HIV affects both mortality and fertility and, consequently, it has important effects on population growth and the sex and age composition of a population. A fuller understanding of the biological and behavioral determinants of HIV transmission would give us the ability to design effective interventions that target specific mechanisms, situations, and people.
For generalized HIV epidemics, where the prevalence among pregnant women is consistently over 1%, Bayesian melding has been used as the basis for probabilistic HIV prevalence projections by UNAIDS and WHO. It combines expert opinion and past trends in inputs and outputs, and provides a framework for assessing uncertainty in mechanistic models in general. The talk will introduce some statistical methods for estimation and short-term extrapolation of HIV/AIDS trends from limited surveillance data, and an Incremental Mixture Importance Sampling (IMIS) algorithm that improves the sampling efficiency of the Bayesian melding while retaining its essential simplicity and transparency.
More about this seminar http://www.healthmetricsandevaluation.org/news-events/seminar/statistical-models-estimating-and-predicting-hivaids-epidemics