R code for Circadian syndrome and adverse ADL and IADL functional trajectories among middle-aged and older adults in China

This repository contains R code supporting the primary analyses and key sensitivity analyses reported in the revised manuscript “Circadian syndrome is associated with adverse ADL and IADL functional trajectories among middle-aged and older adults in China.” The included code covers longitudinal ADL and IADL trajectory modelling, multinomial logistic regression, one-vs-rest causal forest analysis, posterior-probability–based assessment of trajectory-classification uncertainty, inverse-probability-of-attrition weighting, and propensity-overlap diagnostics.CircS is an observational exposure.

The term “causal forest” is retained as the established name of the algorithm, and the corresponding estimates are interpreted as model-estimated conditional probability differences rather than causal treatment effects.Individual-level data from the China Health and Retirement Longitudinal Study (CHARLS) are not included. Access to and redistribution of CHARLS data are governed by the CHARLS data-use requirements.