Tobacco Documents Online

4 EPA WATCH .l lwia~wo*iwwp of.wvaRMrarf AprLate7! *dh'WW -ndnrek'.n dry rhr EPA dw D.p.rae.nr af %weiar. OtSli.t, dw tV/ria Maw th. US~. Cagin .ed/.d.a{ ma ad focsl.awrrs VOL 2 NUMBER 15 AUGUST 15. 1993 EPA'S RISK ASSESSMENT ON ETS CHALLENGED IN CAPITOL HILL HEARINGS I In tension-laden Congressional hearings marked by terse exchanges between committee members and EPA officials, representatives of the Environmental Protection Agency July 21 defended EPA's findings that second-hand smoke is a Group A carcinogen. The fireworks took place in hearings before the House Energy and Commerce Committee's subcommittee on'health and the environment and the House Agriculture Committee's subcommittee on specialty, crops and natural resources. EPA officialswere subjected to a withering cross fire of questions from congressmen challenging the agency's scientific methodology employed in its risk assessment on environmental tobacco smoke (ETS). Released in January, the report, Respiratory Flealth Ef jects ojParsive SmokinF Lung Cancer and Other Disorders, the EPA study concluded that ETS causes between 2,500 and 3,500 lung cancer deaths per year among non-smokers and former smokers. The ETS issue has taken on particular significance in light of recent Congressional initiatives to reform the Federal government's risk assessment process. Indeed, the hearings themselves are but the latest example of growing Congressional unrest over the current State of Federal'risk assessment procedures. The questions raised at the hearings tended to confirm critics' charges that EPA's scientific methodologies are in need of reform (see preceding article). Moreover, tobacco companies, which stand to lose millions in consequence of local smoking bans resulting from EPA's ruling, are taking the agency to court, saying the agenry, used'faulty science to arrive at its conclusions. Confidence Levels Congressmen Charlie Rose (Democrat of North Carolina), Scotty Baesler (Democrat of Kentucky), and Thomas Bliley (Republican of Virginia), all of whom are from tobacco-producing states, led the Congressional assault on EPA's report. The congressmen focused their attention on what has become the most controversial aspect of the EZ"S report, name1y the sgencys use of 90 percent confidence kvels in compiling its results. Traditionally, scfentists, including those at EPA, use a 95 percent confidence level, meaning that there is no more than a 5 percent possibility the results occurred by chance. EPA cancer guidelines for eptdemiologicalstudies require that the results cannot be due to chance, i.e., they must be statistically sfgnificant. EPA used a 95 pettxntconfidence level in its first draft of the report, and its switch to a 90 percent level in the report's final version has raised questions about the scientifrcvalidity of its findings. (Biostatistics textbooks are unambiguous on this point, stressing the ethical imperative of selecting the probability levels at the outset of the experiment rather than at the end of the process.) Morton Lippmann, chairman of EPA s Science Advisory Board (SAB), which reviewed the report, defended the agency's lowering of the confidence level by explaining that EPA decided to go with a"one-tail test" instead of a"two-tail test." He told the Energy and Commerce subcommittee that a 90 percent confidence level on the one-tail test is equivalent to a 95 percent confidence level on a two-tail test. He added that EPA switched during the course of the study, after concluding that a one-tail test was adequate. (Steven Bayard, EPA's project manager on the ETS study, argued in a similar fashion before the House Agriculture subcommittee.) But in response to a question by Congressman Mike Kreidler (Democrat of Washington), Mr. Uppmann conceded tha4 by, lowering the confidence level, EPA had fundamentally altered the outcome of its research. "So the tobacco industry is simply wrong when it argues that the statistical test applied by the EPA to individual studies was improper and did not show an effect at a 95 percent confidence leverr Mr. Kreidler asked. '°Ihey are not wrong in a technical sense," Mr. Lfppmaan answered, "that if they insist on doing a two-tail test, it would fail the significance IeveL~' Internal Disagreement Documents released at the hearings also teveal substantinl' disagreement within the agency over how the ETS issue was handled. EPA staff asked to review the report, complained of having only 11 days to review a 600- page document. They further recommended that ETS be classified as a Group B instead of a Group A carcinogen. Finally, in a devekipmeat that could bear dir" on the outcome of the lawsuit, questions were raised' regarding EPA's exclusion of a study from the 11 studies it used in i nsicts assessment. The so-called Brownson report, published by the National Cancer Institute in November 1992, showed no overall increase in risk for spouses of smokers. Congressatan Bliley and'other members of the Energy and Commerce subcommittee argued that if the Brownson study had been included along with the other studies EPA reviewed, the agency would not have been able to show a statistically significant link between passive smoking and lung cancer. . In response, Mr. Lippmann said the Brownson report had not been peer N reviewed at the time EPA was undertaking its analysis and as such Q could not be included in its report. ~. However, William Farlsnd, EPA's top risk assessment official, admitted W before the Agriculture subcommittee ~ that EPA staff had received a pre- O publication copy of the Brownson 'R studit, and documents released at the L+ hearings showed the agency had been ~..1 receiving data from Dr. Brownson O ciere Lme I9R9 CA Since it willibe several months before the case goes to court„it is difficult to assess how these latest revelations will affect the outcome. For its part, EPA has announced that it is filing a motion to dismiss the case brought against it by the tobacco industry. +