Obstetrics occupies a place of honour in the history of evidence-based medicine, providing some of the earliest examples of clinical audit and practical epidemiology. Unfortunately, one of the first—a study of puerperal fever in Aberdeen by Alexander Gordon (1795)—showed that he and his colleagues were killing their patients by spreading the disease; a fact not widely accepted for another 60 years. His colleagues retaliated by hounding him out of practice (see Wootton R. Bad medicine. Oxford: Oxford University Press, 2006:ch 12): they accused him of excessive bloodletting, which may well have been true. However, even now we don’t know how best to treat iron-deficiency anaemia in pregnancy, according to a Cochrane Review (CD003094). It found 17 trials, mostly small, heterogeneous, and of poor quality, providing insufficient evidence on important basic questions like the harm:benefit ratio for parenteral iron. For several other gestational woes, evidence also tends to be anecdotal and patchy, though an Egyptian group has recently conducted a small randomised controlled trial (RCT) on 40 women with hyperemesis gravidarum bad …