A0354
Title: Semiparametric estimator for the covariate-specific ROC curve
Authors: Pablo Martinez-Camblor - Geisel School of Medicine at Darmouth (United States) [presenting]
Juan-Carlos Pardo-Fernandez - Universidade de Vigo (Spain)
Abstract: The study of the predictive ability of a marker is mainly based on accuracy measures sures provided by the so-called confusion matrix. Besides, the area under the ROC curve, AUC, has become a popular index for summarizing the overall accuracy of a marker. However, the nature of the relationship between the marker and the outcome and the role that potential confounders play in this relationship could be fundamental in extrapolating the observed results. Directed acyclic graphs (DAGs), commonly used in epidemiology and causality, could provide good feedback for learning the possibilities and limits of this extrapolation applied to the binary classification problem. Both the covariate-specific and the covariate-adjusted ROC curves are valuable tools that can help to better understand the real classification abilities of a marker. Since they are strongly related to the conditional distributions of the marker on the positive (subjects with the studied characteristic) and negative (subjects without the studied characteristic) populations, the use of proportional hazard regression models arises in a very natural way. The use of flexible proportional hazard Cox regression models is explored for estimating the covariate-specific and the covariate-adjusted ROC curves. Their large- and finite-sample properties are studied, and the proposed estimators are applied to a real-world problem.
