Monday, November 17, 2008

Massachusetts Study Shows Lower Smoking Prevalence Associated with Reduced Heart Attack Rates, But Doesn't Show Any Immediate Effect of Smoking Bans

After careful examination of the study released last week by the Massachusetts Department of Public Health, it is clear that the study fails to show any immediate decline in heart attacks following passage of the statewide smoking ban. The study does demonstrate, however, that declining smoking prevalence is closely tied to declines in heart attack rates. Thus, while the study documents that interventions - such as smoking bans - which reduce smoking prevalence will eventually result in declines in heart attacks, it does not provide any evidence that smoking bans will immediately reduce heart attacks. This study refutes, rather than supports the conclusion of studies from Helena, Pueblo, Bowling Green, and several other sites.

The Rest of the Story

The study was designed primarily to examine whether the statewide smoking ban, implemented on July 5, 2004, resulted in an immediate drop in heart attack deaths. In the first year after the smoking ban was implemented, there was no significant decline in heart attack deaths in the state. Moreover, there was no decline in heart attack deaths even among just those residents living in towns that did not previously have smoking bans. Thus, this study refutes the conclusions from Helena, Pueblo, Bowling Green, etc. that smoking bans immediately reduce heart attacks by decreasing secondhand smoke exposure.

The study did find a decline in heart attack rates from the first to second year after the statewide smoke-free law was implemented, but it turns out that the magnitude of this decline was not significantly different in towns with or without smoking bans prior to the state law. Thus, the study provides no evidence that the statewide smoking ban was associated with any significant decline in heart attacks, even up to two years after its implementation.

While not the primary purpose of the study, it did incidentally find that heart attack death rates were significantly lower, over the long-term period of 1999-2006, among residents of towns with smoking bans compared to residents living in towns without such bans. The mean smoking prevalence in these towns was substantially lower in these towns (17.6% vs. 20.1%). These data provide evidence that smoking prevalence is (not unexpectedly) closely related to heart attack death rates. From this, one can conclude that the observed declines in heart attack rates over time is due, in part, to the observed declines in smoking prevalence. It should be emphasized, however, that these are long-term effects that occur over a long period of time, not the six months to one year that were touted in the Helena et al. studies.

The rest of the story, then, is that the Massachusetts study provides no evidence that smoking bans are associated with an immediate decline in heart attacks and that it in fact refutes the conclusions of previous studies that heart attack rates decline within the first six months to one year following implementation of smoking bans. The study does, however, demonstrate that declines in smoking prevalence will, over time, result in declines in heart attack deaths.

The lack of solid scientific data that smoking bans lead to immediate reductions in heart attacks is not stopping anti-smoking advocates from making such a claim. In a press release from the Henry Ford Health System, the lead author of a new study on this very topic stated: "If Michigan were to implement a comprehensive smoking ban tomorrow, we would see a 12 percent drop in heart attack admissions after the first year."

Note that this study is simply a review of the previous flawed literature (Helena et al.). It provides no new evidence. Unfortunately, the statement by this researcher is unsupported by the scientific evidence. But as I have noted recently, anti-smoking advocates are now holding themselves to very low scientific standards. What now passes for science is what in the past we would have called an interesting anecdote.

One final point deserves emphasis. The release of the Massachusetts study represents yet another example of "science by press release." I do not understand why the researchers didn't wait until the study was accepted for publication and then release the results when the article was published, which is the usual practice. Since the study has been submitted for publication, it would have made sense to wait until publication to release the findings. In fact, many journals will not allow researchers to release a study to the public while it is under review.

If the study was not being submitted for publication, it would be a different matter (assuming that the full study was made available for public review). But because the study is under review at a journal, it cannot be released publicly. Thus, there is no opportunity for scrutiny and review of the study conclusions.

It is very interesting to me that it is always these heart attack and smoking ban studies that are released prior to scientific review. It almost makes one think that the researchers are subconsciously aware that the conclusions are shoddy and that they want to get press coverage of their hoped-for conclusions before the results are actually held up to high scientific standards. I have to admit that it's a great strategy, because once the results are disseminated through the media, it is too late to retract or clarify them later.

No comments: