Thursday, October 15, 2009

Institute of Medicine Report's Conclusions on Smoking Ban Effects are Defied By Its Own Assertions; Study Conclusions, Press Release Severely Biased

A new report from the Institute of Medicine's Committee on Secondhand Smoke Exposure and Acute Coronary Events, entitled "Secondhand Smoke and Cardiovascular Effects: Making Sense of the Evidence," concludes that smoking bans result in a nearly immediate and significant decrease in heart attacks, not only among smokers but among nonsmokers as well. The committee also concluded that brief exposure to secondhand smoke causes heart attacks and refused to qualify its conclusion by noting that such an effect is substantial only in those with severe existing heart disease.

According to the press release: "Smoking bans are effective at reducing the risk of heart attacks and heart disease associated with exposure to secondhand smoke, says a new report from the Institute of Medicine. ... 'It's clear that smoking bans work,' said Lynn Goldman, professor of environmental health sciences, Johns Hopkins Bloomberg School of Public Health, Baltimore, and chair of the committee of experts that wrote the report. 'Bans reduce the risks of heart attack in nonsmokers as well as smokers.'"

The Rest of the Story

Unfortunately, this report might just as well have been called: "Secondhand Smoke and Cardiovascular Effects: Making Nonsense of the Evidence."

The reason for this assertion is two-fold:

First, the conclusions of the report are completely defied by the committee's own assertions that are presented in the actual report.

Second, the report draws conclusions that are essentially meaningless from an epidemiologic and clinical perspective. What the report does is take important questions and distort them so much that the answers no longer have any meaning.

Let me address each of these problems in turn.

To see what I mean about the conclusions of the report not being consistent with the report itself, consider first what the report concludes about the ability, based on the existing evidence, to estimate the magnitude of the effect of smoking bans on heart attack rates.

The report asserts as follows: "The committee was unable to determine the magnitude of effect on the basis of the 11 studies, because of variability among and uncertainties within them. Characteristics of smoking bans vary greatly among the locations studied and must be taken into account in reviewing results of epidemiologic studies. Those characteristics include the venues covered by the bans (such as offices, other workplaces, restaurants, and bars) and compliance with and enforcement of the bans. Other differences or potential differences among the studies include the length of followup after implementation, population characteristics (such as underlying rates of acute coronary events and prevalence of other risk factors for acute coronary events, including diabetes and obesity) and size, secondhand-smoke exposure levels before and after implementation, preexisting smoking bans or restrictions, smoking rates, and method of statistical analysis. The time between implementation of a ban and decreases in secondhand smoke and acute cardiovascular events cannot be determined from the studies, because of the variability among the studies and indeed the difficulty of determining the precise time of onset of a ban."

The report also asserts: "However, because of the weaknesses discussed above and the variability among the studies, the committee has little confidence in the magnitude of the effects and, therefore, thought it inappropriate to attempt to estimate an effect size from such disparate designs and measures."

In other words, what the committee is saying is they have no confidence in making any estimate of the size of an effect of smoking bans on heart attack rates. Another way to say that is this: the committee has no idea of what the effect of smoking bans on heart attacks is.

If you can't even estimate the magnitude of an effect - if you have no confidence in even providing an estimate - then you are hardly in a position to conclude that there is a significant effect of smoking bans on heart attacks, an effect which exceeds random variation combined with the known secular decline in heart attack rates.

Think about this: we know for a fact that heart attack rates are declining substantially, even in the absence of smoking bans. These declines are in part attributable to improvements in the treatment of coronary disease and also to improved medications, such as the statin drugs which are effective in controlling cholesterol levels. When we see a decline in heart attacks after a smoking ban, we need to determine whether the magnitude of that decline is greater than one would expect in the absence of the smoking ban. In other words, does the observed decline exceed the rate of decline one would expect from the secular changes alone?

In order to make such a determination, one needs to quantify the magnitude of the decline in heart attacks. If we can't even estimate, with any confidence, what the magnitude of the decline in heart attacks is, then we are in no position to conclude that we know that the decline is greater than what would have been observed in the absence of the smoking ban. We can't conclude that the observed decline in heart attacks associated with smoking bans has been due to the smoking ban, rather than to the rather drastic declines in heart attacks that have been occurring anyway due to improvements in medical treatment.

Epidemiology is all about estimating the magnitude of effects. Simply judging whether an association works in one direction or the other is not particularly meaningful, especially in this situation where we know a priori that smoking bans do not increase heart attacks.

