Photo of Bexis

Lord knows, there are a lot of meritless MDLs.  Bexis’ and Michelle’s Bone Screw litigation, prompted by a televised CYA freak-out by then-FDA Commissioner David Kessler over an off-label use that had become the medical standard of care, was one, and it gave us Buckman.  Several MDLs against modern anticoagulants, such as Xarelto, are another example, since all of those second-generation drugs were safer than the older forms of blood thinners they superseded.  Currently, the Taxotere MDL – based on the dubious proposition that the plaintiffs cared more about hair loss than most effectively treating their breast cancer – and Zostavax – where not a single plaintiff can prove causation – come to mind.  Our colleagues defending other MDLS can also be excused for believing that their own litigation should be added to this list.

But for sheer factual baselessness, it would be hard to top the still-ongoing Avandia MDL, which somehow has managed to persist since 2007.  The Avandia litigation is infamously based on a purported increased risk that, over a decade ago, the FDA scientifically determined did not exist.  Briefly, a study erroneously found an increased risk of cardiovascular events with Avandia that led to both a boxed warning and additional studies.  The additional studies debunked that supposed increased risk and the FDA removed the boxed warning:

The Nissen Study concluded that . . . Avandia was “associated with a significant increase in the risk of myocardial infarction and with an increase in the risk of death from cardiovascular causes. . . .”  On July 30, 2007, the FDA convened an advisory committee to evaluate the data on Avandia’s cardiovascular safety and to recommend potential changes to its labeling.  On November 14, 2007, the FDA directed GSK to add information to the Avandia label in a boxed warning. . . .  

*          *          *          *

In a decisional memorandum dated November 19, 2013, the FDA wrote that “the data continue to support no statistically significant difference between [Avandia] and . . . the risk of death or major adverse cardiovascular outcomes. . . .  On May 7, 2014, the FDA approved an updated label that removed the cardiovascular risk and restricted access information from the boxed warning.  By this time, however, Avandia sales had dwindled, and the drug was no longer widely prescribed.

In re Avandia Marketing, Sales Practices & Products Liability Litigation, 2024 WL 4582876, at *3 (E.D. Pa. Oct. 25, 2024) (footnotes omitted).

That 2007 was when both the initial label change adding the boxed warning later found to be baseless, and the creation of the Avandia MDL, occurred is hardly coincidental.  Once the label change was in the works, plaintiffs and their third-party litigation funders went to work soliciting anybody who had taken the drug and suffered any sort of cardiovascular event (which are unfortunately common in the general population).  But an MDL that started out as Godzilla became a zombie once the FDA admitted that the science did not support the boxed warning some seven years later.

All the Avandia personal injury cases have long since bitten the dust – either settled in what turned out to be a tremendous waste of money that could have been better spent on drug development – or dismissed once any pretense at scientific validity had vanished.  What’s left is the true bottom of the barrel:  third-party payors with their hands out, claiming purely economic loss from purchasing a drug with what turned out to be a nonexistent risk.

But at long last, even those dregs appear to be circling the drain.  In the latest Avandia decision, quoted above, all the plaintiffs’ economic loss experts’ opinions were excluded.  Plaintiffs, of course, already had to change their theories in mid-litigation.  Their initial claim, that the defendant purportedly concealed the drug’s “increased cardiovascular risks when compared to alternatives,” went out the window when the FDA determined that the supposed “increase” simply wasn’t.  So, the TPPs switched to a theory alleging improper marketing of the drug as having “better cardiovascular outcomes.”  2024 WL 4582876, at *3.

The two plaintiff experts who performed this legal legerdemain were Meredith Rosenthal and Thomas McGuire, both with doctorates in statistical fantasy, and both claiming they could tease the effects of purported marketing misconduct out of pharmaceutical sales data that had been severely roiled by the aforesaid FDA zigs and zags.

Didn’t work.

Rosenthal fell first.  She claimed that, “based on an econometric model she built using a multiple regression analysis,” she would testify that “41% of Avandia sales” during the relevant period were legally “caused” by (a “substantial contributing factor”) the purported “fraudulent promotion.”  Id. at *5.  She supposedly found “a statistically significant relationship between [defendant’s] marketing and Avandia sales.” Id.  This calculation, of course, is the same kind of thing Rosenthal has been pumping out for years in similar situations.

The road to exclusion, however, was bumpy – because the Avandia decision erroneously employed a pro-admissibility standard that the 2023 amendments to Fed. R. Evid. 702 had declared “incorrect.”  Specifically, Avandia stated:

“As a general rule, the factual basis of an expert opinion goes to the credibility of the testimony, not the admissibility, and it is up to the opposing party to examine the factual basis for the opinion in cross-examination.”  Exclusion is required only where “the expert’s opinion is so fundamentally unsupported that it can offer no assistance to the jury.”

2024 WL 4582876, at *6 (quoting Sterling v. Redevelopment Auth. 836 F. Supp.2d 251, 271-72 (E.D. Pa. 2011), which in turn quoted an Eighth Circuit case, Children’s Broadcast Corp. v. Walt Disney Co., 357 F.3d 860, 865 (8th Cir. 2004)).

As we discussed at length in our posts on the 2023 Rule 702 amendments, the Committee Note to these amendments state just the opposite:

[M]any courts have held that the critical questions of the sufficiency of an expert’s basis, and the application of the expert’s methodology, are questions of weight and not admissibility.  These rulings are an incorrect application of Rules 702 and 104(a).

Rule 702, 2023 Committee Note.  In adopting the amendments, the Civil Rules Committee similarly pointed out, “[t]hese statements misstate Rule 702, because its admissibility requirements must be established to a court by a preponderance of the evidence.”  Committee on Rules of Practice & Procedure, Agenda Book, Tab 7A, “Report to the Standing Committee,” at 871 (June 7, 2022) (available online here).

