AI-powered hybrid self-management coaching for adults with early-onset Type 2 diabetes in Korea: A randomized...

Existing diabetes self-management education and support (DSME/S) programs have predominantly focused on middle-aged and older adult populations with type 2 diabetes (T2D), leaving unclear whether these approaches adequately address individuals with early-onset T2D (diagnosed before age 40). While AI-powered applications show promise for digital-native populations, research indicates systematic resistance to medical AI due to uniqueness neglect-The perception that algorithms cannot account for individual circumstances. The 3IM intervention (3I: Interest, intimacy, interactions; 3 M: Monitoring, measurement, maximization) examines whether augmenting AI-driven coaching with human support enhances outcomes for adults with early-onset T2D.

A randomized controlled trial employing a two-arm parallel-group design will be conducted with eighty-two Korean adults of childbearing age (20-44 years) diagnosed with type 2 diabetes before the age of 40. After a 2-week run-in period for baseline data collection, participants will be randomly assigned to either a comparison group receiving a 12-week coaching program via an AI-powered mobile application only, or an intervention group receiving the same AI coaching supplemented with virtual human support. All statistical analyses will be performed using SPSS (version 26.0+) and/or SAS (version 9.4+). Statistical significance will be defined as a two-sided p-value <0.05. Primary analysis will follow the intention-to-treat (ITT) principle, with a per-protocol (PP) analysis additionally conducted to assess efficacy among study completers. Qualitative data will be analyzed using a triangulation method.

Findings will facilitate the translation of an evidence-based, clinically relevant DSME/S intervention tailored specifically for adults with early-onset T2D.

KCT0010509, Korea Diabetes Control and Prevention agency.