Related Experiment Videos
Does the National Resident Matching Program Rank List Predict Accreditation College of Graduate Medical Education
Steven P Cohen1,2,3, Oludare O Olatoye4, Peter Schumaker5
1Departments of Physical Medicine & Rehabilitation and Anesthesiology, Walter Reed National Military Medical Center, Uniformed Services University of the Health Sciences.
Background:
Training programs seek to match fellows who will perform well during training. The National Resident Matching Program (NRMP) allows programs to choose the most capable trainees, with higher ranked applicants expected to perform better. Pain medicine is one of the most popular non-internal medicine fellowships in the U.S., and multiple specialty pathways for entry make selecting candidates challenging.
Objective:
The aim of this study is to correlate fellow rank order and match number with academic performance.
Methods:
Match information and fellow performance from 4 highly-ranked multispecialty pain programs was obtained with performance evaluated based on the 6 final Accreditation Council for Graduate Medical Education (ACGME) core competencies. The primary analysis correlated each fellow's relative rank-list position with relative subsequent ACGME performance within the same institution and fellowship cohort using cohort-relative percentile ranks, Spearman correlation, stratified permutation testing, and stratified bootstrap confidence intervals. Sensitivity and exploratory analyses used absolute/converted match-number fixed-effects regression, Kendall Tau-B, sex stratification, and domain-specific correlations with false discovery rate adjustment.
Results:
The primary analysis included 93 fellows across 20 institution/cohort strata. Rank-list standing was not significantly associated with ACGME performance (Spearman r = 0.06; 95% CI, -0.19 to 0.32; p = 0.63), and sensitivity, sex-stratified, and domain-specific analyses were also nonsignificant. When stratified according to top-, middle- and bottom-ranked applicants, all domains were significant with bottom-ranked fellows faring worse than other groups.
Conclusions:
Pain medicine rank lists correlate poorly with fellowship performance, though a weak association was noted wherein bottom-ranked applicants performed worse. This suggests that either the metrics used to rank pain medicine applicants and/or those used to measure performance require modification.