Monday, January 16, 2006

Helena Claim of 40% Drop in Heart Attacks is Mathematically Impossible

I may have overstated the degree of plausibility of the claim that a smoking ban could reduce heart attacks by 40% within six months. My argument was that even if all smoking in Helena were completely eliminated, one would expect to see no more than a 50% reduction in heart attacks.

However, according to smoking-attributable risk estimates used by the Centers for Disease Control in its SAMMEC (Smoking-Attributable Morbidity, Mortality and Economic Costs) software, the proportion of heart disease deaths attributable to smoking is far less than 50% - it is estimated to be 17% (see: Mostashari F, Frieden TR. Unrealistic expectations can hinder efforts to expand smoke-free workplace laws. BMJ rapid response, June 9, 2004).

The CDC's estimate is based on the relative risk associated with smoking for death from heart disease in the CPS-II (Cancer Prevention Study - II) cohort study of more than 1,000,000 smokers conducted by the American Cancer Society, as well as on the prevalence of current and former smokers in the United States.

The Rest of the Story

If the CDC's attributable risk estimate of 17% is correct, then if all smoking were eliminated, one would actually expect to see no more than a 17% reduction in heart attacks (and that would probably not occur within six months, but we'll let that point slide by).

Thus, not only is the claim that the Helena smoking ban reduced heart attacks by 40% mathematically impossible, but so is the claim that the Pueblo smoking ban reduced heart attacks by 27%.

The rest of the story is that I was probably understating the case when I suggested that the Helena claim defies common sense and is implausible. The truth is that it is probably mathematically impossible, and so is the Pueblo claim.

As Mostashari and Frieden point out in their letter, we are probably setting ourselves up for failure if we make these types of exaggerated claims, because they are never going to come to fruition, and we don't want to hinder future efforts to promote smoke-free workplace laws by disappointing the expectations of policy makers who have been deceived into thinking that these types of dramatic results are even possible.

No comments: