Abstract Body: Background: Personalized digital health and community-based interpersonal strategies are two promising approaches to improving type 2 diabetes (T2D) management, yet their comparative efficacy is unknown. This study directly compared the effects of a personality-guided mHealth intervention and a community-led interpersonal diffusion strategy on glycemic control in a community-based randomized trial.

Methods:We conducted a 12-month (June 2022 to June 2023), multicenter, cluster-randomized controlled trial. A total of 618 patients with T2D were allocated to one of three groups: a control group (n=207), a personality-tailored mHealth intervention group (n=204), or a community-based interpersonal strategy group (n=207). The mHealth group received smartphone notifications customized according to their Big Five personality trait profiles. The interpersonal group received the identical core content delivered via printed materials and peer-support activities. The primary outcome was the change in glycated hemoglobin (HbA1c) from baseline to 12 months. Data were analyzed using linear mixed-effects models with fixed effects for time, group, and their interaction, adjusted for age, sex, education, income, Household registration, BMI, and Course of disease.

Results:Among the 618 randomized participants (mean age 64.5±7.9 years, 53.6% male, mean baseline HbA1c 7.1±1.1%), 2303 observations were included in the analysis. A significant time effect was observed, indicating a natural rise in HbA1c over the study period in the control group. At 3 months, both intervention groups showed a significantly smaller increase in HbA1c compared to the control (mHealth: β = -0.286, p = .001; interpersonal: β = -0.337, p < .001). This effect persisted at 6 months (mHealth: β = -0.284, p = .001; interpersonal: β = -0.199, p = .023). By 12 months, between-group differences were no longer significant.Moderation analysis revealed that conscientiousness (interaction β = -0.118, p = .025) and neuroticism (interaction β = -0.156, p = .013) significantly moderated the effect of the mHealth intervention at 6 months.

Conclusion:The short-term efficacy of both interventions and the moderating role of personality traits underscore the necessity of moving beyond static diabetes management programs. Our results advocate for a new paradigm of adaptive, personalized care that integrates initial psychological tailoring with proactive long-term support strategies to sustain clinical benefits.