A Blueprint for Increasing Ethnic and Racial Diversity in US Residency Training Programs

Problem People who identify as African Americans, Latinos, or from indigenous backgrounds, are dramatically underrepresented in the U.S. physician workforce. It is critical for academic health centers to recognize racial and ethnic diversity at the residency level and implement changes to enhance diversity among trainees. Approach The Office of Graduate Medical Education (GME) at the University of Pennsylvania Health System (UPHS) developed a multipronged approach to enhance diversity and inclusion (D&I) among residency trainees. The approach included the development of an underrepresented in medicine (UIM) professional network; UIM-focused visiting clerkship programs; holistic review implementation by selection committees; and targeted outreach to UIM candidates, overseen by an associate designated institutional official for UIM Affairs. The authors reported demographic data on residency applicants invited for interviews and matching for all programs at UPHS from 2014-2015 (baseline) to 2020-2021. They also reported data on maximum ranking number programs reached to fill their positions and the average United States Medical License Examination (USMLE) Step 1 scores of matched candidates. Finally, they discussed the implications for leaders who wish to enhance D&I at academic health centers. Outcomes During the baseline year (2014-2015), UIMs represented 12.1% of interviewees and 8.7% of all matched candidates into UPHS residency programs. Over the successive 6 years after incremental implementation of the approach, UIM representation steadily increased. In 2020-2021, UIMs represented 23.2% of interviewees and 26.4% of matched candidates. Programs' maximum rank number to fill and USMLE Step 1 scores of matched candidates remained relatively unchanged. Next Steps The UPHS Office of GME incorporated a purposeful approach to enhance the D&I of its residents. Across 6 years of implementation, UIM representation among resident matches tripled while quantitative program and candidate metrics remained unchanged. Similar efforts should be given further consideration for implementation and evaluation nationwide.