A1441
Title: Tests for comparing ROC curves under the presence of covariates
Authors: Juan-Carlos Pardo-Fernandez - Universidade de Vigo (Spain) [presenting]
Aris Fanjul Hevia - Universidad de Oviedo (Spain)
Wenceslao Gonzalez-Manteiga - University of Santiago de Compostela (Spain)
Abstract: The receiver operating characteristic (ROC) curve is a graphical tool routinely used to evaluate the performance of a binary classification procedure based on a continuous marker. In many practical applications, covariates related to the marker are available. Under these circumstances, it is of interest to evaluate the influence that those covariates might have on the performance of the marker in terms of classification ability by means of the covariate-specific ROC curve, which is defined in terms of conditional distributions. Several tests to compare covariate-specific ROC curves are discussed, including the cases with univariate and multivariate covariates. In practice, these tests would allow to decide if, for a given value of the covariate, the classification capabilities of several markers differ. The proposed methodologies rely on nonparametric estimation of the involved ROC curves and bootstrap resampling plans to approximate the null distribution of the test statistics. The proposed procedures are used to analyze a real data set of patients with pleural effusion.
