摘要：For more than 40 years, the tobacco industry prevailed in lawsuits brought by injured smokers, despite overwhelming epidemiological evidence that smoking caused lung cancer. Tobacco lawyers were able to create doubt about causation. They sought to persuade jurors that “everybody knew” smoking was harmful but “nobody knows” what causes cancer by recreating in court the scientific debate resolved by the 1964 Surgeon General’s Report. The particularistic structure of jury trials combined with the law’s mechanistic view of causation enables a defendant to contest virtually any claim concerning disease causation. Despite judicial efforts to eliminate “junk science” from lawsuits, a well-financed defendant may succeed in persuading jurors of the epidemiological equivalent of the proposition that the earth is flat. IN DAUBERT V MERRILL DOW, 1 the US Supreme Court undertook to eliminate “junk science” from the courtroom. It concluded that the traditional test—general acceptance within the field—was too limited and that the judge also must evaluate whether the witness’s hypothesis has been subjected to peer review, whether the results described as flowing from that theory are replicable, and whether there is a known error rate for the hypothesis advanced. Daubert addressed the question of when a witness will be permitted to testify; it did not address how the witness’s testimony will be evaluated and weighed by the jury. Unfortunately, the very “junk science” that the Supreme Court sought to ban from the courtroom may still appear through the cross-examination and arguments of lawyers who find it in their client’s interest to make good science appear bad. Tobacco litigation provides the telling example. If lawyers can successfully cast doubt in a juror’s mind on the relationship between cigarette smoking and disease, the danger of junk science driving out good science remains in virtually any lawsuit asserting that sustained exposure to a toxin produced the plaintiff’s disease. Proving tort causation through epidemiological evidence presents a conceptual challenge. Epidemiology seeks to answer the question, “Do smokers have a higher incidence of a particular form of lung cancer than non-smokers?” Tort law asks, “Did smoking Marlboros for 20 years cause this particular plaintiff’s lung cancer?” Epidemiological evidence most directly answers a question of general causation—Can exposure to this toxin contribute to the development of this disease?—while tort law demands an answer to a question of specific causation—Did exposure to the toxin cause this plaintiff’s disease? The jury is not asked the question that the evidence might actually permit it to decide accurately—whether the defendant’s conduct increased the risk that the plaintiff would become ill with a particular disease—but is required to answer a question that the evidence cannot possibly enable it to answer definitively 2 —whether the defendant’s product made this particular plaintiff ill. Courts have resolved this tension by permitting the plaintiff to prove causation through the testimony of an expert that “it is more likely than not” (the law’s requirement for proof in a civil case) that it was exposure to the defendant’s product that caused the injury. Although this view of legal causation has been criticized as a matter of both science and policy, 2 courts typically take the position that expert testimony stating that a disease is present in an exposed population at twice the rate that it occurs in the unexposed population is sufficient to permit a finding of cause. They reason that there is at least a 50% chance that the plaintiff’s injury came from exposure to the questioned substance. 3 Evidence can be legally sufficient in the sense that it permits the judge or jury to find for the plaintiff without being persuasive. A jury evaluating legally sufficient evidence must still determine (a) whether the evidence is believable and (b) if believed, whether it establishes the disputed proposition—for example, that exposure to a particular toxin caused the plaintiff’s injury. While it is true that “[u]ntil sufficient evidence of causation is developed, all plaintiffs would lose,” it is not true that “afterwards, assuming that there exists adequate proof of exposure and other necessary elements of the legal theory being pursued, all would likely win.” 3 (p782) Plaintiffs who offer legally sufficient evidence of causation can and do still lose their lawsuits. Translating statements about risk factors into conclusions of legal causation can be (and has been) undermined by the ability of defense lawyers to take the very qualities that may make epidemiology scientific and posit them as reasons for the jury to be skeptical that legal cause has been established. The tobacco litigation over the past 50 years illustrates how the tension between the “commonsense” and scientific understandings of causation will not be resolved simply by permitting experts to say that it is “more likely than not” that the defendant’s activities were the necessary cause of the plaintiff’s injury. When defendants are, as the tobacco companies were, prepared to challenge expert testimony through cross-examination designed to demonstrate the difference between risk factors and causation as the witness understands those terms, jurors may be confused as to what they are required to determine. 4
