Problematic Prostatic Prediction

To the Editor.— Chodak and Schoenberg1address the important issue of efficacy of prostate cancer screening by digital rectal examination but have mislabeled an important parameter of that test. The authors repeatedly refer to a test "specificity of 29%" erroneously calculated as (true test positives/true test positives+false test positives) and compared it with that reported by others. The value they calculated was really the positive predictive value. The specificity of a test should represent the percentage of negative tests among patients free of disease (true test negatives/true test negatives+false test positives). Unless an appropriate gold standard is available to distinguish those individuals who are truly test negative from those with early undetectable disease (false test negatives), the specificity of a screening test must assume a low disease prevalence with a few false-negatives "contaminating" both the numerator and denominator. The high prevalence in older men of microscopic prostate cancers (almost