Semiparametric Inference of the Youden Index and the Optimal Cutoff Point under Density Ratio Models

摘要: The Youden index is a popular summary statistic for receiver operating characteristic curve. It gives the optimal cutoff point of biomarker to distinguish diseased and healthy individuals. In this paper, we propose model distributions individuals in groups via semiparametric density ratio model. Based on model, use maximum empirical likelihood method estimate point. We further establish asymptotic normality proposed estimators construct valid confidence intervals corresponding automatically covers both cases when there no lower limit detection (LLOD) fixed finite LLOD biomarker. Extensive simulation studies real data example are used illustrate effectiveness its advantages over existing methods.