|
Epi Wit & Wisdom Articles
Daubert Decision Seen as a Step
Forward in Reducing Junk Science
By C.R. (Reg) Allen Jr.
[Ed. The January issue of the
Epi Monitor carried a news item about the Supreme Court decision to
reinforce the role of judges as screeners of scientific evidence for
reliability and subsequent admissibility in court. The original
decision which found that judges had this responsibility was the
Daubert decision in 1993 that had epidemiologic evidence at the heart
of the controversy. Epidemiologists and a lawyer interviewed at that
time (See Aug/Sept 1993 issue of the Epi Monitor for comments by
Kenneth Rothman, Susan Rose, Robert Morgan and Shanna Swan) were
unsure about the long term impact of what was hailed by many at the
time as a landmark decision. In the following article, Reg Allen, a
former public health physician turned practicing attorney, gives his
overall assessment of the impact the decision has had.]
Junk science has been a growing
concern in the courts. The tremendous increase in use of expert
witnesses has unfortunately included some individuals willing to offer
opinions of dubious value for a fee. Such testimony can have a
misleading and prejudicial impact on the jury. Lacking a base in
traditional scientific methodology, junk science not only clutters the
courts with unreliable information and wastes time and other
resources, but it can also undermine the credibility and integrity of
the justice system.
Against this backdrop, Daubert
vs. Merrell Dow Pharmaceuticals Inc [509 U.S. 579, 113 A.Ct. 2786, 125
L.Ed. 2nd 469 (1993)] has proved to be a landmark US Supreme Court
case on scientific evidence. This case arose from the consolidation of
two cases in which plaintiffs claimed that mothers’ ingestion of
Bendectin, a prescription drug for nausea, caused birth defects. Both
plaintiffs and defendant proffered experts who based opinions on
evidence which included epidemiologic studies. The procedural history
of the case is complex, but ultimately the exclusion of plaintiffs’
expert testimony was affirmed.
The court interpreted federal
rules of evidence to require the trial court judge to serve as a
gatekeeper to screen scientific, technical or other specialized
knowledge for relevance and reliability. The court decided that the
adjective “scientific” connotes more than subjective belief or
ungrounded speculation. The judge is to focus on the basis for an
expert’s opinions, not the expert’s conclusions per se. To assist the
trial judge in determining the admissibility of proffered scientific
evidence, the court listed some “observations” or reliability factors
which can be considered:
1) whether the scientific
knowledge can be tested
2) whether the theory or
technique has been subjected to peer review and publication
3) the known or potential rate
of error
4) whether there is general
acceptance of a theory
The impact of Daubert has been
large, although difficult to quantify. In the more than four years
since the decision was handed down, it has been cited in hundreds of
subsequent reported cases, and these numbers do not count unreported
trial court cases. It has been the subject of or referred to in dozens
of law review articles, including both praise and criticism.
Perhaps its greatest impact is
an effect very difficult to quantify—-introducing caution into the
selection and use of experts in discovery and trial. Hearings on
motions to exclude an expert may address the reliability factors set
forth in Daubert.
As a US Supreme Court case,
Daubert is controlling authority for federal cases but much of its
impact results from it being widely followed at the state level.
Daubert and its state law progeny have had a major influence on many
areas of law, including negligence law. This is of particular
importance to epidemiologists, since negligence law includes, inter
alia, health provider malpractice, product liability and toxic torts.
Some states have not only
embraced but have extended the reach of Daubert. For example, in
Texas, [Robinson vs. E. I. Dupont de Nemours, 923 S.W. 2d 549 (Tex
1995)] followed Daubert and added two additional reliability factors:
1) the extent to which the
technique relies upon subjective interpretation of the expert
2) the non-judicial uses which
have been made of the theory.
Further, in Merrell Dow
Pharmaceuticals Inc vs. Havner, 953 S.W. 2d 706 (Tex 1997), the Texas
Supreme Court explicitly discussed the role of epidemiologic
information to support a finding of causation (a topic for another
day).
Daubert and its progeny are not
complete or necessarily smooth solutions to the problem of junk
science. Its requirements may seem burdensome, the judge’s
gate-keeping discretion may be an imperfect screen for “good” science,
and excluded experts may feel wronged, to name but a few of the
critiques. Further law and application will no doubt refine the
Daubert approach. Nonetheless, the overall impact has been to foster
closer scrutiny of experts and scientific evidence and to move a step
forward in reducing junk science.
This line of cases has strong
implications for epidemiologists. The growing importance and
recognition of epidemiologic evidence increase the possibility that
epidemiologists may become involved in litigation, most likely as
experts and hopefully not as defendants. As an expert, an
epidemiologist might prepare a report; give deposition testimony;
and/or give live testimony at trial. An important part of an
epidemiologist’s preparation for reports and testimony is
understanding and satisfying the Daubert reliability requirements for
admissible evidence.
