Virtual patients in the acquisition of clinical reasoning skills: does presentation mode matter? A quasi-randomized controlled trial

Background: The objective of this study is to compare two different instructional methods in the curricular use of computerized virtual patients in undergraduate medical education. We aim to investigate whether using many short and focused cases - the key feature principle - is more effective for the learning of clinical reasoning skills than using few long and systematic cases. Methods: We conducted a quasi-randomized, non-blinded, controlled parallel-group intervention trial in a large medical school in Southwestern Germany. During two seminar sessions, fourth-and fifth-year medical students (n = 56) worked on the differential diagnosis of the acute abdomen. The educational tool - virtual patients - was the same, but the instructional method differed: In one trial arm, students worked on multiple short cases, with the instruction being focused only on important elements ("key feature arm", n = 30). In the other trial arm, students worked on few long cases, with the instruction being comprehensive and systematic ("systematic arm", n = 26). The overall training time was the same in both arms. The students' clinical reasoning capacity was measured by a specifically developed instrument, a script concordance test. Their motivation and the perceived effectiveness of the instruction were assessed using a structured evaluation questionnaire. Results: Upon completion of the script concordance test with a reference score of 80 points and a standard deviation of 5 for experts, students in the key feature arm attained a mean of 57.4 points (95% confidence interval: 50.9-63.9), and in the systematic arm, 62.7 points (57.2-68.2), with Cohen's d at 0.337. The difference is statistically non-significant (p = 0.214). In the evaluation survey, students in the key feature arm indicated that they experienced more time pressure and perceived the material as more difficult. Conclusions: In this study powered for a medium effect, we could not provide empirical evidence for the hypothesis that a key feature-based instruction on multiple short cases is superior to a systematic instruction on few long cases in the curricular implementation of virtual patients. The results of the evaluation survey suggest that learners should be given enough time to work through case examples, and that caution should be taken to prevent cognitive overload.