An Approach to Detecting Sex Bias in Residency Recruitment
Jayaprakash N, Marks A, Folt J, Lewandowski C, and Goyal N. 138 An Approach to Detecting Sex Bias in Residency Recruitment. Ann Emerg Med 2019; 74(4):S55.
Ann Emerg Med
Study Objectives: Studies have noted sex disparities in emergency medicine (EM) in which women are less likely to hold leadership positions and advance in academic ranks. The objective of this study was to explore whether women are treated differently from men during the residency recruitment process. Methods: This is a single center retrospective analysis of the recruitment data for Henry Ford Hospital EM (HFHEM) residency from 2015 to 2018. We included all United States (US) graduates who applied to our residency and stratified by sex. Data was abstracted from local databases of applicant scores related to their file score, interview score and composite scores (file score + interview score) as well as data related to their position on the final rank list. The file score is generated by a detailed review of all data in each applicant’s file and primarily reflects their academic performance in medical school. The interview score reflects the applicant’s communication and professionalism skill as assessed by multiple faculty interviewers. We additionally reviewed the Electronic Residency Application Service (ERAS) data regarding sex distribution for 2015 - 2018 for benchmarking. We were primarily interested in the median position on the rank list of women versus men. Secondarily we compared the file, interview, and composite score of women versus men. Results: From 2015 through 2018, 224 (34.5%) women and 425 (65.5%) men were ranked and submitted to the National Residency Matching Program on behalf of our residency program. Nationally, during the same time period amongst all applicants, 5168 women (34.3%) applied to EM via ERAS compared to 9900 (65.7%) men. At HFHEM, taking into account all applicants, we had 1441 (33.5%) women and 2866 (66.5%) men. The median (IQR) position on the HFHEM rank order list for women was 72.5 (37 - 111.5) and for men was 88 (43 - 128), p = 0.013. The average (±SD) interview score for women was 18.7 (±2.2) and for men 18.1 (±2.5), p = 0.01. We did not find a statistically significant difference between the file scores for women compared to men (35.5 ±4.5 vs. 34.5 ±5.2, p = 0.18). Conclusion: The HFHEM residency rank lists reflect similar distributions of women and men compared to the national applicant pool. Our recruitment process generates similar scores for the academic proficiency of both women and men and our scoring at the interview suggests a higher performance for women than men. Ultimately, women rank more competitively on our rank lists compared to men. Other residency programs may consider using this template to assess for bias in their recruitment process.