Leisure lifestyle patterns and multidimensional health outcomes among older adults in China: a latent class analysis of the 2023 Chinese General Social Survey

Leisure activities may relate to health in later life not only as individual behaviours but also as multidimensional configurations of participation. This study aimed to identify leisure lifestyle patterns among older adults in China and to examine their associations with depressive tendency, health limitation, and poor self-rated health. Data were drawn from the face-to-face household component of the 2023 Chinese General Social Survey. The primary latent class analysis included 2,307 adults aged 60 years or older with complete information on 12 ordered categorical leisure-activity indicators. Survey-weighted logistic regression models examined associations between class membership and the three outcomes after adjustment for demographic, family, socioeconomic, social-security, and body mass index covariates. Classification uncertainty was assessed through categorical distal-outcome analysis using the DCAT procedure and a 200-draw pseudo-class analysis. Survey-weighted and available-item latent class models, an alternative definition of poor self-rated health, and sex- and residence-based subgroup analyses were conducted as sensitivity or exploratory analyses. A four-class solution was retained, comprising low-frequency television-oriented (31.79%), moderate-frequency home-based screen-oriented (34.09%), high-frequency socially and physically active (21.47%), and high-frequency diversified (12.64%) patterns. Compared with the high-frequency diversified class, the low-frequency television-oriented class had higher adjusted odds of depressive tendency (OR 2.242, 95% CI 1.383–3.635), health limitation (OR 1.903, 95% CI 1.230–2.945), and poor self-rated health (OR 2.318, 95% CI 1.515–3.544). Corresponding ORs for the moderate-frequency home-based screen-oriented class were 2.037 (95% CI 1.280–3.241), 1.820 (95% CI 1.203–2.754), and 1.812 (95% CI 1.223–2.685), respectively. The high-frequency socially and physically active class had higher adjusted odds of depressive tendency (OR 1.659, 95% CI 1.005–2.738) and poor self-rated health (OR 1.796, 95% CI 1.175–2.744), but not health limitation (OR 1.240, 95% CI 0.791–1.943). However, its association with depressive tendency was attenuated after accounting for classification uncertainty in the pseudo-class analysis. The weighted and available-item latent class models yielded substantively similar class structures, and most exploratory subgroup interactions were not statistically significant. Leisure participation among older adults formed heterogeneous patterns that differed in both frequency and composition. Membership in the low-frequency television-oriented and moderate-frequency home-based screen-oriented classes was associated with less favourable indicators across the psychological, functional, and subjective health dimensions examined compared with membership in the high-frequency diversified class. These cross-sectional findings do not establish temporal ordering or causality. They nevertheless underscore the value of examining leisure activities as configurations and provide a basis for future longitudinal research on the barriers and resources associated with broader and personally meaningful participation.