Screening for prostate cancer with prostate-specific antigen (PSA), a simple blood test, is complex, controversial, intellectually challenging and ethically concerning. Sweden has contributed actively to the knowledge base for PSA screening in the last couple of years and a more informed debate is now possible. I will in this article summarize what we currently know about PSA testing. I discuss the closely related issues of the natural history of early-stage disease and the risks and benefits of radical local treatment. I propose that the mortality reduction following PSA screening is probably modest, whilst substantial harms are well documented. Furthermore, there is growing evidence that the PSA test is profoundly limited for screening purposes. I therefore concur with the growing number of health and professional agencies - notably outside the US - that advise against population screening with PSA. Through PSA screening the medical community has generated a pseudo-epidemic of over-diagnosed non-lethal prostate cancer. Molecular tools to distinguish innocent, over-diagnosed prostate cancer from lethal tumors that deserving curative treatment are necessary to improve screening test performance. To date, extensive attempts to identify molecular predictors of outcome have remained unsuccessful, and no ideal screening test is within sight.