Expert, Novice, and Intermediate Performance: Exploring the Relationship Between Clinical Reasoning Behaviors and Diagnostic Performance

Abstract

Understanding the diagnostic process and the interplay between gathering and interpreting information can reduce the inaccuracies that lead to medical errors. In this study, we examined the relationship between medical students’ (n = 46) performance profiles and the type of clinical reasoning behaviors they executed while diagnosing a clinical patient in the context of an intelligent tutoring system, BioWorld [2]. Performance was measured by efficiency (similarity to an expert solution), confidence, and time. We found three groups: high, low, and intermediate performance. High-performing students were characterized by high efficiency, intermediate students had average efficiency and confidence, and low performing students were more characterized by low confidence rather than their efficiency score. We found that the high performers put more effort in integrating elements of the clinical case, a deep learning strategy. Unexpectedly, the high and intermediate groups additionally selected more information from the patient history, a shallow learning strategy. Our findings contribute to understanding of learning of clinical reasoning skills using an intelligent tutoring system.