Stats are the real cause for concern
The figures used by Alcohol Concern to substantiate its claims are so out of whack it is impossible to have an informed debate, says Phil Mellows.
Drinking these days sometimes seems to be no more than a branch of mathematics. From the elaborate calculations required to work out whether you’ve exceeded your recommended units to population-wide alcohol-related mortality rates, the social and emotional pleasures and dangers of having a drink are drowned by numbers.
That’s because numbers are essential to the public health approach to alcohol. Statistics constitute the evidence base from which “right” policies are supposed to flow. They promise certainties, a firm grounding for your theories and a way of measuring, and modelling, results.
Yet that promise is frequently broken. What statistics actually mean can change according to the angle from which you look at them, how you interpret and use them. The temptation is always there to turn them to the service of assumption and dogma, to create policy-based evidence rather than evidence-based policy. They are not neutral but political tools.
Journalists, too, love numbers. There’s nothing like a statistic from scientific research to hang a story on, and this is what the public health lobby has understood, and exploited, so well in recent years.
Take Alcohol Concern. For nearly 30 years it has been the go-to organisation for comment on anything to do with drinking issues. It describes itself on the website of Eurocare (membership of which it shares, inciden- tally, with prohibitionist organisations) as “the principal information source on alcohol harm”.
It’s probably right. Which is worrying. Because on so many other things, Alcohol Concern is just plain wrong. And by that I don’t mean by that I disagree with its opinions. We can debate those. The problem is that we can’t a have a proper debate when the statistics it uses to support its arguments are so wildly misused.
The most recent, and perhaps most breathtaking, example came in the run-up to this year’s showpiece campaign, Dry January. The Alcohol Concern press release included shock figures designed to catch the eye of the busy journalist. And they’re completely wrong. Let’s have a look at them.
More than four and a half million working days in the UK could be lost to hangovers this January.
This number comes from a six-year-old survey by Prudential Insurance, which asked employees whether they ever went into work with a hangover. That’s right, these people have gone to work. They are not lost working days.
Ignoring this fact, Alcohol Concern goes on to state that:
Around 200,000 thousand (sic) people go to work with a hangover every day costing the economy around £6.4 billion each year.
That £6.4 billion is so far out it’s a bit of a giveaway. It’s nearly a third of the annual £21 billion that total alcohol harm is supposed to cost the UK (another dodgy stat which we’ll deal with later). But where do they get the £6.4 billion, just for hangovers?
It comes from the eight-year-old Labour government Alcohol Harm Reduction Strategy document, which calculates it from the 17 million working days lost each year as a result of alcohol.
This, in turn, is made up of three components: increased sickness, inability to work through unemployment and early retirement and premature death.
Now, I’ve had some pretty horrendous hangovers in my time, when death would have been a blessed relief, but I’ve never actually died. Nor has a hangover forced me into early retirement, however tempted I might have been.
Alcohol Concern is laughably out-of- whack. And that’s not the only example, although it may be the worst. If this organisation is our “principal information source on alcohol harm” we truly are in deep trouble.
Why does it think it can get away with such a shoddy misuse of statistics? Because it can. It knows that news desks are under-staffed and reporters are under pressure. They have neither the time nor the patience needed to check the numbers in a press release.
It’s also true that Alcohol Concern has been going through something of a crisis itself. The withdrawal of Department of Health funding has had a dramatic impact. Staffing has been reduced from 23 to 12 and high profile chief executive Don Shenker has had to go, to be replaced by former chief executive Eric Appleby, on a two-day week.
Income from training and consultancy work has also declined sharply, more effort has had to be put into fundraising and the organisation has had to rein in its activities while maintaining its presence in the media and its influence on alcohol policy – key to its status as a charity.
Indeed, since the cuts started to bite in 2010 Alcohol Concern has managed to produce no fewer than 40 substantial reports, all published online to save money and all free to download to make them easily accessible.
No doubt much of this prodigious output is good quality, but there are occasions when you feel that the charity is winging it to get the headlines, hoping nobody will notice.
