Skip Internal Navigation
EB Research - Methodological Research in Epidemiology
Receiver Operating Characteristic (ROC) Curves
The ROC curve is a useful and widely used analytical tool for assessing a biomarker's usefulness in distinguishing individuals with and without disease. An extensive literature exists for its estimation and application. Researchers here have made significant contributions to this literature in the following areas:
- Estimating the ROC curve and summary measures accounting for various sources of measurement error
- Using maximum likelihood under various distributional assumptions in both univariate and multivariate settings
- Estimating the ROC curve and summary measures from measurements based on a pooling and hybrid design
- Assessing efficiency and robustness of various techniques
This continued research will allow development and analysis of biomarkers under a wide variety of real world limitations and complications.
Enrique F. Schisterman, Ph.D. & Neil Perkins, Ph.D.
- Perkins NJ,Schisterman EF, Vexler A. (2013). Multivariate Normally Distributed Biomarkers Subject to Limits of Detection and Receiver Operating Characteristic Curve Inference. Academic Radiology, 20(7):838–846. PMID:23747152