We’ve discussed recently how a federal statute intended to allow suits against international terrorists has been misapplied as allowing suits against pharmaceutical companies. While there’s still hope for that dangerous deviation to be rectified, another federal statute, the Racketeering Influenced Corrupt Organizations (“RICO”) Act has been so widely abused that it is rarely, if ever used against its congressionally intended target – organized crime – in civil litigation.
Back in 2019, our fourth worst decision was Painters & Allied Trades District Council 82 Health Care Fund v. Takeda Pharmaceuticals Co., 943 F.3d 1243 (9th Cir. 2019) (“PATDC82 I”), because the third-party payor (“TPP”) plaintiffs were allowed to misuse RICO to pursue what was essentially a garden variety inadequate warning case – that the drug Actos increased the risk of bladder cancer.
Indeed, that’s what the label says. According to the drug’s current FDA labeling (mostly for generic versions), taking it for over a year “increased the relative risk of developing bladder cancer in any given year by 40% which equates to an absolute increase of 3 cases in 10,000.” Actos label, Warnings & Precautions §5.5. To put that in context, the Actos label also carries the FDA’s most serious type of warning – a boxed warning – but for congestive heart failure, not bladder cancer. Id. at p.1 (highlights).
Nonetheless, the Ninth Circuit’s PATDC82 I allowed a RICO claim alleging that, between 1999 and 2011, defendants concealed that risk from the FDA and that, as a result, every TPP in the country paid for Actos prescriptions that it otherwise would not have reimbursed.
Plaintiffs seek to recover economic damages under RICO for the payments they made to purchase Actos under the assumption that it was a safe drug, which they allege they would not have purchased had they known that Actos increases a person’s risk of developing bladder cancer.
943 F.3d at 1247. This litigation does not allege that anybody ever actually suffered from bladder cancer (or any other injury) from the drug. Id. (“Plaintiffs do not, however, seek to recover economic or non-economic damages caused by any person’s actual ingestion” of the drug.). What’s worse, the fraud on the FDA claim, if brought under state law, would be preempted by Buckman Co. v. Plaintiffs Legal Committee, 531 U.S. 341 (2001). But by attaching it to the federal RICO statute, it escapes preemption (since preemption only operates against state-law claims).
You can see where this is going. RICO allows plaintiffs to structure their claims so that the actual risk, bladder cancer, becomes irrelevant to damages and is only the purported basis for causation. But as a causal matter, the TPP claims themselves make no sense. The drug is still on the market, and it has other risks that the FDA has found significantly more serious – requiring a boxed warning – than a three-one hundredths-of one-percent (0.0003) increase in absolute incidence of bladder cancer. Yet what the plaintiffs were claiming is that, despite Actos’ undisputed effectiveness as a diabetes treatment, no TPP in the country would have purchased it for that purpose because of the 3/10,000 bladder cancer increased risk. PATDC82 I, 943 F.3d at 1251. But the FDA has concluded just the opposite.
Even the Ninth Circuit seemed rather queasy about the factual basis of these broad-brush TPP claims, but because it was reviewing a Rule 12 motion to dismiss, it kicked the can down the road:
We note that Defendants’ argument that had Plaintiffs not taken Actos, they would have paid for an alternative drug to treat their type 2 diabetes, has not fallen on deaf ears. It seems quite logical that Plaintiffs would have paid for a different drug to treat patients’ diabetes. . . . But at this stage in the proceedings, we take Plaintiffs’ allegations that they would not have bought or paid for Actos as true. Plaintiffs do not allege that they would have paid for an alternative diabetes drug had they known Actos carries an increased risk of causing bladder cancer. Further, if what Defendants argue proves true, Plaintiffs may still be entitled to damages if the alternative drugs they would have paid for cost less than Actos. Plaintiffs have alleged there are at least three less expensive alternatives to Actos, and discovery may prove Plaintiffs were likely to have bought these alternatives. In any event, this is a damages question for another day.
Id. at 1251 n.7.
Some three years later, “another day” arrived – but the PATDC82 madness continues – indeed it got worse. Now, Painters & Allied Trades District Council 82 Health Care Fund v. Takeda Pharmaceutical Co., ___ F. Supp.3d ___, 2023 WL 4191651 (C.D. Cal. May 24, 2023) (“PATDC82 II”), has certified a nationwide class action of TPPs on these same facially incredible RICO economic loss claims, turning this bogus case into a huge deal and thereby “plac[ing] inordinate or hydraulic pressure” on defendants to settle in order to avoid even a small risk of ruinous liability. E.g., Laudato v. EQT Corp., 23 F.4th 256, 260 (3d Cir. 2022) (many other cases recognize this problem with class actions).
How could this happen?
