Diagnostic Tests and ROC Analysis

This workshop will introduce some public health-related examples where statistical classification (i.e., the development of a prognostic or diagnostic model) could be utilized, and some of the basic terminology (e.g., sensitivity, specificity, etc.) and design concepts associated with these problems. Starting with the simple single biomarker case, we will discuss the different methods for developing and evaluating diagnostic/prognostic models in the receiver operating characteristic (ROC) framework.

Building on the simple case, we will discuss expanding models to incorporate multiple biomarkers using linear combination-based approaches (e.g., logistic regression or Fisher’ linear discriminant function) and more network based approaches (e.g., CART and neural networks). While the primary focus of this workshop is on examples discriminating between two classes, general extensions of these methods will also be introduced.

ACCREDITATION: The University at Buffalo Jacobs School of Medicine and Biomedical Sciences is accredited by the ACCME to provide continuing medical education for physicians.

CERTIFICATION: The University at Buffalo Jacobs School of Medicine and Biomedical Sciences designates this live activity for a maximum of 2.0 AMA PRA Category 1 Credit(s)TM.  Physicians should claim only the credit commensurate with the extent of their participation in the activity.

CREDIT: This program is supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1TR001412.