Group Testing in Heterogeneous Populations Using Halving Algorithms

• Activity Number: 282
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): Michael S. Black*+ and Christopher R. Bilder+ and Joshua Tebbs
• Companies: University of Nebraska-Lincoln and University of Nebraska-Lincoln and University of South Carolina
• Keywords: binary response ; group testing ; latent response ; order statistics ; pooled testing ; retesting

Abstract

When a large number of individuals are screened for an infectious disease or other binary characteristic, group testing (also known as pooled testing) is often used to reduce the total number of tests. Traditionally, group testing has assumed each individual is independent with the same risk of positivity. More recently, there is a growing set of literature generalizing the previous work to heterogeneous populations so that each individual has a different risk of positivity. In our presentation, we will examine how this generalization affects the frequently used procedure of halving positive groups. We will show that the heterogeneity does not affect the mean number of tests when individuals are placed randomly into sub-groups. However, when individuals are placed into sub-groups based on their risk probabilities, we show the mean number of tests can be reduced for particular situations.