Estimation of Contextual Exposure to HIV from GPS Data

TL;DR

This paper introduces a statistical framework combining GPS data and HIV prevalence mapping to estimate individual exposure risk based on mobility patterns in rural South Africa.

Contribution

It develops a novel methodology integrating local HIV prevalence estimation with GPS-based activity space analysis for assessing HIV exposure risk.

Findings

Mobility patterns vary systematically with sex and age.

Activity space expansion correlates with increased HIV exposure.

Identifies individuals at higher risk based on their mobility and exposure measures.

Abstract

We present a comprehensive statistical methodological framework for estimating contextual exposure to HIV that includes local (grid-cell level) estimation of HIV prevalence and human activity space estimation based on GPS data. The development of our framework was necessary to analyze HIV surveillance and sociodemographic survey data in conjunction with GPS data collected in rural KwaZulu-Natal, South Africa, to study the mobility patterns of young people. Based on mobility and contextual exposure measures, we examine whether the sex and age of study participants systematically influence the extent and structure of their mobility patterns. We discuss techniques for investigating how the study participants' contextual exposure to HIV changes as their activity spaces expand beyond residential locations, as well as methods for identifying study participants who may be at increased risk of acquiring HIV. KEYWORDS: Contextual HIV exposure; GPS-based mobility analysis; Activity space; HIV prevalence mapping