PC26. Predicting a Successful Plastic Surgery Residency Applicant - The Linear Rank Model

Abstract

PURPOSE: Linear rank modeling (LRM) is an algorithm that uses relevant inputs to model expert judgment. Our institution previously described how LRM improves consistency and fairness in resident selection. We have utilized LRM over the last five years to assist with ranking integrated plastic surgery residency (PRS) applicants. The primary aim of this study was to determine if LRM predicts applicant match success. Our secondary aim was to assess if LRM demonstrates bias against groups of underrepresented gender or self-identified race categories. METHODS: The LRM scores for applicants who applied to a single institution between 2018-2022 were calculated. Data was collected on applicant demographics, Step 1 & 2 scores, number of publications, extracurricular activity scores, letter of recommendation scores, global intuition rank, and match success. RESULTS: The mean LRM score of 231 applicants was 59. With every one-point increase in LRM score, there was a significant 8.3% increase in the likelihood of applicant match success (p <.001). An algorithm was developed to estimate the probability of match success based on the LRM score. No significant difference in average LRM scores was found between gender (p =.285) or self-identified race groups (p =.137). CONCLUSION: LRM is a valuable tool to help create rank lists by providing an objective assessment of residency applicants. This model can rank applicants based on a holistic set of criteria and estimate an applicant’s probability of successfully matching into an integrated PRS residency.