Using United States Medical Licensing Examination(®) (USMLE) examination results to predict later in-training examination performance among general surgery residents.

OBJECTIVE: To examine the relationship between and predictive nature of United States Medical Licensing Examination(®) (USMLE) Step 1, Step 2, and American Board of Surgery In-Training Examination (ABSITE) postgraduate year (PGY) 1-5 scores from 2 general surgery programs from 1999-2009, with a goal of discerning how Step 1 and Step 2 scores should be used in resident selection and screening.
DESIGN: A descriptive, retrospective, correlational study was conducted using data from existing program records.
SETTING: Two accredited Midwestern community teaching hospital general surgery residency programs. PARTICIPANTS: Data were collected from the records of N = 34 residents completing the programs in the years 1999-2009.
RESULTS: Although a statistically significant correlation exists between USMLE Step 1 and ABSITE PGY 3 scores, Step 2 scores were more highly correlated to ABSITE scores from all years of residency training. In both hierarchical and simple regression models, Step 2 scores were superior in predicting ABSITE PGY 1-5 scores, with the strongest predictability for PGY 3 and PGY 5 scores.
CONCLUSIONS: USMLE Step 1 scores showed limited utility in predicting later ABSITE scores whereas Step 2 scores were more predictive of all years of ABSITE scores. These findings should prompt additional research into the relationship between examination performances at different points along the general surgery education continuum. In the mean time, general surgery program directors and faculty might do well to examine the relationships between Step 1 and Step 2 scores and ABSITE scores in their own programs to evaluate the usefulness of considering either score when ranking potential residents for selection into a program.