It is a truism of the festive season that, as the party lights come out, so to do the Temperance Brigade. Last week we had two pronouncements: one from Sir Liar, advising no booze for under fifteen year olds, and another from Alcohol Concern, telling us we grossly underestimate our consumption when answering drink surveys. What Alcohol Concern didn’t say – probably because they hadn’t realised it – is that the study data behind their pronouncements show that the current safe drinking limits of 21 units/week for men and 14 for women – already known to be arbitrary – are also misleadingly low.
How can this be? To answer that question, we need to look at what the Alcohol Concern study showed, how the risk of alcohol related harm is assessed, and at how that assessment is translated into health policy.
The study was remarkably simple. Instead of relying on us to say how much we drink, the researchers approached the question from the other end – how much alcohol is sold, and then compared the average amount of alcohol sold per boozer with how much the average boozer said they drank. Rather unsurprisingly, there was a gap between the amount of alcohol leaving the shops, and how much we said we drank.
Nor was the gap trivial. In surveys, average alcohol consumption is 16 units per week, whereas the amount of alcohol sold suggests the average tippler sinks 26 units per week – equivalent to one bottle of wine per week extra for the average drinker. The real difference is likely to be even greater – the study only looked at alcohol sold here, and so excluded that bought abroad, not to mention Aunt Matilda’s rhubarb wine…
Alcohol nannies naturally jumped on the figures. One warned of ‘gluttonous attitudes’ leading to ‘devastating effects’, while another cautioned ‘If we underestimate our drinking levels, then we’re underestimating the amount of harm we can expect to happen to our families, communities and wider society’ before adding archly ‘Poor survey intelligence can result in misinformed policy’.
How true – but not in the way our nanny meant. Because of the way we do population level studies into the relationship between alcohol consumption and alcohol related harm, we are markedly over-estimating harm – with the result that the limits are misleadingly low.
The anomaly arises because in these studies we relate not what people actually drink, but what people report they drink, to the various harmful outcomes that we are interested in. But, we now know, each group in our study is, on average, drinking a bottle of wine a week more than they say the drink. The effect is that the so-called J-shaped curve gets stretched to the right – you have to drink more to get to the same level of risk. The chap who says he is drinking 21 units per week is more likely actually drinking 31 units a week, and so the risk threshold that we thought attached to 21 units is really attached to 31 units.
And yet ‘policy’ tells us that that level of risk is there at 21 units a week. It is indeed misinformed policy – it is misinformed by an assessment that exaggerates the real risk. But then, ‘policy’ has never had much to do with science, and everything to do with social control.
That is not for a moment to under-estimate the harm that alcohol can do. It is instead to recognise what sensible drinkers have long known: that, for the vast majority of drinkers, that extra glass isn’t going to kill them.