It has been said, with maybe a bit of sarcasm, that a company developing a drug hopes that its drug will become successful enough to attract frivolous lawsuits. OK, so maybe only outside lawyers have offered such an aphorism, but bear with us. Imagine that a drug is developed to treat a really common condition, like high cholesterol, it becomes accepted as a first-line treatment, and becomes one of the most widely used prescription drugs of the last few decades. Imagine then that the patient population that would take this drug has lots of co-morbidities and that, while the drug is being used on a long-term basis, some predictable portion of the drug’s users will develop a new condition, like diabetes. Imagine then that, because the drug has been widely used for so long, there are lots of studies, published and unpublished, that look at measures like blood glucose and new diagnosis of diabetes, among many other things. It should not be hard to imagine that thousands of the patients taking the drug would develop diabetes on the drug in the absence of any relationship and that plaintiff lawyers would round many of them up to sue based on the expectation that the lawyers and their favorite experts could gin up proof of causation that would survive a Daubert challenge if the manufacture did not pay to get rid of the litigation first. This is hardly fancy as we have posted on multiple orders from Lipitor MDL (here, here, and here) that excluded plaintiffs’ causation experts and then granted summary judgment for the manufacturer across the board.
Much like the Zoloft MDL affirmance we lauded last year, all of this went up on appeal to either revive or affirm the end of an entire litigation. We are pleased to say that In re Lipitor (Atorvastatin Calcium) Mkt’g, Sales Pracs. & Prods. Liab. Litig. (No. II), MDL No. 2502, — F.3d —, 2018 WL 2927629 (3d Cir. June 12, 2018), did the latter. Without repeating the history of all of the decisions below that we detailed previously, there were five basic issues on appeal, the admissibility of each of the three experts plaintiffs offered, whether plaintiffs could use other evidence to establish causation without experts, and whether plaintiffs’ responses to show cause orders were sufficient to avoid summary judgment. Each deserves some attention.
Plaintiffs’ statistician was up first. His approach was to re-analyze clinical trial data to suggest that there was an increase in blood glucose levels and infer that as proof of causation for diabetes. The disconnect here is fairly obvious, but the statistician compounded the problem by including both patients with a single instances of elevated glucose levels during the trials and patients with elevated glucose levels before the study began. Id. at *4. Plaintiffs, their statistician, and other experts had agreed that a single increased glucose level did not indicate diabetes. Id. The statistician also agreed, as the MDL court put it, that he “lacked the expertise to opine about any implications that single glucose readings might have about the possibility of new-onset diabetes.” Id. This might have been enough to exclude his opinions, but he also relied on one test of statistical significance after the standard test failed and presented calculations of average blood glucose increases in a misleading and result-driven fashion. Id. at **5-6. He also re-analyzed a study that had found no increased in the rate of diabetes with the drug compared to placebo based on a applying a new definition of diabetes after the fact and by someone who lacked relevant expertise. Id. at **6-8. The Fourth Circuit affirmed the exclusion of his opinion, noting that the MDL court “properly discharged its gatekeeping duty” by weighing “classic concerns regarding reliability and relevance.” Id. at *6.
Plaintiffs also offered an internist to interpret the medical literature and perform meta-analysis of select studies, which he attempted to dress up with a purported application of the Bradford Hill criteria. Noting the importance of dose to these analyses, the MDL court asked the internist to provide an analysis specific to each commercially available dose of the drug. The MDL court ultimately excluded his causation opinion as to all but the highest dose because he acknowledged there was not a statistically significant increased risk of diabetes for the other doses. Id. at *9. On appeal, plaintiffs challenged that their internist could not just lump all the doses together and offer a single causation opinion. Given the facts here—like a 10 mg low dose, a 80 mg high dose, and studies specific to each dose—we do not think this requirement should have been the least bit controversial. The Fourth Circuit, however, while holding that the MDL court did not “abuse its discretion in asking the expert to produce a dose-by-dose analysis,” cautioned that this was not a new requirement for all cases. Id. at *10. The more serious, and recurring, issue was whether statistical significance was required for a causation opinion based on epidemiologic evidence and the Bradford Hill criteria. Again, we think the Fourth Circuit could have gone a little farther—like you always or almost always need epidemiologic evidence as a starting point for causation in a product liability case and epidemiologic evidence must be statistically significant (with multiple studies with an increased risk greater than 2.0) to count—but its conclusion that the MDL court had not abused its discretion on the record here was good enough. Specifically, the internist’s purported application of the Bradford Hill criteria and failure to establish that reliance on non-statistically significant results was accepted by epidemiologists were enough for the court to find his causation opinions unreliable. Id. at *12.
Plaintiffs also offered another internist to opine on specific causation for one of the bellwether plaintiffs. While the plaintiffs touted that this expert had used a differential diagnosis to come to her opinion, the expert herself did not say that she did and claimed to use a methodology for her opinion in litigation that she had never used in her own practice. Id. at *13. She also could not rule out other causes like the plaintiff’s weight and weight gain and relied too heavily on the post hoc ergo prompter hac fallacy. Id. at *15. Again, this exclusion was within the MDL judge’s discretion.
Like the MDL court, the Fourth Circuit did not a bright line rule on whether general medical causation for product liability cases involving a pharmaceutical could ever be established without any expert testimony from the plaintiff. We think the better approach, as spelled out in some state’s law, is to require expert testimony on these issues. However, the Fourth Circuit’s conclusion that the specific non-expert evidence offered by plaintiffs—principally snippets from U.S. and foreign labeling—was not enough to establish causation is fine by us. The causation issues are “complex and manifold” and the non-expert evidence from plaintiffs “isn’t especially strong.” Id. at *17 (contraction in the original). So, the bottom line was more than fine by us: “To hand to the jury the evidence here and ask it to reach a conclusion as to causation with any amount of certainty would be farcical and would likely result in a verdict steeped in speculation.” Id. Put another way, if a court is supposed to be the gatekeeper for expert evidence on key issues, it cannot just allow dubious non-expert evidence to suffice on an issue that would require an expert under Rule 702.
The last issue for the Fourth Circuit to address was whether the MDL court could require the remaining plaintiffs to come forward with evidence showing they could prove specific causation after the Daubert and summary judgment orders. Plaintiffs’ argument on this was essentially that the MDL should have remanded all the cases rather than fulfilling the mission of the MDL court to decide common pretrial issues. This argument was a bit disingenuous, because the plaintiffs surely would have been comfortable with summary judgment or Daubert motions being denied across the board had the rulings on the bellwether cases gone their way. “Here, it was the district court’s prerogative to determine whether it could dispose of the cases before it on the merits.” Id. at *18. We may not always be a fan of the direction MDL courts have taken in the last decade or so, but they are supposed to do what the MDL court did here. At the end of the day, this MDL court “discharged [its] duties meticulously and thoughtfully” an ended a litigation as it should have been – with the manufacturer winning without facing the uncertainty of jury trials or succumbing to the pressure of a large number of pending cases.