Multidimensional Social Isolation and All-Cause Mortality Among Middle-Aged and Older Adults in China: Gender and Rural-Urban Differences

This study examines whether multidimensional social isolation is associated with all-cause mortality and explores whether these associations vary by gender and rural-urban residence in China. Using nationally representative longitudinal data from 12,686 adults aged 45 and older in the China Health and Retirement Longitudinal Study (2011-2020), we estimated Cox proportional hazards models examining social isolation as structural disconnectedness across five domains: partnership status, contact with parents, contact with children, contact with friends, and social participation. Social isolation showed a graded association with mortality, with moderately and severely isolated individuals experiencing higher mortality hazards than socially integrated adults. Being unpartnered and not participating in social activities were also associated with higher mortality. Stratified analyses suggested stronger associations with overall isolation among rural than urban residents. Findings highlight the value of examining both cumulative social disconnectedness and its distinct relational domains in understanding mortality risk in China.