For one thing, plaintiffs changed their tune. Now, instead of claiming the none of the TPPs would have reimbursed any Actos prescription, they claim that, because “Actos sales dropped . . . precipitously” after the 2011 label change adding the bladder cancer language, a “significant number” of claims would not have been reimbursed by TPPs nationwide. PATDC82 II, 2023 WL 4191651, at *3. Instead of any direct evidence of causation, the plaintiffs relied on the usual statistical mumbo-jumbo we see in these cases: an expert’s “econometric regression model” supposedly showing that, had the 3/10,000 bladder cancer increased risk been on the label “from the beginning,” “TPPs would have paid for 56% fewer Actos prescriptions during the class period.” Id. at *3 (footnote omitted). The class period is between 1999 and 2010. Id. at *4.
We searched PATDC82 II for the word “generic” – it does not appear. Apparently, plaintiff’s statistician, in considering the post-2011 decline in Actos sales, did not deem it important that generic versions became available in August 2012. We’re not statisticians, but it’s our impression that the availability of cheaper generic drugs tends to depress sales of the original branded drug (indeed, TPPs drive this shift). So, in considering the decline in branded prescriptions in this case, there’s good reason to believe that post hoc doesn’t ergo with propter hoc.
Nonetheless, a nationwide RICO class action was certified, but not a statewide California class action on state causes of action that require proof of reliance.
First, the class plaintiff got away with destruction of evidence. It preserved no records for nine of the eleven years of the class period. Id. at *6 (plaintiff “has documents and formularies reaching back to only 2009”). In dramatic contrast to the underlying Actos personal injury litigation, where juries were allowed to consider far less significant spoliation claims than missing records for 85%+ of the class period, in PATDC82 II the loss of ten years of TPP transactions was no big deal “because it [plaintiff] did not destroy or dispose of any documents − Prime Therapeutics [plaintiff’s pharmacy benefit manager] did.” Id. That happened only a couple of paragraphs after PATDC82 II accepted plaintiffs’ claim that “the relationship” that resulted in this destruction of evidence was “typical” for purposes of class certification:
[T]he manner in which [class plaintiff] administers its benefits − i.e., via a pharmacy benefit manager − mirrors the approach of TPPs across the country. According to Plaintiffs’ expert . . ., the relationship between [plaintiff] and its pharmacy benefit manager . . . is typical of such relationships in the United States, and it includes usual and customary services.
Id. at *5 (citations, footnotes and quotation marks omitted). Apparently, massive loss of evidence was merely a “usual and customary service” of pharmacy benefit mangers nationwide.
Second, as for superiority, PATDC82 II admitted that a class trial would face “enormous logistical hurdles,” but nonetheless found a nationwide class action “superior.” 2023 WL 4191651, at *8. Otherwise “there could be hundreds or thousands of individual lawsuits.” Thus, “[o]ne supposed ‘nightmare’ trial is preferable to many hundreds of shorter ones.” Id. That two paragraphs hardly strikes us as persuasive analysis. The same sky-is-falling argument could be said about any litigation with sufficiently “numerous” class members, so it’s rather tautological. More importantly, the underlying claims are transparently meritless for reasons already discussed by not only us, but by the Ninth Circuit, so the sky is unlikely to fall. As individual claims, these would not pass the red-face test.
These facile findings favoring of class certification only got worse when “predominance,” where the true trench warfare occurred, was addressed. Id.
Third, PATDC82 II finds predominance because “evidence of conduct all stems from the behavior” of the defendants. Id. That sweeping generalization ignored any differences between the behavior of hundreds or thousands of individual TPPs, as if they all acted in lockstep – which they simply don’t. See, e.g., In re Thalomid & Revlimid Antitrust Litigation, 2018 WL 6573118, at *4 (D.N.J. Oct. 30, 2018); In re Skelaxin (Metaxalone) Antitrust Litigation, 299 F.R.D. 555, 565 (E.D. Tenn. 2014) (both complete with diagrams). And, as already mentioned in the spoliation discussion, one hopes there are significant differences in how they and their pharmacy benefit managers approach their record-keeping.
Belying the supposed “rigorous” analysis class certification requires, PATDC82 II kicked the can further down the road. Individual factual differences were:
immaterial at this stage of the litigation. Whether the evidence will support Plaintiffs’ claims is a matter for trial or summary judgment; it is peripheral to the question of whether (or not) the issue is common to the class.
Id. at *9. If actual factual differences were “immaterial” to class certification, then common issues would predominate in virtually every class, which is plainly not the law. Class certification would be the norm, rather than a clear outlier position in prescription medical product litigation. At the degree of abstraction taken in PATDC82 II, any damages class becomes certifiable. The lack of anything approaching “rigorous analysis” was glaring.
