On July 17, 1998, United States District Judge William L.
Osteen vacated Chapters 1-6 of and the Appendices to EPA's Respiratory
Health Effects of Passive Smoking: Lung Cancer and other Disorders,
EPA/600/6-90/006F (December 1992).
Judge
Osteen wrote, "In conducting the ETS Risk Assessment, disregarded
information and made findings on selective information; did not disseminate
significant epidemiologic information; deviated from its Risk Assessment
Guidelines; failed to disclose important findings and reasoning; and left
significant questions without answers. EPA's conduct left substantial holes in
the administrative record. While so doing, produced limited evidence, then
claimed the weight of the Agency's research evidence demonstrated ETS causes
cancer."
The reason science enjoys such weight and credibility is
that it has generally been structured so as to encourage a wide scrutiny of
methods, data and findings, peer review and a healthy debate from all
viewpoints. In the end, faulty data and fanciful theories are laid to rest, and
truth emerges the winner. Usually.
Scientists, like everyone else, are
subject to personal bias. They can set out to prove a pet theory, they can
ignore data which contradicts a favored hypothesis, and they can read into data
facts which simply aren't there. In addition, and perhaps more importantly,
their employers can be biased. Employers can put pressure on researchers or
even research firms to validate a preordained position or to produce a desired
result. Not that they really have to, since it is much easier to hire or retain
investigators who agree with them to begin with. And even that is not
necessary, since the employer gets to write conclusions and recommendations.
Scientists who dissent can effectively be buried simply by not publishing what
they have to say.
In an open and honest world, all of this would amount
to little, since all sides would have their say and we could count on
disinterested observers to wade through the fancy and glean the facts. In the
case of secondhand smoke, for example, it is near impossible to find anyone who
is both disinterested and of sufficient standing to be heard. What we have
instead is a war of credibility over conflicting interpretations of biased
reports. What we have is politics.
On one side are anti-smokers and the EPA. On the other,
smokers and the tobacco companies. At this moment, the antis and the EPA
clearly have the upper hand, enjoying as they do a credibility which is just as
clearly undeserved. No one is surprised that the tobacco companies have a
financial agenda. The same people, on the other hand, express astonishment at
the suggestion that the government might actually have a political agenda. Or
at the fact that the government, too, is an employer, with more power than most
to select and pressure the scientists it employs and to bury the ones that
dissent. Nowhere is this power more abused than on the subject of ETS. Last
year's EPA report on ETS, the cornerstone of the anti-smokers' arguments, is
worse than just bad science.
It is corrupt.
Corrupt
science has two salient characteristics. First, instead of starting with a
hypothesis and data and deriving from that a conclusion, it does just the
opposite: starting with a desired conclusion, it then selects data in order to
support the hypothesis. Second, it stifles dissent by excluding dissenters from
the process of review and by using ad hominem arguments to question their
character and motives. The EPA is guilty on both counts.
Of the 30 studies on spousal smoking referred to in the
EPA report, only 6 found any statistically significant association between ETS
and cancer in nonsmokers married to smokers, and none found a strong relative
risk. The studies actually used by the EPA were limited to 11 studies done in
the United States. Using the EPA's own Guidelines for Carcenogenic Risk
Assessment, none of these showed a statistically significant risk. These
guidelines call for a 95% Confidence Interval. By lowering it to 90%, only one
of the 11 studies showed a statistically significant risk. More importantly,
the two largest and most recent studies, one of which was partially funded by
the National Cancer Institute, were omitted from consideration altogether. Had
these two been included, no statistically significant risk would have been
found even after lowering the Confidence Interval to 90%. Even after violating
its own guidelines, in other words, the EPA could still show no statistically
significant risk without selecting data to fit its hypothesis. This cooked data
is the EPA's only basis for declaring ETS to be a "Group A" carcinogen. ("Group
A", incidentally, does not mean "extra deadly". It simply means
"human".)
The EPA's studies on ETS operate under a "zero threshold"
hypothesis, or the assumption that if huge quantities of something are
dangerous, then microscopic quantities are dangerous also. The data they used,
however, fails to bear this out: virtually all of the studies used either found
no risk at all or a risk so weak that it would not be considered significant if
applied to other subjects.
A "strong" risk is one with an "odds ratio"
of 5 to 20 - reflecting an incidence of the problem five to twenty times higher
in a group that was exposed to something than in a control group that wasn't. 5
in a population of 100,000, say, compared to 1. (Or 500 vs 100, or 5 in a
million vs 1; since it is a ratio, it indicates nothing about the size of the
risk itself. Only the relative risk.) A ratio of 1 indicates no risk at all.
