A crucial point about alcohol research

There are still so many things to say about the junk science that has been created to support the new alcohol guidelines, but a crucial point that is often overlooked was made in the Guardian's letters page last week.

What is not clear about the new guidelines for alcohol consumption is whether the risks are calculated on what people actually drink or what they report that they drink. In 2013, the Institute of Alcohol Studies used market research data and assessed the average consumption of all those above 15 in the UK at 11.7 litres a year or 1,170 units. That’s 22.5 units per week on average. The same publication gives reported rates of consumption that seem to be about half this level. If the damage from alcohol is mapped against reported consumption – and rules are drawn up on this basis – these rules will be overly strict for those who assess their alcohol consumption honestly.

Robert East
Professor emeritus, Kingston University

The answer is that the guidelines and the epidemiology are both ultimately rooted in research that takes people's self-reported consumption at face value. We know that people only report a half to two-thirds of what they actually drink. If the epidemiology shows that 14 units offers peak protection, the real figure is probably more like 21 or 28 units. If the epidemiology shows that 30 units is safe, we can assume it's more like 45 units.

You only ever hear about under-reporting when campaigners are saying 'OMG! We're drinking more than we thought'. This is not really true since we know how much is sold from tax receipts, but under-reporting has a profound impact on what we believe about safe and harmful drinking because epidemiologists do not have their subjects' tax receipts. As the letter above suggests, under-reporting can only lead to an exaggerated perception of risk.