Please join us for a Joint GWU/CTSI-CN/Washington DC VA Informatics Seminar taking place Wednesday, September 3, 11:00am - 12:00pm Eastern Daylight Time. Dr. Liz Workman will present:
More than the Sum of its Parts: Applying Topic Modeling and Explainable AI to Deep Learning in Understanding Problematic Opioid Use
Problematic opioid use is a major crisis, especially among Veterans of the U.S. Military. In previous work we developed a natural language processing tool to identify problematic opioid use in Veterans Affairs clinical notes. In this work, we developed an application that identifies clinical notes associated with patients who were documented as experiencing problematic opioid use only in clinical notes, as compared to patients who had received a relevant ICD code for problematic opioid use. The application used topic and topic word output from topic modeling to train deep neural network models. The models performed well, achieving area under the curve values of 82% or above, exceeding the performance of baseline models using only topics or topic words. We also computed impact scores, an explainable artificial intelligence method that identifies critical features in training deep learning models. The impact scores extended additional understanding to the data and outcomes.
Dr. Workman’s research interests mainly address how people obtain, process, and communicate biomedical information, and how discoveries in these areas can lead to improved patient outcomes, especially for underserved populations. She received a PhD in Biomedical Informatics from the University of Utah in 2011, where she developed a novel algorithm that adaptively summarizes abstracts from the biomedical literature. Her postdoctoral work at the National Library of Medicine/National Institutes of Health addressed several research areas, including information seeking behavior, automated knowledge discovery, data visualization, retrieval algorithms, and the semantic analysis of text. Dr. Workman’s current research activities include applying methods in natural language processing, machine learning and generative A.I. to clinical text. This has led to new methods and findings in clinical text, including generative AI - enabled symptom extraction, and insights into problematic opioid use documentation.