International Journal of Nursing Studies Advances · Jun 28, 2026
Longitudinal dynamic relationships among social frailty, depression, cognitive function, and physical frailty in community-dwelling older adults: A random intercept cross-lagged panel model study
By Wu A, Sun S, Huang X et al.
BACKGROUND: Social frailty is a critical indicator of social health in older adults. While its links to physical and mental health are recognized, longitudinal evidence integrating social frailty, depressive symptoms, cognitive function, and physical frailty is limited. OBJECTIVE: To examine the time-ordered and longitudinal predictive associations among these four domains in urban community-dwelling older adults. DESIGN: A longitudinal study using a random intercept cross-lagged panel model (RI-CLPM). SETTINGS: Urban communities in China. PARTICIPANTS: 3503 community-dwelling adults aged ≥60 years from the China Health and Retirement Longitudinal Study (2013-2018). METHODS: Social frailty was assessed using a proxy measure. RI-CLPM was applied to estimate between-person associations and within-person cross-lagged associations across three time points. RESULTS: The model showed good fit (CFI = 0.968, TLI = 0.923, RMSEA = 0.049, SRMR = 0.026). At the between-person level, social frailty was positively correlated with depressive symptoms (r = 0.352, p < .001) and negatively correlated with cognitive function (r = -0.622, p < .001); depressive symptoms were positively correlated with physical frailty (r = 0.370, p < .001). At the within-person level, autoregressive effects differed across time intervals-for example, social frailty T1→T2 (β = 0.074, p = .008), depressive symptoms T1→T2 (β = 0.054, p = .047), cognitive function T1→T2 (β = 0.068, p = .012), and physical frailty T2→T3 (β = 0.120, p < .001). Cross-lagged paths indicated that depressive symptoms at T1 were associated with higher physical frailty at T2 (β = 0.045, p < .05), and physical frailty at T2 predicted higher depressive symptoms at T3 (β = 0.051, p < .05). In addition, depressive symptoms at T2 predicted lower cognitive function at T3 (β = -0.047, p < .05), and cognitive function at T2 predicted lower physical frailty at T3 (β = -0.041, p < .05). Several within-wave residual correlations were significant, indicating concurrent associations among the constructs within the same time point beyond the modeled longitudinal pathways. (e.g., depressive symptoms with physical frailty and cognitive function). These were mainly observed between depressive symptoms and physical frailty and between depressive symptoms and cognitive function (e.g., T2: r = 0.260 and r = -0.110; T3: r = 0.231; all p < .001). CONCLUSIONS: Social frailty is primarily associated with psychological and cognitive status at the between-person level, whereas depressive symptoms and physical frailty exhibit bidirectional time-ordered associations. Community health management should adopt a multidimensional risk-profile perspective. REGISTRATION: Not registered.
doi.org/10.1016/j.ijnsa.2026.100615Open accessPubMed 42437170Populations