HIV Prevalence in Key Populations: a Semiparametric Bayesian Hierarchical Model for Scarce and Imbalanced Data

• Activity Number: 259 - SPEED: Missing Data and Causal Inference Methods, Part 2
• Type: Contributed
• Date/Time: Monday, July 29, 2019 : 3:05 PM to 3:50 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Amy Zhang* and Le Bao and Michael Daniels
• Companies: Pennsylvania State University and Pennsylvania State University and University of Florida
• Keywords: semiparametric model; Bayesian hierarchical model; HIV; data sparsity

Abstract

HIV/AIDS prevalence has decreased dramatically for most countries since the initial outbreak of the epidemic 30 years ago. However, certain populations, such as female sex workers, their clients, and intravenous drug users are rarely directly included in countries’ response to the HIV/AIDS epidemic, leading to disproportionately high prevalence of HIV/AIDS among these high risk groups. Moreover, data on these groups are often sparse due to their hard-to-reach nature. To improve estimation of HIV prevalence among high risk groups, we propose a semiparametric Bayesian hierarchical model, which will efficiently pool information together across the risk groups and allow for better estimation of HIV prevalence within risk groups with scarce data.