Comparative values of medical school assessments in the prediction of internship performance.

BACKGROUND: Multiple undergraduate achievements have been used for graduate admission consideration. Their relative values in the prediction of residency performance are not clear. This study compared the contributions of major undergraduate assessments to the prediction of internship performance.

METHODS: Internship performance ratings of the graduates of a medical school were collected from 2012 to 2015. Hierarchical multiple regression analyses were used to examine the predictive values of undergraduate measures assessing basic and clinical sciences knowledge and clinical performances, after controlling for differences in the Medical College Admission Test (MCAT).

RESULTS: Four hundred eighty (75%) graduates' archived data were used in the study. Analyses revealed that clinical competencies, assessed by the USMLE Step 2 CK, NBME medicine exam, and an eight-station objective structured clinical examination (OSCE), were strong predictors of internship performance. Neither the USMLE Step 1 nor the inpatient internal medicine clerkship evaluation predicted internship performance. The undergraduate assessments as a whole showed a significant collective relationship with internship performance (ΔR2  = 0.12, p < 0.001).

CONCLUSIONS: The study supports the use of clinical competency assessments, instead of pre-clinical measures, in graduate admission consideration. It also provides validity evidence for OSCE scores in the prediction of workplace performance.