Supposedly, Einstein said that, “insanity is doing the same thing over and over again and expecting different results.”
If that’s the case, then these marketing-based class actions seeking refunds of the purchase price (or some fraction of it) for prescription drugs are at the Einsteinian extreme. As we’ve mentioned before, class certification in these sorts of cases has been denied again, and again, and again. We could add a lot more “agains” – but you get the drift. Just click on the “class action” label in the right hand margin for all our posts on the subject (warning: there are a bunch).
And the other day it happened again – for the second time – in the Neurontin off-label use litigation.
There’s a lot of preliminaries, about commonality, numerosity, and typicality, that involves some complicated subclassing. But these cases always come down to one insurmountable barrier – predominance – and one irreducible fact – that when you’ve got learned intermediary physicians making prescription decisions, the basis for those decisions will always be individualized. Call it “reliance,” or call it “causation” in cases involving information about prescription drugs, the result will be the same. Doctors apply their knowledge and experience to engage in patient-specific analyses of drug risks and benefits when prescribing drugs. That’s their job. Prescription decisions aren’t “caused” by anything on a classwide basis. And that goes double for off-label use, where by definition doctors have to rely on information sources other than drug labeling.
And, in the end, that’s what the Neurontin court held. No “shortcuts” are possible. Slip op. at 43.
Given the limitations in [statistical] analysis, [the expert’s] report does not suffice to defeat the predominance challenge. As is discussed above, without this statistical proof to support a presumption of causation, plaintiffs cannot meet Rule 23(b)(3)’s predominance requirement.
Slip op. at 43-44. Predominance did it again.
So why did the plaintiffs even try a second time in Neurontin? Maybe it’s Einsteinian insanity, but we have to think that it’s because the court left the door slightly ajar the first time around by allowing that a class might be permissible if only a “de minimus” amount of off-label use existed before the defendant’s supposed misrepresentations, and such use increased several orders of magnitude (or something like that) afterwards. In re Neurontin Marketing & Sale Practices
Litigation, 244 F.R.D. 89 , 113-14 (D. Mass. 2007).
But the Neurontin court recognized that four decisions handed down in the interim closed the door on this “de minimus” exception. First, in International Union of Operating Engineers Local No. 68 Welfare Fund v. Merck & Co., Inc., 929 A.2d 1076, 1085-86 (N.J. 2007), the court both took the NJ consumer fraud act off the table by broadly rejecting fraud-on-the-market type theories of proving causation. Second, McLaughlin v. American Tobacco Co., 522 F.3d 215, 223-25 (2d Cir. 2008), reversed the primary precedent cited the first time around in Neurontin, and equally broadly rejected fraud-on-the-market causation proof. Third, in In re St. Jude Medical Inc. Silzone Heart Valve Products Litigation, 522 F.3d 836, 840 (8th Cir. 2008), the court recognized that, even where a consumer fraud statute does not require affirmative proof of “reliance,” a defendant’s individualized evidence of non-reliance as a defense to causation precludes class certification. Fourth, another judge in the same district as Neurontin – In re TJX Cos. Retail Security Breach Litigation, 246 F.R.D. 389, 395 (D. Mass. 2008), broadly rejected presumptions of reliance outside of federal securities actions.
Slip op. at 24-29.
The Neurontin court also rejected the plaintiff’s stab at “statistical” proof. “Plaintiffs’ theory is nothing if not novel; they ask the Court to permit a statistical analysis to function as common proof of causation for millions of disparate and varied human interactions that resulted in off-label prescriptions for Neurontin.” Slip op. at 33. That model failed utterly to distinguish between mere off-label marketing and “fraudulent” marketing. Rather, the expert put the rabbit in the hat – “assum[ing] on instruction of counsel that all detailing during the class period was both off-label and fraudulent.” Slip op. at 39.
Why are we not surprised? If you uniformly assume what you’re supposed to be trying to prove, then you’re likely to be able to show that it can generally be proven. A classic case of “lies, damn lies, and statistics” – or “garbage in, garbage out.” The court thought so too – even to the point of using its own cliche:
The core assumption in the [expert’s] model is that off-label prescriptions caused by detailing expenditures were necessarily caused by a fraud, that is, that off-label promotion was the same as fraudulent promotion. As Madison Avenue would have predicted, [the expert] finds a strong correlation between expenditures on Neurontin promotion and the number of prescriptions written for the drug.
Slip op. at 39. Try as they might, it’s impossible for plaintiffs to reduce individualized physician prescription decisions to a classwide statistical model:
[T]he record in this case demonstrates why the use of spending on fraudulent off-label detailing as a means to ascertain the number of prescriptions subject to the fraud is flawed. Significantly, the testimony of the prescribing physicians for the . . . subclass representatives indicates that only one of them . . . was ever detailed by defendants about Neurontin. . . . Even if this hurdle could be overcome, [the expert’s] analysis does not take into account any other factors that may have led doctors’ to prescribe Neurontin for off-label indications. The deposition testimony of the doctors for the . . . class representatives shows that their decisions to prescribe Neurontin resulted from a wide variety of influences unrelated to the three components of defendants’ alleged fraud.
Slip op. at 40-41.
And the third-party payor class failed as well – for the reasons stated by the New Jersey Supreme Court in Operating Engineers – because “TPPs exhibit a great degree of heterogeneity.” Slip op. at 46. Tautological statistics assuming what plaintiffs purported to prove failed again, and plaintiffs could “not present the court with an acceptable form of common proof. Slip op. at 50.
Thus, even where plaintiffs had pleaded “blatantly illegal off-label promotion activities for which [defendants] have been criminally sanctioned,” slip op. at 50, individual litigation, and not class actions, were the means for prosecuting any legitimate claims that might be out there.
It doesn’t take an Einstein to conclude where further pursuit of this sort of class action is headed.