Monday, January 09, 2006

Helena Findings Defy Common Sense; Strongest Argument Yet For Prematurity of Study Conclusions

While I have presented several lines of reasoning (line 1 line 2 line 3) to explain why I think the conclusions from the Helena and Pueblo studies are premature (these studies concluded that there was a 40% and 27% reduction, respectively, in heart attack admissions within 6-18 months after a bar and restaurant smoking ban was implemented), I have yet to present what I think is the strongest evidence to date that the conclusions being widely disseminated regarding these studies are unwarranted and far too premature.

First, let's stipulate, for the purposes of this argument, that the smoking ban in Helena was so effective that it immediately resulted in the complete elimination of both smoking and secondhand smoke in Helena.

The question is: by what percentage would the number of observed heart attacks in Helena fall in six months?

I think the maximum possible proportion by which heart attacks could drop due to the complete elimination of all smoking is about 50% (and this is being liberal).

Consider the following:

1. After one year of quitting smoking, the risk of heart attacks in the most susceptible smokers (those who have already had a heart attack) drops by 50%.

2. It is estimated that among all people under 50, the proportion of heart attacks attributable to smoking is 50%. The proportion would be lower if the entire population were considered.

3. Smoking is estimated to cause approximately 170,000 heart attacks each year, out of a total of about 865,000 heart attacks, or a proportion of 20%. Even assuming that secondhand smoke causes a whopping additional 30% of all heart attacks, the proportion of heart attacks caused by smoking and secondhand smoke is not higher than about 50%.

I think it is relatively safe to say that if all smoking were eliminated, and therefore all secondhand smoke as well, we could at best expect to see a 50% decline in heart attacks.

Never have I seen any data that suggests that the majority of heart attacks are due to smoking and secondhand smoke. I think the highest plausible figure is about 50%.

The Rest of the Story

OK then. If the Helena smoking ban resulted in the complete and immediate elimination of all smoking, then at most, we would expect to see a 50% reduction in heart attacks.

So how can it possibly be the case that just banning smoking in bars and restaurants in Helena resulted in a 40% reduction in heart attacks?

Not only is the claim not supported by sufficient data that rules out the possibility that the observed decline in heart attacks was largely attributable to random variation in the data, but the claim flies in the face of common sense and plausibility.

I simply don't see any way that the elimination of secondhand smoke in bars and restaurants in Helena and any effect this had on reducing cigarette consumption could possibly have reduced heart attacks by 40% when the complete and total elimination of smoking and secondhand smoke exposure in Helena would at most reduce heart attacks by 50%.

Common sense (or more technically, plausibility) is one of the most important criteria that we use to judge the validity of scientific conclusions. In this case, it is simply not plausible, even under the most extreme of potential assumptions, that the smoking ban in Helena could have resulted in a 40% decline in heart attacks.

In other words, it is not plausible that 40% of the heart attacks that were occurring in Helena were attributable to either secondhand smoke exposure in bars and restaurants in the city or to smoking that was reduced or eliminated after the smoking ban.

There is really only one way in which this plausibility problem could be overcome. If the study (or a study) were to document some sort of unprecedented and unbelievable drop in smoking rates in Helena due to the smoking ban, then it would become plausible to observe this magnitude of an effect. For example, if it turned out that 75% of smokers in Helena quit smoking after the smoking ban, then the observed findings would become believable. Even if 50% quit, I think it would be believable.

But I am not aware that the study or any other study has indeed documented such drastic reductions in population smoking rates in Helena (or anywhere else that a smoking ban was implemented). In fact, I'm not aware of any data that has documented that there was any reduction in smoking in Helena following the smoking ban.

The lack of plausibility of the Helena claim is, I want to emphasize, not the only reason why I believe it is premature, if not completely invalid. Rather, it is the combination of the lack of plausibility with the plausible alternative explanations that have been offered to account for the findings in light of the severe limitations in the study design, in particular, the possibility that the observed finding was simply reflecting random variation in the heart attack trends.

Finally, I want to re-iterate that I am a proponent of smoking bans, and I do not opine that the invalidity of the Helena and Pueblo conclusions means that smoking bans are not justified. In fact, I don't think it has anything to do with that question, because I wouldn't expect to see a dramatic decline in heart attacks in association with any smoking ban (for the reasons outlined above).

The reason I think this is important is because it goes to the very credibility of the tobacco control movement. If we continue to go around spouting out these claims, eventually the public is going to catch wind of the fact that we are blowing a bunch of wind, especially in light of the common sense conclusion that there is no way a smoking ban could reduce heart attacks by 40%. As a result, our credibility will be hurt.

And the problem is that in the future, I'm afraid the public is not going to be able to discern whether it is the case that we are making exaggerated claims or solid ones. Once we lose our credibility, there may not be an easy way to regain it. And that will jeopardize and undermine our future efforts to promote smoking bans because no one will believe any of the claims we make, even the ones that are indeed justified and well-supported by the scientific evidence.

No comments: