Gender Bias in Surgical Oncology Fellowship Recommendation Letters: Gaining Progress

OBJECTIVE: Gender bias has been identified in letters of recommendation (LOR) in many different surgical training fields. Among surgeons, women comprise over 30% of the full-time faculty positions nationally and surgical oncology is one of the most gender diverse surgical subspecialties. We sought to determine if bias existed in LOR submitted to a Complex General Surgical Oncology (CGSO) fellowship. DESIGN: LOR for the CGSO fellowship were retrospectively analyzed from applicants at a single institution over an 8-year period (2013-2020). The linguistic content of the letters was analyzed using Linguistic Inquiry and Word Count (LIWC2015), a validated text analysis program. Using multivariable analysis, LOR were compared by gender of both applicant and letter writer to explore the association between gender and the characteristics of the applicants and letter writers. SETTING: University of North Carolina at Chapel Hill (UNC), Division of Surgical Oncology and Endocrine Surgery. PARTICIPANTS: Applicants interviewed for the CSGO fellowship program at the UNC from 2013 to 2020 as well as all applicants from the 2018 application cycle, regardless of interview status. RESULTS: About 841 letters from 219 interviewed applicants throughout the 2013 to 2020 surgical oncology fellowship application cycles were included. No difference in authenticity, clout, analytic thinking, or emotional tone of the letters was seen when comparing men and women applicants. Of the 41 word categories analyzed, only "references to achievement" in LOR written for women was significantly higher when compared to LOR written for men (p = 0.01). Interestingly, significantly more women applicants had at least 1 LOR written by a woman (p = 0.04). A subset analysis of all applicants regardless of interview status from the 2018 cycle included 294 LOR from 77 applicants. With the inclusion of noninterviewed applicants, LOR for men had more analytic tone than LOR for women (p = 0.02), otherwise there were no significant differences between the groups. CONCLUSIONS: Very few differences in LOR were found for applicants at a CGSO fellowship program based on applicant or letter writer gender. The lack of gender bias demonstrates progress within the field of surgical oncology, likely a result of recent work and educational effort in this area. Efforts to expand this progress into other surgical sub-specialties are necessary. (C) Y2020 Published by Elsevier Inc. on behalf of Association of Program Directors in Surgery.