Related Experiment Video
Updated: Jul 7, 2026
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
Nonparametric sequential evaluation of diagnostic biomarkers.
Aiyi Liu1, Chengqing Wu, Enrique F Schisterman
1Biometry and Mathematical Statistics Branch, National Institute of Child Health and Human Development, Department of Health and Human Services, 6100 Executive Blvd., Rockville, MD 20852, U.S.A. Liua@mail.nih.gov
Statistics in Medicine
|February 5, 2008
Summary
This study introduces new statistical methods for evaluating biomarker diagnostic accuracy using nonparametric group sequential testing. These methods enable robust comparisons of biomarker performance, even with correlated data from repeated measurements.
Area of Science:
- Biostatistics
- Medical Diagnostics
- Statistical Methods
Background:
- Evaluating diagnostic accuracy of biomarkers is crucial in clinical research.
- Continuous test outcomes, especially from repeated measurements, present statistical challenges due to potential correlations.
- Existing methods may not adequately handle correlated continuous biomarker data for comparative accuracy assessments.
Purpose of the Study:
- To develop and present nonparametric group sequential testing procedures for evaluating and comparing the diagnostic accuracy of biomarkers.
- To address scenarios involving continuous test outcomes, including independent or paired (correlated) data.
- To enable the application of Brownian motion-based design methods for biomarker evaluation.
Main Methods:
- Development of nonparametric group sequential testing procedures.
- Utilizing a two-dimensional statistic proposed by Whitehead (1999) to facilitate analysis.
- Application of Brownian motion-based design methods for statistical inference.
- Handling of both independent and paired (correlated) continuous test outcomes.
Main Results:
- The proposed procedures provide a framework for robust evaluation and comparison of biomarker diagnostic accuracy.
- The methods are applicable to continuous biomarker data, accommodating correlations from repeated measures.
- The reliance on Whitehead's statistic allows for the use of established Brownian motion-based design principles.
Conclusions:
- The developed nonparametric group sequential testing procedures offer a valuable tool for assessing and comparing biomarker accuracy.
- These methods are particularly useful when dealing with correlated continuous biomarker outcomes.
- The study facilitates more rigorous and efficient clinical evaluations of diagnostic biomarkers.
