Tuesday 22 December 2015

Lying about alcohol and A & E attendances

There is so much anti-booze hysteria today, but this stood out...

Three in four people in A&E at weekend are there because of alcohol

It refers to a study published in Emergency Medicine which is actually quite useful. It took two sets of data from two different years at the Newcastle upon Tyne Hospitals NHS Foundation Trust and estimated the proportion of A & E attendances that were alcohol-related.

The Telegraph reports that...

The alcohol-related attendance rate varied substantially from four per cent to 60 per cent on weekdays, but rose to 70 per cent at the weekend.

While focusing on the biggest number, the newspaper fails to mention the average. The study actually concluded that 'the prevalence rates of alcohol related attendances were 12% and 15% for the retrospective and prospective cohorts, respectively'.

Is 12-15 per cent a lot? Well, it's a damn sight lower than the 35 per cent estimate that underpins every single 'cost of alcohol' study the government has issued in the last fifteen years. As I explained in a recent IEA report, the 35 per cent figure is based on a long forgotten MORI survey from 2000. It was not based on any clinical evidence whatsoever. Studies like the one published today which bother to breath-test attendees typically arrive at a much lower estimate.

When I calculated the cost of alcohol to the public purse, I used an estimate that more closely reflects real evidence: 14 per cent. Given that a study in Newcastle—a notorious city for binge-drinking—has arrived at a similar estimate, I suspect that 14 per cent is still too high for the country as a whole. (Incidentally, none of the newspapers that covered this study bothered to mention that Newcastle was the city in question.)

Based on the figures, researchers calculated that it costs each emergency unit around £1 million a year just to treat drink related problems and injuries.

£1 million might seem a lot of money but to put it in context, the Newcastle upon Tyne Hospitals NHS Foundation Trust has an annual budget of £1,006,000,000. Alcohol-related A & E attendances therefore make up 0.1 per cent of its expenditure.

As for the idea that 'three in four people in A&E at weekend are there because of alcohol', this claim is made (more or less) in the text of the study...

On weekend days, over 70% of attendances were alcohol related...

But it is not borne out by the data. Table 2 of the study shows that alcohol-related attendances made up 17.4 per cent of Saturday's total in the retrospective study and 23.9 per cent of Saturday's total in the prospective study. For Sundays, the figures are 16.9 per cent and 19.7 per cent.

In other words, alcohol-related attendances make up less than 20 per cent of the weekend total, not 'three in four'. This is further confirmed by a graph showing the hourly rates at the weekend.

As you can see, the only time when attendances reach 70 per cent is between 2am and 4am when the total number of attendances are very low because most people are in bed. During the daytime, the figure is well below 10 per cent.

With a little scrutiny of the data, it becomes clear that 'On weekend days, over 70% of attendances were alcohol related' should read 'On weekend days, alcohol related attendances reached a peak of 70% at 3 am'. The sloppy wording of that sentence led the journal to issue a press release titled 'Almost three quarters of weekend emergency care caseload linked to booze' and the newspapers took it from there. It will not be long before this ridiculous factoid becomes the conventional wisdom.

You can't entirely blame journalists for this lie being spread (although it would be nice if health journalists could read academic studies). The lie is right there in the abstract and it was repeated in the press release. Incompetence or deliberate deceit? That, as ever, is the question.

No comments: