Beyond the USMLE: The STAR Algorithm for Initial Residency Applicant Screening and Interview Selection

Background: Efficient, nonbiased methods for screening residency candidates are lacking. The purpose of this study is to highlight the design, implementation, and impact of the Selection Tool for Applicants to Residency (STAR), an objective approach to selecting candidates to interview for residency selection purposes. Materials and methods: Single-institution retrospective cohort study of medical student applicants and current residents of a single otolaryngology residency program from 2008 to 2015 was performed. STAR was introduced to the selection process in 2013 with no USMLE cutoff score needed to receive an interview. Single-institution review of otolaryngology residency program applications from 2008 to 2015 was performed. STAR was introduced in 2013. In addition to applicants, we analyzed characteristics of residents who successfully matched into our program. Prealgorithm residents (n = 16) and postalgorithm residents (n = 12) were compared to assess the impact of this approach on characteristics of successfully matched residents at the program. Results: Three hundred sixty-five applications were analyzed. Applicant pools before and after algorithm displayed similar characteristics. Interestingly, while there was no USMLE "cutoff," scores significantly increased after algorithm. There was no significant difference in the proportion of women (P = 0.588) or underrepresented minorities (P = 0.587) invited to interview pre- and post-STAR. The algorithm significantly decreased the time needed to review applications and interview residency candidates without impacting the overall composition of the interviewee pool. Conclusions: Traditional application review methods can be time consuming and may not ensure effective screening of applicants. STAR, or similar objective tools, may be a viable alternative to evaluate applicants, reduce evaluative time, and potentially decrease the impact of unconscious bias. (C) 2018 Elsevier Inc. All rights reserved.