The association between quality of connections and diagnostic accuracy in student-generated concept maps for clinical reasoning education with virtual patients.

Abstract

Concept maps are a learning tool that fosters clinical reasoning skills in healthcare education. They can be developed by students in combination with virtual patients to create a visual representation of the clinical reasoning process while solving a case. However, in order to optimize feedback, there is a need to better understand the role of connections between concepts in student-generated maps. Therefore, in this study we investigated whether the quality of these connections is indicative of diagnostic accuracy. We analyzed 40 concept maps created by fifth-year medical students in the context of four virtual patients with commonly encountered diagnoses. Half of the maps were created by students who made a correct diagnosis on the first attempt; the other half were created by students who made an error in their first diagnosis. The connections in the maps were rated by two reviewers using a relational scoring system. Analysis of covariance was employed to examine the difference in mean connection scores among groups while controlling for the number of connections. There were no differences between the groups in the number of concepts or connections in the maps; however, maps made by students who made a correct first diagnosis had higher scores for the quality of connections than those created by students who made an incorrect first diagnosis (12.13 vs 9.09; p=0.03). We also observed students' general reluctance to use connections in their concept maps. Our results suggest that the quality, not the quantity, of connections in concept maps is indicative of their diagnostic accuracy.