Friday 26 August 2011

A million alcohol-related hospital admissions?

Last night, I was on BBC Five Live discussing the latest shock-horror alcohol admissions figures for UK hospitals (on the iPlayer from 51.00). Actually, as Mark Wadsworth noticed, there's nothing new about them. They were given to the media back in February, and then released officially in May and have now been revived in the Silly Season for no other reason than that the temperance lobby wanted to milk them one more time. Apparently it sells newspapers.

The big (old) news is that there were more than a million alcohol-related hospital admissions last year in the UK despite (and this is rarely mentioned) alcohol consumption having been in decline for the last eight years, and despite (this is never mentioned) Britain having the third highest alcohol taxes in the world. This is a doubling in admissions in less than a decade. A truly remarkable phenomenon, if true.

In my naïvité, I always imagined that it was doctors and nurses who decided whether a hospital admission was alcohol-related, but then I read this post at the always informative Straight Statistics, which explains that hospital admissions data are divided up according to various assumptions. For example, it is assumed that 20% of all stomach cancer admissions are alcohol-related, half of epilepsy admissions are alcohol-related and a quarter of admissions for extreme cold are alcohol-related.

These assumptions are based on individual epidemiological studies which may or may not reflect the true risks. Whether true or not, if a man goes to hospital complaining of hypertension, he will make up a quarter of an alcohol-related admission, even if he is a teetotaler. If he goes to hospital 8 times, he will have added two alcohol-related admissions to the statistics—remember, these are admissions, not different individuals. The system is called 'alcohol-attributable fractions' and you can read all about it here should you wish.

Having read many an epidemiological study, I am not convinced that this is the very best way to get a handle on alcohol-related harm. The NHS's figures have already been distorted improved beyond recognition in recent years. The methodology has changed, for one thing, as the Office of National Statistics acknowledges:

These figures use a new methodology reflecting a substantial change in the way the impact of alcohol on hospital admissions is calculated. Previously the calculation counted only admissions for reasons specifically related to alcohol. The new calculation, for which the methodology is described in the report, includes a proportion of the admissions for reasons that are not always related to alcohol, but can be in some instances (such as accidental injury).

The number of ailments that are categorised as 'alcohol-related' has jumped from 7 to 14 to 20 in the space of eight years:

Finished admission episodes are identified where an alcohol-related diagnosis is recorded in any of the 20 (14 from 2002/03 to 2006/07 and 7 prior to 2002/03) primary and secondary diagnosis fields in a Hospital Admission Statistics record.

As a consequence of these changes we are seeing an enormous rise in admissions that are 'partially attributable' to alcohol which would not have been considered alcohol-related in the past. And of these partially attributable admissions, the biggest growth has come from diseases of old age which are blamed on alcohol by the system of attributable fractions.

As the Portman Group has pointed out in a letter that probably won't be published, hospital admissions are rising in general, largely as a result of the ageing population. According to them, the proportion of hospital admissions that are 'alcohol-related' has risen from 1.32% to 1.34% in five years, which does not sound quite like the 'epidemic' you may have read about.

Take hypertension, for example. There are many hundreds of thousands of admissions for hypertensive disease every year. The system of attributable fractions assumes that 28% of them are due to alcohol (13% for women), and, as a result, the number of 'alcohol-related' hypertension admissions has risen from 136,000 to 383,900 since 2002. There are also a huge number of admissions for heart palpitations, of which 182,300 are blamed on alcohol (up from 87,000 in 2002).

I am not necessarily saying that these ailments were not caused by alcohol in some way, but these two conditions make up more than half of the million admissions that have filled the headlines. And those headlines are invariably accompanied with images of young people out on a Saturday night. The unmistakeable impression is that a million people are being rushed to hospital with alcohol poisoning or liver cirrhosis. They're not. The admissions that are wholly attributable to alcohol make up less than a quarter of the total.

I suppose images of people in their eighties repeatedly being admitted to hospital with heart problems wouldn't sell the temperance lobby's agenda so effectively, but it is these pensioners, rather than the teen bingers, who are the real face of this moral panic. Hospital admissions for heart failure, hypertension, stroke, stomach cancer etc. will almost certainly continue to rise in the future because of the ageing population. As a consequence, the number of partially attributable alcohol-related hospital admissions will also rise. It has to, because although the number of admissions change, the underlying assumptions do not.