Odds ratios under 3 are usually considered so low as to be the possible result
of random variation or margin of error. The odds ratio of contracting cancer
from chlorinated tap water, for example, has been calculated by the EPA to be
1.5 - not enough to worry about, in other words. Yet the EPA used an odds ratio
of 1.19 - considerably smaller than that of chlorinated tap water - to classify
ETS as a human carcinogen.
One of the largest and most recent studies
of ETS is the Brownson study, partially funded by the National Cancer
Institute. This study found odds ratios varying from .7 in non-smoking spouses
of smokers exposed for fewer than 40 years, to 1.3 in those exposed for over 40
years. .7 is a negative correlation, meaning that those exposed to ETS
for less than 40 years experienced fewer cancers than the control group.
Since the implication that ETS actually protected those subjects from cancer is
biologically implausible, the only other conclusion that can be drawn is that
the study's margin of error, caused by random variation, is .3 or higher. This
means that the 1.3 figure is equally suspect. The total risk for all groups
averages out to exactly 1, or no risk at all.
The Brownson study was
available to the EPA, but was not used in its report. Had it been included, the
conclusions would have had to have been revised downwards to show no
risk.
Though the EPA claims a "19% increase in risk", the actual order
of magnitude is less than intimidating. Even if the odds ratio of 1.19 is
eventually found to be accurate and not the result of random variation, what
this would mean is that instead of the 6 lung cancers per year normally
expected in a population of 100,000, we might find 7. Put another way, the
average lifetime risk of dying from exposure to ETS (as opposed to eventually
dying of something else) would be about 1 in 700. Furthermore, the risk (if
real) is concentrated among the non-smoking spouses of smokers exposed to ETS
(a heavier exposure than that found elsewhere) for upwards of 40 years, and
cannot be extrapolated to include the general population. By way of comparison,
the December, 1989 study by the Department of Transportation estimated the
probability of contracting cancer from the cosmic radiation at cruising
altitude as being some 2,000 times greater than the risk posed by ETS while on
the aircraft. The question is, is this sufficient grounds to justify the
subjugation of 50 million people? Before you answer, consider this: driving
across town to your favorite restaurant entails a risk many times higher - 17
to be exact - than inhaling any secondhand smoke once you get there. If you
include the risks associated with the food and drink you are likely to consume
there, that multiple rises to over 450. And if you choose one of the many
non-smoking restaurants provided by a free market, you get no exposure at all.
A look at the people who worked on this study explains this twisted
methodology. Leading anti-smokers pervade both the EPA and the Science Advisory
Board responsible for reviewing its analysis, and some of the work was
contracted to the founder of a prominent anti-smoking group. It should come as
no surprise that the EPA's Workplace Policy Guide was written well before the
ETS risk study was completed.
The underlying philosophy at work here is
that the end justifies the means; that since a smoke-free society is deemed to
be a worthy goal, manipulation of data and even lies are permissible in
achieving it. The EPA official responsible for the report admitted in "Science"
magazine (7/31/92) that "she and her colleagues had engaged in some fancy
statistical footwork" in order to arrive at the indictment of ETS. The opinions
of dissenting EPA scientists were ignored, and references to works not in
accordance with the EPA position were omitted from the bibliography of the
report. Finally, those gainsayers who actually manage to make themselves heard
are refuted not with data and reason but with character assassination.
Corrupt scientists have always been eager to ride the bandwagon of popular
opinion. Early in this century, Russian scientists drew diagrams purporting to
show that the skulls of Jews were smaller than average, reflecting a smaller
brain and therefore a lowered intelligence and moral sense. This "justified"
the pogroms. In our own country, studies claiming a lower IQ among African
Americans (referred to in the studies as "Negroes") "justified" segregation.
The damage done by these studies took decades to undo, while millions suffered.
Both conclusions share an insidious and horrifying circularity: if the subject
of the slur has the temerity to question the evidence, the response is "You
can't believe him, he's stupid. Look: my charts prove it!" And that is exactly
what is being repeated today. Using corrupt science, the EPA has manufactured a
conclusion which states, in effect, that smokers and the tobacco companies are
killing 3,000 Americans each year and are by implication morally corrupt. When
anyone disagrees, the anti-smokers have the perfect counter argument: "You
can't believe them, they're morally corrupt. Look: the EPA report proves it!"
"Political science" has taken on a whole new meaning.
Copyright (c) 1994, 1995 by Joe Dawson.
All rights reserved.