Skip to main content

  • Residency Applications and Implicit Bias

    By Jean Shaw
    Selected and Reviewed by Neil M. Bressler, MD, and Deputy Editors
    Add to My Bookmarks

    Journal Highlights

    JAMA Ophthalmology, December 2021

    Download PDF

    For ophthalmology residency appli­cants, does the redaction of identifi­ers that might trigger reviewers’ bias have an impact on application scores? Pershing et al. compared redacted and unredacted application reviews at their academic institution. They found that the distribution of application scores was similar for redacted versus unre­dacted applications, with no differences based on an applicant’s sex, race/eth­nicity, or country of origin.

    For this quality improvement study, 46 faculty members reviewed randomized sets of 462 redacted and unredacted applications during the 2019-2020 ophthalmology residency application cycle. Redacted identifiers included name, sex or gender, race and ethnicity, and race- or gender-associ­ated groups or activities (e.g., student associations). Each application was reviewed in redacted form by one reviewer and in unredacted form by another. All reviewers received implicit bias training, and each faculty mem­ber reviewed a mix of redacted and unredacted applications. Applications were assigned a score from 1 (best) to 9 (worst). Linear regression was used to evaluate the adjusted association of redaction, self-reported applicant char­acteristics, and reviewer characteristics with application scores.

    Of the 462 applicants, 185 (40%) were female, 277 (60%) were male, 71 (15.4%) self-identified as underrepre­sented in medicine (URiM) individuals, and 47 (10.2%) were international med­ical graduates (IMGs). Of the 46 faculty members, 14 (30.4%) were female, 32 (69.6%) were male, and two (.4%) were URiM individuals.

    The mean application review score was 4.83. The distribution of scores was similar for redacted versus unredacted applications, with no difference based on sex, URiM status, or IMG status. Applicants were more likely to receive better scores if they had attended a top 20 medical school (as ranked by U.S. News & World Report), held an additional advanced degree, and had higher United States Medical Licensing Examination (USMLE) Step 1 and Step 2 scores.

    Overall, the authors said, they be­lieve that these results are encouraging. However, they noted that metrics such as medical school rankings and USMLE scores may be subject to con­founding and bias. They recommended other initiatives to improve diversity, such as pipeline programs, implicit bias training for faculty, and targeted out­reach to underrepresented applicants. (Also see related commentary by O’Rese J. Knight, MD, Elise V. Mike, MD, PhD, and Angela R. Elam, MD, in the same issue.)

    The original article can be found here.