THE RELATIONSHIP OF CLINICAL KNOWLEDGE TO MONTHS OF CLINICAL-TRAININGAMONG MEDICAL-STUDENTS

Purpose. To assess the correlation of the number of months of clinical training with clinical knowledge, as measured by the United Stares Me dical Licensing Examination (USMLE) Step 2. Method. The total number o f months of clinical training and percentile scores on USMLE Step 2 we re determined for 217 Stanford University School of Medicine graduates from 1992 through 1994. Percentile scores on each subsection of the M edical College Admission Test (MCAT) and the National Board of Medical Examiners Part 1 or USMLE Step 1 (Part 1/Step 1) for the graduates we re also determined. For some analyses the graduates were separated int o three groups according to the duration of clinical training. The Pea rson product-moment correlation coefficient was used to quantify and d efine the significance of correlations. The Jonckheere-Terpstra nonpar ametric test was used to assess trends across the three groups. A mult iple linear regression model was used to test the effects of confoundi ng variables. Results. The total numbers of clerkship months ranged fr om 12 to 23; the median was 18. A highly significant correlation was f ound between increasing months of clinical training and increasing sco res on Step 2 (p = .002); a weaker significant correlation was found w ith scores on Part 1/Step 1 (p = .03). The correlation for Step 2 scor es did not diminish appreciably (p = .004) when scores for Part 1/Step 1 and each MCAT subsection were introduced into the regression model. Conclusion. A highly significant correlation was found between the am ount of clinical training and the acquisition and utilization of clini cal knowledge. In the current climate of concerns about the rising cos ts of medical education, the impulse to solve these concerns by decrea sing the amount of medical students' clinical training should be appro ached with caution.