[Skip to Content]
[Skip to Content Landing]
August 23, 1995

The Utility of Prostate-Specific Antigen for Detecting Prostate Cancer-Reply

Author Affiliations

Northwestern University Medical School Chicago, Ill
Brigham and Women's Hospital Boston, Mass

JAMA. 1995;274(8):608-609. doi:10.1001/jama.1995.03530080023024

In Reply.  —Drs Brunswick, Marshall, and Meyerhoff calculate predictive values from our data. We agree that predictive values are important, but chose not to present them in our article because their computation and interpretation, particularly given our study design, are not straightforward. Since predictive values are dependent on disease prevalence and factors affecting the occurrence of false-positive results, they can vary substantially in different study populations.1 Moreover, in a nested case-control study, the artificial prevalence represented by the cases and controls alone (a prevalence determined by the case-control ratio) is obviously a problem. If the controls are an unbiased sample of the total nondiseased members of the cohort, as Brunswick, Marshall, and Meyerhoff assume, one can estimate predictive value. However, we used an "incidence density" approach for sampling controls (ie, the controls are randomly selected from similar-aged men at risk when each case is diagnosed), so our controls are

First Page Preview View Large
First page PDF preview
First page PDF preview