Specifically, this erroneous reliance on Eighth Circuit language for incorrect “general rule” and “fundamentally unsupported” statements traces back to Loudermill v. Dow Chemical Co., 863 F.2d 566, 570 (8th Cir. 1988) – as proven in Hose v. Chicago Northwest Transportation Co., 70 F.3d 968, 974 (8th Cir. 1995), which was quoted in the Children’s Broadcast case.  Loudermill, of course, predates even Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), and its holdings, quoted indirectly in Avandia, do not reflect current Rule 702 standards and are simply no longer the law.

Thus, it was error for Avandia (or any other court) to rely on the Loudermill language.  Amended Rule 702 provided the applicable standard, and “[a]ll laws in conflict with such rules shall be of no further force or effect after such rules have taken effect.”  28 U.S.C. §2072(b).  Bizarrely, Avandia nowhere acknowledged, let alone addressed, the effect of the 2023 amendments to Rule 702.

But ultimately this error in Avandia was harmless.  It only made the route to exclusion longer and more involved than it had to be.  As to Rosenthal, the decision could have ended on page *7, with her reliance on statistically insignificant data being decisive.

Instead, Rosenthal’s opinions failed due to  her bogus regression analysis.  That turned out to be as phony as a three-dollar bill:

As an illustration, [defendant’s] responsive expert . . ., conducted a falsification analysis wherein he replaced the input variable in Dr. Rosenthal’s model with datasets having no connection to Avandia and bearing no resemblance to Avandia promotional expenditures − specifically, monthly beef production, Colorado River flows, U.S. carbon emissions, and Hershey’s expenditures.  In each case, [the defense expert] employed Dr. Rosenthal’s methodology (including her use of two depreciation rates) and found a positive and statistically significant relationship with Avandia sales.

2024 WL 4582876, at *8.  Forget the details.  They don’t matter.  The fundamental problem here was that the purported regression analysis was rigged from the beginning to produce “statistically significant” results no matter what inputs were used.  Rosenthal “has not identified any datasets that would fail to result in a statistically significant relationship with Avandia sales.”  Id. (emphasis original).  As Avandia concluded:

Rosenthal’s efforts to identify the depreciation rates that “best fit” the data, in the context of a single model involving all marketing and all sales with respect to a single drug, are suggestive of a methodology that is results-oriented and which inverts the scientific method.  Because these individual steps diverge from established methods in the community, and because the results-oriented nature of the model is further confirmed by [defendant’s] falsification analysis . . ., the Rosenthal Model is fundamentally unreliable and does not withstand scrutiny under Rule 702.

Id. at *10 (footnotes omitted).

Further, her methodology “deviate[d] from the literature in substantial and important ways,” was not supported by “a single study that implements [it],” and “not supported by the relevant community of economists or published research in the field.”  Id. at *8-9 (footnotes omitted).  In short, Rosenthal stands exposed as an academic fraud, peddling “regression” analyses pre-ordained to find liability no matter what.  We can only hope that future courts preclude her from ever darkening any courtroom’s doors again.

Unlike Rosenthal, McGuire was offered solely as damages expert on “quantity effect,” that being “the difference between what Avandia coverage cost and the cost of coverage of cheaper, safer drug.”  2024 WL 4582876, at *11.  He advanced “two methods,” id.,  the first of which relied on Rosenthal’s bogus analysis.  That fell with Rosenthal, and need not be discussed further.  Id. *13.  Alternatively, McGuire used something called a “step-down adjustment.”  That was quite a step; yielding purported damages in excess of a billion dollars.  Id.  The adjustment was based on an unsupportable “assumption” – that drug sales otherwise would have been zero for the entire relevant period:

The assumption that, absent [defendant’s] alleged fraudulent promotion, sales of Avandia would have been zero across its entire time on the market is untethered from the record. . . .  McGuire essentially assumes that cardioprotective promotion is the only reason why any sales of Avandia were ever made.  There is no factual basis for such an assumption. . . .  Given the complete lack of support in the record for [this] assumption, [McGuire’s] opinions on damages under [this scenario] must also be excluded.

2024 WL 4582876, at *15 (footnotes omitted).  In other words, the basis for McGuire’s opinion was so flimsy that it failed even under the erroneously pro-admission Loudermill standard that the decision applied.

McGuire offered a second scenario, involving substitution of a generic drug, that generated even more ridiculously inflated damages numbers.  Id.  But it was only a disguised, more extreme version of the same invalid assumption that Avandia sales would have been zero.

A simple calculation of the difference between the price of Avandia and the price of [the generic], [defendant] contends, assumes that there would have been no sales of Avandia at all during an eight-year period, and therefore it suffers from the same flaws as [the] analysis under [the other scenario].  The Court agrees.

*          *          *          *

McGuire’s [generic drug] calculation is premised on the same kind of counterfactual assumptions as identified with respect to [the prior scenario]. . . .  McGuire must assume that all Avandia prescriptions were [generic] prescriptions instead, and further that [the generic] would have been the only alternative that was prescribed.  There is no basis in the present record for such assumptions.  Accordingly, these calculations must be excluded.

Id. (footnote omitted).

Rosenthal and McGuire were the only experts that the Avandia plaintiffs presented.  They have both been excluded, so that plaintiffs have no basis whatever for proving any supposed damages.  This particular Avandia decision arose in the context of class certification, id. at *3, which presumably will be denied.  However, the defendant also has a pending summary judgment motion, id., and with their experts excluded, plaintiffs will have a hard time with even their own individual cases – particularly if the MDL judge can be persuaded to apply the proper, post-2023 standards for expert admissibility under Rule 702.  Thus, we hope that this decision spells curtains for the long-running and uniquely meritless Avandia litigation.