A Parametric Investigation of Mortality at All Ages in a Rural, South African Population

David J. Sharrow, University of Washington
Samuel J. Clark, University of Washington

Throughout the 1990s and into the 21st century, southern Africa experienced increasing adult mortality, which researchers have linked to escalating HIV prevalence over this period. In this paper, we utilize a parametric model, the Heligman-Pollard law of mortality, estimated via a Bayesian Melding procedure, to assess rapidly changing mortality schedules for men and women in a rural population of South Africa. Despite having been used in a variety of other, mostly economically developed contexts, the Heligman-Pollard model has been under utilized in the analysis of mortality in high HIV-prevalence settings. The ability of parametric models to easily summarize changes in mortality over time via elegant demographic parameter interpretations, make these methods especially valuable for researchers concerned with dynamic mortality profiles in the presence HIV/AIDS. This paper also reports on the relationship between growing HIV prevalence in this region and the change in the Heligman-Pollard parameters over this time period.