A warning about Wales

The Daily Post reports that new figures soon to be released by Wales’ Chief Medical Officer will show that heart rates have dropped in the region during the second year of its smoking ban. I look forward to seeing those figures, but the close of the article is particularly striking:

There was a 12.5% fall in the number of patients admitted to hospital with a heart attack between October and December last year, compared to the same period in 2006, before the ban on smoking in enclosed public spaces was introduced in Wales: some NHS trusts have seen the number of heart attacks fall by up to 40% in the same period.

See anything suspicious there? To my knowledge there isn’t anyone promoting the theory that smoking bans only reduce heart attacks in the final three months of the year. So why focus on just that time period?

The answer is that cherry-picking those 3 months supports ban proponents’ predetermined conclusion. As Christopher Snowdon has shown with the government’s own data, the rate of AMI actually rose slightly during the year the ban took effect and there were more heart attacks in the 5 months immediately following the ban than there were in those same months the year before. Overall there was practically no difference in the AMI rate between the two years, which is surprising since Wales’ rate of heart attacks had been on a downward trend. A fair reading of all the data is that the ban had no discernible effect, yet the story told by activists and mindlessly repeated by reporters is that the ban caused a reduction in heart attacks.

Such a weird statistic should have immediately raised flags for a journalist. A quick search on Google for the words “Wales,” “smoking ban,” and “heart attacks” would have revealed why the figure was selected. Unfortunately Post reporter Tom Bodden took it at face value, so readers don’t get the full story and the misleading statistic is propagated even more widely.

As Snowdon notes, the Welsh rate of AMI has been falling in recent years by as much as 10% per year. The fact that it didn’t fall in the year smoking was banned is an anomaly. My guess is that in the report about to be issued the rate will have resumed its previous decline and, in a classic case of conflating correlation with causation, researchers will credit the drop solely to the smoking ban.

It will be up the to journalists covering the report to treat it with skepticism, ask tough questions, and at the very least seek out a critical source before phoning in their stories. I’m not optimistic that they will.

Update 12/7/09: My prediction seems to be too optimistic. Chris decided to check the data again today and it turns out that the heart attack rate in Wales actually increased last year. So how will the upcoming report show that it declined? Well, that will be interesting to find out!

Update 12/9/09: Here’s the story and here’s Snowdon on why it’s just another example of junk science.

Previously:
More lazy tobacco reporting

Comments

  1. Well spotted, Jacob. I’ve gone back to the Welsh data and there’s a new development that really puts the last nail in the coffin.

    http://velvetgloveironfist.blogspot.com/2009/12/welsh-miracle-is-not-dead-it-just.html

  2. Ed says:

    There will never be a study that completely proves a correlation between smoking ban and heart attacks.
    But less people smoking = less smoking-related diseases.

Leave a Comment

*