Evaluation and computation of diagnostic tests: a simple alternative

Methods of evaluating and comparing the performance of diagnostic tests are of increasing importance in medical science. When a test is based on an observed variable that lies on a continuous scale, an assessment of the overall value of the test can be made through the use of a Receiver Operating Characteristic (ROC) curve. The ROC curve descibes the discrimination ability of a diagnosis test for the diseased subjects from the non-diseased subjects. The area under the ROC curve (AUC) represents the probability that a randomly chosen diseased subject will have higher probability of having disease than a randomly chosen non-diseased subject. For comparing two diagnostic systems, the difference between AUCs is often used. In this paper we have investigated various methods of the comparison of equality of two AUCs and proposed a McNemar test for the comparison of two diagnostic test procedures. The proposed test is based on an optimal cut-off point that discriminates the individuals in actually positive or actually negative cases for which we have a 2£2 contingency table where we can apply the McNemar test. The operating characteristics of the proposed test are evaluated using extensive simulation over a wide range of parameters.