Psychosocial resources, sleep quality, and self-management behavior in patients with type 2 diabetes mellitus: a latent profile and mediation analysis

Abstract

Background Type 2 diabetes mellitus (T2DM) requires sustained self-management, yet adherence remains suboptimal. This study aimed to identify heterogeneous self-management behavior profiles in T2DM patients and explore the mediating role of sleep quality between integrated psychosocial resources and self-management. Methods A cross-sectional study recruited 474 hospitalized T2DM patients from a tertiary hospital in Jinan, China. Self-management behaviors were assessed using the Diabetes Self-Care Scale (DSCS), while psychosocial resources integrated scores from the Perceived Social Support Scale, Heartland Forgiveness Scale, and Self-Regulatory Fatigue Scale(reverse-scored). Sleep quality was measured via the Pittsburgh Sleep Quality Index (PSQI). Latent Profile Analysis (LPA) classified self-management patterns, and structural equation modeling tested sleep quality’s mediation effect using bootstrap sampling (5,000 resamples). Results LPA identified four distinct self-management profiles: Low (17.93%), Medium-Low (28.48%), Medium-High (19.83%), and High (33.76%). Significant intergroup differences emerged in sleep quality (PSQI), social support, forgiveness, and self-regulatory fatigue (P < 0.001). Psychosocial resources correlated strongly with self-management profiles (β = 0.398, P < 0.001) and were significantly associated with better sleep quality (β=-0.836, P < 0.001). Sleep quality partially mediated the relationship, accounting for 52.39% of the total effect (indirect β = 0.438, 95% CI [0.343, 0.526]). Conclusions T2DM patients exhibit heterogeneous self-management patterns. Psychosocial resources were positively associated with self-management behaviors directly and indirectly through sleep quality. Interventions targeting psychosocial support and sleep improvement may foster sustainable self-management behaviors. Trial registration : Not applicable (observational study).