JMIR Publications

ABSTRACT

Background:

Background:

Improving self-management behaviours is crucial for better treatment outcomes in type 2 diabetes mellitus (T2DM). Mobile health (mHealth) interventions can overcome the limitations of traditional community management by facilitating behavioural change. However, most models focus on predicting behaviour initiation, with limited insight into mechanisms of initiation and long-term maintenance. To address this, we developed Artificial Intelligence-based Health Education Accurately Linking System (AI-HEALS), a WeChat-based mHealth intervention grounded in an extended Multi-Theory Model (MTM) of Health Behavior Change, to explain diabetes self-management processes.

Objective:

Objective:

To examine how the extended MTM explains the impact of AI-HEALS on self-management behaviours in adults with T2DM.

Methods:

Methods:

Building on a randomized controlled trial (RCT) in which the intervention group received the AI-HEALS mobile app, we conducted a mixed-methods study. Phase 1 validated an extended MTM-based scale and assessed MTM constructs, diabetes-specific skills, and self-management behaviors using structural equation modeling (SEM). Phase 2 comprised longitudinal semi-structured telephone interviews, analyzed using thematic analysis.

Results:

Results:

The extended MTM scale demonstrated good reliability and validity. SEM revealed that changes in baseline social environment (β = 0.226, P = 0.003), physical environment (β = 0.248, P = 0.001), and diabetes-related skills at 6 months (β = 0.162, P = 0.013) directly predicted self-management behaviours at 12 months. Additionally, baseline changes in social environment indirectly influenced behaviour confidence at 6 months via participatory dialogue (disadvantages) at 3 months. Participatory dialogue (advantages) at 3 months positively influenced both behavioural confidence and emotional transformation at 6 months. Thematic analysis of qualitative data identified three key domains, environmental, cognitive, and attitudinal/skill-related factors, that influenced behavioural change. The findings from quantitative and qualitative strands were consistent and complementary.

Conclusions:

Conclusion: Both quantitative and qualitative findings demonstrate the theoretical and practical value of the extended MTM in explaining and promoting self-management behaviours among individuals with T2DM. As an interactive digital health intervention, AI-HEALS facilitated behavioural change through a multidimensional approach, enhancing environmental accessibility, providing cognitive support, and reinforcing skill-building. The extended MTM framework holds promise for broader application across diverse populations and behavioural contexts. It offers a robust theoretical foundation for designing more targeted and sustainable interventions, ultimately empowering patients, improving digital health literacy, and advancing personalised chronic disease management. Clinical Trial: Registration: Chinese Clinical Trial Registry Identifier: ChiCTR2300068952.