Fourth, similar factual problems were shoved under the rug with respect to the purported “pattern of racketeering activity” – allegations of “mail fraud and wire fraud.” PATDC82 II, 2023 WL 4191651, at *10. Changes over time didn’t matter, the opinion held, because any changes “would have changed for all members of the class at the same time.” Id. The federal rules don’t treat fraud that way. Rather they require that it be pleaded with particularity, Fed. R. Civ. P. 9(b), precisely because attempts at fraud affect everyone differently.
Again, PATDC82 II ignored reality. There are so many TPPs, and related pharmacy benefit managers, that the decision’s unspoken assumption that each and every one of them was affected by every purported misrepresentation at the same time and the same way simply beggars belief. That assumption trumped plaintiffs’ burden of proof. How individual class members encountered (or did not) various alleged misstatements was essentially ignored. The unstated premise here was fraud on the market – an assumption that every class member always has perfect information. Fraud on the market is not allowed outside of securities law, and even there has been restricted.
Fifth, while courts have dumbed down RICO by abandoning any connection to organized crime, the statute still “require[s] both proximate and but-for causation.” Id. at *11 (citation omitted). Thus, the “whole class” must “suffer damages traceable to the same injurious course of conduct.” Id. (citation and quotation marks omitted). The Ninth Circuit’s footnote, quoted above, mentions several factual reasons why that can’t possible be true here, but PATDC82 II danced around all of those problems, coming to grips with none of them.
PATDC82 II first characterized defendants’ argument as asserting that some TPPs would have “paid for Actos prescriptions that would have been written anyway, fraud be damned.” Id. Nothing is being “damned.” Rather disparate causation issues derived from: (1) plaintiffs never claiming that the drug didn’t effectively treat diabetes, a widespread and serious condition; (2) the FDA never removed the drug from the market; and (3) use of the drug not being deterred by other risks, see the boxed warning, much more significant than the alleged 3/10,000 increased bladder cancer risk. Rather than putting plaintiffs to their proofs, the opinion simply took their expert’s statistical “analysis at face value,” id. at *12 – hardly “rigorous” analysis – without determining if that the analysis included the confounding factor of the 2012 advent of generic competition in assessing the purported rate of sales decline after the bladder cancer risk became common knowledge. The data were then rotely applied to all the earlier years.
Prescription medical product liability litigation has existed for decades. Such litigation has an accepted way to analyze the question of “prescribe or not to prescribe” – what individual prescribing physicians actually do with “new” information. This time-tested type of evidence is mostly absent from the analysis in PATDC82 II – as in Neurontin, the only actual prescriber testimony belied plaintiffs’ position. 2023 WL 4191651, at *17 (discussing “excerpts of two depositions” of “prescribing physician[s]”). No evidence is cited in PATDC82 II that any actual prescriber did what plaintiffs claim statistically happened on a classwide basis – decide not to prescribe a safe and effective diabetes treatment because of a 3/10,000 increased risk of bladder cancer. Rather,
Plaintiffs’ approach is a textbook example of how the use of “computer records, clerical assistance, and objective criteria” can obviate the need for an evidentiary hearing on each claim. The Court is persuaded that common questions of fact still predominate.
Id. at *13 (citation omitted). Actually, “plaintiffs’ approach” is a “textbook example” of using statistics to evade their burden of proof. Without any consideration of what actual doctors did for actual patients with actual diseases, all that was proven was the maxim “garbage in, garbage out.” Once again, the analysis more closely resembled improper fraud on the market than anything else.
Sixth, a confounding factor that the Ninth Circuit identified was that many (some 30%, apparently) patients allegedly taken off of the drug were switched to more costly alternative medications – meaning that the TPPs would have saved money with Actos. Id. at *13. Here, PATDC82 II submerged that issue by flipping (rather than merely ignoring) the burden of proof. Instead of conducting anything resembling a “rigorous” analysis of the question, the opinion asserted that the defendants did not prove “how many TPPs” were involved. Id.
Defendants were accused of “only regurgitat[ing]” (use of that word, alone, betrays pro-plaintiff bias) that 30% of all patients were switched to higher cost medications. Id. Not only is the necessary “rigor” nowhere to be seen, but this refusal to investigate stands in sharp contrast to the statistics used earlier, id. at *11-12, where an effect on slightly less than 57% of prescriptions supposedly impacted 98.5% of all TPPs. We’re not statisticians, but we would guess that similar analysis of this 30% (a little more than half of the prior 57%) of all prescriptions would have generated a figure in the neighborhood of 90% of TPPs. But that analysis would tend to defeat class certification, so it was simply omitted in this portion of PATDC82 II.