Now this is where my 2nd observation comes in. By answering the question: do smoking bans reduce or increase heart attacks, the report is actually making nonsense out of the evidence. Of course smoking bans don't increase heart attacks. The question is: what is the magnitude of the effect.

The committee recognizes that the existing studies are so seriously flawed that one has no confidence in being able to judge the effect size. But instead of concluding that the evidence is insufficient, they go ahead and conclude that smoking bans significantly reduce heart attacks anyway.

A second example is the press release's conclusion about whether the observed reductions in heart attacks occur in smokers or nonsmokers. The report asserts: "Only two of the studies distinguished between reductions in heart attacks suffered by smokers versus nonsmokers." Later, it emphasizes this point: "In most of the studies, the portion of the effect attributable to decreased smoking by smokers as opposed to decreased exposure of nonsmokers to secondhand smoke cannot be determined."

Clearly, this is not sufficient evidence to draw a conclusion about whether the observed reductions in heart attacks are due to reduced active smoking or reduced tobacco smoke exposure among nonsmokers.

Nevertheless, the press release states: "Bans reduce the risks of heart attack in nonsmokers as well as smokers."

So much for requiring evidence before drawing a conclusion.

Perhaps the problems I am discussing are most evident in the report's conclusion regarding the effects of brief secondhand smoke exposure on heart attack risk. Based on the evidence, no one would deny that a brief exposure might trigger a heart attack in a person with severe existing coronary artery disease.

But the report goes beyond that in its conclusion. It states that brief secondhand smoke exposure may trigger heart attacks, but without qualifying that statement to make it clear that it refers specifically to people who have coronary disease. Instead, it makes it sound like a healthy person could walk into a smoky bar, sit down for 20 minutes, and keel over from a heart attack.

Why is this qualification not added to the study conclusion?

I believe it's because the report aims to be more sensationalistic and scare people into thinking that they could drop dead from a heart attack from a brief tobacco smoke exposure, even if they are healthy.

But failing to qualify the statement turns the conclusion from being accurate to being inaccurate, from being truthful to being misleading.

What it really means is that a political goal, not a purely scientific one, is driving the report's conclusion regarding the acute cardiovascular effects of tobacco smoke exposure. I find this unfortunate because it really taints the scientific integrity of the tobacco control movement.

I should probably add that if you read the report carefully, it actually makes the assertion that brief secondhand smoke exposure can appreciably increase the risk of heart attack among healthy people. The report states: "The data provide evidence that it is biologically plausible for secondhand smoke to be a potential causative trigger of acute coronary events. The risk of acute coronary events is likely to be increased if a person has preexisting heart disease."

I read this as asserting that brief secondhand smoke exposure triggers heart attacks among people with and without existing heart disease, but that the risk is higher for those with existing heart disease. I do not believe there is any evidence to suggest that such an assertion is true. The report provided no evidence that a healthy person may suffer a heart attack from a mere 20 to 30 minutes of secondhand smoke exposure.

There is one other major problem with the report that deserves mention, especially since I think it indicates a bias of the report.

The report claims to have reviewed unpublished data and to have attempted to identify unpublished studies that might have found no effect of smoking bans on heart attacks. The report states that "no such studies were identified." I find this difficult to believe, especially since I was a reviewer of the report and I made the committee aware of several unpublished analyses which documented no significant effect of smoking bans on heart attacks. Such studies were conducted in England, Scotland, and Wales. Furthermore, a large but unpublished study of all communities in the United States reported no effect of smoking bans on heart attacks, but this study was ignored by the report.

Note that the latter study, the largest of its kind, concluded that: "In contrast with smaller regional studies, we find that workplace bans are not associated with statistically significant short-term declines in mortality or hospital admissions for myocardial infarction or other diseases."

It is unfortunate that this study was ignored. I don't see how the review can be considered to be comprehensive if it threw out or ignored all the studies that failed to find an effect, but included, without question, all studies that found an effect, even if these studies failed to include a comparison group which is crucial to being able to infer whether the observed decline in heart attacks was attributable to the smoking ban.

Finally, I want to make it very clear that I am not impugning the integrity of the committee or any of its members. I don't think they've done anything wrong. I just think that the report is biased and that subconsciously, there was some sort of pressure operating which led to the report drawing conclusions that were not appropriate given the report's own assertions and review of the evidence. I also think this bias led to the report distorting the questions which it asked and failing to directly answer the questions (rather than distorting them so that the "answer" came out more "favorably").

No comments: