Fostering Diagnostic Accuracy in a Medical Intelligent Tutoring System

Diagnostic classification is an important part of clinical care, which is often the main determinant of treatment and prognosis. Clinicians’ underor over-confidence in their performance on diagnostic tasks can result in diagnostic errors which can lead to delay in appropriate treatment and unnecessary increase in the cost of medical care. This paper presents a version of SlideTutor aiming to reduce pathologists’ and dermatopathologists’ bias in diagnostic decision-making. This is accomplished by frequently prompting them to make metacognitive judgments of confidence, presenting them with the expert diagnostic solution path for each case, and de-biasing them by making them conscious of their metacognitive biases. This paper describes and summarizes the functionalities of SlideTutor, its cognitive training, tutoring phase, expert feedback, metacognitive intervention, and the open learner model.