Even before the virtual demise of hormone replacement therapy, soy was promoted as a “natural” alternative on the basis that Japanese women eat a lot of it and allegedly don’t get menopausal symptoms. I’m not sure whether it is also alleged that Japanese women have less osteoporosis, but this Italian trial looked at the effect of a pure soy derivative, genistein, on bone mineral density in postmenopausal women. Apparently genistein likes the oestrogen receptor-β which is abundant in bone, while it is relatively uninterested in the oestrogen receptor-α, present in the reproductive tract. This is borne out in this double-blind trial: it increased bone density without increasing endometrial thickness. Whether it may decrease fractures, we do not yet know: we need a longer trial with hard end-points, not just chalky ones.

Almost every day, we explain the benefit of some risk-reducing treatment to patients, so naturally we are experts on the subject. All right then, tell me the difference between an odds ratio and a hazard ratio. While you are thinking about that, I’ll tell you about a study of how people respond to different ways of expressing the benefit of treatment. If you put it in terms of immediate benefit, everybody wants it; if you put it in terms of a number-needed-to-treat (NNT) a lot of people want it; and if you put it in terms of postponement of a condition, people waver. But these were healthy volunteers, not typical patients. Now, about that hazard ratio?

It’s nice when a big systematic review confirms something you felt must be true: that warfarin is much more protective than aspirin against stroke in patients with non-valvular atrial fibrillation. It’s even better when a large prospective trial proves the same thing in the actual population you are treating (elderly people in the middle of England) – and this will be confirmed when The Lancet carries the results of the BAFTA trial in a week or two.

Every two months I get sent between 30 and 70 papers to look through for a column called Evidently… which I write for Evidence Based Medicine. For some reason, an awful lot of them seem to be about acupuncture. The message of all the studies is quite clear: putting needles into people at random with an air of confidence makes many of them feel better, whether or not you put them in the places and manner dictated by the Ancient Wisdom of China. This meta-analysis shows that for osteoarthritis of the knee, sham acupuncture is probably the cheapest, safest and most effective intervention yet discovered. It requires no skill, just a few needles and the plausible manner that many of us have sought to acquire over the years.