Pooled testing for HIV prevalence estimation: exploiting the dilution effect

Abstract

We study pooled (or group) testing as a method for estimating the prevalence of HIV; rather than testing each sample individually, this method combines various samples into a pool and then tests the pool. Existing pooled testing procedures estimate the prevalence using dichotomous test outcomes. However, HIV test outcomes are inherently continuous, and their dichotomization may eliminate useful information. To overcome this problem, we develop a parametric procedure that utilizes the continuous outcomes. This procedure employs a hierarchical pooling model and estimates the prevalence using the likelihood equation. The likelihood equation is solved using an iterative algorithm, and a simulation study shows that our procedure yields very accurate estimates at a fraction of the cost of existing procedures.