Friday, 17 April 2015

Tim Stockwell: merchant of doubt

The campaign to pretend that moderate alcohol consumption isn't good for you continues this month in the pages of Addiction and it's no surprise to find Tim Stockwell taking up the cudgels. Stockwell is one of the world's leading neo-temperance activists, the author of the Canadian Minimum Pricing Miracle study, and is determined to return us to the days of Scientific Temperance Instruction when alcohol was deemed unsafe at any level.

The methods of these people are becoming clear. They find heterodox evidence, exaggerate its importance, ignore the wider literature and declare victory. They are particularly fond of using what Deborah Arnott once described as "literally a confidence trick: the appearance of confidence both creates confidence and demoralises the opposition." They behave as if the benefits of moderate consumption were a bizarre urban myth and sporadically declare that the conventional wisdom which, they claim, has long been teetering has finally collapsed.

Stockwell did this in 2012 and got delightfully smacked down by two scientists. He did it again in 2013 in an op-ed titled 'Another serious challenge to the hypothesis that moderate drinking is good for health?' Mike Daube did it a few months ago in the BMJ in an op-ed titled 'Alcohol’s evaporating health benefits' which said the consensus view was based on "outdated evidence and wishful thinking". And Stockwell is at in once more in Addiction this month with an op-ed titled 'Has the leaning tower of presumed health benefits from ‘moderate’ alcohol use finally collapsed?'

Perhaps they think that the scientific consensus can be defeated by opinion pieces in journals? Where exactly is their evidence? The study Stockwell was getting excited about in 2013 found that moderate drinkers had a 24-46 per cent lower rate of death from cardiovascular disease and that "never alcohol users also had a greater risk of death than lifetime light users". It included this graph showing a clear J-Curve.

The study that Mike Daube treated like a game-changer in February used a crude statistic trick to downplay clear evidence in its own data that drinkers had lower mortality rates than teetotallers. It's going to take more than this piffle for the deniers to overwhelm the mass of evidence supporting the alcohol J-Curve.

In a characteristic move, Stockwell brings up the old 'sick quitter' hypothesis (first mooted way back in 1988) and talks about possible confounding. He even cites a study from 2005 which raised this question, but he fails to cite any of the studies that have tested the hypothesis by excluding people who were (a) sick, and/or (b) quitters. These studies found that moderate alcohol consumption was still associated with lower rates of cardiovascular mortality and/or lower overall mortality—eg. here, here, here and here. Stockwell doesn't mention them. I wonder why.

The more you study these people's behaviour, the harder it is to make the obvious comparison. They ignore the majority of the evidence, they focus on outlying studies that support their argument and they persist in arguing about statistics long after the arguments have been resolved. These are the tactics that got the tobacco industry labelled 'merchants of doubt' in the twentieth century. And that is what neo-temperance activists are when it comes to the alcohol J-Curve in the twenty-first century.

No comments: