To the Editor.
—Dr Johansson and colleagues1 present data from Örebro County in Sweden with a very favorable 15-year outcome for untreated prostate cancer which, if they are correct, would have tremendous effects on the treatment of localized prostate cancer. However, the study suffers from a bias of unknown nature resulting in a survival twice as high as expected.Because of good access to health care in Sweden, almost all patients who later die from prostate cancer are diagnosed during their lifetime; ie, they emerge with clinical disease in different stages before death. A high diagnostic activity can be expected to diagnose more patients in early clinical stages while a low diagnostic activity leads to diagnosis of more advanced tumors. When curative treatment is not practiced, the absolute number of patients who die from prostate cancer within 1 geographic area is largely independent of the diagnostic activity.From official