In December 2023, back when the ink was still drying on the amendments to Federal Rule of Evidence 702, the Southern District of New York excluded all five general causation experts proffered by plaintiffs in the In re Acetaminophen MDL whose cases had been selected for trial.  We discussed that impressively detailed and painstaking decision here. Observing that all 50 states require proof of general causation in product liability cases and that the plaintiffs now had no experts to carry that water, the court instituted an order to show cause process that led to the dismissal of approximately 550 cases. However, over the defendants’ objections, the court permitted a small handful of new arrivals to the MDL to name their own general causation expert and try again. In yet another impressively meticulous opinion, the court recently doubled down on its Rule 702 “gatekeeping” role and excluded the new plaintiffs’ backup general causation expert in In re Acetaminophen – ASD-ADHD Prod. Liab. Litig., No. 22MC3043 (DLC), 2024 WL 3357608 (S.D.N.Y. July 10, 2024).

The plaintiffs in In re Acetaminophen allege that prenatal use of acetaminophen caused attention deficit hyperactivity disorder (ADHD) and/or autism spectrum disorder (ASD) in offspring. The plaintiffs’ first five proffered general causation experts were excluded in part because the court took issue with the experts’ “transdiagnostic” approach — combining literature of ADHD and ASD into a single analysis — because it merely “obscured limitations in the scientific literature.”   The new plaintiffs’ expert — number six overall — sidestepped the “transdiagnostic” issues and sought to opine solely on a supposed link between prenatal acetaminophen use and ADHD.  But the court, still mindful of acetaminophen’s critical role as the only pain reliever and fever reducer indicated for use during pregnancy and the FDA’s multiple reviews of its safety, remained vigilant in its Rule 702 obligation.

The court’s detailed analysis began by assessing the new expert’s qualifications. Despite a distinguished career as an epidemiologist, the expert had “limited professional experience with psychiatry, toxicology, and neurology.” Her “engagement with these issues” only began when she was retained for the litigation, and she had “little relevant expertise other than her substantial credentials as an epidemiologist” and no prior “familiarity with the pertinent literature.” Although the court concluded that the expert met Rule 702(a)’s qualification requirement, it observed that “the fact that her opinion was developed for litigation” was reason “to undertake a particularly careful examination of the opinion.”

This “particularly careful examination” found the expert’s causation analysis unreliable for multiple reasons. First, the court was troubled by the expert’s handling of genetic confounding (i.e., the idea that genetic factors could increase both acetaminophen use during pregnancy and, separately, the incidence of ADHD in offspring). The expert acknowledged genetic confounding was “very concerning” and that genetics “could be a true confounder.” Yet she dismissed the available large-sample sibling control studies due to their supposed limitations while prioritizing more limited studies, failing to “seriously engage with [the] findings.” The court took particular issue with the expert’s failure to address one particular study — “a sophisticated large-scale study funded by the NIH” which found “the apparent association between exposure to acetaminophen and ADHD disappears altogether when genetic confounding is accounted for.” Overall, the court found the expert’s approach not reflective of “the rigor required to render an admissible opinion on causation.”

Second, and independently, the court excluded the expert’s Bradford Hill analysis because it “display[ed] result-oriented reasoning.” Bradford Hill analyses comprise nine criteria used by epidemiologists to assess causality after observing an association. The plaintiffs’ new expert touted just three factors: consistency, temporality, and dose-response. But, following a careful study-by-study analysis, the court took issue with the expert’s treatment of all three of those factors. The expert’s consistency analysis failed to “contend with the FDA’s ultimate finding that ‘there are still study limitations and inconsistent study findings that prohibit causal interpretations of the association between [acetaminophen] exposure and functional neurobehavioral outcomes.’” The expert’s temporality analysis was plagued by “flagrant cherry-picking” of the literature. And, contrary to the expert’s opinion, the dose-response factor did not weigh “heavily” in favor of causality because “it is undercut by the limited data on which it is based and her failure to consider the impact of confounding by indication and genetics on which she relies most heavily for this factor.” Finding that the expert either ignored or conceded most of the remaining factors and found no strong support in any of them, the court rejected the expert’s general causation analysis as “result driven” and neither “an objective [n]or rigorous application of scientific methodology.”

In many ways, the court’s rejection of the plaintiffs’ sixth general causation expert mirrors its exclusion of the first five. It acknowledged the expert “confronted long odds in offering her opinion on causation” in light of the available studies, which the court reviewed in impressive detail. Indeed, at one point, the plaintiffs took issue with the court’s familiarity with the science and its close parsing of a key study, prompting the court to “decline[] to blinker its assessment . . . simply because plaintiffs prefer that the [c]ourt not consider the study.” This new opinion is even further illustration of a court committed to fulfilling its Rule 702 gatekeeping role because, as noted in the court’s initial Rule 702 ruling, “[i]t matters to get this right.”