[Transition patterns of frailty states among Chinese aged 50 years and above: a sequence analysis]

Objective: This study aims to use sequence analysis to identify frailty state transition patterns in longitudinal measurements of Chinese people aged ≥50 years, and then analyze the associations of demographic, socioeconomic, and behavioral factors with these patterns. Methods: This study used four waves of data from the China Health and Retirement Longitudinal Study, wave 1 to wave 4 longitudinal data, consisting of 32 variables to construct a frailty index, which were categorized into five states based on their magnitude, and then constructed a sequence of frailty states for each individual. Second, this study uses optimal matching for clustering sequences. Finally, we used multinomial logistic regression to estimate the associations of demographic, socioeconomic, and behavioral factors with the different patterns of frailty sequence. Results: This study included 8 670 survey respondents aged ≥50 years with complete data across all four waves. More than half of the respondents (68.20%) who were in a robust state at baseline transitioned to a pre-frail state during the 8-year follow-up period (2011-2018). Men spent a longer mean time in a robust state than women (2.62 years vs. 1.59 years), a slightly shorter mean time in a pre-frail state (4.22 years vs. 4.52 years), and a significantly shorter mean time in a frail state (1.01 years vs. 1.74 years). Sequences of frailty state transitions for all respondents were clustered to obtain four patterns. The results of multinomial logistic regression analysis showed that the transition from a state of frailty to an unfavorable state pattern (patterns 2, 3, and 4) was generally associated with poorer sociodemographic characteristics. Conclusions: Our study revealed distinct patterns and durations of frailty transitions across different gender and age groups. A range of demographic, socioeconomic, and behavioural factors influenced the transition patterns of frailty states, underscoring the necessity of personalised interventions.