Yesterday, I discussed a new study published online ahead of print in the journal Archives of Internal Medicine which concluded that smoke-free restaurant and bar ordinances in Olmsted County, Minnesota produced a 34% decline in heart attacks and a 17% decline in sudden cardiac deaths. I pointed out that there was no comparison group, so that one cannot determine whether the observed declines in Olmsted County differ from secular declines that may have occurred elsewhere in Minnesota during the same time period. However, I presented data from the HCUP database which showed that in Minnesota as a whole, there was a 26% decline in heart attacks during the identical time period. This calls into question the validity of the study's conclusion that the decline in heart attacks was due to the smoking ban.
In addition, I pointed out that the observed decline in sudden cardiac deaths was not statistically significant and that the authors appeared to be biased in their dismissing a similarly non-significant 17% increase in sudden cardiac deaths which occurred after the restaurant smoking ban.
Today, I present data from the CDC's National Center on Health Statistics (NCHS) compressed mortality file, which allows to examine heart attack death rates in Olmsted County during the study period.
I thank Dr. Brad Rodu of the University of Louisville for running and sharing these analyses with me.
The Rest of the Story
According to the CDC data, between 1999 and 2001 (prior to the restaurant smoking ban), the age-adjusted rate of heart attack deaths in Olmsted County dropped from 95.9 to 61.8, a decline of 35.6%.
Between 2001 and 2003 (the first two years of implementation of the restaurant smoking ban), the heart attack death rate in Olmsted County declined from 61.8 to 59.4, a decline of 3.9%. Even if one goes out to 2005 to allow a longer time for the death rate to drop, the decline from 2001 is only 24.9%, less than the drop in heart attack rates prior to the restaurant smoking ban.
Thus, these data refute the conclusion that the restaurant smoking ban led to a decline in heart attack death rates in Olmsted County.
Between 2006 and 2009 (the first three years of the bar smoking ban), the heart attack death rate in Olmsted County increased from 35.0 to 41.7, an increase of 19.1%.
Thus, these data refute the conclusion that the bar smoking ban led to a decline in heart attack deaths.
Overall, from 2001 to 2009, the decline in the heart attack death rate in Olmsted County was 32.5%, which is a slower rate of decline than occurred in the pre-smoking ban period of 1999-2001.
Thus, it appears that overall, there was a deceleration of the existing high rate of decline in heart attacks in Olmsted County during the study period. These data provide no evidence that the smoking bans led to a decline in heart attack deaths and therefore cast further doubt on the conclusions of the Annals of Internal Medicine study.
Here is a graph of these data, produced by Dr. Rodu:
The pattern of decline in heart attack deaths in the state of Minnesota as a whole was similar to that in Olmsted County. Thus, my overall impression is that Olmsted County smoking bans probably had no significant impact on heart attack deaths.
The rest of the story is that these data do not support the study's conclusion that the smoking bans in Olmsted County produced a dramatic decline in heart attacks, as the authors of the study concluded.
Again, I have to emphasize that I would like nothing more than to find a dramatic effect of smoking bans on heart attacks in the short-term because I have devoted a substantial portion of my career to promoting smoke-free laws. However, if the data do not support such a conclusion, they don't support such a conclusion and we need to refrain from drawing the conclusion that we favor simply because we want to see that our work is having an effect. I believe that bans on smoking in bars, restaurants, and workplaces, are justified regardless of whether they have an immediate effect on heart attacks. In fact, I would not expect these policies to have dramatic, immediate effects on heart attacks because it takes many years for the atherosclerotic process to occur. So I would not expect to see an impact on heart disease rates until many years following the implementation of these laws. What we would expect to see more rapidly are changes in respiratory disease and respiratory symptoms, as a number of studies have demonstrated.
The danger here is that if we lose our scientific integrity, we lose the public's trust. If we exaggerate or distort the science, we may undermine our public reputation. Once that is lost, it is almost impossible to regain. It seems a shame to sacrifice the public's trust because we want to be able to present more data that is favorable to our cause, even though I believe the cause is a legitimate and important one. In my view, the ends do not justify the means if those means involve sacrificing our scientific integrity.