PhD candidate Michelle Bak successfully defended her dissertation, "Promoting a Healthy and Comprehensive Diet through Theory-Driven Large Language Models-based Agents," on July 14.
Her committee included Assistant Professor Jessie Chin (chair); Associate Professor Dong Wang; Ian Brooks, iSchool research scientist and director of the Center for Health Informatics; Affiliate Associate Professor Jana Diesner; and Suma Bhat, assistant professor in the Department of Electrical and Computer Engineering.
Abstract: This dissertation explores how integrating behavior change theories into Large Language Model (LLM) can improve digital health counseling for individuals ambivalent about dietary change. General purpose LLMs often fail to adapt responses to users' motivational readiness, particularly in early stages of change. To address this gap, this study integrates the Transtheoretical Model (TTM) and Motivational Interviewing (MI) into prompt engineering to meet the information needs of these individuals in adopting a healthy and comprehensive diet. The improved LLM demonstrates potential in encouraging cognitive and affective self-image assessment in relation to health behaviors to reduce ambivalence and strengthen commitment to change through targeted, psychologically grounded interactions. The improved LLM also significantly increased behavioral intention without altering knowledge or risk perception, aligned with characteristics of individuals in the contemplation stage. Through MI-based strategies like reflective listening and affirmations, the improved LLM helped participants reduce ambivalence and consider actionable steps to dietary change. This research lays a foundation for digital health solutions that enable personalized interventions and long-term maintenance of health behaviors.