Underrepresented Minorities are Underrepresented Among General Surgery Applicants Selected to Interview.

Objective: Diversity is an ill-defined entity in general surgery training. The Accreditation Council for Graduate Medical Education recently proposed new common program requirements including verbiage requiring diversity in residency. "Recruiting" for diversity can be challenging within the constraints of geographic preference, type of program, and applicant qualifications. In addition, the Match process adds further uncertainty. We sought to study the self-identified racial/ethnic distribution of general surgery applicants to better ascertain the characteristics of underrepresented minorities (URM) within the general surgery applicant pool.

Design: Program-specific data from the Electronic Residency Application Service was collated for the 2018 medical student application cycle. Data were abstracted for all participating programs' applicants and those selected to interview. Applicants who did not enter a self-identified race/ethnicity were excluded from analysis. URM were defined as those identifying as Black/African American, Hispanic/Latino/of Spanish origin, American Indian/Alaskan Native, or Native Hawaiian/Pacific Islander-Samoan. Appropriate statistical analyses were accomplished.

Setting: Ten general surgery residency programs-5 independent programs and 5 university programs.

Participants: Residency applicants to the participating general surgery residency programs.

Results: Ten surgery residency programs received 10,312 applications from 3192 unique applicants. Seven hundred and seventy-eight applications did not include a self-identified race/ethnicity and were excluded from analysis. The racial/ethnic makeup of applicants in this study cohort was similar to that from 2017 to 2018 Electronic Residency Application Service data of 4262 total applicants to categorical general surgery. Programs received a median of 1085 (range: 485-1264) applications each and altogether selected 617 unique applicants for interviews. Overall, 2148 applicants graduated from US medical schools, and of those, 595 (28%) were offered interviews. The mean age of applicants was 28.8 ± 3.8 years and 1316 (41%) were female. Hispanic/Latino/of Spanish origin, Black, and American Indian/Alaskan Native/Hawaiian/Pacific Islander-Samoan applicants constituted 12%, 8%, and 1% of total applicants, but only 8%, 6%, and 1% of those selected for interview. Overall, 29% of applicants had United States Medical Licensing Examination (USMLE) Step 1 scores ≤220; 37 (6%) of those selected for interviews had a USMLE Step 1 score of ≤220. A higher proportion of URM applicants had USMLE scores ≤220 compared to White and Asian applicants. Non-white self-identification was a significant independent predictor of a lower likelihood of interview selection. Female gender, USMLE Step 1 score >220, and graduating from a US medical school were associated with an increased likelihood of being selected to interview.

Conclusions: URM applicants represented a disproportionately smaller percentage of applicants selected for interview. USMLE Step 1 scores were lower among the URM applicants. Training programs that use discreet USMLE cutoffs are likely excluding URM at a higher rate than their non-URM applicants. Attempts to recruit racially/ethnically diverse trainees should include program-level analysis to determine disparities and a focused strategy to interview applicants who might be overlooked by conventional screening tools.