Like most primary care doctors, I have spent a working lifetime prescribing antibiotics for respiratory tract infections, knowing that in most cases they make no difference at all. Until now, a plausible excuse for such behaviour has been that we cannot reliably distinguish between viral and bacterial infection, but recently, procalcitonin has been suggested as a blood marker that does just that. But is it reliable, and will it really change doctor behaviour? A well-conducted trial from Switzerland (Arch Int Med 2008;168:2000–7) attempts to answer these questions in a real-life setting: 53 primary care physicians recruited 458 adult patients with respiratory tract infections that they thought might need an antibiotic. Patients were centrally randomised to be treated as usual or to have their therapy guided by blood procalcitonin measurements, with a turn-around time of 2–4 hours. The physicians who entered the study were already low antibiotic prescribers, but their antibiotic use dropped almost 4-fold when guided by procalcitonin, while patient outcomes were identical. I hope that …