Further Applications and Clarification of the Multistate Life Table Decomposition Method

Tianyu Shen , Vienna Institute of Demography, Austria Academy of Sciences
Iñaki Permanyer, Center for Demographic Studies

Understanding why populations differ in life expectancy and related health measures remains a central question in demography and population health. Decomposition methods provide a valuable framework to disentangle these differences by quantifying the contributions of population composition and transition dynamics. While multistate life tables offer a more complete representation of population processes than Sullivan’s method, existing decomposition approaches often overlook the explicit role of mortality transitions and are limited to characteristics fixed at baseline. This study extends current decomposition methods within the multiple multistate method (MMM) framework to explicitly incorporate mortality as part of the decomposition and to accommodate both time-fixed and time-varying subgroup characteristics. The proposed approach enables researchers to quantify the contribution of mortality alongside other state transitions, offering a more flexible and interpretable tool for analyzing disparities in multistate outcomes. We demonstrate the method using data from the Health and Retirement Study across three applications: a three-state model with mortality as an absorbing state, a model stratified by educational attainment, and a multiple multistate model jointly considering morbidity and disability. The results illustrate how the extended decomposition clarifies the relative contributions of population structure, transition dynamics, and mortality to differences in health expectancy and longevity.

See extended abstract

 Presented in Session 39. Innovations in Life Table and Multistate Modeling