Wednesday, September 28, 2011

State-Wide Smoking Bans: Little to No Impact on Heart Attack Deaths

My research group just published a study documenting that the implementation of state-wide smoking bans in California, Utah, Delaware, South Dakota, New York and Florida had little or no immediate measurable effect on deaths from heart attack (acute myocardial infarction, AMI). The study was published in the Journal of Community Health (here).

Nicholas Peiper, Phil Cole and I studied the AMI death rate in those states before and after smoking bans were implemented, comparing those rates to historical and national trends. The results are illustrated in the figure.

The state-wide bans had no effect in California, Utah, Delaware or South Dakota (the AMI death rate actually increased almost 9% in South Dakota during the target year). But after bans in Florida and New York in 2003, AMI death rates declined 9% and 12% respectively, which exceeded the expected declines based on historical trends in those states. However, they were not significantly different from the 10% decline in the 44 states that did not have smoke-free ordinances in 2004.

Since 2004 several published reports have claimed that smoking bans in small cities result almost immediately in reductions in heart diseases. For example, in Helena, Montana hospital admissions for acute myocardial infarction (AMI) declined 40%, from 40 before to 24 after implementation of a smoke-free ordinance (here). In Pueblo, Colorado, AMI admissions dropped 27%, from 399 before to 291 over a 1.5-year period following a ban, and the authors reported that the decline occurred “within months” of implementation (here). In Bowling Green, Ohio, hospital admissions for ischemic heart disease and heart failure fell from 36 before to 22 after a ban (here), and in Monroe County Indiana, hospital admissions for AMI among non-smokers declined from 17 before to 5 after a ban (here).

But there are two problems with these claims. First, it’s dangerous to make big claims based on tiny numbers. Phil Cole and I presented evidence that the Helena and Pueblo findings “are consistent with random variation because of the small number of observations on which they are based.” (here).

Second, none of the previous reports accounted for the long-term downward trend in heart disease in the U.S. In our study smoking bans might have taken the credit for the 9% decline in Florida and the 12% drop in New York if the historic drop in AMI that occurred throughout the U.S. in 2004 was ignored. The figure above shows the impressive decline in American heart attack deaths since 1991.

We concluded that “Smoke-free ordinances may serve public health objectives by providing non-smokers with indoor environments that are free from irritating and potentially harmful pollutants. However, this study does not provide evidence that these ordinances result in a measurable immediate reduction in AMI mortality of the magnitude claimed by reports based on very small incident numbers.”

No comments: