Enrollment in clinical trials according to patients race: experience from the VA Cooperative Studies Program (1975–2000)

Abstract

Background: Racial distribution of clinical trial participants is important because results from these studies serve to define evidence-based practice. This report summarizes the experience of the VA Cooperative Studies Program (CSP) in enrolling white, black and Hispanic patients. Methods: An analysis of enrollment in randomized controlled trials conducted by VA CSP between 1975 and 2000. A standardized enrollment ratio for each trial was calculated by dividing the observed number of enrolled white patients in the trial by the expected number of eligible white patients based on the proportion of white patients hospitalized at the enrolling VA Medical Centers. Results: 138 VA CSP clinical trials were initiated between 1975 and 2000, 83 contained information on race for 71,463 patients. Overall, 76% of enrolled patients were white, 20% were black, and 4% were Hispanic. Based on standardized enrollment ratios, 60 of the 83 trials had 95% confidence intervals that excluded 1.0. Of these, 32 studies enrolled more white patients than expected and 28 enrolled more Black and/or Hispanic patients than expected based on the racial distribution of patients hospitalized at sites involved in the trials. When trials were separated by intervention type, 13 of the 19 trials that had an invasive arm enrolled fewer minority patients than expected. In trials that targeted diseases that affect minority populations to a greater degree than whites (diabetes, hypertension and end stage renal disease), 11 of the 14 trials enrolled more minority patients than expected. Conclusions: There were several trials that enrolled either more or less minority patients than expected based on patients hospitalized at study sites. Trials that included an invasive arm enrolled fewer minority participants than expected. Trials that involve invasive therapies may wish to adopt special recruitment strategies to reach minority populations.