Full Longitudinal Mediation Modeling Analysis of Frailty, Depression, and Cognitive Impairment

Abstract Frailty, depression, and cognitive impairment (CI) are interrelated geriatric syndromes that significantly impact the functioning and quality of life of older adults. In the context of a super-aging society, a clear understanding of their long-term interactions is essential for timely intervention and evidence-based policymaking. This study aimed to identify the directionality, mediating pathways, and critical time points for intervention by analyzing the dynamic interactions among them over time using the four-wave Korean Frailty and Aging Cohort Study (N = 1,377). Data were collected from a nationally representative sample of older adults in Korea from 2016 to 2023. Descriptive statistics characterized baseline levels and wave-specific changes. Pearson’s correlation coefficients assessed cross-sectional associations, while autoregressive and cross-lagged panel models evaluated temporal stability and causal directionality. Full Longitudinal Mediation Modeling (FLMM) explored potential mediation effects. All three variables showed significant temporal stability (β = .60–.75, p < .001). Depression in the initial stage significantly predicted increased frailty (β =.18, p < .01) and CI (β = –.12, p < .05) in subsequent waves, while the reverse effects were weak or insignificant. FLMM showed that depression mediated the effect of baseline cognitive function on future frailty, explaining approximately 35% of the total effect. Depression was identified as a significant mediator and longitudinal predictor of late-life physical and CI. Early screening and intervention for depressive symptoms can mitigate the progression to frailty and CI. These findings support integrated and timely strategies in community-based geriatric care and provide practical guidance for the effective allocation of preventive resources.