Setting a minimum price of 45p a unit for alcohol in Scotland will cut alcohol-related deaths by 300 by year ten, according to the Scottish Government, basing its claims on the recent update of the University of Sheffield’s model-based appraisal of the policy.

That’s fine, but how will they tell? Alcohol-related deaths in Scotland have already fallen by nearly as many (228) since their peak in 2006, in the absence of any such policy. That’s 228 in four years (2006-10), not ten.

Why? Optimists would say that’s because consumption also fell over the same period, as measured by the Scottish Health Survey. As the Sheffield report itself acknowledges, the 2010 survey shows “a general reduction of alcohol consumption at the population level for all four modelled beverage types”, the proportion of harmful and hazardous drinkers is lower than in 2008 and the mean consumption has decreased for all drinker and gender subgroups, except for male harmful drinkers whose consumption has increased.

The report also acknowledges that alcohol-related deaths have fallen. For those that are 100 per cent alcohol attributable the fall is 3 per cent for acute and 12 per cent for chronic conditions, and there has also been a 6 per cent fall in deaths related to partially attributable chronic conditions such as liver disease and various cancers.

The obvious conclusion to draw would be that drinking has declined, and so have its undesirable consequences. But that isn’t the view of the Sheffield team, who question the health survey results in comparison with what they call the “gold standard” – alcohol consumption as measured by Neilsen, the information and measurement company.

The Neilsen data shows alcohol consumption flat or even slightly increasing since 2003, which implies not only that people misrepresent their consumption when answering the survey, but that their misrepresentations have been growing. In 2003 the survey showed the Scottish population consuming 13.87 units a week on average, 63 per cent of the Neilsen sales data that recorded consumption at 21.99 units a week. By 2010 the survey registered 11.58 units, only 51 per cent of Neilsen’s 22.72 units.

Doctors have long suspected – and usually assume – that their patients drink twice as much as they claim to. And under the pressure of media and public health attention on drinking, it’s quite plausible that respondents are progressively shaving more off their claimed consumption. A major reason for the discrepancy is that measures of spirit poured at home are typically two units rather than the one assumed by the survey.

(Incidentally, since studies of health damage are normally based on what people claim they drink, rather than what they actually drink, if you do stay within recommended limits you’re probably safer than you thought you were.)

The degree of under-reporting suggested in the Sheffield report is consistent with other research, including a report for Alcohol Concern produced in 2009 by a team at Liverpool John Moores University. On the other hand, data from HM Revenue and Customs does show a fall in total UK consumption of alcohol between 2008 and 2010 of 5.6 per cent (from 8.9 litres of pure alcohol per head in 2008 to 8.4 litres in 2010). That fall is roughly in line with the fall in Scottish alcohol-related deaths.

But if the Neilson data is right and drinking is steady or even rising, why are alcohol-related deaths falling? Better treatment of disease is one possible cause; so are public health and industry-led programmes aimed at reducing dangerous drinking. The Scottish Health Survey results for 2008 and 2010 show falls in the proportion of drinkers who drink at harmful or hazardous levels. In 2008, for example, 7.1 per cent of men fell into the harmful category, but in 2010 that had fallen to 5.0 per cent. (Of course, this could simply represent a growing reluctance to admit to excessive drinking.)

There’s no reason to doubt the general proposition that increasing prices, whether through taxes or by minimum price legislation, would cut consumption and reduce health damage. It would also, at 45p per unit, put £103 million per year into the hands of retailers, some of whom are held responsible for harmful drinking by their aggressive pricing policies, and take £10.4 m per year out of tax revenues because duty will decrease by more than VAT increases.

That £10.4 m is, by coincidence, exactly the same as was spent in Scotland in 2007-08 for preventative activities to discourage the abuse of both alcohol and drugs, according to Audit Scotland. The bulk of this money went on drug programmes. So the minimum pricing policy stands to lose considerably more tax revenue than was spent in total on the discouragement of excessive drinking by the Scottish Government as recently as 2007-08.

Should the policy become law - its legality is still in question, despite assurances by the Scottish Health Secretary Nicoal Sturgeon yesterday that it is compatible with EU law -  it will be interesting to compare alcohol-related deaths in England with those in Scotland. That may be the only handle we will have on whether the policy is working as intended. 

Update

NHS Health Scotland have been in touch to point out that plans do exist to evaluate minimum pricing, should the policy become law. I have attached a word file that summarises the plans: more details are available here.