CMStatistics 2022: Start Registration
View Submission - CMStatistics
B0342
Title: Covariate adjustment methods for the evaluation of biomarkers in multi-class setting Authors:  Duc Khanh To - University of Padova (Italy) [presenting]
Gianfranco Adimari - University of Padua (Italy)
Monica Chiogna - University of Bologna (Italy)
Abstract: The statistical evaluation of a biomarker plays an important role in medical research, but the evaluating process is often done marginally, i.e., by using the biomarkers values only. In some cases, however, there are covariates, for instance, age, gender, and smoking status, that can influence the biomarker behavior, and, therefore, also impact its accuracy. Thus, in practice, the evaluation of such possible effects is needed. Recently, various methods have been developed to address possible covariates' effects on the evaluation of biomarkers. Most proposed methods focus on a two-class setting, whereas a multi-class setting is very scarcely considered in the statistical literature. We will introduce our new proposed methods to evaluate the accuracy of biomarkers in the presence of covariates, and will provide some discussion on the future development of this topic.