This study aims to construct age-stratified parametric survival models of disability risk based on intrinsic capacity (IC) using nationwide longitudinal data and to establish a simplified IC grading system.

Using longitudinal data from CHARLS 2011—2020 for individuals aged 60 years and older, IC scores were calculated using six commonly used methods, and five parametric survival models were constructed within each age group. The optimal model was selected based on the concordance index (C-index) and Akaike Information Criterion (AIC), and reference thresholds for IC grading were determined according to the quartiles of predicted disability risk.

A total of 3 272 older adults were included, with a median age of 65.00 (62.00-70.00) years. IC scores calculated by the 100-score scaling method combined with the log-normal parametric survival model showed the best predictive performance in most age groups, particularly for the 60-69 years (AIC=13 497.80, C-index=0.648 7), 70-79 years (AIC=5 306.87, C-index=0.624 8), and ≥80 years (AIC=804.73, C-index=0.643 9) age groups. The survival risk model revealed that IC grading thresholds decreased with increasing disability risk, with variations observed across age groups, sexes and numbers of chronic diseases. Older adults, females and those with more chronic diseases had lower IC thresholds. For instance, at the median disability risk level among those without chronic diseases, IC thresholds were 55.80 and 51.60 for males and females aged 60-69 years, respectively, decreasing to 45.10 and 34.80 for those aged ≥80 years.

IC scores calculated using the 100-score scaling method combined with the log-normal parametric survival model showed the best predictive performance for disability risk among Chinese older adults. The established IC grading thresholds may provide a reference for health assessment and graded interventions among Chinese older adults.