Cancer surveillance series: Interpreting trends in prostate cancer - Part I: Evidence of the effects of screening in recent prostate cancer incidence, mortality, and survival rates

Background: The prostate-specific antigen test was approved by the U.S. Foo d and Drug Administration in 1986 to monitor the disease status in patients with prostate cancer and, in 1994, to aid in prostate cancer detection. Ho wever, after 1986, the test was performed on many men who had not been prev iously diagnosed with prostate cancer, apparently resulting in the diagnosi s of a substantial number of early tumors. Our purpose is to provide insigh t into the effect of screening on prostate cancer rates, Detailed data are presented for whites because the size of the population allows for calculat ing statistically reliable rates; however, similar overall trends are seen for African-Americans and other races. Methods: Prostate cancer incidence d ata from the National Cancer Institute's Surveillance, Epidemiology, and En d Results Program and mortality data from the National Center for Health St atistics were analyzed, Results/Conclusions: The following findings are con sistent with a screening effect: 1) the recent decrease since 1991 in the i ncidence of distant stage disease, after not having been perturbed by scree ning; 2) the decline in the incidence of earlier stage disease beginning th e following year (i.e,, 1992); 3) the recent increases and decreases in pro state cancer incidence and mortality by age that appear to indicate a calen dar period effect; and 4) trends in the incidence of distant stage disease by tumor grade and trends in the survival of patients,vith distant stage di sease by calendar year that provide suggestive evidence of the tendency of screening to detect slower growing tumors, Implications: The decline in the incidence of distant stage disease holds the promise that testing for pros tate-specific antigen may lead to a sustained decline in prostate cancer mo rtality. However, population data are complex, and it is difficult to confi dently attribute relatively small changes in mortality to any one cause.