What question did this study set out to answer?

September 1, 1988Biometrics

Comparing the Areas under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach

View Full Paper

Ask AI

Bookmark

Authors

EDElizabeth R. DeLongGeneral Cardiology DDDavid M. DeLongUniversity of North Carolina at Chapel Hill DCDaniel L. Clarke‐PearsonUniversity of North Carolina at Chapel Hill

Discussion

Member takes

Implication

Statistical methodology study demonstrates a nonparametric framework for comparing correlated ROC curves using generalized U-statistics, highlighting improved evaluation of paired diagnostic tests.

Key Points

Develop a robust nonparametric method to evaluate and compare the areas under two or more correlated receiver operating characteristic (ROC) curves derived from diagnostic tests performed on identical subjects.
Formulated a nonparametric statistical framework utilizing the theory of generalized U-statistics.
Constructed an estimated covariance matrix to account for correlation inherent in paired diagnostic data across continuous or graded scales.
Established a distribution-free approach to test differences between correlated empirical ROC areas without requiring specific parametric distribution assumptions.
Demonstrated that large-sample asymptotic properties of generalized U-statistics enable direct hypothesis testing and confidence interval generation for correlated diagnostic endpoints.

Cite This Study

DeLong et al. (1988) studied this question.

synapsesocial.com/papers/6985d5ffd5ebcb2b6d595491 https://doi.org/10.2307/2531595

View Full Paper

Ask AI

Bookmark

Implication

Statistical methodology study demonstrates a nonparametric framework for comparing correlated ROC curves using generalized U-statistics, highlighting improved evaluation of paired diagnostic tests.

Comparing the Areas under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach

Discussion

Member takes

Implication

Key Points

Cite This Study

Comparing the Areas under Two or More Correlated Receiver Operating Characteristic Curves: A Nonparametric Approach

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider