On the optimum selection of diagnostic tests in computer-aided differential diagnosis

A new method with multivariate individual selection of diagnostic tests for computer-aided differential diagnosis is presented. It has been developed on the basis of new results of our own in the field of mathematical decision theory. Its application to the differential diagnosis of four congenital heart diseases by means of ECG features proves our method to be superior to the usual multivariate statistical analysis. Establishing the individual optimum sequence of diagnostic tests for each patient our diagnostic procedure needs only two-thirds of the diagnostic tests used by the methods until now without an increase of the error rate. The error rate can be further reduced by additional introduction of the so-called reserved diagnostic statement. The managing of our procedure may be compared with the diagnostic behavior of a physician.