Tuesday 7 February 2017

New data on alcohol-related deaths

Public Health England's alcohol report was based on a false narrative created by activists in the neo-temperance lobby. Its fundamental deceit was to claim that alcohol consumption and alcohol-related deaths were rising until the duty escalator was introduced in 2008, after which they fell.

The culprits were Nick Sheron and Ian Gilmore of the Alcohol Health Alliance who have repeated this claim again and again and who also claimed that alcohol-related deaths would rise after the beer and wine escalator was scrapped in 2013 and the spirits escalator was scrapped in 2014. The whole point of this charade is to portray price, consumption and mortality as inextricably linked and, therefore, to present endless price rises as the solution.

Sheron and Gilmore used some basic statistical tricks which PHE should have seen through. As I have previously discussed, if you measure alcohol consumption properly (ie. per capita), it peaked in 2004, not 2008. This seems a minor point but it is important since the peak in alcohol-related deaths and the introduction of the duty escalator occurred four years later. The correlation claimed by Sheron and Gilmore does not exist.

The Office for National Statistics published its latest data on alcohol-related deaths today and it makes for interesting reading. There has been essentially no change since 2012, but there has indeed been a decline since 2008.


Although the decline in alcohol-related mortality does not exactly correlate with the decline in alcohol consumption (and is nowhere near as sharp as the 18 per cent decline in alcohol consumption), this graph could be used as prima facie evidence that the decline in alcohol affordability which came about as a result of the recession and duty escalator had an impact on mortality.

Perhaps it did, but if you dig a little deeper, that hypothesis looks shaky. Looking at the mortality rates in the UK's four countries reveals that the decline has largely been a Scottish phenomenon and it clearly began before 2008.

Alcohol-related mortality rate (males)
Alcohol-related mortality rate (females)

In Scotland, the alcohol-related death fell by more than a third between 2003 and 2012, from 47.7 per 100k to 29.9 per 100k for men, and from 19.6 per 100k to 12.5 per 100k for women.

But in England, there has been little change since 2003 despite a sharp decline in alcohol consumption and the introduction of policies that were supposed to make things better or worse, such as so-called '24 hour drinking' and the alcohol duty escalator. There was a clear rise between the early 1990s and early 2000s, but the trend has been largely flat ever since (rather like childhood obesity, as it happens).

If campaigners are going to attribute these changes to the affordability of alcohol, they need to explain why the trends have been radically different in Scotland than in England, despite both countries having the same tax rates.

While they're at it, they should explain why alcohol-related mortality is higher now than it was in 2001 despite alcohol consumption falling by nearly a fifth. Advocates of the total consumption model can no longer use the 'lag effect' as an excuse. The lag is not that long.

And they could tell us why there has been no rise in alcohol-related mortality since 2012 despite the alcohol duty escalator being ditched and rising incomes making alcohol more affordable.

So far, they have avoided these questions. Instead, Sheron and Gilmore have continued to make predictions for England and Scotland that are all about price:

Alcohol related deaths are most likely set to increase in England as incomes outstrip rises in taxation, argue experts in The BMJ today.

Meanwhile, the number of alcohol related deaths will likely continue to decrease in Scotland if legislation on minimum unit pricing for alcoholic drinks is implemented.

Time will tell, but there should be some acknowledgement that the total consumption model and the price-harm hypothesis have been poor predictors of outcomes so far, and that the decline in alcohol-related deaths has been most apparent in Scotland where it began several years before alcohol became less 'affordable'.

As an interesting side note, it seems likely that the statisticians are gearing up to create another epidemic overnight:

Upcoming changes to this bulletin 

In the coming months we will be working to review our definition of alcohol-related deaths. This work is being conducted to improve the consistency of outputs on deaths related to the abuse of alcohol produced by different government departments across the UK. We will be holding a consultation on this definition in the summer of 2017 with the view to using an improved definition in our next release. 

This is an opportunity for the anti-drinking cabal of quakademics such as Mark Bellis, Mark Petticrew and Ian Gilmore to get the ONS to abandon its proper definition of 'alcohol-related' (meaning 'caused by alcohol') in favour of a dubious system of partially-attributable fractions based on junk epidemiology. 

There has been growing interest in the use of partially-attributable conditions to provide a truer burden of alcohol consumption on population health and use of health services associated with these conditions. 

I bet there has. These people have already hi-jacked the Chief Medical Officer's office and Public Health England. They will be delighted to add the ONS's scalp to their collection.

No comments: