The Zoloft MDL has finally collapsed. We started to see the writing on the wall in June 2014, when the MDL judge ruled that none of the plaintiffs’ experts could testify that taking Zoloft while pregnant could cause birth defects. That is to say, the plaintiffs had no admissible expert opinion establishing general causation. That’s a major problem, and the other shoe finally dropped on Tuesday, when the district court granted summary judgment in all but a handful of cases in the MDL. Or, as the district court put it, “At the end of the day, Plaintiffs have failed to raise a jury question on the necessary predicate to success in any case: that Zoloft was capable of causing their injuries.” In re Zoloft Prods. Liab. Litig., No. 12-MD-2342, 2016 WL 1320799, at *11 (E.D. Pa. Apr. 5, 2016) (emphasis added).
We don’t often see entire MDLs go down like this. The District of Minnesota famously dismissed the entire Sprint Fidelis defibrillator MDL on preemption grounds in 2009, which was particularly noteworthy because it was an order on the pleadings. The Accutane MDL ended a few years ago in Florida with an order granting summary judgment in all but one case on the basis that the drug’s manufacturer adequately warned about the risk of bowel injury. Our only beef with that result was that it took so long—many years of tortured litigation, although several good orders came out of that MDL on experts and otherwise.
So the district court’s order showing the Zoloft plaintiffs the way to the courthouse exit is not unprecedented. But still, essentially ending an MDL with a stroke of a pen is big news, and the order shows not only that the plaintiffs did not have the goods, but also that the district court showed remarkable patience in allowing the plaintiffs multiple opportunities to develop what they needed.
Here is what happened. The JPML created the Zoloft MDL in 2012 and noted that general medical causation was among the common issues. Id. at *1. The MDL judge therefore issued a pretrial order in November 2012 scheduling discovery from the drug’s manufacturer, expert discovery on general causation, and Daubert hearings to follow. Id. The plaintiffs therefore came forward with four experts “on the issue of general causation in an effort to establish that Zoloft, when used at therapeutic dose levels during human pregnancy, is a teratogen capable of causing a range of birth defects.” Id.
This is where it started to go sideways for the plaintiffs. On June 27, 2014, the district court entered an order excluding the general causation opinions of the plaintiffs’ epidemiologist, who had failed to base her opinions on scientifically valid methodology and reasoning. Id. We wrote about that order here, and the main problem was that the expert eschewed accepted statistical models in favor of a “trend” analysis that, in a nutshell, was applied by the expert in a way likely to confirm her own a priori hypothesis. The court followed that order with a ruling that the plaintiffs’ other three general causation experts—none of them a medical doctor—also could not testify that Zoloft caused birth defects in humans. Id. at *2. That order caused us to wonder, “Is anything left”?
The plaintiffs then requested and received permission to offer a whole new expert, biostatistician Nicholas Jewell. Id. We wrote on that “do-over” here, but Dr. Jewell did not fare so well either. The district court issued an order on December 2, 2015, excluding his opinions because he failed to consistently apply scientifically accepted methods. Id. at *3. The main issue was that he failed to consider all the epidemiological studies, particularly larger and more recent studies that did not support his conclusions. He also improperly considered internal company documents, whose use would potentially be misleading to the jury. Id. It was a sophisticated and well-reasoned order that became one of our ten best drug and medical device decisions of 2015.
We recount this history for a reason: This week’s order granting summary judgment was not a kneejerk reaction. It was the result of a case plan that the MDL judge established at the outset of the proceeding and implemented patiently over nearly four years. Plaintiffs had every opportunity—including expensive discovery—to develop expert opinions, and when they failed, the court even gave them a second chance.
Indeed, the theme of “we have extensively litigated this” runs throughout the order granting summary judgment, and the district court is sometimes critical of the plaintiffs for acting as though they were writing on a clean slate. As a substitute for admissible expert opinion, the plaintiffs offered adverse event reports, internal company documents, new reports from already-excluded experts, a “differential diagnosis,” and the opinion of former FDA commissioner David Kessler.
To this, the district court’s initial observation was that
[a]though the Court has considered Plaintiffs’ evidence in full, the Court will not engage in an ad hoc third round of Daubert proceedings, as to do so would provide Plaintiffs “with an open-ended and never-ending opportunity to meet a Daubert challenge until [they] ‘get it right.’” Plaintiffs essentially attempt to proceed as if general causation has not already been litigated extensively . . . .
Id. at *4. Enough was enough. On the merits, the court made several rulings that we would highlight. First, although the court stopped short of holding that epidemiological evidence is always required to prove general causation, it ruled that “epidemiology is the best evidence of general causation in a toxic tort case” and that “where epidemiology is available, it cannot be ignored.” Id. at *5. This rule is at the crux of the court’s orders excluding plaintiffs’ experts, who consistently failed to account for a great deal of epidemiological research that did not support their conclusions.
Second, the district court was having none of the plaintiffs’ attempted resubmission of previously excluded opinions under the guise of submitting “specific causation” opinion. Simply put, “The Court has ruled that neither [expert] can testify that Zoloft . . . can cause birth defects in humans. As this opinion in inadmissible as to people in general, it must inadmissible as to any particular Plaintiff.” Id. at *7.
Third, the district court correctly rejected the “differential diagnosis” of another expert because he did not properly “rule in” Zoloft as a cause. We’re not sure how anyone could, in light of the court’s prior rulings that the plaintiffs could not prove general causation. Neither could the court, which observed that the new expert added “nothing new” and that the “ruling out” process by itself “cannot establish causation.” Id. at *8.
Fourth, with regard to Dr. Kessler, he was not an expert at all. He had “not conducted the analysis that the Court has explained in its earlier opinions that Daubert requires,” and he made no effort to reconcile purportedly contrary studies. What then did Dr. Kessler do? He attached attorney-prepared summaries to his report and blithely stated that he would “leave it to other epidemiologists to discuss the strengths and weaknesses of each study.” Id. at *9. What other epidemiologists? The court already excluded their opinions. Dr. Kessler was a mere conduit, and the court quickly determined that he failed under Daubert.
Finally, the plaintiffs submitted case reports, adverse event reports, and company documents. But anecdotal evidence is only reported data, “not scientific methodology.” Id.. As for company documents, those are “not typical of documents that experts would generally rely upon in a causation analysis, in part because “[t]he cited studies are a better source.” In any event, the company documents merely showed that employees raised questions and considered label changes, which at most showed the company’s knowledge, not causation. Id.
We like this order because of its deep and systematic consideration of all the evidence. In the end, the court rejected the plaintiffs’ request to delay summary judgment to allow for case-specific discovery. What would be the point? With no admissible proof of general causation, none of the plaintiffs could survive summary judgment regardless of their individual situations. Id. at *10. The court also denied the plaintiffs’ request for dismissals without prejudice. We can see why the plaintiffs made this request: The alleged injury was birth defects, and they wanted to take a second bite (or third or fourth) at the apple if new epidemiology later emerged. The court, however, had seen enough. There is extensively epidemiology for Zoloft developed over 20 years, and the issue of general causation has been thoroughly litigated “with the result that Plaintiffs cannot prevail despite having multiple opportunities to produce the required expert testimony.” Id. at *11.
Multiple opportunities. That is all that Plaintiffs could expect, and that is what they received. With opening day just behind us, we will close with a baseball reference—three strikes, and you’re out.