Skip to content

Broadening eligibility criteria and diversity among patients for cancer clinical trials


March 2024


Kaur M, Frahm F, Lu Y, et al. Broadening eligibility criteria and diversity among patients for cancer clinical trials. NEJM Evidence. 2024.

Our summary

Despite substantial U.S. investments in clinical research for health improvement, certain populations are still underrepresented in clinical trials, limiting the generalizability of research advancements. There is a clear and urgent need for greater diversity in clinical research. To address this issue, the FDA released draft guidance on Race and Ethnicity Diversity Action Plans in clinical trials, also mandated by 2022 federal legislation. Previous FDA guidance suggested broadening eligibility criteria to enhance inclusivity, but the actual impact of broadening eligibility criteria on sociodemographic diversity remains unclear.

This study aimed to evaluate how broadened eligibility criteria influence the sociodemographic diversity of eligible participants among cancer patients, utilizing a nationwide electronic health record-derived de-identified database.

Why this matters

Analyzing clinical trial eligibility criteria across 22 cancer types, this study demonstrates the application of real-world data to broaden trial eligibility criteria. It finds that commonly applied strict criteria can exclude over 50% of patients and disproportionately affect historically underrepresented groups. The study demonstrates that broadening eligibility criteria can increase the absolute number and representation of eligible patients from these groups while only modestly decreasing real-world overall survival. This research challenges prevailing approaches to strict exclusion of patients with comorbidities and is one of many critical paths to increasing diversity in clinical trials. Overall, this work contributes to the importance of broadened eligibility criteria and advances our understanding of their effects on outcomes and health equity across diverse cancer types.

Read the research