Any health benefits provided by low alcohol consumption are likely to be slight and limited to older women, says a new study published in the British Medical Journal. Previous work touting the beneficial effects of drinking has been skewed by poor sampling techniques, says Craig Knott, lead author of the new research, most notably the habit of grouping together individuals who have never drunk alcohol with those who have stopped drinking.

Knott and his colleagues came to their conclusion after looking at ten years of health data covering more than 50,000 individuals older than 50 from England. The team found that comparing drinkers with individuals who had never drunk before showed that low alcohol consumption provided only slight health benefits for two groups of light drinkers: men aged 50 to 64 and women aged 64 and over. The alcohol intake for these groups was small — less than a small glass of wine a day for women or less than a large glass a day for men — and the researchers stressed that even these results might still be the result of selection biases. These could include surveying individuals who are in professional employment and easier to contact, but are also healthier overall than the general population.

These sorts of biases are exactly what have led to a flurry of articles in recent years all touting the supposed health benefits of drinking — these studies are catnip to reporters because they’re counterintuitive and easily shared. Arguing over the slight benefits of small amounts of alcohol may seem trivial (especially as the damage caused by serious drinking is not in doubt), but Knott says this isn’t the case. Studies painting drinking in even a slightly positive light trickle into the public conscious, giving people an excuse, no matter how silly, for grabbing another drink. And the studies also affect legislation. An editorial in the BMJ accompanying the new research points out that "politicians [have] used evidence on possible benefits to justify the failure to act on reducing harms" while the alcohol industry "assiduously promoted this evidence, both to reassure customers and to argue against ... policies."

However, it’s biases in the original studies that are creating the misinformation in the first place. Knott says that he and his team looked at a meta-analysis of 34 studies and found that studies showing the positive benefits of alcohol often made the mistake of straightforwardly comparing drinkers and non-drinkers. "Unfortunately, non-drinkers are far from a homogeneous group," says Knott via email, with the category including "both former drinkers and never drinkers."

Because former drinkers have increased rates of depression, report worse health on average, and have increased health risks in general, lumping them in with lifelong teetotalers skews the results significantly, says Knott. What’s more, of the studies he and his team examined, only two broke their information down into age groups that included over-64s — a glaring omission, as older individuals are less able to metabolise alcohol and are also more likely to have chronic illnesses. Not separating this group out from younger individuals could skew the results — one of the reason that the new study focused on older age groups.

"Given the harms attributed to alcohol use, it is not surprising that reports showing possible mortality benefits for low level users attracted enthusiasm among consumers, the media, and the alcohol industry," says the editorial, adding that "in health as elsewhere, if something looks too good to be true, it should be treated with great caution."