This study focused on factors that impact on the clinical efficacy of diagnostic laboratory tests. A new graphical method (cost-error plot) was developed for incorporating prevalence and the costs of misclassification on test performance, two important factors that affect the practicality of diagnostic tests. Unlike receiver operating characteristic plots, the clinical efficacy of laboratory tests can be readily compared by cost- error plots over a range of possible prevalence and misclassification costs. Future work will be done developing software to facilitate production of cost-error plots.
