The Use of Facial Recognition Software and Published Manuscripts to Examine Trends in Surgical Editorial Board Diversity

Introduction: Recent social justice movements have highlighted the need for improved diversity and inclusion. These movements have emphasized the need for inclusivity of all genders and races in all sectors including surgical editorial boards. There is currently not an established, standardized method to assess the gender, racial, and ethnic makeup of surgical editorial board rosters, yet artificial intelligence is a method that can be utilized to determine gender and race in an unbiased manner. The aim of the present study is to determine if recent social justice movements correlate with an increase in diversity-themed articles published and if there is an increase in the gender and racial makeup of surgical editorial boards determined by artificial intelligence software. Methods: Impact factor was used to assess and rank highly regarded general surgery journals. The website of each of these journals was examined for pledges of diversity in their mission statements and core beliefs of conduct. To determine the number of diversity-themed articles that were published during 2016 and 2021, each surgical journal was analyzed for diversity-themed articles using 10 specific keywords in PubMed. To determine the racial and gender makeup of editorial boards in 2016 and 2021, we obtained the current and the 2016 editorial board roster. Roster member images were collected from academic institutional websites. Betaface facial recognition software was used to assess the images. The software assigned the gender, race and ethnicity of the image supplied. Betaface results were analyzed using a Chi Square Test of Independence. Results: We analyzed 17 surgical journals. Only four of 17 journals were found to have diversity pledges on their website. For diversity themed publications, 1% of articles in 2016 and 2.7% in 2021 were published specifically about diversity. There was a significant increase in the amount of diversity articles/journal published per year in 2016 (6.59) compared to 2021 (25.94, P < 0.001). There was no correlation between impact factor and articles that publish diversity keywords. 1968 editorial board member images were analyzed using Betaface software to determine gender and race in both time periods. There was no significant increase in diversity of editorial board members regarding gender, race, and ethnicity temporally from 2016 to 2021. Conclusions: In the present study, we found that although the number of diversity-themed articles has increased over the last 5 y, however the gender and racial makeup of surgical editorial boards has not changed. Further initiatives are needed to better track and diversify the gender and racial composition of surgical editorial boards.