Measuring adult mortality in the era of HIV/AIDS: Estimates from census age distribution

Robert G. Mswia, University of North Carolina at Chapel Hill
Philip W Setel, University of North Carolina at Chapel Hill
Tukufu Zuberi, University of Pennsylvania

In resource-poor settings, researchers apply various techniques to estimate mortality and other demographic outcomes. We employ Preston-Bennett technique to illustrate how one can estimate levels of adult mortality in the era of HIV/AIDS. Here we use data on population age distribution from two consecutive censuses in the United Republic of Tanzania. This method indirectly takes into account the impact of HIV/AIDS on the population that would not be captured by other estimation techniques. We find high levels of mortality between ages 15–60 years in the intercensal period. The timing and age-patterns of mortality strongly suggest a major demographic impact of HIV/AIDS. We also find significant sex differentials on mortality levels: younger women more affected than men. In the absence of a vital statistics system that produces good quality, representative data on mortality levels and causes of death, census-based age-distribution methods can be useful for investigating the HIV/AIDS impact on mortality.

  See paper

Presented in Session 22: Adult mortality in Africa