Group Testing for Multiple Infections with Application to the Infertility Prevention Project

• Activity Number: 63
• Type: Topic Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: Biometrics Section
• Author(s): Joshua Tebbs*+ and Christopher S. McMahan and Christopher R. Bilder
• Companies: University of South Carolina and Clemson University and University of Nebraska-Lincoln
• Keywords: Correlated binary data ; Dorfman retesting ; Efficiency ; EM algorithm ; Latent response ; Pooled testing

Abstract

This paper considers group testing for multiple infections, motivated by chlamydia and gonorrhea testing for the Infertility Prevention Project (IPP). For specificity, we consider a pooling algorithm currently used by the University of Iowa Hygienic Laboratory to perform testing for the IPP. We derive the operating characteristics of the algorithm for classification (i.e., case identification), and we also propose a new way to estimate population-level probabilities of infection with group testing data. This is the first paper to consider both classification and estimation for group testing with multiple infections. We use simulated data and real IPP data to illustrate our methods.