Letters of recommendation for OB/GYN residents do not show gender, racial biases

Letters of recommendation in OB/GYN residency applications did not show the same gender and racial biases that have been previously seen in other medical specialties, according to recent findings.

Researchers noted that letters of recommendation (LOR) are an important factor in selecting applicants for residencies and fellowships, and studies in multiple specialties have identified gender and racial biases in LOR through text-analysis software programs.

“Few studies examine bias in LOR in OB/GYN training; none examine racial bias,” Oluwateniola E. Brown, MD, a urogynecologist and pelvic reconstructive surgeon at Northwestern Medical Group, and colleagues wrote in the American Journal of Obstetrics and Gynecology. “This study aims to determine if gender and racial biases exist in LOR for medical students applying to OB/GYN residency. Specifically, we sought to determine if there are differences in language used in LOR based on applicant gender and if the applicant was underrepresented in medicine.”

The researchers performed a linguistic analysis on 3,060 letters written for 834 applicants that were submitted to a single obstetrics and gynecology residency program during the 2019-2020 application cycle. Slightly more than half the letters (53%) were written by men. Of the applicants, 721 (87%) were women, 198 (24%) of whom belonged to an underrepresented group in medicine — including Black/African Americans, Mexican Americans, Native Americans and mainland Puerto Ricans. Specifically, Brown and colleagues used Linguistic Inquiry Word Count, or LIWC — “a widely used, validated text analysis software program,” they said. LIWC has an internal dictionary that classifies words into preselected word categories. At the same time, the researchers conducted a qualitative analysis on the overall tone of the letters, which consisted of assessing their analytic thinking, authenticity, emotional tone and clout — or “writing that is authoritative, confident and indicative of leadership,” the researchers said.

However, they found no differences in word categories in letters written for men and women applicants, as well as no word category differences between underrepresented applicants and white and Asian applicants.

“A possible explanation for the lack of gender differences in our LIWC analysis may be the representation of women in OB/GYN compared to other surgical specialties,” Brown and colleagues wrote. “Women are underrepresented in most surgical training programs; OB/GYN has historically been the exception. In 2019, women represented 83.8% of OB/GYN residents and fellows and 58.9% of active physicians in the specialty, making OB/GYN the specialty with the highest percentage of women trainees and the specialty with the second-highest percentage of active women physicians.”

However, in their qualitative analysis, the researchers identified what they called “noteworthy themes.” For example, compliments on work ethic, leadership and surgical skills were more frequently found in LOR for non-underrepresented applicants, and phrases that may have raised doubts about an applicant were more common in LOR for those who were underrepresented in medicine.

“A commitment to advancing women and underrepresented individuals in our field necessitates examining processes that we rely on through the lens of equity,” Brown and colleagues wrote. “This commitment to reflection and careful refinement is critical in creating an environment of equal opportunity and advancement. Subtle, easily dismissible linguistic differences may have implications for the perceived competence of individuals from specific identity groups and may contribute to disparities in opportunities and academic rewards.”