Title: ROC surface analysis in presence of verification bias
Authors: Duc Khanh To - University of Padova (Italy) [presenting]
Abstract: Receiver operating characteristic (ROC) surface and volume under ROC surface (VUS) are frequently used for assessing accuracy of a diagnostic test when disease status has three categories (e.g., non-diseased, intermediate and diseased). A problem encountered in studies that evaluate a new diagnostic test is that often not all patients undergo disease verification, due to the expensiveness and invasiveness of the gold standard test. Using only the verified patients to estimate the ROC surface and its volume would result in verification bias. We present several verification bias-corrected estimators of the ROC surface and the VUS, based on imputation and re-weighting techniques. These estimators work when the true disease status is missing at random . In addition, we present an R package and the corresponding Shiny web application, which support researchers in carrying out the ROC surface analysis in presence of verification bias.