Well this has been an interesting experience for me. Thanks to this post by Ben Goldacre, my “Behind the Smokescreen” post has received around 50 times the traffic I usually get in my quiet little corner of the blogosphere. For the sake of new readers, I should clarify that I am not Ben Goldacre. I think this was clear enough from his own post in which he (quite reasonably) made the disclaimer that he hadn’t fact-checked my fact-checking, but some of the other places around the web that have picked this story up have made this ambiguous. I don’t really care about getting credit, it’s more a case of not wanting my perspective misattributed to someone who might not 100% agree with me or my interpretation.
Speaking of other places around the web that have picked this up… all most people seem to care about is the “23 times” issue. To me, that’s actually the least interesting bit. Yes it was an error, and needed correction, but it’s been made by so many other people that I don’t think the BMA deserve to be pilloried for that, especially when they promptly corrected it. What bothered me more about the report were two things. Firstly, the big gaps in between the stated facts (i.e. X proportion of smokers smoke in cars, Y proportion of children’s journeys are in cars, ergo children are at “significant” (but unquantified) risk of exposure to second hand smoke). This seems to me to be both a general and ecological fallacy. The second is the misuse of data about in-vehicle smoking as a driving hazard, which took a bunch of papers about one thing and claimed they showed evidence about something else.
Nowhere have I seen any defence of the use of the evidence in this report, only the argument that ‘we know enough about the harms of smoking already to take this action’. My view remains that knowing that smoking is harmful is not a justification to recommend a resource-heavy intervention that might have little health benefit, when those resources might be better deployed on an alternative, more effective, intervention. Either way, I dislike seeing something packaged as evidence based when the evidence has been misreported.