Association of Women Leaders with Women Program Director and Trainee Representation Across US Academic Internal Medicine

Background: Women are underrepresented within internal medicine (IM). Whether women leaders attract women trainees is not well explored. Objective: To characterize leader and trainee gender across US academic IM and to investigate the association of leader gender with trainee gender. Design: Cross-sectional study. Participants: Leaders (chairs, chiefs, program directors (PDs)) in 2018 and trainees (residents, fellows) in 2012–2016 at medical school-affiliated IM and seven IM fellowship programs. Exposure: Leadership (chair/chief and program director; and, for resident analyses, fellow) gender. Main Measures: Our primary outcome was percent women trainees (IM residents and, separately, subspecialty fellows). We used standard statistics to describe leadership and trainee gender. We created separate multivariable linear regressions to evaluate associations of leader gender and percent women fellows with percent women IM residents. We then created separate multivariable multilevel models (site as a random effect) to evaluate associations of leader gender with percent women subspecialty fellows. Key Results: Our cohort consisted of 940 programs. Women were 13.4% of IM chairs and <25% of chiefs in each fellowship subspecialty (cardiology: 2.6%; gastroenterology: 6.6%; pulmonary and critical care: 10.7%; nephrology: 14.4%; endocrinology: 20.6%; hematology-oncology: 23.2%; infectious diseases: 24.3%). IM PDs were 39.7% women; fellowship PDs ranged from nearly 25% (cardiology and gastroenterology) to nearly 50% (endocrinology and infectious disease) women. Having more women fellows (but not chairs or PDs) was associated with having more women residents (0.3% (95% CI: 0.2–0.5%) increase per 1% fellow increase, p<0.001); this association remained after adjustment (0.3% (0.1%, 0.4%), p=0.001). In unadjusted analyses, having a woman PD (increase of 7.7% (4.7%, 10.6%), p<0.001) or chief (increase of 8.9% (4.6%, 13.1%), p<0.001) was associated with an increase in women fellows; after adjustment, these associations were lost. Conclusions: Women held a minority of leadership positions in academic IM. Having women leaders was not independently associated with having more women trainees.