Group Testing Regression Models with Fixed and Random Effects

Abstract

SummaryGroup testing, where subjects are tested in pools rather than individually, has a long history of successful application in infectious disease screening. In this paper, we develop group testing regression models to include covariate effects which are best regarded as random. We present approaches to fit mixed effects models using maximum likelihood, investigate likelihood ratio and score tests for variance components, and evaluate small sample performance using simulation. We illustrate our methods using chlamydia and gonorrhea data collected by the state of Nebraska as part of the Infertility Prevention Project.

A regression model for pooled data in a two-stage survey under informative sampling with application for detecting and estimating the presence of transgenic corn

“…Appropriate grouptesting methodologies for a complex survey can result in substantial savings without significant loss of precision and can be used for estimating the prevalence of a rare attribute, such as transgenic corn or human diseases. For this reason, the aim of the present paper is to bring together the ideas of Chen et al (2009) and Grilli and Pratesi (2004), which means generalizing the group-testing methodology to take into account the weights when a two-stage survey is performed, and we perform an application for detecting and estimating the presence of transgenic corn in Mexico. Researchers use complex sampling schemes (e.g.…”

“…When the proportion of samples with the attribute is less than 10%, group testing is very attractive because it produces significant savings in the number of diagnostic tests required and time expended, and helps to preserve the anonymity of the tested subjects. First used by Dorfman (1943) for detecting soldiers with syphilis during the Second World War, group testing has been used to estimate the prevalence of a wide variety of diseases in humans, animals and plants (Cardoso et al, 1998;Kacena et al, 1998;Verstraeten et al, 1998;Muñoz-Zanzi et al, 2000;Tebbs and Bilder, 2004;Chen et al, 2009). It has also been used for analysing biomarker data (Delaigle and Hall, 2012), detecting drugs (Xie, 2001), solving problems in information theory (Wolf, 1985) and even in science fiction (Bilder, 2009).…”

“…Chen et al (2009) presented group-testing regression models for imperfect diagnostic tests (with sensitivity and specificity of less than 1) with fixed and random effects, which produce the most accurate estimates when the sampling process is in clusters. More recently, McMahan et al (2013) also provided group-testing regression models for mixed effects in the presence of dilution effects.…”

“…Group-testing regression methods are available for fixed and mixed (fixed + random) effects (Farrington, 1992;Vansteelandt et al, 2000;Chen et al, 2009). Chen et al (2009) presented group-testing regression models for imperfect diagnostic tests (with sensitivity and specificity of less than 1) with fixed and random effects, which produce the most accurate estimates when the sampling process is in clusters.…”

Optimal sample sizes for group testing in two-stage sampling

“…Dorfman's method has been extended in many ways. For example, there are group testing regression models for fixed effects, for mixed effects, for multiple-disease group testing data, with imperfect diagnostic tests [with sensitivity ðS e Þ, specificity ðS p Þ , 1, or with dilution effect], and non-parametric group testing methods, among others (Yamamura and Hino, 2007;Hernández-Suárez et al, 2008;Chen et al, 2009;Zhang et al, 2013).…”