Monday, 4 June 2012

A quarter of a pint of beer a day?

My take on the media reporting of our recent paper on alcohol and health

Our paper modelling the costs and benefits to health of alcohol consumption in England came out last week.   It was widely covered by the media including the News Quiz on Radio 4 where the panellists had a good laugh at the findings of the research.   Our paper shows that the optimum average consumption level for drinkers in this country is about 5g/day (about half a unit) and this would avert or delay about 4600 deaths per year.   I should emphasise that this paper describes what is and makes some tentative conclusions about what should be the case.   There has been little criticism of the methods for, and results of, the paper but much derision at the supposed conclusions.

There is a line in the paper that reads:  “On this basis, we recommend that the public health target for alcohol consumption in England should be to reduce median alcohol consumption to half a unit per day for both men and women.”    We should, perhaps, have been clearer about what we meant by ‘a public health target’.   We meant a target level for the whole population of England and NOT a target for any one individual.   By median intake we meant the average for everyone not the individual’s average intake. 

The paper got reported in the press with such headlines as ‘Cut alcohol intake to just a quarter pint of beer a day, experts advise’ (Guardian) and ‘Don't drink more than quarter of a pint a DAY’ (Daily Mail).   It might sound quibbling but we weren’t advising people to drink just quarter of a pint a day.   What we were trying to say was that if people ON AVERAGE drank quarter of a pint a day the whole population would be healthier.   And if the Government was to advise people about their drinking then they should tell them the truth i.e. that alcohol is good for them in small amounts but that anything over quarter a pint is (on average) harmful.  

The Guardian and Mail clearly failed to understand what we were getting at but some other newspapers did seem to understand.   For example the Telegraph’s headline was ‘ Sticking to a quarter pint of beer a day would save thousands of lives’  and BBC Online’s headline was ‘Reducing alcohol to half a unit a day saves lives’.  These headlines –if somewhat imprecise - better capture what our paper says and both wisely avoid the issue of what the public should be advised on the basis of our research.  

Why did the Guardian claim that we – the expert authors of the paper - were advising people to drink just a quarter a pint of beer a day?   We didn’t speak to the journalist in question and this was just sloppy journalism.   The Telegraph article, for example, doesn’t anywhere suggest that we were advising the public anything.  

However it would be disingenuous of us to pretend that we weren’t hoping that our research would have some effect on the public and their view of alcohol.  And we did say in the paper that, “we recommend that the public health target for alcohol consumption in England should be to reduce median alcohol consumption to half a unit per day for both men and women.”   Furthermore there is clearly some relationship between what would be good for the population and what individuals should individually drink if their health is to be optimised.

But the paper does not address the question of how our epidemiological research should be translated into advice to the public.   Perhaps we should have been clearer in that regard.   How results such as ours should be used to inform advice to the public should be the subject of future investigation – not necessarily by us.   But there are two things that might be said in a preliminary sort of way. 

Firstly, as some commentators have pointed out, a median intake of half a unit a day is even theoretically compatible with the Government’s ‘recommended daily limits’ of 3-4 units a day for men and 2-3 units a day for women - for, as Andrew Sales points out in the comments on our paper at BMJ Open ‘at this level of population consumption [an average of half a unit a day] the model…allows for individual consumptions at much higher level.’   (If the median is half a unit then 50% of people can afford to drink more than a half a unit.) 

Note that the Government does not says that 3-4 units for men/2-3 units for women is optimal or even safe as, for example, the Daily Mail describes the Government’s ‘recommended daily limit’.  In fact the Government says that ‘There's no guaranteed safe level of drinking, but if you drink less than the recommended daily limits, the risks of harming your health are low’.   Well it depends on what you mean by ‘low’.   I would argue that the risk of drinking 2-4 units is hardly ‘low’ but my ‘low’ may not be their ‘low’.   Furthermore I question whether 2-4 units should be ‘recommended’.   What is to be ‘recommended’ about this ‘limit’?

Secondly it is not clear – as David Spiegelhalter has pointed out in a rather good commentary on our paper – that what the public should be advised is what would absolutely minimise their risk of ill health.   For some other bits of public health advice the Government doesn’t attempt to advise what would be ideal but what would be realistic.   For instance eating 5 portions of fruit and vegetables is not the best we could do.   In Greece they eat around 10 portions of fruit and vegetables (on average) each day and in consequence Greece’s cardiovascular disease rates are lower than ours.   The Government’s 5–a-day message was designed to be a realistic target for individuals and the population not what we should ideally eat.  On the other hand (contra Spiegelhalter) with other public health advice the Government does recommend to the public what would absolutely minimise their risk.   Take smoking for instance.  Here the advice is to quit not just cut down.

So has our paper and the media coverage helped or hindered the confusion about the effects of alcohol consumption on health and how much we should drink.   I am hoping that when the dust has settled, its overall contribution e.g. to the Government’s currentreview of its alcohol guidelines, might be helpful.  As the Commons Science and Technology Committee says these are confusing.   On a personal note I am not pleased by the results of our research.   I was hoping that alcohol might have been better for my health.