I think I found the study they’re talking about thanks to this article. I might take a look at it—if the methodology is literally just ‘smoking was banned, then the heart attack rate dropped’, that sucks.
(Edit to link to the full study and not the abstract.)
Just skimmed it. The methodology is better than that. They use a regression to adjust for the pre-existing downward trend in the heart attack hospital admission rate; they represent it as a linear trend, and that looks fair to me based on eyeballing the data in figures 1 and 2. They also adjust for week-to-week variation and temperature, and the study says its results are ‘more modest’ than others’, and fit the predictions of someone else’s mathematical model, which are fair sanity checks.
I still don’t know how robust the study is—there might be some confounder they’ve overlooked that I don’t know enough about smoking to think of—but it’s at least not as bad as I expected. The authors say they want to do future work with a better data set that has data on whether patients are active smokers, to separate the effect of secondhand smoke from active smoking. Sounds interesting.
I think I found the study they’re talking about thanks to this article. I might take a look at it—if the methodology is literally just ‘smoking was banned, then the heart attack rate dropped’, that sucks.
(Edit to link to the full study and not the abstract.)
Just skimmed it. The methodology is better than that. They use a regression to adjust for the pre-existing downward trend in the heart attack hospital admission rate; they represent it as a linear trend, and that looks fair to me based on eyeballing the data in figures 1 and 2. They also adjust for week-to-week variation and temperature, and the study says its results are ‘more modest’ than others’, and fit the predictions of someone else’s mathematical model, which are fair sanity checks.
I still don’t know how robust the study is—there might be some confounder they’ve overlooked that I don’t know enough about smoking to think of—but it’s at least not as bad as I expected. The authors say they want to do future work with a better data set that has data on whether patients are active smokers, to separate the effect of secondhand smoke from active smoking. Sounds interesting.