Published April 1998 v
Epidemiologists and Journalists
Gain Insight at Boston University Symposium
A symposium was held at Boston
University in early October to provide an opportunity for journalists,
particularly those covering nutrition topics, to meet with
epidemiologists and obtain a better grasp of what epidemiology is all
about. As a result of the meeting, both journalists and
epidemiologists gained new insights about one another’s fields,
according to Ellen Ruppel Shell, co-director of the BU program on
science and journalism which hosted the meeting. An account of the
meeting appeared in the New York Times on October 11.
According to Ruppel Shell, the
scientists in attendance came to better understand that journalists do
not write their stories primarily to educate their readers, but to
report the news. This helps explain why journalists do not present all
of the caveats that scientists may attach to their findings. This
practice has appalled some scientists, according to Ruppel Shell.
The journalists in attendance
came to understand that scientists can have an agenda beyond financial
gain, said Ruppel Shell. Most often, reporters are looking for how a
report may be tainted because of financial considerations. In fact,
dogma and strongly held beliefs may pose more of a potential “conflict
of interest” for scientists than financial matters.
What Journalists Like About
Epidemiology
Ruppel Shell told the Epi
Monitor that epidemiology is fascinating to journalists because:
• It deals with populations
• Its pronouncements appear to
have immediate consequences for people as a whole (in contrast to
clinical studies which involve only a few subjects)
• Its findings relate to humans
(in contrast to basic science studies which tend to involve only
animals)
• Its findings frequently appear
in the New England Journal of Medicine
• It appears easy to understand
(at the surface at least)
• The jargon is limited compared
to molecular biology and other sciences.
Published November 1995 v
Epi Critic Speaks at ISEE
Meeting
Gary Taubes, the freelance
journalist whose Science article on the “Limits of Epidemiology” irked
so many epidemiologists back in 1995 because of his negative portrayal
of epidemiology, was an invited speaker at the recent International
Society for Environmental Epidemiology meeting in Boston. Speaking in
a session on “The Media As A Link To The Community,” Taubes’
presentation was entitled “Epidemiology in the Press: Telling Time by
the Second Hand.”
Taubes prefaced his talk by
informing the audience that much of his career has been spent
investigating bad science. For example, he recently completed a book
on cold fusion. He referred to his earlier article in Science on the
“Limits of Epidemiology” and summarized it by saying that he had
criticized epidemiology for looking for subtle effects without the
right tools. He likened the efforts of epidemiologists to find health
effects to those investigators who looked for canals on Mars in the
days before the right telescopes were invented. In a more recent
Science article, “The (Political) Science of Salt,” (August 14) Taubes
again criticized a large body of epidemiologic work.
Taubes began his talk by
revealing his main message, namely that the reason for the persistent
conflict between the media and science is that the demands of the
scientific process and the demands of the media are incompatible. Put
another way, Taubes emphasized that the requirements for a good story
and the requirements for good science are at odds.
One criterion for making the
news is to claim something at variance with the conventional wisdom.
This is the classic “man bites dog” story. In science, consistency is
more prized. Bad science is more sensational than good science and the
press favors this because it is a better story despite the likelihood
of being proven wrong over the long run. Taubes quoted Australian
physicist and historian of science John Ziman about how 90% of what is
in physics textbooks is true, whereas 90% of what is in physics
journals is false. Extending this to epidemiology and assuming that
only 50 - 75% of the reports in epidemiology journals will later prove
to be false, the problem is that these reports are likely to be picked
up by the press, said Taubes. In physics, reports are ignored because
the public is less interested. Not so in epidemiology.
In an interesting coincidence,
Taubes referred to a New York Times story on August 18, 1998, the same
day as his talk at ISEE, by Jane Brody detailing “Health Scares That
Weren’t So Scary.” In this article, Brody describes an American
Council On Science and Health report “Facts Versus Fears” which
reviews the greatest unfounded health scares of the last five decades
(www.acsh.org/Publications/reports/ factsfears.html). Among the
unfounded scares she describes are those about alar, electric
blankets, cellular phones, asbestos in schools, and coffee and
pancreatic cancer. This last scare was caused by a 1981 epidemiology
report which could not be confirmed subsequently by the same
investigators.
Taubes offered no solutions for
the limitations he perceives in epidemiology and in the press, nor for
reducing their tendencies to conflict. Thus, he closed his talk by
saying that trying to arrive at the truth about epidemiology and
health from newspapers is like trying to establish the correct time by
looking at the second hand on a clock.
Taubes and other speakers in the
session took questions at the end. One participant took advantage of
the opportunity to chastise Taubes for his 1995 article critical of
epidemiology. Taubes responded by saying that numerous epidemiologists
had read his article in draft before it was published. He claimed, “I
did not write anything in the article that you epidemiologists do not
know in your hearts to be true.”
Published August/September 1998
v
|
