Risk Reassessment

"The effects of secondhand smoke have long been controversial," noted The Washington Post in a front-page story on Sunday. That simple statement of fact must have startled a lot of people.

Since 1993, when the Environmental Protection Agency announced that secondhand smoke causes 3,000 lung cancer deaths each year, public health officials and anti-smoking activists have hammered home the idea that lighting up around other people is a form of negligent homicide. Most press coverage of the issue has implied that no one outside the tobacco industry questions this view.

But now that a federal judge has overturned the EPA's decision to classify what it calls "environmental tobacco smoke" as a known human carcinogen, skepticism about the hazards of ETS will be harder to dismiss. In response to an industry lawsuit filed five years ago, U.S. District Judge William L. Osteen agreed that the "EPA publicly committed to a conclusion before research had begun" and "adjusted established procedure and scientific norms to validate the Agency's public conclusion."

The details highlighted by Osteen's decision are damning. To begin with, the EPA started writing a policy guide recommending workplace smoking bans four years before it officially declared ETS a hazard. William Reilly, then the EPA's administrator, conceded that "beginning development of an Agency risk assessment after commencement of work on the draft policy guide gave the appearance of…policy leading science."

So did the manner in which the EPA conducted its risk assessment. When it pooled data from various epidemiological studies of ETS and lung cancer, the EPA initially used the conventional definition of statistical significance. It later switched to a less rigorous standard because the results were not impressive enough.

"Using standard methodology, EPA could not produce statistically significant results," Osteen concluded. "EPA changed its methodology to find a statistically significant association."

Since the EPA excluded studies that would have undermined its findings, even the weak, barely significant association it found after changing the rules probably would not hold up if all the relevant data were analyzed. "If even a fraction of Plaintiffs' allegations regarding study selection or methodology is true," Osteen said, "EPA cannot show a statistically significant association between ETS and lung cancer."

The judge also touched upon the most important flaw in the evidence, which has to do with the limitations of epidemiology in dealing with low levels of exposure: When the association between a disease and a risk factor is weak, it may be impossible to rule out alternative explanations.

The closest the EPA came to addressing this basic problem was its suggestion that it did not really need information about the risk of lung cancer among people exposed to ETS because it already had plenty of data about the hazards of smoking. In essence, the EPA argued that if a lot of tobacco smoke causes cancer, so does a little. But this simply assumes what the epidemiological studies were supposed to prove.

Such criticism, once dismissed as tobacco industry sophistry, has gained credibility because of Osteen's ruling. "Secondhand smoke involves such a low concentration of carcinogens that a strong cancer link is hard to establish," The Washington Post reported, adding that "many epidemiologists say" a risk increase of the size estimated by the EPA "is too low to constitute convincing proof."

Granted, the Post's ETS coverage was more balanced than that of other major news news organizations even before Osteen's decision. But the Los Angeles Times, which usually sits at the front of the anti-tobacco bandwagon, carried the Post article on its front page, above the fold.

In its story, the Associated Press noted that "the extent of the hazard posed to nonsmokers from smoke exhaled by other people remains controversial." Even The New York Times was moved to acknowledge that the EPA report had been "questioned by some independent authorities, like the Congressional Research Service."

Tobacco's opponents were disturbed by this turn of events. "There is no legitimate question in the scientific community that secondhand smoke is a Class A carcinogen," insisted Stanley Rosenblatt, a plaintiffs' attorney who handles secondhand smoke cases.

Despite the fact that it was Osteen who upheld the Food and Drug Administration's jurisdiction over cigarettes last year, Rosenblatt suggested that the judge is biased in favor of the tobacco industry because he sits in North Carolina. Rosenblatt, of course, has no ax to grind.