Behavioral pathways to hospitalization: physical exercise, social participation, and healthcare utilization in China

For middle-aged and older adults, everyday health behaviours — physical exercise in particular — are closely tied to how much they rely on the healthcare system, yet evidence on how exercise and different forms of social participation relate to intensive healthcare use remains limited. This study examines how physical exercise and several forms of social participation are associated with hospitalization frequency. Using five waves of the China Health and Retirement Longitudinal Study (CHARLS, 2011–2020), we analyse 46,129 person-wave observations from 22,400 respondents. The primary specification is a nested negative binomial model in correlated-random-effects form with year fixed effects. Conditional fixed-effects Poisson and Poisson models serve as supporting and sensitivity analyses. Additional sensitivity analyses examine vigorous, moderate, and light exercise and self-rated health as an alternative outcome. Heterogeneity by gender, age, and education is examined with subgroup models and formal interaction tests. Physical exercise is negatively associated with hospitalization frequency, and the association is stable across specifications \(\:\left(\beta\:\:=\:-0.257,\:robust\:SE\:=\:0.054,\:95\%\:CI\:-0.362\:to\:-0.151,\:p\:<\:0.001\:in\:the\:full\:model\right)\) . Playing mahjong, chess, or cards, or visiting a community activity room is negatively associated with hospitalization frequency, as is participation in educational or training activities, although the latter is uncommon. Going out and interacting with friends, organizational participation, and volunteer/charitable or non-coresident caregiving activities are not statistically significant. In sensitivity analyses, vigorous, moderate, and light exercise are each negatively associated with hospitalization frequency, while self-rated health shows a directionally favorable pattern that is not uniformly significant across thresholds. Exercise coefficients are negative across gender, age, and education subgroups, but formal interaction tests show no evidence of heterogeneity. Physical exercise and selected forms of social participation are associated with lower intensive healthcare utilization among middle-aged and older adults. The findings are broadly supported by alternative exercise-intensity measures and, less uniformly, by self-rated health as an alternative outcome. Given the observational design, these results should be interpreted as associations rather than evidence of causal reductions in hospitalization.