Ultimately, PATDC82 II, deep-sixed the issue altogether – holding (1) that “the act of paying for fraudulently induced prescriptions − even when the alternatives are more expensive − is an injury”; (2) “whether the net economic loss is zero (or negative) is a question of damages”; and (3) “damages question for another day.” Id. at *14 (emphasis original). Abracadabra! Neat trick to make both the issue, and any semblance of “rigorous analysis,” disappear.
Seventh, the necessary rigor was also absent from the causation analysis. Plaintiffs, of course, sought to base causation entirely on 50,000-foot-level expert statistical opinions. As we said in a prior post on the same topic, “Yes, there are travelling econometricians who can regress their way to any conclusion you might want.” The defense contended that this was a prescription drug warning case, even if decked out in RICO garb, the critical role of prescribing doctors (as recognized by all 50 states) cannot be ignored. PATDC82 II, 2023 WL 4191651, at *15-16. We addressed this issue at length in our discussions of the competing Neurontin (pro-plaintiff) and Sidney Hillman (pro-defendant) appellate decisions, so we won’t delve deeply here.
While at least acknowledging this issue, PATDC82 II again made it go away it by flipping the burden of proof – converting an essential causation step into a mere affirmative defense. First, the acknowledgement:
[There is] a key flaw with Plaintiffs’ predominance argument. Namely, [a defendant] could still depose individual prescribing physicians to contest Plaintiffs’ theory of but-for causation. . . . And even if Plaintiffs present evidence that such testimony is “unreliable,” a trier of fact could nonetheless rely on the physicians’ testimony to qualify, discredit, or reject Plaintiffs’ common evidence of but-for causation. . . . Since the number of testifying physicians would likely increase with the number of TPPs in the class, and that testimony would be linked to specific TPPs, such evidence would constitute individualized evidence . . . [and] a real and significant risk exists that individualized factual determinations would swamp common ones on the question of but-for causation.
Id. at *16 (citations omitted). Second, the vanishing act:
At this point, one might conclude that individualized questions of fact predominate over common questions − at least, when it comes to but-for causation. But that conclusion is premature. It is not clear that [defendants] will − or even can − avail themselves of a TPP-by-TPP causation defense using doctor-by-doctor testimony. To sustain an affirmative defense, a defendant must have evidence. . . . While the Court could speculate whether [defendants] will depose (or even can depose) many prescribing physicians, it is not this Court’s role to make decisions on conjecture.
Id. at *17. Apparently, to counter one statistician, defendants must exhaust themselves with hundreds of physician depositions simply to defeat class certification. We remember deposing over 500 surgeons in a couple of months (40 simultaneous tracks) in the Bone Screw MDL litigation, and that wasn’t even a class action. It was a huge expenditure of time and effort. To defeat predominance, it shouldn’t take many examples of prescriber testimony to establish that warning causation is not capable of being proven – or even affirmatively defended – on a classwide basis.
In sum, on the basis of a “regression model, internal email conversations, academic studies, data regarding physician information requests, and the results of [a defendant’s] internal investigations,” PATDC82 II concluded that: (1) more than a decade after the fact and with actual purchase records for only one of eleven years, plaintiffs could establish (2) that every TPP in the country, regardless of how structured, (3) would have declined to reimburse a specific (to three significant digits) percentage of (4) on-label uses of an FDA-approved diabetes treatment that, (5) then and now (with 20 more years of clinical experience) the FDA considers safe and effective, (6) all because of a 3/10,000 increased risk of bladder cancer.
PATDC82 II thus reconfirmed the adage that “there are three kinds of lies: lies, damned lies, and statistics,” while ratcheting up settlement pressure in that litigation exponentially. The “rigor” that the Supreme Court requires to prevent reliance on questionable evidence in class certification was entirely absent from PATDC82 II – with individual issues repeatedly glossed over or avoided by shifting the burden to the opponents of class certification.
Almost incidentally, PATDC82 II denied certification of a California state-law class because those claims required individual proof of causation/reliance on clamed false statements. 2023 WL 4191651, at *25-26 (“the importance of materiality to the element of reliance − which traverses the [California state law] CLRA, UCL, and FAL claims . . . amplifies the importance of any evidence related to that inquiry” which creates “a potentially far greater need for individualized testimony”). That takes us back to the beginning of this post – the fundamental problem being rampant abuse of the RICO statute in areas having nothing to do with organized crime, in order to pursue claims that, under traditional legal theories, require individualized proof of essential elements.
Perhaps the Supreme Court should employ its recently invented “major questions” doctrine, cf. West Virginia v. EPA, 142 S. Ct. 2587, 2607-08 (2022), to hold that Congress must say so expressly if it either (1) wanted RICO to apply beyond organized crime, or (2) expected affirmative proof of reliance to be broadly abolished in fraud cases.
But don’t hold your breath.