Barriers in Ophthalmology Residency Applications for Students Identifying as Underrepresented in Medicine: A San Francisco Match Analysis.

OBJECTIVE: There is a significant lack of ophthalmologists who self-identify as underrepresented in medicine (URiM) in the physician workforce. Prior literature has revealed bias in traditional metrics for selection relied on by resident programs such as United States Medical Licensing Examination (USMLE) scores, letters of recommendation (LOR), and induction into medical honors societies such as Alpha Omega Alpha (AOA). The purpose of this study was to elucidate race-based differences in word usage within ophthalmology residency letters of recommendation that may disproportionately affect URiM applicants. DESIGN: This was a retrospective, cohort study. SETTING: This was a multicenter study across the Wilmer Eye Institute at Johns Hopkins, the University of California San Francisco, and the University of North Carolina at Chapel Hill. PARTICIPANTS: San Francisco (SF) Match applications submitted to three ophthalmology residency programs between 2018 and 2020 were reviewed. URiM status, USMLE Step 1 score, and AOA membership were recorded. Letters of recommendation were analyzed using text analysis software. T-tests and chi-squared or Fisher's exact tests were used to compare continuous and categorical variables, respectively. Frequency of word/summary term usage in letters of recommendation were the main outcome measures. RESULTS: Relative to non-URiM applicants, URiM applicants had lower USMLE Step 1 scores (mean difference=7.0; p<0.001). Non-URiM letters of recommendation were more likely to describe applicants as "dependable" (p=0.009) and highlight "research" (p=0.046). URiM letters were more likely to describe applicants as "warm" (p=0.02) and "caring" (p=0.02). CONCLUSIONS: This study identified potential barriers for URiM ophthalmology residency applicants which can help guide future interventions to increase workforce diversity.

All Other Things Being Equal: Exploring Racial and Gender Disparities in Medical School Honor Society Induction.

PURPOSE: A large body of literature has demonstrated racial and gender disparities in the physician workforce, but limited data are available regarding the potential origins of these disparities. To that end, the authors evaluated the effects of race and gender on Alpha Omega Alpha Honor Medical Society (AOA) and Gold Humanism Honor Society (GHHS) induction. METHOD: In this retrospective cohort study, the authors examined data from 11,781 Electronic Residency Application Service applications from 133 U.S. MD-granting medical schools to 12 residency programs in the 2014-2015 application cycle and to all 15 residency programs in the 2015-2016 cycle at Yale-New Haven Hospital. They estimated the odds of induction into AOA and GHHS using logistic regression models, adjusting for Step 1 score, research publications, citizenship status, training interruptions, and year of application. They used gender- and race-matched samples to account for differences in clerkship grades and to test for bias. RESULTS: Women were more likely than men to be inducted into GHHS (odds ratio 1.84, P < .001) but did not differ in their likelihood of being inducted into AOA. Black medical students were less likely to be inducted into AOA (odds ratio 0.37, P < .05) but not into GHHS. CONCLUSIONS: These findings demonstrate significant differences between groups in AOA and GHHS induction. Given the importance of honor society induction in residency applications and beyond, these differences must be explored further.

Plagiarised letters of recommendation submitted for the National Resident Matching Program.

OBJECTIVES: Residency is a critical step in the professional development of physicians. Given the high stakes of the process, it is not surprising, although nonetheless troubling, that professional misconduct may occur: multiple studies have, rightfully, condemned applicants for plagiarising personal statements or misrepresenting publications. To date, however, no studies have examined whether faculty members may engage in similar behaviours. METHODS: Software was used to evaluate 3864 unique applications containing 13 617 letters of recommendation submitted through the Electronic Residency Application Service (ERAS) to the Yale University Department of Psychiatry Residency Program in recruitment cycles from 2013-2014 to 2016-2017. The software identified pairs of letters written by different authors with at least 75 words of shared language, counting only words present in contiguous phrases of at least eight words (Shared Wording Across People, SWAP). Independent reviewers determined whether the shared language reflected acceptable forms of common usage (e.g. describing a setting of care) or was intended to convey unique attributes of an applicant, thereby representing plagiarism. RESULTS: We found that 5.1% of letters contained SWAP and 4.1% of letters contained plagiarism. In total, 11.8% of all applications included at least one plagiarised letter. By comparison, 2.6% of applicants' personal statements contained plagiarism. CONCLUSIONS: The present data demonstrate a surprising prevalence of plagiarism in letters of recommendation written for residency applicants. These data call into question both the relative weight that should be accorded to letters of recommendation and, more broadly, beg further discussion of how we conceptualise professionalism in medicine.

Differences in words used to describe racial and gender groups in Medical Student Performance Evaluations.

PURPOSE: The transition from medical school to residency is a critical step in the careers of physicians. Because of the standardized application process-wherein schools submit summative Medical Student Performance Evaluations (MSPE's)-it also represents a unique opportunity to assess the possible prevalence of racial and gender disparities, as shown elsewhere in medicine. METHOD: The authors conducted textual analysis of MSPE's from 6,000 US students applying to 16 residency programs at a single institution in 2014-15. They used custom software to extract demographic data and keyword frequency from each MSPE. The main outcome measure was the proportion of applicants described using 24 pre-determined words from four thematic categories ("standout traits", "ability", "grindstone habits", and "compassion"). RESULTS: The data showed significant differences based on race and gender. White applicants were more likely to be described using "standout" or "ability" keywords (including "exceptional", "best", and "outstanding") while Black applicants were more likely to be described as "competent". These differences remained significant after controlling for United States Medical Licensing Examination Step 1 scores. Female applicants were more frequently described as "caring", "compassionate", and "empathic" or "empathetic". Women were also more frequently described as "bright" and "organized". CONCLUSIONS: While the MSPE is intended to reflect an objective, summative assessment of students' qualifications, these data demonstrate for the first time systematic differences in how candidates are described based on racial/ethnic and gender group membership. Recognizing possible implicit biases and their potential impact is important for faculty who strive to create a more egalitarian medical community.

A quantitative experimental paradigm to optimize construction of rank order lists in the National Resident Matching Program: the ROSS-MOORE approach.

PURPOSE: As part of the National Resident Matching Program, programs must submit a rank order list of desired applicants. Despite the importance of this process and the numerous manifest limitations with traditional approaches, minimal research has been conducted to examine the accuracy of different ranking strategies. METHOD: The authors developed the Moore Optimized Ordinal Rank Estimator (MOORE), a novel algorithm for ranking applicants that is based on college sports ranking systems. Because it is not possible to study the Match in vivo, the authors then designed the Recruitment Outcomes Simulation System (ROSS). This program was used to simulate a series of interview seasons and to compare MOORE and traditional approaches under different conditions. RESULTS: The accuracy of traditional ranking and the MOORE approach are equally and adversely affected with higher levels of intrarater variability. However, compared with traditional ranking methods, MOORE produces a more accurate rank order list as interrater variability increases. CONCLUSIONS: The present data demonstrate three key findings. First, they provide proof of concept that it is possible to scientifically test the accuracy of different rank methods used in the Match. Second, they show that small amounts of variability can have a significant adverse impact on the accuracy of rank order lists. Finally, they demonstrate that an ordinal approach may lead to a more accurate rank order list in the presence of interviewer bias. The ROSS-MOORE approach offers programs a novel way to optimize the recruitment process and, potentially, to construct a more accurate rank order list.