The multistate life table method has been
proposed by Rogers and colleagues in order to take into account the
recovery of lost functions and return to a state of good health.
Data comes from longitudinal surveys which provide
requested probabilities over a couple of years:
- onset of disability or other health problems
- recovery from disability or other heath problems
- probability of dying for people who had initially disability and for people who were initially healthy.
Based on this set of probabilities, multistate life tables can be computed by simulating from age to age the risks of entering disability, recovering and dying and deriving from it the person-years with and without disability.
The advantage of this method - based on transitions between
states of health - is that it gives a period
indicator that takes into account the reversibility of
disability. The specific drawback of the multistate life table
method arises from the scarceness of adequate
data. Data requirements for multistate methods are
considerable and there are very few countries where national data
are available or likely to be available for some time. Biases are
introduced when the gaps between successive waves of longitudinal
studies are too long, thus failing to capture a part of the flows
between health states during the inter-survey.
- Rogers A, Rogers RG, Branch LG. A multistate analysis of active life expectancy. Public Health Rep 1989;104:222-225.
On this webpage you can find documents and links toward different methods to compute multistate life tables:
- The Interpolated Markov Chain method
Lièvre A., Brouard N. and Heathcote Ch. Estimating health expectancies from cross-longitudinal surveys. Mathematical Population Studies 2003;10(4):211-248
- The Gibbs Sampler for Multistate Life Tables Software-GSMLT: Brown S, Lynch S.
- The SPACE program:
Cai L., Hayward M., Saito Y., Lubitz J., Hagedorn A., Crimmins E.M. Estimation of multi-state life table functions and their variability from complex survey data using the SPACE Program. Demographic Res. 2010;22(6):129-158.