It would be unfair to only accuse Alcohol Concern of misusing statistics, though. Dodgy figures are endemic in the public health lobby, and in government departments, too, repeated over and over in what’s not so much a game of Russian whispers but Russian shouts.
The loudest of all is the £21 billion total annual cost of alcohol harm in the UK. Quoted in the opening of the Government Alcohol Strategy, it’s the kind of large number that gives permission for all sorts of policies to be pursued in order to bring it down.
Yet the whole idea of sticking a total price tag on alcohol harm is highly questionable and the object of a fierce debate by leading academics in the field sparked last year by the rubbishing of such studies by the respected Professor Klaus Makela.
In his paper Makela homes in on Sheffield University’s seminal work which calculated the potential impact of minimum unit pricing. The results of this research have been so powerful that it has pretty much formed the sole evidence base for the government’s decision to adopt it as policy.
According to Sheffield the biggest gain from minimum pricing comes from a reduction in people losing their job through drink, which accounts for as much as 75% of the national cash benefit of a 40p minimum price.
Yet, as Makela points out, you can’t calculate the cost to society of people with alcohol problems becoming unemployed because someone else takes the job. It’s bad news for the person who’s thrown on the dole, but there’s no loss of productivity to the nation unless the job itself goes. And that’s the economy, stupid. Not alcohol.
This, and other criticisms, sent a shock wave to the very top of the academic establishment that has, over 40 years, formulated the theoretical underpinning of the kind of restrictive alcohol policies that are being pushed today worldwide.
And the gurus don’t seem to have much of an answer, running from “the best we’ve got” kind of defence to the astonishing response of Thomas Babor, lead author of the key text Alcohol: No Ordinary Commodity.
Babor accepts that cost-of-alcohol studies like Sheffield’s are “not scientifically credible” but “serve other functions”.
“In a democracy, politicians and policy- makers often need to be shamed into doing the right thing, and costs to society have the ability to shame, blame and even defame,” he says. “It is the simple, single monetary figure that captures public attention more than anything else.”
Probably Babor wasn’t expecting this to be read beyond the obscure Scandinavian journal in which he wrote it. But we must be grateful to him for being so candid: it’s rubbish science, but that doesn’t matter.
The alcohol question is bigger than that. The stats don’t have to be right, they just have to support the moral-political case for greater controls and higher prices.
And Alcohol Concern is merely a clumsy exponent of this strategy of policy-based evidence.
● Alcohol Concern declined to provide comment for this analysis.
One on Every Corner, Alcohol Concern’s 2011 report into off-licence density and alcohol harm in young people, finds a correlation between the number of off-licences in an area and cases of alcohol poisoning.
The exception is in London, which is then excluded from the overall findings “because young people in London consistently consume less alcohol than the average in England and with a lower frequency”, which pretty much scuppers your case if you’re trying to prove that greater availability means greater consumption.
Ploughing on regardless, the report declares that “the analysis showed that nearly 10% of all alcohol specific hospital admissions in England, excluding London, are directly attributable to off-licence density,” blatantly confusing correlation with causation.
It’s not that Alcohol Concern is ignorant of this basic principle of statistical analysis. On the following page it says: “This study does not set out to establish cause and effect”. But it does it anyway.
With the admirable aim of making the case for increased investment in alcohol treatment, a 2010 Alcohol Concern report included the headline-grabbing statistic that 1.6 million people in the UK are moderately or severely dependent on alcohol, 24% up and costing society £3.7 billion.
However, the declared source of the 1.6 million, which has been frequently repeated elsewhere, is a 2007 survey by Leicester University, which has the figure as 0.5% of the population or a more sober 300,000 – fewer than in a comparable survey in 2000.
The Government Alcohol Strategy of March 2012 declares that “over the last decade we have seen a culture grow where it has become acceptable to be excessively drunk in public and cause nuisance and harm”, citing “almost 1 million alcohol-related crimes” after generously rounding up the figure in the British Crime Survey.
Check the source and you find that over the 10 years alcohol-related crime, defined by the victim believing the offender was under the influence of drink, has fallen by 25%, and has been declining even faster than crime rates overall.