Regression models for group testing data with pool dilution effects

摘要: Group testing is widely used to reduce the cost of screening individuals for infectious diseases. There an extensive literature on group testing, most which traditionally has focused estimating probability infection in a homogeneous population. More recently, this research area shifted towards individual-specific probabilities regression context. However, existing approaches have assumed that sensitivity and specificity pooled biospecimens are constant do not depend pool sizes. For those applications, where assumption may be realistic, these can lead inaccurate inference, especially when sizes large. Our new approach, exploits information readily available from underlying continuous biomarker distributions, provides reliable inference settings pooling would beneficial does so even larger We illustrate our methodology using hepatitis B data study involving Irish prisoners.