Race/Ethnicity in Medical Education: An Analysis of a Question Bank for Step 1 of the United States Medical Licensing Examination

Phenomenon: There is growing concern over racial/ethnic bias in clinical care, yet how best to reduce bias remains challenging, in part because the sources of bias in medical education are poorly understood. One possible source is the routinized use of race/ethnicity in lectures, assessment, and preparatory materials, including question banks for licensing examinations. Because students worldwide use question banks to prepare for the United States Medical Licensing Examination, we examined how race/ethnicity was used in one of the most commonly recommended question banks. Approach: We analyzed the use of race/ethnicity in all 2,211 questions in a question bank for Step 1 of the United States Medical Licensing Examination for the following: the frequency of mentions of racial/ethnic groups, whether the use of race/ethnicity was merely descriptive or was central to any part of the question, and whether the question associated race/ethnicity with genetic difference. Findings. In sum, 455 of the 2,011 (20.6%) of the questions in the question bank referred to race/ethnicity in the question stem, answer, or educational objective. The frequency of mentions of racial/ethnicity was disproportionate to the U.S. population: 85.8% referred to White/Caucasians, 9.70% to Black/African Americans, 3.16% to Asian, 0.633% to Hispanics, and 0.633% to Native Americans. No cases referred to Native Hawaiians/Pacific Islanders. The proportion of mentions of race/ethnicity classified as either a routine descriptor or central to the case varied by racial/ethnic category. The association between genetics and disease in cases also varied by racial/ethnic category. Insights. The routinized use of race/ethnicity with no specific goal in preparation materials, such as question banks, risks contributing to racial bias. The implications of routinized use extend to assessment in medical education. Race/ethnicity should be used only when referring to social experiences of groups relevant to their health, not as a proxy for genetics, social class, or culture.