Thursday, May 20, 2010

New Study Concludes that Smoking Ban in Arizona Decreased Heart Attacks, Despite Increase in Heart Attacks in Most of the State

A new study published online ahead of print in the American Journal of Public Health concludes that the statewide smoking ban in Arizona caused a reduction in acute myocardial infarction [AMI] (heart attack) admissions in the state (see: Herman PM, Walsh ME. Hospital admissions for acute myocardial infarction, angina, stroke, and asthma after implementation of Arizona's comprehensive statewide smoking ban. American Journal of Public Health 2010).

The study methods were as follows: "We compared monthly hospital admissions from January 2004 through May 2008 for these primary diagnoses and 4 diagnoses not associated with SHS [secondhand smoke] (appendicitis, kidney stones, acute cholecystitis, and ulcers) for Arizona counties with preexisting county or municipal smoking bans and counties with no previous bans. We attributed reductions in admissions to the statewide ban if they occurred only in diagnoses associated with SHS and if they were larger in counties with no previous bans."

The results were reported as follows: "Statistically significant reductions in hospital admissions were seen for AMI, angina, stroke, and asthma in counties with no previous bans over what
was seen in counties with previous bans."

The study concludes: "Arizona’s statewide smoking ban decreased hospital admissions for AMI, stroke, asthma, and angina."

The authors also conclude that the smoking ban resulted in a cost savings of $16.8 million during its first 13 months due to reduced hospital admissions for cardiovascular disease and asthma.

The Rest of the Story

Although it would be a wonderful thing if the smoking ban led to such a drastic decline in cardiovascular disease admissions, the problem is that for 83% of the state's population, there was a significant increase in heart attacks associated with the implementation of the statewide smoking ban, according to the study.

For the counties in which a workplace smoking ban, but not necessarily a bar or restaurant smoking ban, was in place in at least one city during the period 2004-2007 -- which make up 83% of Arizona's population -- the study fails to find any decline in heart attacks. In fact, it reports a significant increase in heart attacks associated with the implementation of the statewide smoking ban in 2007-2008 (see Figure 1a). In these counties, the rate of heart attack admissions - which had been steadily dropping prior to the smoking ban - actually started to increase after the statewide smoking ban. And again, these counties cover the overwhelming majority of the state's population.

While the investigators appear to dismiss this problem by arguing that the statewide smoking ban would not be expected to lower heart attacks in the "ban counties" because they already had smoking bans in place, this explanation is not plausible. The ban counties were those which merely had to have one city with just a workplace smoking law. These are not counties which had stringent, comprehensive smoking bans that included all hospitality establishments, like restaurants and bars. Exposure to secondhand smoke in typical workplaces (non-hospitality workplaces) is quite low and one would not expect that a workplace smoking ban that excluded bars and restaurants would have much effect on reducing secondhand smoke exposure.

Moreover, the extension of these bans to include bars and restaurants would - according to the conclusions of anti-smoking groups - result in a significant decline in heart attacks.

By including any smoking bans, not just meaningful laws that eliminate smoking in bars and/or restaurants, in the comparison group, the study managed to produce an "intervention" group that was terribly unrepresentative. These were communities that were so far behind the times (and probably, with a large proportion of smokers in their populations) that not a single city in their jurisdictions had even enacted a law to protect workers in office workplaces.

In fact, the study acknowledges that the division of counties into "ban counties" and "non-ban counties" was essentially a division of counties into "urban" versus "rural" regions, respectively. There are so many other differences between urban and rural regions that one cannot validly use urban regions as a comparison group to estimate the secular trends in heart attacks in rural regions. But this is exactly what this study did.

The flaw of the study, then, is that it used a comparison group that was not comparable. It used urban regions as a comparison group to estimate the baseline trends in heart attacks in rural regions. The lack of similarity between these groups renders the study methodology flawed.

There are a number of reasons why rural regions would be expected to lag behind urban regions in terms of the secular decline in cardiovascular disease that has been observed in recent years. Differences in access to treatment, higher smoking rates, and differences in other risk factors for cardiovascular disease would all be expected to play a role. What we are possibly observing in this study -- and a plausible, alternative explanation for the findings that the study cannot rule out -- is that cardiovascular disease improvements that had already occurred in urban areas started to finally spread to rural regions. The main point is that the study is not capable of ruling out this possibility; thus, it cannot validly conclude that the observed decline in cardiovascular disease in the rural areas was attributable to the statewide smoking ban.

What is interesting to me is that although the study attributes the decline in heart attacks in rural regions to the smoking ban, it refuses to attribute the increase in heart attacks observed in urbane regions to the smoking ban. Conveniently, it offers another explanation: there must be some other factor involved.

But if there is some other factor that explains the increase in heart attacks in urban regions, could not that same other factor explain the decrease in heart attacks in rural regions. You can't set forward a method and that reject that method when the findings don't come out the way you would like. But that's exactly what this study does.

The study argues: "The statistically significant increase seen in AMI would be even harder to explain as being caused by the statewide ban. ... These changes are likely because of an increase in causal factors for AMI or for its diagnosis other than the statewide ban."

In other words, the observed effects in rural counties (favorable) are attributable to the smoking ban, but the observed effects in urban counties (unfavorable) are not. This is clearly post-hoc reasoning that is not consistent with the methods set out in the paper.

The rest of the story is that this study failed to find any effect of the statewide smoking ban on heart attacks among 83% of the population of Arizona, despite the fact that the overwhelming majority of that population was not covered by smoke-free bar and restaurant ordinances prior to the statewide law. The study is thus inconsistent with the conclusions of anti-smoking groups that smoke-free restaurant laws result in an immediate and dramatic decline in heart attacks.

Once again, I would love nothing more than to see documentation of an immediate decline in heart attacks due to smoking bans, since I have devoted a significant portion of my career to working for such laws. However, the existing evidence simply doesn't support such a conclusion. And this study is no exception. In fact, if anything, this study provides evidence that no such effect occurred in response to the Arizona smoking ban.

Of course, this doesn't mean that the smoking bans are not justified. It simply means that anti-smoking advocates should not make false promises about the immediate declines in heart attacks that will result from such policies.

No comments: