Reporting of race and ethnicity in breast cancer research: room for improvement.

Health disparities in breast cancer outcomes according to race/ethnicity are well documented. Randomized clinical trials (RCT) offer an opportunity to evaluate differences in disease biology and response to therapy that may contribute to disparities. We conducted a PubMed search to identify all English language original reports of breast cancer RCT from October 2001 to October 2006. The primary outcomes of interest were reporting of accrual and results by race or ethnicity of trial subjects. We evaluated the correlation between study characteristics and reporting of race/ethnicity. A total of 197 eligible trials were identified among 29 journals. Accrual was reported by race in 17% of studies and results analyzed by race in only 2%. Reporting of race was associated with National Cancer Institute funding (38 vs. 13%, P = 0.001), US cooperative group trials (52 vs. 13%, P < 0.0001), trials with US sites (43 vs. 5%, P < 0.0001), and trials enrolling > 500 subjects (24 vs. 12%, P = 0.055). Pharmaceutical industry funding, # of centers, stage of disease, nature of experimental intervention and study outcomes were not associated with reporting of race. Among US studies reporting trial accrual by race/ethnicity, the mean accrual distribution was 81% white, 7.6% black, 9.6% Asian, and 7.2% Hispanic subjects. The majority of breast cancer RCT fail to report the race/ethnicity of participants. Low accrual of black subjects and failure to report accrual and outcomes by race in RCT may contribute to difficulty in understanding and overcoming health disparities in breast cancer.