Using Topic Modeling to Identify Patients’ Educational Needs
Abstract: With the increase in chronic diseases in the US as the population ages, there is a need to facilitate the selection of relevant educational materials for patients. Augmenting electronic medical record systems so that they can recommend educational materials is consistent with their “meaningful use” as mandated in recent legislation related to healthcare. Recommending appropriate educational materials to patients is especially necessary because the outcomes of chronic diseases are highly dependent on patient self-management. This paper presents a method for assessing patient educational needs based on the patient’s medical record. The method we present is based on topic modeling and we report on several techniques for improving results when using topic modeling in this context.
Presider: Pasquina Campanella, University of Bari