Using Educational Data-Mining Techniques to Identify Medical Student Misconceptions in a Virtual Patient Environment

• In Event: Technology in Teaching

Abstract

This study aimed to discover misconceptions in medical diagnostic reasoning by mining user interactions in MedU, a web-based learning environment with virtual patient cases. Data from 13,000 attempts at a single case were extracted from the MedU database. A subgroup discovery method was applied to discover patterns in learner-generated annotations and answers to multiple-choice items pertaining to the diagnosis and management of acute myocardial infarction (heart attack). A two-step supervised approach was found to significantly increase prediction precision, uncovering four common misconceptions at a rate greater than 70%. These findings inform the design of an adaptive system that tailors the delivery of formative feedback to address the specific needs of learners who exhibit different misconceptions in medical diagnostic reasoning.

Authors

• Eric G. Poitras, University of Utah

• Laura Naismith, University Health Network

• Tenzin Doleck, McGill University

• Susanne P. Lajoie, McGill University