Group testing regression models with dilution submodels

摘要: Group testing, where specimens are tested initially in pools, is widely used to screen individuals for sexually transmitted diseases. However, a common problem encountered practice that group testing can increase the number of false negative test results. This occurs primarily when positive individual within pool diluted by ones, resulting pools negatively. If goal estimate population-level regression model relating disease status observed covariates, severe bias result if an adjustment dilution not made. Recognizing this as critical issue, recent binary approaches have utilized continuous biomarker information acknowledge effect dilution. In paper, we same overall but take different approach. We augment existing models (that assume no dilution) with parametric submodel pool-level sensitivity and all parameters using maximum likelihood. An advantage our approach it does rely on external data, which may be available surveillance studies. Furthermore, unlike previous approaches, framework allows one formally whether present based data. use simulation illustrate performance estimation inference methods, apply these methods 2 infectious data sets.