31 January 2012

Unhealthful News 199 - Good study, expected boring result, no coverage (smoking bans and heart attacks)

My friend Michael Marlow just published a solid (i.e., economics quality rather than public health quality) analysis of the effect of implementing place-based smoking bans (which is mostly restaurants and bars for this time period) on heart attacks.  He used data from the entire USA and a long period, which avoids the cherrypicking problem that is common in this literature.  He found that the bans had no detectable effect.  This is not surprising to anyone expert in the area, since all of the high-quality studies on the topic have found the same thing.  (If you wish to become expert, I suggest reading back through Chris Snowdon's posts on the topic -- his blog is probably the best repository of information and analysis on the topic.)

However, this result presumably would presumably be surprising to most of the people involved in political discussions about bans -- implementing new ones, posturing about the importance of existing ones, trying to justify going even further, etc.  It might tend to deflate some of the hype that is still influencing policy.

So with that in mind, what is striking is the news coverage of Marlow's article: as far as I can tell, none whatsoever.  Striking, but not surprising.  This is, of course, a perfect storm for not appearing in the news: a workaday analysis that the author did not try to overhype (in particular, he was incredibly modest and restrained in suggesting the better-known results suggested there was an effect, explaining how those authors engaged in the worst behaviors of of politicized junk science without ever actually saying so); a non-dramatic result -- in other words, the real core of science rather than some wild flight of fancy; and a result that displeases those who control or influence most of the press.

Indeed, this observation about the unhealthfulness of the the press is almost so boring that it makes for probably the most boring and shortest #UnhealthfulNews post ever.  But someone needed to mention this.

2 comments:

  1. "almost so boring that it makes for probably the most boring and shortest #UnhealthfulNews post ever"

    Funny and true!

    Help me understand Carl, did someone or some organization anticipate the smoking ban would reduce the incidence of heart attacks?

    Mike Bluestone

    ReplyDelete
    Replies
    1. Mike,

      The incredulity that I detect in your question is justified. I always have to decide how much background to provide, based on guesses about what my readers have already read. This one was a little thin, of course, so let me fill in.

      It is certainly the case that no honest person with a clue about the science predicted that. Even if you imagined a scenario where second-hand smoke exposure dropped from 1970s levels to zero, in order to detect the effect you would have to (a) believe the worst-plausible-case scenarios about the effects of SHS and (b) believe that you could design an instrument that would detect the resulting <1% predicted change. To detect the effect of merely reducing exposure in restaurants and bars, assuming there is any effect at all, is obviously impossible.

      I am not sure how aggressive the claims were by those campaigning for the first of these bans (as I mentioned, Snowdon is the go-to source for more information). But now the claims are common. This is because the anti-smoking activists started putting out "research" that claimed miracle-level effects. The early tactics were (a) to pick one small population that happened to have a random drop in heart attacks, and then choose start and stop times for the data that further exaggerated the change, and (b) to look at changes over a longer time and attribute the changes after the ban to the ban (ignoring the fact that heart attack rates have, for various reasons, dropped hugely, most everywhere in the West, over the last couple of decades). When the calls of "bullshit" became too much for that, they started concocting statistical models that basically were designed to "predict" that the drop in heart attacks would have been less were it not for the ban.

      These "studies" were utter junk science, and I don't just mean by the standards of good science. They were low quality, even by the standards of public health, and they were dishonest, even compared to typical anti-tobacco work.

      While several good studies (and thousands of analyses of the analyses) have debunked the nonsense, the claims big effects still are common in the political realm.

      Delete

Note: Only a member of this blog may post a comment.