On statistical comparison of two diagnostic tests

A one-sample hypothesis testing method is presented for comparing the accuracy of two diagnostic tests. Costs and inconveniences related to the false positive and false negative diagnostic test results, to the extent that they are quantifiable, are considered as measures of the importance of avoiding misclassification. The use of the proposed method is illustrated by comparing the effectiveness of aspiration biopsy cytology and frozen section in detecting malignant thyroid disease.