The two preceding letters reflect the controversy over how aggressive clinicians should be in recommending for or against early detection efforts for prostate cancer, particularly with PSA determinations.

Dr. Ashley and Mr. Knuth argue that the predicted benefits of early detection of prostate cancer would be greater at any given age if healthier men with longer life expectancies were selected for screening. We agree. However, we do not agree that clinicians screen only healthy men for prostate cancer (particularly when "healthy" is defined as having no significant illnesses, normal results on basic laboratory tests, and not smoking), and no evidence is cited to support this assertion. In a 1995 nationwide survey of primary care physicians, for example, we found that 55%, 66%, 65%, and 58% of physicians, "almost always" perform routine PSA tests on men in their 50s, 60s, early 70s, and late 70s, respectively, despite the obvious fact that many of these men are being seen for medical problems [1]. Therefore, we believe that the life expectancies we used in our model, which represent an "average" burden of comorbidity, are appropriate for making policy decisions about screening. We agree, however, that clinicians should consider each man's age and health status when advising them individually about screening.

Dr. Winters would prefer an even stronger injunction against PSA testing for men in their 70s, citing the "slippery slope" toward aggressive intervention that may await men who are found to have prostate cancer at screening. Although we generally agree that our estimates of risks and benefits would favor a nonscreening decision for most men in their 70s, we cannot be sure that a healthy man in his early 70s who decides to accept the risks of screening and is fully cognizant of this slippery slope is being unreasonable in doing so. Only time, and the results of clinical trials, will tell. We do agree that a discussion of the downstream consequences of PSA testing, the known risks, and the uncertain benefits of aggressive treatment should be conducted with men in this age group before a test is ordered.