Bridging Social Determinants of Health and Multidimensional Frailty Among Older Adults: A Domain-Specific Latent Class and Network Analysis.
Abstract
Background
Social determinants of health (SDoH) are increasingly recognized as key contributors to frailty, yet their multidimensional associations remain underexplored. This study examined how multiple SDoH domains and specific variables relate to distinct frailty domains among Chinese community-dwelling older adults.
Methods
We analyzed 10,654 adults aged ≥ 60 years in the 2020 China Health and Retirement Longitudinal Study. Five SDoH domains, defined by the Healthy People 2030 framework, were assessed: Economic Stability, Education Access and Quality, Health Care Access and Quality, Neighborhood and Built Environment, and Social and Community Context. Latent class analysis was used to operationalize domain-level SDoH groupings. Frailty was measured using a 30-item multidimensional frailty index. Mixed graphical models were applied to estimate domain- and variable-level networks. Centrality and bridge strength identified influential and bridging nodes. Complete-case and post hoc edge removal sensitivity analyses were conducted.
Results
At the domain level, Education Access and Quality and Economic Stability were the main bridge domains linking SDoH to frailty, with the strongest association observed between Education Access and Quality and cognition deficits (edge weight = 0.688). At the variable level, education, access to healthcare, and employment status were key bridging variables linking SDoH to frailty. Cognition consistently showed the highest bridge strength within the frailty community. The strong education-cognition association should, however, be interpreted cautiously.
Conclusions
Education-, economic-, and healthcare-related disadvantage were key structural links to frailty, with cognition serving as the main frailty bridge. These findings provide insight into important pathways through which social disadvantage may contribute to frailty risk. CLINICAL TRIAL NUMBER Not applicable.