Modelling multi-state health transitions in China: a generalised linear model with time trends

Katja Hanewald, Han Li, Adam W. Shao

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)


Rapid population ageing in China has urged the need to understand health transitions of older Chinese to assist the development of social security programmes and financial products aimed at funding long-term care. In this paper, we develop a new flexible approach to modelling health transitions in a multi-state Markov model that allows for age effects, time trends and age-time interactions. The model is implemented in the generalised linear modelling framework. We apply the model to evaluate health transitions of Chinese elderly using individual-level panel data from the Chinese Longitudinal Healthy Longevity Survey for the period 1998-2012. Our results confirm that time trends and age-time interactions are important factors explaining health transitions in addition to the more commonly used age effects. We document that differences in disability and mortality rates continue to persist between urban and rural older Chinese. We also compute life expectancies and healthy life expectancies based on the proposed model as inputs for the development of aged care and financial services for older Chinese.

Original languageEnglish
Pages (from-to)145–165
Number of pages21
JournalAnnals of Actuarial Science
Issue number1
Early online date2 Jul 2018
Publication statusPublished - Mar 2019


  • Generalised linear models (GLMs)
  • Multi-state health transitions
  • Long-term care
  • Healthy life expectancy
  • China


Dive into the research topics of 'Modelling multi-state health transitions in China: a generalised linear model with time trends'. Together they form a unique fingerprint.

Cite this