In Daubert v. Merrell-Dow Pharmaceutical, General Electric. v. Joiner, and Kumho Tire v. Carmichael the U.S. Supreme Court sought to bring principles for reviewing expert testimony in line with the Federal Rules of Evidence. The opinions sought to ensure that legal arguments would better comport with the pertinent science needed for the legal cases at issue. To achieve this goal the court gave trial judges a greter duty to review expert testimony for relevance and reliability before plaintiffs could bring their case to a jury. Despite these goals, lower courts have struggled with reviewing scientific testimony and evidence. Some courts so restricted expert testimony and its scientific foundation that scientists found it difficult to present basic scientific evidence about the toxicity of chemicals in a courtroom.
An outstanding decision by the First Circuit Court of Appeals this March in Milward v. Acuity Specialty Products (639 F.3d 11 (2011)) contravened numerous mistaken views about scientific methodology, evidence evalualtion and constraints on testimony that had emerged from defendants’ presentations and other courts over time. This decision is precedent for courts in the First Circuit and stands as an example of an excellent analysis of scientific testimony and the role of trial courts for other appellate jurisdictions. As one who had argued in the scholarly literature for the conclusions the court adopted and who served as the scientific methodology expert in this case, it is especially encouraging to see this change in the law.
Prior to Milward courts had created a number of misleading or mistaken doctrines for addressing scientists’ testimony in court. First, the Supreme Court itself in Joiner endorsed a study-by-study assessment of scientific data as to whether each piece individually and reliably supported an expert’s inference of causation, something no responsible scientist would do. Scientists would integrate all scientifically relevant evidence for assessing the toxicity of a chemical substance and draw their conclusions from that. Second, some trial and appeallate courts had required human epidemiological studies in support of expert testimony, privileging such evidence. Some courts went further, insisting that testimony be supported with epidemiological studies showing a relative risk of at least twice the background rate of disease. Third, and often as a consequence of the previous point, other courts had excluded animal and other studies as inappropriate support for testimony, particularly when human studies are not available. Yet scientists often conclude that a cheimcal is toxic precisely on the basis of such data. Fourth, courts tended to treat “no effect” epidemiological studies as definitive evidence of no adverse effect, even though scientists would not accept this simple inference. Finally, “weight of evidence” arguments – or, as I will call them, “inferences to the best explanation” – the essence of scientific reasoning, were rejected outright by some courts, usually leading to the exclusion of reasonable scientific arguments. Plaintiffs facing such court views could easily lose without presenting their cases to juries.
The Milward court rejected these views and more. Plaintiff Brian Milward alleged that he developed acute promyelocytic leukemia (APL) caused by his exposure to benzene in products manufactured by defendants. Defendants sought to exclude the “general causation” testimony by plaintiffs’ expert, Dr. Martyn Smith of the University of California, Berkeley, that benzene exposure is capable of causing APL, a sub-variant of acute myelogenous leukemia. The trial court granted defendants’ motion to exclude Dr. Smith’s testimony and entered judgment for the defendants. Plaintiffs appealed. The First Circuit reversed the trial court’s exclusion of Dr. Smith’s testimony. Its holdings go a long way toward opening the courthouse doors a bit wider so injured parties can present their cases to juries.
The First Circuit explicitly recognized a strict line between the role of judge as gatekeeper of reliable testimony and the jury’s role as finder of fact on scientific issues. It said that courts may not exclude experts whose testimony falls within the range of matters about which reasonable experts can disagree. In their gatekeeping role courts may not decide which of competing scientific views is correct. The First Circuit held that the trial court “took sides on questions that are currently the focus of extensive scientific research and debate—and on which reasonable scientists can clearly disagree. In this, the court overstepped the authorized bounds of its role as gatekeeper.” (p. 22)
The appellate court rejected an “atomistic” study-by-study assessment of the scientific basis of expert testimony. “The district court erred in reasoning that because no one line of evidence supported a reliable inference of causation, an inference of causation based on the totality of the evidence was unreliable.” (22-23)
It also noted, citing National Cancer Institute scientists, that “there should be no such hierarchy” of evidence for carcinogenicity as between epidemiological and some other kinds of evidence (17).
The court held that for inference to the best explanation or “weight of the evidence” methodologies “No serious argument can be made that the weight of the evidence approach is inherently unreliable.” (18-19) Moreover, “in this mode of reasoning, the use of scientific judgment is necessary.” (18 (emphasis added)) On a related point the court held that the nine “viewpoints” identified by famed epidemiologist and statistician Austin Bradford Hill for guiding judgments about toxicity conclusions had “no algorithm” for application. (18)
Following the Restatement of Torts it further held that “reasonable scientists may come to different judgments about whether such an inference about toxicity is appropriate.” (Id. § 28 reporters’ note comment. c(4)) If as the Restatement indicates, when causation conclusions involve substantial scientific judgment, courts should not interfere or second-guess. This appears to leave a more appropriately limited role for court gatekeeping of expert testimony.
The First Circuit’s rulings on epidemiological evidence are particularly important. “The district court read too much into the paucity of statistically significant epidemiological studies. The absence of peer-reviewed epidemiological studies does not, as defendants contend, make it “almost impossible” for Dr. Smith’s opinion to be admissible. Epidemiological studies are not per se required as a condition of admissibility regardless of context … Rather, this is a case in which there is a lack of statistically significant epidemiological evidence, and in which the rarity of APL and difficulties of data collection in the United States make it very difficult to perform an epidemiological study of the causes of APL that would yield statistically significant results. . . . The court erred in treating the lack of statistical significance as a crucial flaw.” (24)
Milward represents a new and much better day for legitimate science to be practiced in the courtroom as it is in the lab. It permits more wronged individuals or parties with legitimate scientific claims for injuries to go to juries. Scientists, lawyers and injured parties have good reason to applaud the decision.