practices.
The evidence is particularly clear on three points. First, community health worker partnerships are the single most impactful intervention and should be considered a standard component of diversity enrollment strategy for oncology trials in underserved communities. Second, multilingual consent — beyond interpreter-only services — is a prerequisite for equitable enrollment in linguistically diverse regions. Third, diversity improvement requires sustained investment at the site level; aspirational targets without corresponding resources produce aspirational results.
The regulatory environment, the scientific imperative for generalizable data, and the ethical obligation to ensure equitable access to experimental therapies all converge on the same conclusion: diversity in oncology trials is no longer optional. The data in this brief provide a practical roadmap for translating that imperative into operational reality.
Want to Learn More?
Contact our team to discuss diversity enrollment strategies and how our network benchmarking data can inform your clinical trial design.