Estimating life expectancy using an age-cohort model in Taiwan.

Abstract

OBJECTIVES: Life expectation is a valuable summary index in public health and actuarial science. The life expectancies published in the vital statistics, however, are derived from the "current" rather than from the "cohort" life table. The former is based on a strong assumption of constant mortality in the population, whereas the latter calls for a recording of the mortality experience of a group of individuals, which is often an impossible task. Thus, a method of calculating cohort life expectancy without actual follow up is much needed. METHODS: Estimation of cohort life expectancy was based on an age-cohort model. Mortality data for the male population in Taiwan from 1951 to 1990 are used to illustrate the methodology. RESULTS: The increment of life expectancy over time in Taiwan is actually steeper than was previously thought using the current life table technique. CONCLUSIONS: The method is easy to implement and the data required are the usual age and period cross classified mortality data. It warrants further investigation.