RESUMO
INTRODUCTION: In France, at the end of the sixth year of medical studies, students take a national ranking examination including progressive clinical case-based multiple-choice questions (MCQs). We aimed to evaluate the ability of these MCQs for testing higher-order thinking more than knowledge recall, and to identify their characteristics associated with success and discrimination. METHODS: We analysed the 72 progressive clinical cases taken by the students in the years 2016-2019, through an online platform. RESULTS: A total of 72 progressive clinical cases (18 for each of the 4 studied years), corresponding to 1059 questions, were analysed. Most of the clinical cases (n=43, 60%) had 15 questions. Clinical questions represented 89% of all questions, whereas basic sciences questions accounted for 9%. The most frequent medical subspecialties were internal medicine (n=90, 8%) and infectious diseases (n=88, 8%). The most frequent question types concerned therapeutics (26%), exams (19%), diagnosis (14%), and semiology (13%). Level 2 questions ("understand and apply") accounted for 59% of all questions according to the Bloom's taxonomy. The level of Bloom's taxonomy significantly changed over time with a decreasing number of level 1 questions ("remember") (P=0.04). We also analysed the results of the students among 853 questions of training ECNi. Success and discrimination significantly decreased when the number of correct answers increased (P<0.0001 both). The success, discrimination, mean score, and mean number of discrepancies did not differ according to the diagnosis, exam, imaging, semiology, or therapeutic type of questions. CONCLUSION: Progressive clinical case-based MCQs represent an innovative way to evaluate undergraduate students.