Friday, 12 January 2018

Fake statistic of the day

There's going to be a PR stunt in Parliament on January 22nd when Dr Sarah Wollaston uses the health select committee to campaign for minimum pricing. Wollaston is in charge of the committee and has been advocating for this regressive policy for years. There's no word yet on who will be appearing at the 'inquiry', but I expect we will see the usual faces from the Sheffield fantasy modelling club plus the likes of the UK Temperance Alliance (AKA the Institute of Alcohol Studies).

The Lancet's Marxist editor, Richard Horton, has given some publicity to Wollaston's kangaroo court in this week's issue. In addition to making the absurd claim that '[t]he science supporting minimum unit pricing seems overwhelming', he says this:

Chaired by independent-minded Conservative Member of Parliament and former general practitioner, Sarah Wollaston, the committee will review evidence for and against minimum unit pricing at a moment when liver disease is on a trajectory to become the biggest cause of death in England and Wales.

Even by Horton's standards, this is nonsense on stilts. Liver disease is nowhere near being the top cause of death. Here are the figures from the Office for National Statistics for men in England and Wales. Liver disease is tenth on the list, causing half as many deaths as prostate cancer and an eighth as many deaths as heart disease.

For women, liver disease doesn't even make the top fifteen:

If we look at the age-standardised mortality figures from Public Health England, the story is much the same. Here are the men. You'll see the rate of liver disease right at the bottom in blue.

And here are the women. Once again, liver disease doesn't make the cut.

And - without wanting to labour the point - here are the proportion of all deaths in England attributed to each major 'killer'. Liver disease just about sneaks into the top ten for men with 1.9%, but not for women.

If you look at Public Health England's graph for men (above), you can see the trend in liver disease deaths. Although Horton claims that 'liver disease is on a trajectory to become the biggest cause of death', the graph shows that the trajectory is basically flat. This is also the case for alcohol-related diseases in general, which rose in the 1990s but have been broadly flat for more than a decade.

So what the hell is Horton talking about? It can be no slip of the pen because he has been repeating the claim on social media:

The source for the claim seems to be a study he published in the Lancet last month. The Independent reported this study with the headline 'Liver disease to become biggest killer by 2020 with alcohol and obesity to blame' but that's not what the study said.

The study was, in practice, a briefing for anti-alcohol lobbyists written by the usual neo-temperance suspects (Ian Gilmore, Petra Meier etc.). Nick Sheron was one of the authors and he gave a quote to the media when it was published. We have seen before that Sheron likes to ignore the size of the population when he does his calculations, and that trait is in evidence in the Lancet study which says:

Alcohol consumption in the UK, which peaked at around 5642000 hL (hectolitres) in 2008–09, decreased when the duty escalator was introduced to around 4843000 hL in 2013–04, and increased again to 5126000 hL in 2016–17 after the duty escalator was withdrawn.

Sheron and Gilmore used this sleight of hand in Public Health England's risible alcohol policy evidence review. The appropriate measure here is per capita alcohol consumption. Per capita alcohol consumption actually peaked in 2004 and has been falling ever since. Sheron likes to pretend that the decline began in 2008 because it enables him to link it to the alcohol duty escalator. As he says in the study:

These changes show how responsive population alcohol consumption is to small changes in taxation and further support the Commission’s recommendation for an increase in overall alcohol taxation.

Total cobblers. There was a bigger fall in consumption in the four years before the duty escalator was introduced than in the four years afterwards, but since this doesn't fit the temperance narrative they rewrite history.

They then say that...

Alcohol-related deaths in England and Wales decreased from a peak of 7312 in 2008, when the alcohol duty escalator was introduced, to 6999 by 2012, but increased to 7630 in 2016 after abolition of the alcohol duty escalator in 2013.

Again, the authors are using absolute numbers when they should be using death rates per 100,000 people. Reading their statement, you would get the impression that deaths peaked in 2008 and then fell before reaching new heights following the abolition of the alcohol duty escalator. But this, again, disregards population growth. The rate of deaths per 100,000 men in England was 15.5 in 2008 and was 14.5 in 2016. For women, the rate was 7.0 in 2008 and 6.8 in 2016. Rates in Wales were also lower for both sexes in 2016.

Sheron et al. create a false narrative of record levels of liver disease mortality which are rising rapidly (and which can only be reduced by raising prices). But even in this fantasy, it is obvious that the 7,630 alcohol-related deaths mentioned in the study (of which only a proportion involve liver disease) make up less than two per cent of the 500,000 deaths that occur each year in England and Wales.

In what parallel universe, therefore, is liver disease poised to become Britain's biggest killer? It turns out that this isn't what the Lancet study claims at all. It doesn't make any claims about the number of deaths. Instead, it looks at the much more specific issue of lost years of life. This makes liver disease appear more important because it is less of a disease of old age than dementia, heart disease and cancer.

But although Quality Adjusted Life Years (QALYs) are a conventional measure in public health, they do not make liver disease look significant enough for Sheron et al. and so they switch to the rather less conventional measure of lost years of working life before the age of 65. This helpfully disregards all the people who die after the age of 65 - ie. the vast majority of the British population - and results in this graph.

I can't vouch for the figures in this graph because they have been created by the authors of the study and are not available in any of the routine statistics that are published by the NHS or ONS. They could be rubbish, but even if we assume that they are correct, it is clear that the 'trajectory' has been flat for the last decade. If liver disease overtakes heart disease under Sheron's bespoke measure, it will be because the number of lost working-age years before the age of 65 from heart disease falls, not because the number of lost working-age years before the age of 65 from liver disease rises.

None of this has any bearing on the number of people dying from liver disease. Horton's claim that 'liver disease is on a trajectory to become the biggest cause of death in England and Wales' not merely untrue, it is a million miles from the truth. Liver disease is responsible for well under two per cent of all the deaths in England and Wales. Mortality rates have not been rising and they show no sign of rising.

Nevertheless, you can expect this fake statistic to be trotted out for years to come with the once respected Lancet journal cited as the source.

No comments: