Ensuring equal access: Language as exclusion criteria among clinical trials in gynecologic oncology

• Published in: Journal of clinical oncology Vol. 40; no. 16_suppl; p. e18587
• Main Authors: Lee, Sarah S., Dinicu, Andreea, Arthurs, Likolani, Seda, Danielle, Pothuri, Bhavana, Lightfoot, Michelle DS
• Format: Journal Article
• Language: English
• Published: 01-06-2022
• ISSN: 0732-183X; 1527-7755
• DOI: 10.1200/JCO.2022.40.16_suppl.e18587

e18587
Background: Enrolling a diverse patient population into clinical trials may help mitigate disparities and ensure generalizability of results. We aimed to assess clinical trial eligibility criteria that may adversely impact enrollment of underrepresented groups. Methods: We searched clinicaltrials.gov for gynecologic cancer studies between 1997 and 2021. Studies were excluded if they included non-gynecologic or multiple sites of disease; hereditary gynecologic cancer syndromes were included. The inclusion and exclusion criteria of each study were reviewed to determine if demographic factors were used for enrollment screening. Our outcome of interest was the characterization of inclusion and exclusion factors that may adversely impact enrollment of historically underrepresented groups. Data were analyzed using chi-square tests, logistic regression, and Cochrane-Armitage test of trends. Results: A total of 1614 studies were included; 902 (56%) were ovarian, 357 (22%) were cervical, and 355 (22%) were uterine cancer studies. The majority (86%, n = 1394) were therapeutic trials, 38% (n = 608) were phase 2, and 8% (n = 123) were phase 3. Ovarian cancer trials were most likely to be industry-sponsored (252, 28%), compared to only 12% of cervical and endometrial cancers (n = 44 each, p < 0.001). Overall, 190 gynecologic cancer studies (12%, n = 190) excluded patients based on language. Cervical cancer studies were the most likely to exclude patients based on language (n = 67, 19%), compared to endometrial cancer (n = 45, 13%) and ovarian cancer (n = 78, 9%, p < 0.001). Compared to investigator-initiated trials, industry-sponsored trials (adjusted OR [aOR] 0.07, 95% confidence interval [CI] 0.02-0.29) and organizational/government-sponsored trials (aOR 0.37, CI 0.21-0.66) were less likely to exclude based on language. Compared to drug/device trials, behavior/quality of life studies (aOR 31.63, CI 18.50-54.09) and prevention/diagnosis trials (aOR 5.14, 2.81-9.41) were more likely to exclude based on language. During the study period, the number of studies using language as exclusion criteria increased over time (p < 0.001). Conclusions: Over the last three decades, one in ten gynecologic cancer trials excluded patients based on language. Studies focused on cervical cancer, investigator-initiated, and behavior/quality of life studies were most likely to exclude on the basis of language. Eliminating language as an exclusion criterion in clinical trial eligibility could improve understanding of the health-related quality of life among all patients with gynecologic cancer and help achieve equity in clinical trial enrollment.