Tuesday, May 30, 2006

New Study Shows that Helena Conclusions About Effects of Smoking Ban on Heart Attacks are Scientifically Implausible

A new study published in the current issue of the journal Tobacco Control estimates the percentage of heart disease deaths that are attributable to smoking in 38 countries in the World Health Organization's Western Pacific and South East Asian regions (see: Martiniuk ALC, Lee CMY, Lam TH, Huxley R, Suh I, Jamrozik K, Gu DF, Woodward M. The fraction of ischaemic heart disease and stroke attributable to smoking in the WHO Western Pacific and South-East Asian regions. Tobacco Control 2006; 15:181-188).

For males, the estimated proportion of heart disease deaths attributable to smoking among males ranged from 13-33% and in females, it ranged from 1-28%.

The Rest of the Story

The overall prevalence of male smoking in these countries is substantially higher than it is in the U.S., and the overall prevalence of female smoking in these countries is considerably lower. Thus, the proportion of heart disease deaths in the U.S. is almost certainly lower for males and higher for females than that reported in the study.

However, using the highest reported attributable fractions for each sex, which correspond to a smoking prevalence of 82% for men and 80% for women, the maximum attributable fraction for heart disease deaths in the U.S. (where smoking prevalence is about four times lower) would be 30%. This appears to be a conservative assumption for what the maximum attributable fraction is (it would likely be lower than this in the U.S. because smoking prevalence is so much lower).

This means that at most, about 30% of heart disease deaths in the U.S. are attributable to smoking. If all smoking were eliminated completely, one would expect to see about a 30% decline in heart disease deaths.

Now here's my point: if eliminating smoking completely would only reduce heart disease deaths by 30%, how could a simple smoking ban in bars and restaurants reduce heart attacks by 40%? It just isn't scientifically plausible.

Yet this conclusion is being widely disseminated by anti-smoking groups, and forms part of the basis of a national anti-smoking group campaign to heighten the emotional appeal of secondhand smoke messages in order to support our agenda.

While I support the agenda, I do not support the use of shoddy science to promote that agenda. And this is shoddy science. If it's scientifically implausible that an effect can occur in the first place, then it's certainly shoddy science to conclude that the effect did occur, especially based on the results of a small study in which there is only one data point following the smoking ban and in which the entire conclusion is based on observing 16 fewer heart attacks.

No comments: