In a provocative editorial in the New England Journal of Medicine by Gilbert Welch and colleagues from Harvard, the authors claim that the reduction in prostate cancer mortality seen in the US in the last twenty years is more apparent than real. Dr. Welch is a famous naysayer; he stands virtually alone in discounting the value of screening mammography against all conventional and evidence-driven wisdom. This time it’s prostate cancer. He coins a term unfamiliar to this author: sticky diagnosis bias. By that he means that since the incidence of prostate cancer went up initially with PSA screening, more men were recognized to have the disease, and more deaths were attributed to it — correctly or not. The diagnosis stuck to the men and created bias. Hence any reduction in mortality would have to be judged in the context of an inflated death rate to begin with.
The bar graph shows what should be a real phenomenon: a lower death rate with PSA screening. In fact Dr. Welch attributes 
the entire death reduction to better use of palliative hormonal therapy. One may or may not argue that point, but one cannot dispute that we are overdiagnosing prostate cancer — finding perhaps a hundred cases to save a life. In the meantime a substantial percentage of those hundred men get overtreated. Read the entire article to see where this goes. The article is here but may not be available for long; Dr. Stark can email you a PDF version if you request it in the form at the right. The twists and turns in the article are worthy of a mystery novel. Dr. Welch strikes again.