(It should be mentioned that drinking has prevented many thousands of hospital admissions because of the protective effect of alcohol on the heart. The NHS figures cannot be expected to include these, but any honest anti-alcohol campaigner would calculate how many admissions have been saved and debit it from the total.)

However, beneath all this noise, there is evidence of a rise in genuine alcohol-related diseases. The number of deaths from alcoholic liver disease—a diagnosis that is less prone to statistical fiddling—has risen from just over 3,000 to just over 4,000 in the last decade. This is much less dramatic than the temperance lobby likes to claim (and the rate fell last year) but it is a rise and it is compatible with falling consumption because overall consumption is a poor predictor of chronic alcoholism.

And chronic alcoholism is the issue. Along with those who commit alcohol-related violence, it is the chronic alcoholics who are the real problems. They need help and support while the thugs need punishment. The broad population-based policies of the temperance movement will not address either issue. They are going down the same dead-end of denormalisation that the tobacco control mob have gone down. They will punish the many for the sins of the few while doing nothing for the minority of people who abuse alcohol in a damaging way. Like the imbeciles of the anti-smoking lobby, they are obsessed with an imaginary war with the drinks industry which will do nothing to reduce harm or save lives. If these people were only idiots they would be tolerable, but they are dangerous idiots because their broad brush policies ignore the real problems.


Anonymous said...

Tony should have stopped the Liver Trust anti alcohol woman from constantly interrupting you.

Anonymous said...

Like with tobacco, the doctors "just say" alcohol dunnit as they keep trying to maintain the quota to substantiate the statistical "facts", like with the tick-boxes on death certificates, alcohol or tobacco related, ticked even if it's not, just to meet the quota, to "prove" something already stated without facts to back it up.

Mark Wadsworth said...

Ta for link.

Yes, the figures are completely made up, but at least they are colluding on what the 'official' figure is. Until a couple of years ago, each fakecharity or NHS body would issue wildly different figures, forecasts etc.

But that 1 million is a such a huge number that personal experience would appear to disprove it. How many times have you, or anybody you know, been an 'alcohol related admission'??

It might not the sort of thing that people shout about, but I seriously cannot think of a single friend to whom it's ever happened.

And I'm talking dozens of people I know at any one time, multiplied by a quarter of a century of being on the piss. Although I've been to A&E a few times because of one act of stupidity or another, none of them happened while I was pissed.

Statistically, if one in sixty people go to hospital for booze related reasons each year, then I ought to be able to remember ten or twenty instances within my social/family circle, and I can't think of a single one, nothing, nix, nada.

Pat Nurse MA said...

Mark - your reasoning could apply to tobacco related harm too - I have never seen any of my friends fall foul and trot regularly off to hospital because of their smoking.

Ian R Thorpe said...

The important word is 'related'. It could mean someone who passed a pub on the way to hospital is an alcohol related admission. Someone who choked on a peanut while enjoying a half of shandy is an alcohol related admission. Someone walking quitely home after a couple of pints when they were mugged by a gang of chavs would be an alcohol related admission ...

Always watch for the vagueness.

BrianB said...

Chris - I posted a comment on this very subject back on 27 May when you last wrote a piece about it. I won't repeat the whole thing, but just so no-one is under any illusions here: those mythical 384k admissions 'for alcohol-related hypertension' were 384,000 admissions with high blood pressure, not 384,000 admissions for high BP.

The difference is not merely semantic, but is extremely significant, as there are up to 20 different diagnoses recorded against each admission, but very few admissions (<4% from memory), where high BP was mentioned, had high BP as the primary diagnosis - which is the actual reason for the admission, of course.

There were less than 40,000 admissions for high BP (primary diagnosis) in total(1), before any application of 'attributable fractions', so how come 28% of 40k can become 384k? Sleight of hand, or just plain dishonesty?

The increase has clearly got nothing to do with alcohol. It is all down to doctors including more 'additional factors' diagnoses in the patients' episode records, coupled with a general increase in measurements of high BP in the population as a whole.

It's more an indicator of the 'measurement' behaviour of GPs and hospital doctors, quite frankly, and I strongly doubt that general population-wide BP is any higher, on average, than it was eight years ago.

(1) Source: hesonline

Christopher Snowdon said...

Thanks Brian. Very interesting.