BACKGROUND AND PURPOSE: The relationship between osteopathic (COMLEX) and allopathic (USMLE) exam scores in not well described. We sought to describe the relationship between COMLEX and USMLE scores for osteopathic medical students who reported scores for both exams during application to internal medicine residency. METHODS: Cross sectional study of 588 matched USMLE/COMLEX Step 1 scores and 241 matched Step 2 scores. Means, standard deviations, and pairwise correlations for matched scores were calculated. RESULTS: USMLE and COMLEX paired scores resulted in a Pearson's correlation of 0.85 for Step 1 scores and 0.79 for Step 2 scores. COMLEX means were 560 and 561 for Level 1 and 2 and USMLE means were 209 and 215 for Step 1 and 2. CONCLUSIONS: USMLE and COMLEX scores are strongly related for internal medicine residency applicants who took both exams. Our sample performed approximately 0.8 standard deviation higher that the national COMLEX means on both Level 1 and 2 but at the national mean on USMLE Step 1 and 2. The variation in scoring methodology and potential differences between the study sample and individual applicant characteristics suggesteject caution should be taken in using a given COMLEX score as a predictor of performance on the USMLE.
BACKGROUND AND PURPOSE: The relationship between osteopathic (COMLEX) and allopathic (USMLE) exam scores in not well described. We sought to describe the relationship between COMLEX and USMLE scores for osteopathic medical students who reported scores for both exams during application to internal medicine residency. METHODS: Cross sectional study of 588 matched USMLE/COMLEX Step 1 scores and 241 matched Step 2 scores. Means, standard deviations, and pairwise correlations for matched scores were calculated. RESULTS: USMLE and COMLEX paired scores resulted in a Pearson's correlation of 0.85 for Step 1 scores and 0.79 for Step 2 scores. COMLEX means were 560 and 561 for Level 1 and 2 and USMLE means were 209 and 215 for Step 1 and 2. CONCLUSIONS: USMLE and COMLEX scores are strongly related for internal medicine residency applicants who took both exams. Our sample performed approximately 0.8 standard deviation higher that the national COMLEX means on both Level 1 and 2 but at the national mean on USMLE Step 1 and 2. The variation in scoring methodology and potential differences between the study sample and individual applicant characteristics suggesteject caution should be taken in using a given COMLEX score as a predictor of performance on the USMLE.