Purpose: To examine the predictive validity of the National Board of Osteopathic Medical Examiners' Comprehensive Osteopathic Medical Licensing Examination of the United States of America (COMLEX-USA) series with regard to the American Board of Family Medicine's (ABFM's) In-Training Examination (ITE) and Maintenance of Certification for Family Physicians (MC-FP) Examination.
Method: A repeated-measures design was employed, using test scores across seven levels of training for 1,023 DOs who took the MC-FP for the first time between April 2012 and November 2014 and for whom the ABFM had ITE scores for each of their residency years. Pearson and disattenuated correlations were calculated; Fisher r to z transformation was performed; and sensitivity, specificity, and positive and negative predictive values for the COMLEX-USA Level 2-Cognitive Evaluation (CE) with regard to the MC-FP were computed.
Results: The Pearson and disattenuated correlations ranged from 0.55 to 0.69 and from 0.61 to 0.80, respectively. For MC-FP scores, only the correlation increase from the COMLEX-USA Level 2-CE to Level 3 was statistically significant (for Pearson correlations: z = 2.41, P = .008; for disattenuated correlations: z = 3.16, P < .001). The sensitivity, specificity, and positive and negative predictive values of the COMLEX-USA Level 2-CE with the MC-FP were 0.90, 0.39, 0.96, and 0.19, respectively.
Conclusions: Evidence was found that the COMLEX-USA can assist family medicine residency program directors in predicting later resident performance on the ABFM's ITE and MC-FP, which is becoming increasingly important as graduate medical education accreditation moves toward a single aligned model.