The diagnosis of prostate cancer has undergone an unprecedented recent increase, while mortality has increased much more slowly. We examined new prostate cancer diagnoses from 1987-1992 in a nationwide prospective cohort study of 51 529 men enrolled in the Health Professionals Follow-up Study, a population likely to be medically sophisticated and thus early to adopt medical innovations. The age-adjusted incidence of prostate cancer rose approximately 2(1/2) fold during the study period. Nearly all of the increase occurred among organ-confined tumors, with a smaller increase for regionally-advanced tumors and none for metastatic tumors. Using a Poisson regression model of newly-diagnosed cancers, we found organ-confined cancers rose abruptly by 86% (95% CI: 36-256%; P=0.0001) and regionally-advanced tumors by 73% (95% CI:12-267%; P=0.01) after March, 1991, when a study advocating screening using the prostate specific antigen (PSA) was published in the New England Journal of Medicine. The recent increase in the incidence of prostate cancer is probably due to the increased PSA screening of asymptomatic men, resulting in the diagnosis of large numbers of men with early-stage disease from 1990 onward. Despite ongoing debate over the value of PSA screening, the rate of diagnosis accelerated sharply after the publication of a well-publicized but inconclusive study advocating screening. In a context of growing disease awareness, well-publicized research reports may result in unexpectedly amplified acceptance into medical practice.