Age patterns and sex ratios of adult mortality in countries with high HIV prevalence

Bruno Masquelier, Université Catholique de Louvain
Jeff Eaton, Imperial College London
Patrick Gerland, United Nations Population Division
François Pelletier, United Nations Population Division

Adult mortality rates are poorly estimated in countries with high HIV prevalence due to the lack of comprehensive vital registration. A common approach consists in adding the number of AIDS-related deaths derived from epidemiological modeling on top of AIDS-free mortality rates. The latter are often obtained by anchoring model schedules of mortality to child survival. With such a counterfactual approach, age patterns and sex ratios of adult mortality are largely driven by model parameters. Here we compare UNAIDS all-cause mortality rates and orphanhood prevalence with survey and census data to detect significant deviations between empirical and model-based estimates. We also assess the sensitivity of model outputs to changes in the sex ratio of incidence and age patterns of infection. Both sibling and orphanhood data suggest larger sex ratios than in model-based estimates. Age patterns of mortality also have a larger AIDS hump in model outputs than in surveys and censuses.