Friday, 29 January 2010

Alcohol consumption falls again

The Office for National Statistics released the latest drinking figures yesterday. Encouraged by the ONS press release, most of the media have focused on a rise in alcohol-related deaths of 3.5%. As this is a fairly pedestrian rise, the gentlemen of the press have turned the clock back 16 years to take a long-view; hence headlines like Alcohol-related deaths in Britain rising fast (Reuters) and British boozing deaths double (The Sun).

From the latter:

According to the Office of National Statistics, 4,023 people were killed by booze in 1992, rising to 9,031 in 2008.

There have been very substantial changes to the way "alcohol-related" deaths have been defined since 1992, but let's not get sidetracked by that right now. Instead, let's look at the amount people are drinking. 2009 saw a rise in how much women drink, but there was a greater decline in 2008 so it's as you were for the ladies.

The stats for males, however, should objectively be the big news story. Alcohol consumption has declined at by far the greatest rate in recent history, as this graph indicates.

But even this graph does not tell the whole story. 2007 saw the ONS bring in a major change to the way they calculated alcohol consumption. They insisted that everyone was drinking stronger beverages from larger glasses. Overnight, the amount men drank rose by 45% for women and 23% for men. In effect, this statistical 'correction' created the whole 'binge-drinking epidemic'.

The massive spike you see in the graph above reflects the statistical change, not any real change in consumption. It's an apples an oranges comparison. Whatever you think of the idea of glasses getting bigger, they didn't suddenly become so in 2007. If you want to see the real trend you need to use the same method throughout. The graph below shows alcohol consumption under the original method.

As you can see, there has been a fairly continuous drop in consumption. The picture for women is less clear, but still a decline is evident.

For a long time, Britain's falling alcohol consumption has been a well-kept secret. It doesn't fit the narrative of boozy Britain and doesn't support the neo-prohibitionist's scare-mongering. It is now so obvious that consumption has fallen (under either method of calculation) that this truth is now being whispered. The BBC, for example, gives it a brief mention:

The trend to less consumption began in 2002...

Indeed it did. So much for the riding tide of alcohol abuse.

...and is unlikely to be reflected in figures on the alcohol deaths for some years.

Herein lies the problem. We have 13 years of falling consumption and yet deaths have doubled in the last 16. What will it take for patterns of death to start following patterns of consumption?

There is no answer to this because overall consumption is not a predictor of alcohol-related death. It's a red herring. If you want to know how many people are going to get liver cirrhosis or pancreatitis you need to know how many people are drinking at chronic levels. Fewer than 10,000 die from alcohol-related deaths, even after the definition has been broadened enormously. This is not a problem to be solved at the population level and consumption amongst the general population tells you almost nothing.

Policy-makers are so wedded to population-level solutions that this important point gets lost. It is quite possible to have serious alcohol-related disorder and death even as general consumption falls. So focused is the government on its ridiculous 28-unit-a-week minimum that it ignores those drinking over 100 units who are genuinely at risk.

As long as this continues, the many will be punished for the sins of the few while those at most risk will slip through the net.


David Strange said...

Excellent article which illuminates the actual UK drinking patterns rather than just quoting data selectively to underline the neo-prohibitionist agenda. We are drinking less, whether this is a good thing or not is debatable, but it clearly shows the fallacy of how figures for alcohol 'related' deaths have been redefined and manipulated. If we have been drinking smaller amounts then it doesn't take the sharpest tool in the box to realise that the death rate should also be decreasing. When we see dubious numbers like this increasing alcohol related death rate we should point out the flaws in how they were derived and not be brow-beaten into towing the neo-pro, alcohol de-normalisation line about how anyone who has the odd drink is ruining their own lives and placing an unreasonable burden on the state.

Anonymous said...

It is possible to find figures for pure alcohol consumption. I can't remember where. Could be Customs and revenue or a drinks industry website. These are accurate because duty is paid on just about all alcohol consumed in the UK. This gets rid of the change in measurement problem and all the unit nonsense.

Anonymous said...

Wow - looks like everyone missed the "interview" on 5 Live this morning where the figures were artfully spun to once again bemoan the steady rise in alcohol consumption and how something must be done about it, and how "great" it was that the police now have powers to simply confiscate alcohol from anyone under 18 etc. etc.

TheFatBigot said...

Did I read this correctly: "28 units a week"?

Oh dear, I always thought it was 28 units a day.

Anonymous said...

Thank you Chris for the information.
My Step-Mother has pancreatitas and has been in and out of hospital from Oct 2009 and is now back in there.

I checked out a few thing because I had never heard of it, this misinformation is unbelievable.
This link shows "smoking" top of the list, but I would not expect anything else from that site!

My Stepmothers whole family has type 2 diabetes, 2 smokers, 3 non-smokers. Non of them drink even the made up so called safe units.

As we all know diabetes 2 is rising and here is some info I looked into before Xmas. I cannot find the link for diabetes higher risk for pancreatitas though.
Excessive alcohol use is often cited as the most common cause of acute pancreatitis, yet gallstones are actually the most common cause. Less common causes include hypertriglyceridemia (but not hypercholesterolemia) and only when triglyceride values exceed 1500 mg/dl (16 mmol/L), hypercalcemia, viral infection (e.g., mumps), trauma (to the abdomen or elsewhere in the body) including post-ERCP (i.e., Endoscopic Retrograde Cholangiopancreatography), vasculitis (i.e., inflammation of the small blood vessels within the pancreas), and autoimmune pancreatitis. Pregnancy can also cause pancreatitis, but in some cases the development of pancreatitis is probably just a reflection of the hypertriglyceridemia which often occurs in pregnant women. Pancreas divisum, a common congenital malformation of the pancreas may underlie some cases of recurrent pancreatitis. Pancreatitis is less common in pediatric population.

snip~ "Acute hepatic porphyrias, including acute intermittent porphyria, hereditary coproporphyria and variegate porphyria, are genetic disorders that can be linked to both acute and chronic pancreatitis. Acute pancreatitis has also occurred with erythropoietic protoporphyria.
Conditions that can lead to gut dysmotility predispose patients to pancreatitis. This includes the inherited neurovisceral porphyrias and related metabolic disorders. Alcohol, hormones and many drugs including statins are known porphyrinogenic agents. Physicians should be on alert concerning underlying porphyrias in patients presenting with pancreatitis and should investigate and eliminate any drugs that may be activating the disorders".

Drugs (prescription) seem to have a play in it.


Christopher Snowdon said...

I flagged it, because I don't think you meant to publish it like this ... ;)

Christopher Snowdon said...

It's all spam. I'll get rid.