Friday, 24 April 2015

Irish vaping mystery solved

On Monday, I mentioned the Irish Cancer Society's claim that 5 per cent of smokers had used an e-cigarette before they started smoking. This, they say, is evidence of a possible 'gateway' effect—an effect that has not been seen in any other country. For the avoidance of doubt, this is what they say on their website:

A concerning fact highlighted by our research was that 5% of current smokers used e-cigarettes before they started smoking. 

I have now found the survey results and there is a simple, predictable explanation for this finding. It's ain't true.

The five per cent statistic actually doesn't refer to all smokers, but to smokers who had used or were using e-cigarettes. The relevant data is below.

56 per cent of the smokers had ever used e-cigarettes so we're actually talking about half of one twentieth of one fifth of the population. Actually, you can't even halve the five per cent figure because the e-cigarette users are not representative of smokers as a whole. They are younger, for a start. If more old smokers tried e-cigarettes, the percentage would decline further.

The research did not find that "5% of current smokers used e-cigarettes before they started smoking". It found that 5% of an unrepresentative subset of smokers who had ever used e-cigarettes said that they had used e-cigarettes before they started smoking. To state the obvious, it was not a survey of every smoker in Ireland and the numbers are tiny. Although the survey started out with 1,000 people, there were only 176 able to answer this question. Of those 176 people, only eight gave the answer that formed the basis of the Irish Cancer Society's claim.

Eight! You cannot draw conclusions from such a tiny sample. You can't even be sure that these respondents used an e-cigarette before they started smoking. When dealing with numbers this small, you can't discount people misreading the question, not reading the question or ticking boxes at random. But even if they answered correctly, you cannot infer—as the authors do above—that it is an indication that "e-cigarettes are a [sic] probably gateway to smoking for this cohort."

Contrast those eight people with the thirty people who used to smoke but now only use e-cigarettes. That's thirty people from a smaller sample of 86 current e-cigarette users, meaning that 35 per cent of e-cigarette users used to smoke but no longer do so. Is that not more consistent with the view that e-cigarettes are a gateway from smoking than a gateway to smoking?

And yet the Irish Cancer Society seems to have little interest in the people who quit smoking, a strange oversight for a health group. Instead, their press release led with the claim that 'Two in three e-cigarette users are also smoking tobacco'. This, of course, means that one in three are not still smoking tobacco. Can the pharmaceutical companies that sponsor the Irish Cancer Society claim the same about users of their nicotine products?

No comments: