A recent Third Circuit decision is a useful and quotable reminder that temporal correlation is not enough to prove causation.
Causation is an indispensable element of any product liability action. It is the same in pharmaceutical third-party payor (TPP) class actions, where it is often the element most likely to defeat class certification. Unsurprisingly, plaintiffs often employ creative workarounds in an effort to sidestep obvious causation problems. In In re Avandia Marketing, Sales Practices and Products Liability Litigation, — F.4th —, 2026 WL 2093904 (3d Cir. July 21, 2026), the Third Circuit made clear that there are no shortcuts.
According to the plaintiff TPPs, the defendant pharmaceutical manufacturer had marketed the medication at issue as not only treating diabetes but offering cardiovascular benefits too. The plaintiffs claimed that the promotional campaign caused physicians to prescribe the medication at inflated rates, forcing the plaintiffs to reimburse prescriptions they would not otherwise have covered.
As evidence in support of that claim, the plaintiffs pointed to the defendant’s internal studies, a causation opinion by the plaintiffs’ litigation expert, and a 2007 study that had raised questions about the medication’s cardiac profile and led to a black-box warning, followed by a drop in prescriptions. The district court certified a class of TPPs.
On appeal, the Third Circuit focused on problems in class-wide causation. The opinion’s opening paragraph not only sets the tone but provides a useful quote for all cases in which causation is contested:
We do not presume in law that x caused y merely because x happened first. The connection might be causal. But it might be coincidental. Or some z might be responsible for x and y alike. As statisticians emphasize, correlation alone does not prove causation.
The panel first rejected the notion that evidence of an alleged “common scheme to deceive” could, by itself, support an inference of class-wide reliance on the alleged misrepresentations. A physician’s prescribing decisions involve dozens of variables, and the district court’s characterization of those decisions “did not cite a single page of the several-thousand-page record.” Even a properly demonstrated fraud scheme is not a free pass to predominance. Causation — in this case a causal relationship between the alleged fraud scheme and the TPPs’ alleged overpayment — cannot be assumed but must be proven.
Turning to the plaintiffs’ proffered evidence on causation, the panel began by noting that the defendant’s internal marketing studies did not isolate the impact of the messaging about the medication’s cardiac profile from the rest of the company’s promotional efforts. Thus, regardless of what the studies might prove about marketing in general, they could not show that the specific alleged misrepresentations had any impact on prescribers. The plaintiffs’ causation expert likewise failed to support causation, as her “results-oriented” analysis had been excluded on Rule 702 grounds by the district court in an order that was not appealed. And for good reason — her regression model was so prone to false positives that it found a statistically significant relationship between prescription volume and obviously irrelevant variables such as US carbon emissions.
That left only the drop in prescriptions that followed the publication of the 2007 study. The plaintiffs argued this was a “natural experiment” proving that earlier disclosure of the medication’s true risks would have caused the same decline and saved the plaintiffs from reimbursing for prescriptions over the interim years.
As an aside, the argument was a bold one for the plaintiffs to make in the 2020s. A subsequent study allayed the concerns that the 2007 study had raised and, by 2013, “provide[d] considerable assurance regarding the cardiovascular safety” of the medication according to the FDA. See, e.g., In re Avandia Mktg., Sales Pracs. & Prods. Liab. Litig., No. 07-MD-1871, 2024 WL 4582876, at *2–3 (E.D. Pa. Oct. 25, 2024). In effect, the plaintiffs asked the Third Circuit to ignore that the alleged cardiovascular risk of the medication had been disproven and merely to assume that, had that nonexistent risk been disclosed sooner, prescribers would have stopped prescribing the medication sooner, and TPPs would have saved money.
The Third Circuit largely obliged the first half of that ask, as it did not mention the FDA’s comment about cardiovascular risk and limited its discussion of the ensuing adjustment to the medication’s black-box warning to one brief footnote. But the panel was not willing to merely assume that a temporal correlation — even a stark one — can establish causation without controlling for confounding variables. With the plaintiffs’ expert’s analysis excluded, there was no rigorous statistical evidence to bridge the gap between correlation and causation. Thus, the panel vacated the class certification and remanded for further proceedings.
The specific holding is that TPPs in a pharmaceutical fraud RICO action may prove but-for causation with class-wide statistical evidence only if it is “sufficiently rigorous to show causation, not just correlation” by “distinguish[ing] the causal significance of the variable at issue and justify[ing] the rejection of competing explanations.” The Third Circuit now joins the First, Second, and Ninth Circuits on that point. In a larger sense, the opinion reinforces — in a very quotable way — that a litigant cannot prove that “x caused y merely because x happened first.”
The material contained in this communication is informational, general in nature and does not constitute legal advice. The material contained in this communication should not be relied upon or used without consulting a lawyer to consider your specific circumstances. This communication was published on the date specified and may not include any changes in the topics, laws, rules or regulations covered. Receipt of this communication does not establish an attorney-client relationship. In some jurisdictions, this communication may be considered attorney advertising.
