Reality or simulation? Performance analysis of medical students through different practical assessments

Abstract: Introduction: Assessing clinical skills represents a challenge in medical training. The heterogeneity in the selection of patients associated to a lack of objective criteria has resulted in a methodological shift toward using the objective structured clinical examination (OSCE). Objective: This study aimed to identify whether the OSCE results in a Gaussian frequency distribution of performance assessment marks compared to the traditional model. Method: 239 students from the Semiotics module of the medicine course between 2016 (traditional model) and 2017 (OSCE) were analyzed using the two-dimensional Kolmogorov-Smirnov test and Student’s t-test for correlation with the performance coefficient (PC). Results: The test scores in the traditional model (p<0.0001; KS = 0.1881) are more discrepant in relation to the normal distribution than the OSCE test scores (p=0.0010; KS=0.1134) and show more correlation with the PC (p <0.0001; r = 0.45) than the OSCE model (p = 0.31; r=0.06). Conclusion: OSCE might provide a more accurate assessment of a student’s performance during medical internship.