Authors: Dinesh Gottipati, MS
Primary Advisor: Trevor Cohen, MBChB, PhD
Committee Members: Jiajie Zhang, PhD
Masters thesis, The University of Texas School of Biomedical Informatics at Houston.
Abstract:
Given the prominent and direct role played by non-expert physicians in clinical care, and the differences between expert and novice comprehension, a support system that provides support by compensating for these differences is desirable. The cognitive literature on expertise emphasizes experts’ distinctive use of intermediate constructs, which are clinically relevant meaningful clusters of signs and symptoms during medical problem solving. Based on these findings, researchers have pursued the goal of developing a cognitive support system that assumes some of the cognitive burden of expert comprehension, namely the use of intermediate constructs, in order to support the clinical comprehension of trainees. In this paper, we evaluate one such cognitive support system, which is based on a previously validated simulation of expert organization of psychiatric narrative into diagnostically relevant higher level knowledge structures. Propositional analysis and Latent Semantic Analysis are the fundamental methods used to measure the effects of the system on clinical comprehension. Results indicate that the system supports the generation of facet-level hypothesis for simple clinical cases, and enhances the coherence of clinical summaries produced by trainees. These results contribute to our understanding of the use of facets as a design concept for cognitive support systems, and indicate directions for further research.