Since 2003, when activists began claiming that smoking bans immediately reduce heart attacks by as much as 47 percent, they have faced two serious credibility problems. First, based on what we know about heart disease rates in current and former smokers, such effects are wildly implausible. Second, given the large number of jurisdictions with smoking bans, it is easy to create a misleading impression by focusing on the places where heart attacks happen to fall.
Judging from the first nationwide study of the question, reported in the November Journal of Policy Analysis and Management, that is exactly what ban boosters have done. "In contrast with smaller regional studies," write RAND Corporation researcher Kanaka Shetty and three co-authors, "we find that smoking bans are not associated with statistically significant short-term declines in mortality or hospital admissions for myocardial infarction or other diseases." In fact, "An analysis simulating smaller studies using subsamples reveals that large short-term increases in myocardial infarction incidence following a smoking ban are as common as the large decreases reported in the published literature."
In other words, although heart attacks do decline in some places with smoking bans, there are just as many places where they rise. On average, the difference between jurisdictions with smoking bans and jurisdictions without smoking bans is essentially zero.
So how could a 2009 report from the Institute of Medicine find "consistent" support for a link between smoking bans and short-term reductions in heart attacks? Shetty et al. suggest that publication bias accounts for the illusory association cited in the report: Data indicating decreases in heart attacks were apt to be reported, while data indicating increases were ignored, even though the two patterns were equally common. For reasons that are unclear, the 2009 report's authors did not consider Shetty et al.'s study, which was available at the time as a working paper from the National Bureau of Economic Research.