Process for Teaching and\tEvaluating Cognition in Medical Decision Making Through Computer Based Simulation

Date: March 25, 2004 Time: (All day)
Event Type: Lecture

We present a learning system that allows a medical trainee to managesimulated patient problems at a high level of decision-making. It allows thetrainee to rehearse problem solving in patient management, to request helpduring the simulation, to receive mentoring relative to the areas that needimprovement, and to demonstrate increasing competence using objectivemeasures with a goal of autonomy with expertise. Within the learning system,the trainee interacts with an ontology which contains general medical andsurgical knowledge and can be supplemented with opinionated as well asevidence-based knowledge. The architecture allows different learning approaches to be employed andallows the trainee to manage a patient from initial contact throughcompletion of management while demonstrating the cognitive decision-makingskills required to step through educationally important aspects. The traineeis able to make a wide choice of actions in patient management at any timein the simulation using uncued selection mechanisms. The outcome of eachaction has consequences, either negative or positive, and is notpre-scripted to result in a single outcome. Help is a combination of virtualmentor feedback and highly relevant literature. The tailored response isinfluenced by the path taken by the trainee, the current state of theinteraction, and projections into the next state of the interaction. Thesystem retrieves the literature component through knowledge gateways (e.g.PubMed) using a search strategy implied by the trainee interaction with theontology and ontology-based reasoning. The trainee can modify this searchfurther to reflect need more accurately.In this presentation, we will describe a) an initial proof-of-conceptsystem, its interaction environment and help capabilities that utilizePubMed; and b) some of the underlying knowledge and language technology - our approach to ontology building and ontology-based text processing thatincludes descriptions of complex processes in diseases, diagnostics andtreatment. We will also suggest how our system can both benefit from and beinstrumental in enriching UMLS and MeSH. \t\t