Predicting A Successful Plastic Surgery Residency Applicant - The Linear Rank Model
Ellen C. Shaffrey, MD, Steven Moura, MA, Pradeep K. Attaluri, MD, Peter J. Wirth, MD, Armin Edalatpour, MD, Alyssa Schappe, BS, Venkat K. Rao, MD, MBA
Division of Plastic and Reconstructive Surgery, University of Wisconsin, Madison, WI
BACKGROUND
•Residency applicant selection is highly competitive and complex
•Prior efforts include personality assessments and standardized letters of recommendation1-3
•Individual metrics and standardized letters may perpetuate the underrepresentation of disadvantaged groups4,5
•Linear Rank Model (LRM) uses aggregate data to model human judgment6
PROBLEM
•While the AAMC advocates for holistic evaluation, due to the growing number of applicants, programs resort to convenient screening methods
AIMS
Primary: Assess if the LRM is predictive of applicant match success for applicants
Secondary: Examine if there are differences in LRM scores between the diversity, equity, and inclusion (DEI) categories of gender and self-identified race
METHODS
•Collected data to screen and rank applicants
•Screened = 617 applicants (2021 – 2022)
• Ranked = 231 applicants (2019 – 2022)
•Standardized rubric applied to screen all applicants and assign a numerical rating (Table 1)
Statistics:
•T-test to evaluate differences between LRM score of race and gender groups
•Logistic regression and Area Under the Curve (AUC) to evaluate predictiveness of the LRM
RESULTS
•No significant difference in LRM score between gender or race for all groups except screened race group (p= 0.047)
•LRM score was significantly associated with match success for both screened (OR 1.11, p< 0.001) and interviewed (OR 1.08, p< 0.001) applicants
•AUC modeling demonstrated that LRM score is significantly more predictive than other traditional application metrics for screened applicants (Figure 1)
•An equation was derived to estimate the probability of an applicant’s match success into plastic surgery based on LRM score (Figure 2)
CONCLUSIONS
•The Linear Rank Model is a standardized tool to efficiently perform a holistic screening of a residency applicant pool
•The Linear Rank Model is a more predictive screening tool for match success than USMLE Scores or research productivity
•How the LRM accounts for race merits further investigation.
REFERENCES
1.Kimple AJ, McClurg SW, del Signore AG, Tomoum MO, Lin FC, Senior BA. Standardized letters of recommendation and successful match into otolaryngology. In: Laryngoscope. Vol 126. John Wiley and Sons Inc.; 2016:1071-1076.
2.Love JN, Smith J, Weizberg M, et al. Council of emergency medicine residency directors’ standardized letter of recommendation: The program director’s perspective. Academic Emergency Medicine. 2014;21(6):680-687.
3.Sarac BA, Janis JE. Matching into Plastic Surgery: Insights into the Data. Plast Reconstr Surg Glob Open. 2022;10(5):e4323.
4.Jarman BT, Kallies KJ, Joshi ART, et al. Underrepresented Minorities are Underrepresented Among General Surgery Applicants Selected to Interview. J Surg Educ. 2019;76(6):e15-e23.
5.Edmond MB, Deschenes JL, Eckler M, Wenzel RP. Racial bias in using USMLE step 1 scores to grant internal medicine residency interviews. Acad Med. 2001;76(12):1253- 1256.
6.Elmaraghi S, Rao VK, Christie BM. Linear Modeling to Reduce Bias in Plastic Surgery Residency Selection. Plast Reconstr Surg. Published online 2021:538-544.
