I mproving workforce representation is an important component of addressing gaps in healthcare, as physician-patient concordance in language, demographics and culture has been correlated with greater patient satisfaction and adherence. Currently, ophthalmology faculty at U.S. medical schools remain among the least heterogeneous compared to other specialties, with prior studies reporting gender-based differences in academic promotion, with men achieving higher promotion rates than women. ^, Similar gender-based differences are also observed at the trainee level, with the proportion of female residents in ophthalmology stagnant at around 40% over multiple decades. ^, These trends are discordant with the increasing proportion of female graduates from U.S. medical schools and the demographics of the U.S. population. ^,

As an important milestone in a physician’s career, oral board examinations must be equitable for all examinees. Yeo and associates reported that resident demographics and family status at internship were observed to be associated with American Board of Surgery (ABS) pass rates. In response to prior findings, the ABS is addressing the potential for unconscious bias among board examiners by increasing representation and adding implicit bias training. Similarly, improving representation among examiners can help mitigate implicit bias towards examinees while also promoting a more experienced team that enhances organizational problem-solving and decision-making capacity.

The purpose of our study was to evaluate characteristics of the American Board of Ophthalmology (ABO) oral board examiners and examinees, as well as the trends in examiner and examinee gender representation over time. Establishing baseline data can help inform and promote efforts to enhance representation and ensure fairness for examinees.

This retrospective cohort study was determined to be exempt by the Institutional Review Board of the Johns Hopkins University School of Medicine. All data were de-identified, and the study adhered to the tenets of the Declaration of Helsinki.

Data of the oral board examiners and examinees were obtained from the ABO. Examiner data were available for the years 2011 through 2024, while the examinee data were available from 2013 to 2024. For examiners, demographic information included gender and race/ethnicity. Professional characteristics included year of initial certification, number of years since certification, years of service as an examiner, total number of exams administered, total number of candidates examined, and average examiner score. Examiners were defined as individuals who participated in administering at least one ABO oral examination during the study period. The average examiner score was defined as the average proportion of candidates passed by each examiner across all oral examinations administered during the study period. For examinees, demographic variables included age at first attempt, gender, and race/ethnicity. Additional variables included years since residency graduation and the first attempt pass/fail status.

Due to incomplete reporting of demographic information for both examiners and examinees, the ABO data was linked with the physician demographic information from the Association of American Medical Colleges (AAMC) to improve the completeness of this information and minimize misclassification of demographic variables.

The primary outcome was examiner and examinee gender representation. Secondary outcomes included average examiner scores, examiner characteristics, and examinee first-attempt oral board examination pass status.

Descriptive statistics were used to summarize examiner and examinee characteristics stratified by gender. Pearson’s chi-squared tests were used to cf categorical variables. For continuous variables, Student’s t-tests were used. Trends in examiner and examinee gender between 2011 and 2024 were assessed using the logistic regression and Wald tests to evaluate changes from reference years. Missing data were reported but not imputed. Data analysis was performed using the Stata version 18.0, and the significance level was set at p <.05. To account for multiple comparisons, a Bonferroni correction was applied. Statistical significance was defined as p <.007 for examiner characteristics (0.05/7 variables) and p <.0125 for examinee characteristics (0.05/4 variables).