Abstract
BACKGROUND: Although intrinsic capacity (IC) has been constructed in older populations, whether IC retains the same structure over time has not been formally examined, nor have the factors associated with the changes in IC over time been thoroughly investigated. This study aimed to establish that the structure of IC remains unchanged over time by testing its longitudinal measurement invariance and to investigate factors that influence the longitudinal change of IC over time. METHODS: Data came from 7,271 participants aged 60 and older from the China Health and Retirement Longitudinal Study in 2011 (Wave 1) and 2015 (Wave 3). Bifactor confirmatory factor analysis (CFA) was used to construct IC with its domains, and the longitudinal measurement invariance of IC between Waves was tested. RESULTS: Bifactor CFA fitted the data well at both Waves and showed good construct validity. Partial scalar invariance was supported with non-invariant intercepts for delayed word recall, math, and close vision. Decreases in IC were associated with increasing age, being female (-0.030, 95% CI: -0.045, -0.016), living in rural areas (-0.019, 95% CI: -0.030, -0.009), BMI < 18.5 (-0.019, 95% CI: -0.035, -0.003), and hypertension (-0.012, 95% CI: -0.022, -0.001). Increases in IC were associated with higher education (primary school: 0.012, 95% CI: 0.001, 0.024; lower secondary school: 0.023, 95% CI: 0.005, 0.041) and drinking ≥4/week (0.019, 95% CI: 0.003, 0.034). Stratifying the sample by gender, the protective effect of education was observed only in women. CONCLUSIONS: The bifactor structure of the IC construct was valid and retained its meaning over time. Longitudinal changes in IC were associated with various sociodemographic factors, lifestyle, and health conditions, confirming the need to monitor IC for timely intervention, particularly in those with risk factors for IC decline.