Latent profiles and associated factors of self-management behaviours among middle-aged and older adults with type 2 diabetes

Abstract

Type 2 diabetes requires sustained self-management across multiple behavioural domains, yet middle-aged and older adults may not follow a single high-versus-low pattern. This study aimed to identify latent profiles of self-management behaviours among middle-aged and older adults with type 2 diabetes and to examine factors associated with profile membership. A hospital-based cross-sectional study was conducted at two hospitals in China from January to June 2026. A total of 706 adults aged 45 years or older with type 2 diabetes were included. Diabetes self-management behaviour was assessed using the Chinese version of the Diabetes Self-Care Scale (DSCS), and the six DSCS dimension scores were used as manifest indicators in latent profile analysis (LPA). Health literacy, social support, self-efficacy, illness perception, sociodemographic characteristics, clinical characteristics, and health-related variables were also assessed. One- to six-profile models were fitted and compared using information criteria, entropy, likelihood ratio tests, profile size, parsimony, and interpretability. Multinomial logistic regression was used to examine variables related to latent profile membership, with the high self-management profile as the reference category. A five-profile solution was selected and showed high classification quality (entropy = 0.920). The profiles were labelled low self-management (15.7%), medium–low self-management with monitoring/foot care deficits (13.2%), exercise-deficit self-management (13.7%), medium–high self-management with good medication adherence (40.4%), and high self-management (17.0%). All six DSCS dimensions and the DSCS total score differed significantly across profiles (all P < 0.001). Lower educational attainment, insulin-containing therapy, and higher systolic blood pressure were mainly associated with membership in lower or domain-specific self-management profiles. Higher health literacy, self-efficacy, and illness perception scores were associated with lower odds of membership in the non-reference profiles, whereas social support showed profile-specific associations. Middle-aged and older adults with type 2 diabetes showed distinct patterns of self-management behaviour across dietary control, exercise regularity, blood glucose monitoring, medication adherence, foot care, and management of hypo-/hyperglycaemia. A profile-based understanding of self-management may help nurses and clinicians identify patients who require comprehensive support and those who need targeted assistance in specific domains.