Tuesday 3 March 2020

Another minimum pricing prophecy fails

In yesterday's blog post I mentioned that the number of alcohol-related admissions to Scottish hospitals rose slightly in 2018/19. This was the first year of minimum pricing, with the policy being in force for eleven of the twelve months.

The number of hospital admissions rose from 38,199 to 38,370. The age-standardised rate rose from 668.8 per 100,000 to 668.9 per 100,000.


These are small changes, to be sure. There hasn't been much change in the rate since 2014/15 when it began to level out after a substantial decline. Still, one can't help but notice that the figures are going up rather than down.

So let's just take a moment to reflect on what was promised by the Sheffield Alcohol Policy Model. The most recent Scottish version of the computer model, which was hopelessly flawed but so influential that it had its own study written about it, predicted that minimum pricing would make the number of alcohol-related hospital admissions plummet by 1,299 in the first year alone.

This factoid has been widely circulated. Here's the SNP's Shona Robson speaking in the Scottish Parliament in 2017, for example:

The benefits of minimum unit pricing will be substantial. As an illustration, last year, Sheffield University modelled that a price per unit of 50p would lead to 58 fewer alcohol-related deaths in the first year, with a cumulative total of 392 fewer alcohol-related deaths within the first five years. The reduction in alcohol-related hospital admissions at that price would be similarly substantial. In the first year, a price of 50p would lead to 1,299 fewer alcohol-related hospital admissions, with a cumulative total of 8,254 fewer alcohol-related hospital admissions within the first five years.

 Here's the Scottish government's consultation document from December of the same year:

The modelling has consistently shown that Minimum Unit Pricing will have most impact on those who drink the most alcohol.

Most importantly the modelling estimates there will be a fall in the number of people admitted to hospital and who die from alcohol-related illnesses.
The model estimated that if a Minimum Unit Price of 50 pence was introduced in the first year there would be:
  • 58 fewer alcohol-related deaths
  • 1,299 fewer alcohol-related hospital admissions

The figure also appeared in the government's impact assessment and in various news stories and 'fact sheets' from pressure groups.

But it didn't happen, did it? Didn't come close.

The model is, was and always will be worthless, but that won't stop the contracts coming in. Despite - or perhaps because of - incontrovertible evidence showing that they changed their methodology when Public Health England paid them to, the Sheffield team were contracted by the Australian government to do the modelling for its drinking guidelines review. And they were recently given a wodge of taxpayers' money to play Numberwang with their model to promote minimum pricing in various English regions.

We see this time and time again in 'public health'. Excitable promises are made loudly and persistently, but when the prophecies fail there is only tumbleweed. Failure brings no consequences.

No comments: