Estimation of Hard-to-Reach Populations at Risk for HIV Infection in Concentrated Epidemics Using Network Scale-Up and Other Methods

• Activity Number: 229
• Type: Topic Contributed
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Tyler McCormick and James W Harmon*+ and Adrian Raftery
• Companies: University of Washington
• Keywords: network ; Bayesian ; epidemic ; HIV ; AIDS

Abstract

Estimating the populations at risk for HIV infection poses some significant challenges. We need accurate information in order to make good policy decisions regarding treatment and prevention. However, some populations of interest are "hard-to-reach" due to the illegal nature of the population (such as drug users) or due to societal prejudices against the population (such as men who have sex with other men). Various methods such as network scale-up or the coefficient method have been used recently to try to reach these populations. Network scale-up gets around the sampling difficulties by asking respondents about their network, thus removing much of the embarrassment about answering personal questions. Our method combines both of these methods and others into a single Bayesian hierarchical model which we use to estimate the sizes of these hard-to-reach populations. We demonstrate the effectiveness our model on a collection of datasets regarding various hard-to-reach populations in the Ukraine that are at risk for HIV infection.