July 31, 2005
An article on Fox News website on 7/30/05 claims the effects of secondhand tobacco smoke (ETS—environmental tobacco smoke) on the heart are “rapid and large.” It says ETS increases the risk of heart disease by about 30 percent. That may be true, but 30 percent of almost nothing is still almost nothing.
A 30 percent increase means a relative risk (also known as risk ratio, or RR) of 1.3. (On the risk scale, zero risk is set at 1.0, not 0.0.) Actually, the American Heart Association website lists the following RRs: 1.25 for cardiovascular heart disease, 1.18 for ischemic heart disease, and 1.13 for arrhythmic heart failure or coronary arrest mortality. The risk of getting cancer from drinking municipal tap water that tens of millions of Americans drink every day is 2.0 to 4.0. So why be concerned about a relative risk of 1.3 to the heart from ETS? It so happens that 1.3 is the exact RR for shortening your life by drinking three cups of coffee per WEEK. That will give you some perspective on the severity of the alleged heart “danger.”
Both the World Health Organization and the American Cancer Society have clearly stated that RRs below 2.0 are too low to be relied upon. And a report by the independent health consulting firm Littlewood and Fennell characterized RRs less than 2.0 as “dancing on the tiny pinhead of statistical insignificance.” Compare this to the claim of “rapid and large” harmful effects from an RR of 1.3. A wealth of published literature dismisses relative risks less than 2.0 (100%) as being insignificant. And Dr. Eugenia Calle, Director of Analytic Epidemiology at the American Cancer Society, has stated the RRs below 1.3 are too low even to be realistically identified, much less be dangerous.
Why aren’t RRs less than 2.0 significant when they can represent impressive sounding percentage increases? The main reason is confounding variables. There are at least 20 of these that have been identified for ETS and heart disease, including: heredity, consumption of fat, consumption of fruits and vegetables, exercise and physical activity, type of employment, ethnic background, cholesterol, socio-economic class, etc. Any one of these could account for an impressive percentage increase in disease, yet no study of ETS has ever come close to controlling for even a large share of these variables. And there could be others that haven’t yet been identified.
It is no wonder, therefore, that Dr. Marcia Angell, editor of the New England Journal of Medicine, one of the world’s leading medical journals, says, “As a general rule, we are looking for a relative risk of 3.0 or more.” Dr. Robt. Temple, director of drug evaluation for the FDA, says, “My basic rule is if the relative risk isn’t at least 3 or 4, forget it.” And the EPA declined to regulate high-voltage power lines because it said the RRs seldom exceeded 3.0.
So how did we go from regarding RRs of 3.0 to 4.0 as the bare minimum for even potential concern to where anything 1.0 to 2.0—which was formerly regarded as insignificant—is now claimed to be a health risk? It goes back to EPA’s classification of ETS as a carcinogen. The agency publicly declared ETS a carcinogen before it even began any research. But its subsequent efforts failed to find evidence to back its position—despite fudging its analyses with statistical contortions that evoked condemnation from the Congressional Research Service, the General Accounting Office, an investigation by the U.S. House of Representatives, and America’s best known scientific journal, Science. Rather than admit it was wrong about ETS being a carcinogen and had wasted taxpayers money for two years trying to prove it, EPA simply “deviated from generally accepted scientific standards” and “deliberately abused and manipulated the scientific data,” as the U.S. Congressional Investigation put it. EPA now claimed that ETS had a RR of 1.19 and said that was health danger. How many cases of lung cancer do you think would be associated with this 19 percent risk increase among 100,000 nonsmokers exposed to ETS? The answer is 1.9 cases. Fewer than 2 cases in 100,000 people! That’s two-thousandths of one percent! You can scare a lot more people by shouting “the risk increases 19 percent!” than by saying this corresponds to an increase in the number of cancer cases of two-thousandths of one percent. But they mean the same thing. Now, do you think differences is heredity, diet, exercise, type of employment etc. could account for 1.9 cases of lung cancer among 100,000 people? You're darn right they could. They could account for a lot more than that.
Also, the standard for biological studies is a 95% Confidence Level, meaning the possibility that the outcome results from pure chance is no more than 5%. EPA could not achieve this with its RR of 1.19 for ETS. So it moved the goal posts, so to speak, to make it easier for ETS to “score” as a carcinogen. It doubled the possibility of a chance outcome to 10%. It had never done this before or since. (For more on EPA and secondhand smoke, see the posting of my testimony to Meeker County below.)
Other government agencies were quick to note the success of EPA’s scientific corruption and to imitate it. The Surgeon General’s reports, NIOSH and other government agencies reiterated EPA’s fraudulent figure of 3,000 deaths from lung cancer due to ETS and adopted the degraded statistical standards and corrupt practices that EPA used to produce it. Now the American Heart Association and other once highly respected medical organizations are doing the same thing. They have deviated from the generally accepted scientific standards to which they so long adhered.
Deaths estimated from adverse health conditions are derived from their risk ratios. If the risk ratios aren’t meaningful, calculations of the numbers of deaths are meaningless. A risk ratio tells you that there is a statistical association between a risk factor and a disease—and that’s all it tells you. It does NOT tell you there is a cause and effect relationship. The higher the risk ratio, the more likely it is that there is a causal relationship, but further research is needed to establish that. Suppose a study shows a statistical association between, say, lung cancer and people being lefthanded—which could happen. Does that mean being lefthanded causes lung cancer? Of course not. We would be jumping to conclusion if we said that. Similarly, we would be jumping to conclusion if we said that municipal water supplies and drinking coffee daily were causing hundreds of thousands of death annually based on calculations from their risk ratios. But that is exactly what many “researchers” are doing regarding many diseases or health conditions. They have jumped to the conclusion that statistical associations—and very weak ones (even statistically insignificant ones)—are proof of cause and effect. That is irresponsible. But it is effective in scaring the public who has no understanding of how meaningless these numbers are, and it provides cover for the political advocate/activists.
“Results of chemical analysis, animal experiments, and human studies are…found not to support claims of an association between workers exposed to environmental tobacco smoke and occupational coronary heart disease.”—Toxicological Pathology, Vol 57, No. 3.
“According to the scientific method, the only justifiable conclusion is that available data continue to falsify the hypothesis that ETS is a CHD [coronary heart disease] factor.”—Environmental Tobacco Smoke and Coronary Heart Syndrome: Absence of an Association, Regulatory Toxicology and Pharmacology 21, 281-295
[ If you wish to make a comment on this essay go to the Blog entry ]