Identifying predictors of equitable gender representation among hematology and oncology fellowship programs.

11055 Background: While gender disparities have largely been bridged among hematology and oncology fellows at the national level, there exist many individual fellowship programs that still have marked gender disparity. Our study is the first to examine gender differences among fellows at an institutional level in order to model characteristics of hematology and oncology programs that lead to more equitable gender representation. Methods: For each of the 148 hematology or oncology fellowship programs listed in the Fellowship and Residency Electronic Interactive Database, we collected data on gender, faculty academic rank, and fellow postgraduate year as listed on individual program websites. Further, a program was identified as an academic center if it had an affiliated medical school. Zip code data was used to identify geographic region, local population density, and racial diversity of each program’s surrounding neighborhood. Multiple linear regression was used to model which factors contributed to a program having a higher percentage of female fellows. Results: 3600 faculty and 1488 fellows were included in the analysis. While women were well represented nationally among fellows (46.4%), 52.4% of fellowship programs had fewer than 50% women, and 28.6% of programs had fewer than 35% women. An individual program was more likely to have more female fellows if it had a higher percentage of female assistant professors (p<0.01), associate professors (p<0.01), assistant professors overall (p<0.01), a female program director (PD) or associate program director (APD) (p=0.02), and a designation as an academic center (p<0.01). Notably, the percentage of female full professors had no significant impact (p=0.11) on whether a program was likely to have more female fellows. Conclusions: While women are well represented nationally among hematology and oncology fellows, nearly one in three fellowship programs still have significant gender disparity. Our multivariate model identified several significant factors that predict whether a program was likely to have gender parity among fellows: a higher percentage of female assistant professors, associate professors, assistant professors overall, a female PD or APD, and a designation as an academic center. Whether these factors together highlight the importance of accessible mentorship or direct female faculty role models requires further study. However, by beginning to identify these significant factors, we hope to aid fellowship programs focus efforts when trying to achieve equitable gender representation.[Table: see text]