Abstract
OBJECTIVE: This study compares performance data from physicians completing 3-year versus 4-year emergency medicine residency training programs. Currently, there are 2 training formats and little is known about objective performance differences. METHODS: This was a retrospective cross-sectional analysis of emergency residents and physicians. Multiple analyses were conducted comparing physicians' performances, including Accreditation Council of Graduate Medical Education Milestones and American Board of Emergency Medicine In-training Examination (ITE), Qualifying Examination (QE), Oral Certification Examination (OCE), and program extensions from 3-year and 4-year residency programs. Some confounding variables were not or could not be considered, such as rationale for medical students to choose one format over another, as well as application and final match rates. RESULTS: Milestone scores are higher for emergency medicine 3 residents in 1-3 programs (3.51) versus emergency medicine 3 residents in 1-4 programs (3.07; P < 0.001, d = 1.47) and highest for emergency medicine 4 residents (3.67). There was no significant difference in program extension rates (emergency medicine 1-3, 8.1%; emergency medicine 1-4, 9.6%; P = 0.05, ω = 0.02). ITE scores were higher for emergency medicine 1, 2, and 3 residents from 1-3 programs and emergency medicine 4 residents from 1-4 programs scored highest. Mean QE score was slightly higher for emergency 1-3 physicians (83.55 vs 83.00; P < 0.01, d = 0.10). QE pass rate was higher for emergency 1-3 physicians (93.1% vs 90.8%; P < 0.001, ω = 0.08). Mean OCE score was slightly higher for emergency 1-4 physicians (5.67 vs 5.65; P = 0.03, d = -0.07) but did not reach a priori statistical significance (α < 0.01). OCE pass rate was also slightly higher for emergency 1-4 physicians (96.9% vs 95.5%; P = 0.06, ω = -0.07) but also non-significant. CONCLUSIONS: These results suggest that although performance measures demonstrate small differences between physicians from emergency medicine 1-3 and 1-4 programs, these differences are limited in their ability to make causal claims about performance on the basis of program format alone.