BackgroundSystemic racial and ethnic inequities continue to be perpetuated through scientific methodology and communication norms despite efforts by medical institutions. We characterized methodological practices regarding race and ethnicity in U.S. research published in leading medical journals.MethodsWe systematically reviewed randomly selected articles from prominent medical journals: Annals of Internal Medicine, BMJ, JAMA, The Lancet, and NEJM within five periods: 1995-99, 2000-04, 2005-09, 2010-14, 2015-18. Original human-subjects research conducted in the U.S. was eligible for inclusion. We extracted information on definitions (conceptualization), measurement/coding (operationalization), use in analysis (utilization), and justifications. We reviewed 1050, including 242 (23%) in analyses.FindingsThe proportion of U.S. medical research studies including race and/or ethnicity data increased between 1995 and 2018. However, no studies defined race or ethnicity. Studies rarely delineated between race and ethnicity, frequently opting for a combined “ethno-racial” construct. In addition, most studies did not state how race and/or ethnicity was measured. Common coding schemes included: “Black, other, White,” “Hispanic, Non-Hispanic,” and “Black, Hispanic, other, White.” Race and/or ethnicity was most often used as a control variable, descriptive covariate, or matching criteria. Under 30% of studies included a justification for their methodological choices regarding race and/or ethnicity.InterpretationDespite regular efforts by medical journals to implement new policies around race and ethnicity in medical research, pertinent information around methodology was systematically absent from the majority of reviewed literature. This stymies critical disciplinary reflection and progress towards equitable practice.FundingFunding was provided through training grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development [T32 HD091058] and the Department of Sociology, UNC Chapel Hill. Carolina Population Center provided general support [P2C HD050924, P30 AG066615]. NRS received additional support from the National Cancer Institute [T32 CA057711].