Probabilistic Population Projections for Countries with Generalized HIV Epidemics

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

Population projections tell us about the future size and age composition of a population. Governments, international organizations, and researchers use these projections for planning, monitoring development goals and research purposes. Typically, these are done deterministically, but there is extensive interest in probabilistic projections. Where probabilistic projections are available, they are limited to countries without substantial levels of HIV prevalence because current mortality projection methodology is inapplicable due to the high number of adult deaths resulting from a generalized HIV epidemic. We propose a method for making probabilistic population projections for countries with generalized HIV epidemics by incorporating the future trajectory of the epidemic in the mortality component of the projection. Our method takes into account uncertainty about future levels of mortality and fertility, the major drivers of population change, as well as uncertainty about the trajectory of HIV prevalence.