In Reply.—We very much appreciate Dr Johnson's careful examination of our report. He correctly pointed out that our method for calculating the PP of misclassification gives the PP for all values below (or above) the specified value. However, this does not represent a conceptual error, but rather an intentional and explicitly defined procedure.
The difference between our approaches is one of perspective. The laboratory engaged in testing and attempting to determine appropriate cutoff levels must make decisions about a range of values. Consequently, the laboratory needs to know the probability of misclassification for a range of test results, ie, for all test results at or below (or above) the cutoff level, rather than just the PP at a specific level. While the PP at a given test result (as Dr Johnson describes) may be useful for other purposes (eg, genetic counseling), we believe our approach to be more useful for defining