Abstract:
|
Correlation when comparing diagnostic test performance is a focus of current research. This work develops the required theory to account for correlation between continuous biomarkers on confidence intervals (CIs) comparing the optimal performance of two diagnostic tests with multiple outcomes (2 or more). We define the optimal point using Bayes Cost, a metric that sums the weighted classifications within a diagnostic test using a cost/benefit structure. Modification to the delta and generalized methods are accomplished and, through simulation, compared to results using a paired bootstrap with respect to CI coverage and width under varying diagnostic test accuracy, sample size, cost/benefit structures, correlation levels, and possible misspecification. We provide updated formulas and methodologies to include the effects of correlation for CI calculations to compare optimal performance between two diagnostic tests using Bayes Cost and demonstrate our methods through a biomedical data application.
|