Gender Bias in Simulation-Based Assessments of Emergency Medicine Residents

Background: Gender-related disparities persist in medicine and medical education. Prior work has found differences in medical education assessments based on gender.

Objective: We hypothesized that gender bias would be mitigated in a simulation-based assessment.

Methods: We conducted a retrospective cohort study of emergency medicine residents at a single, urban residency program. Beginning in spring 2013, residents participated in mandatory individual simulation assessments. Twelve simulated cases were included in this study. Rating forms mapped milestone language to specific observable behaviors. A Bayesian regression was used to evaluate the effect of resident and rater gender on assessment scores. Both 95% credible intervals (CrIs) and a Region of Practical Equivalence approach were used to evaluate the results.

Results: Participants included 48 faculty raters (25 men [52%]) and 102 residents (47 men [46%]). The difference in scores between male and female residents (M = -0.58, 95% CrI -3.31-2.11), and male and female raters (M = 2.87, 95% CrI -0.43-6.30) was small and 95% CrIs overlapped with 0. The 95% CrI for the interaction between resident and rater gender also overlapped with 0 (M = 0.41, 95% CrI -3.71-4.23).

Conclusions: In a scripted and controlled system of assessments, there were no differences in scores due to resident or rater gender.