Impact of different scoring methods on the clinical skills assessment of internal medicine students

The aim of this study was to determine how different scoring methods could influence the results achieved in clinical competence assessment examinations at the end of the medical school curriculum. Fifty-six final-year medical school students were examined using a programme of multimedia computer-based simulated cases (Procuste programme). For generating results, the following scoring methods were applied: (1) the method used for the Internal Medicine official certification examination which includes both weighted and unweighted items; (2) a method with only unweighted items;(3) a method which includes single-item analysis. Each time an absolute pass/fail cut-off standard ( S 60/100) and a relative pass/fail cut-off (mean minus 1 standard deviation) were applied.Taking as a reference the first method, the effect of using different scoring methods on the percentage of student failure was evaluated. Furthermore, correlation studies between the results obtained with these scoring methods and the grades reported by the examinees during the medical school curriculum were performed. It was found that the first method generated a more realistic pass/fail rate compared with the two other methods. Furthermore this method, which has been used for the last 10 years for administering internal medicine examinations, showed better correlation coefficients with the curriculum scores.