RESUMO
OSCE for evaluating clinical competence still remains limited in France. This study presents the results of the first experimental use of an OSCE as a formative assessment of French general practice trainees. Fifty trainees rotated through a circuit of 15 standardized patient-based OSCE cases. Differences in scores were determined by analysis of variance. Reliability was calculated with the coefficient alpha. Pearson correlations were used to determine the relationship between station scores and overall OSCE score. Written questionnaires based on Likert-type scales were used to measure OSCE feasibility. No difference was found between total session scores. Significant item-total score correlations were found for 12 of the 15 clinical problems. The reliability of the examination was 0.58. Most participants agreed that the clinical situations were realistic, simulated patients were believable and sampling of cases was representative of general practice. These data confirm the feasibility of OSCE in assessing performance of general practice trainees. Optimal training of observers should improve examination reliability. This study warrants further development to confirm its usefulness as a summative assessment tool.