A STRATEGY FOR INCREASING CLINICAL TRIAL DIVERSITY USING REAL-WORLD DATA TO INFORM PATIENT SELECTION CRITERIA

Clinical trials often exclude patients with select comorbidities to reduce the risk of adverse events or noise from confounding variables. This study shows how electronic medical records can be used to examine the impact of selection criteria on different racial groups and one approach for increasing the eligible patient pool. This retrospective study identified adults (≥18 years old) with hypertension at any time using the TriNetX Platform and Dataworks USA Network. We stratified patients by race and captured diagnoses of type 1 or 2 diabetes and the results of the most recent corresponding disease monitoring lab: hemoglobin A1c (HbA1c), fasting glucose, or glucose tolerance test (2 or 3 hours). Among the 11.8 million adults with hypertension, 67.5% were White, 17.2% were Black, 2.0% were Asian and 13.3% were other/unknown. Removing patients with any history of the diabetes excluded 29.5% of White patients, 35.9% of Black patients and 37.7% of Asian patients. Overall, 60.7% of patients with hypertension and diabetes, including 65.7% of Black patients, had a relevant lab result. Among patients with a test result, 39.0% of White patients, 40.3% of Black patients and 35.5% of Asian patients most recent test result indicated disease control (HbA1c <6.5%, fasting glucose ≤125 mg/dL, or glucose tolerance ≤200 mg/dL], a potential recovery of 19.1%-26.5% of excluded patients. Use of broad diagnosis-based exclusion criteria may disproportionately exclude non-White clinical trial participants. Electronic medical record data can be used to explore alternative exclusion strategies, such as using lab values to exclude patients based on level of disease control. By broadening the potential participant pool, these approaches may improve the likelihood of recruiting a trial population that better matches the real-world population.