Monday, May 7, 2012

More Thoughts On Bartlett

Our initial post about Bartlett v. Mutual Pharmaceutical Co., ___ F.3d ___, 2012 WL 1522004 (1st Cir. May 2, 2012), was more or less a crie de coeur over what we saw as an essentially absurd result:  that while a simple warning claim involving a generic drug is indisputably preempted under PLIVA, Inc. v. Mensing, 131 S. Ct. 2567 (2011), a claim much more fundamentally in conflict with FDA approval of generic drugs – that state tort law can impose liability for not removing an FDA-approved product off the market entirely – supposedly is not.


Plain and simple, the preemption result in Bartlett makes no sense.

Today we take a closer look at the Bartlett opinion and the reasoning, if it can be called that, that brought the court to its highly questionable result.

Two fundamental facts underlie Bartlett:  First, the plaintiff had (at least in theory) lost her warning-based claim for lack of causation – the prescribing physician did not read or otherwise rely on the allegedly inadequate warning, rendering the warning non-causal.  Second, for whatever reason, the defendant at trial made the strategic decision (which with the benefit of 20-20 hindsight now seems disastrous) not to offer any affirmative defense at all.
The plaintiff’s lack of a warning claim itself had two effects.  One, plaintiff was damn lucky, because warning claims against generic drugs were precisely what the Supreme Court held to be preempted in Mensing.  A verdict in Bartlett based upon a warning claim – the most common theory – would undoubtedly have been overturned before the First Circuit ever saw the case.  Two, since both Mensing and Wyeth v. Levine, 555 U.S. 555 (2009), involved warning claims (again, they’re the most common), neither case was 100% on point.

Bartlett initially demonstrates why “design defect” simply doesn’t fit prescription drugs (some medical devices are different).  As the First Circuit mentioned, you can’t change the “design” of something that “is a one-molecule drug; and the variations in [it] as sold consist of inactive ingredients that ordinarily do not have significant pharmacological effects.”  2012 WL 1522004, at *2.  The ordinary constraint on design defect claims – that there be an alternative design – does not exist where the product is a chemical molecule, let alone one that would require FDA re-approval if it were changed in any way that affected its safety or efficacy.

“Design defect” in the absence of any available alternative design amounts to absolute liability.  It’s not a tort, but rather a judicially-imposed no-premium insurance policy.  For that reason, the Third Restatement and an increasing majority of states (Wisconsin and South Carolina, at least, have changed during the lifespan of this blog) require proof of an alternative design as an essential element of all design-based claims.

New Hampshire, however, isn’t one of those states.  In Vautour v. Body Masters Sports Industries, Inc., 784 A.2d 1178, 1183 (N.H.2001), the New Hampshire Supreme Court rejected the Third Restatement because of the alternative design issue.  See Bartlett, 2012 WL 1522004, at *3.  Vautour, however, involved exercise equipment, and the plaintiff offered alternative design evidence (albeit rather weak).  It was not a case where there was no alternative design possible, and thus an attempt to impose absolute liability.  Nor did Brochu v. Ortho Pharmaceutical Corp., 642 F.2d 652 (1st Cir. 1981) – where the First Circuit allowed a design defect claim involving a prescription drug despite no state-court authority (contrary to proper Erie principles) – involve absolute liability, since that drug had two active ingredients, and the plaintiff argued that a better design would have had less of one of them.  Id. at 655.

The only New Hampshire precedent involving a design defect claim where no alternative design was possible is Buckingham v. R.J. Reynolds Tobacco Co., 713 A.2d 381 (N.H.1998), a tobacco case where the plaintiff sought to impose absolute liability.  The court rejected product liability for the design characteristics of a design that could not be changed.  “When the plaintiff cannot allege that something is ‘wrong’ with the product, strict liability should not be used as a tool of social engineering to mandate that manufacturers bear the entire risk and costs of injuries caused by their products.”  Id. at 384.

Bartlett, however, purported to follow Vautour because it was more recent than Buckingham. 2012 WL 1522004, at *3 (Vautour “followed after Buckingham”).  In reality, however, the court did something quite different.  For the first time under New Hampshire law, Bartlett permitted absolute liability – “design defect” liability for a design that the court admitted could not be changed.  Bartlett reached an unprecedented result by applying Vatour, which did not involve the impossibility of an alternative design, in an impossibility situation.  It did so, we would argue, in direct contravention of the holding in Buckingham rejecting “social engineering” through product liability litigation.

There was one final bulwark against absolute liability in Bartlett – Restatement (Second) of Torts §402A, comment k (1965), the “unavoidably unsafe product” doctrine, recognized in Brochu.  Comment k precludes liability for the “unavoidably unsafe” attributes of a product as long as it bears adequate warnings and isn’t contaminated in some way.  Here, however, the court ruled that the defendant slit its own throat by not putting on any defense.  Comment k is an affirmative defense in New Hampshire, so by not presenting a defense, waived it.  2012 WL 1522004, at *3 (defendant “could still have avoided liability by proving [the comment k defense, but] abandoned that defense on the eve of trial”).  To that extent, we can’t blame the court.

The First Circuit did something else unusual in Bartlett – applying a pure “consumer expectation” test to an alleged defect in a prescription medical product.  2012 WL 1522004, at *3 (“an ordinary consumer would hardly know without further warning that [the drug] or any other ordinary analgesic carries a risk of the kind of ill effects and suffering that [plaintiff] encountered”).  But that’s just not the law in most places, and Bartlett doesn’t cite anything to suggest that New Hampshire would choose this tangent.  Prescription only drugs simply aren’t available to “ordinary consumers.”  They require a physician’s prescription.  They’re the epitome of a “complex” product as to which an “ordinary consumer” would have no expectations.  In practically every other jurisdiction either the physician is considered the “ordinary consumer” (since they control use of the product) or “ordinary consumers” have no expectations other than what is stated in the product’s labeling, in which case plaintiffs are limited to warning defect claims.  As stated recently:
The consumer expectations test considers the same factors at issue in a failure-to-warn claim, and so its application is barred by Mensing. An ordinary consumer forms her expectations regarding the safety of drugs from her doctor or from the drug’s label. Thus, if [the drug] is dangerous beyond the expectations of the ordinary consumer, that can only be a symptom of [defendant’s] failure to update its label or communicate effectively with doctors.

Grinage v. Mylan Pharmaceuticals, Inc., ___ F. Supp.2d ___, 2011 WL 6951962, at *6 (D. Md. Dec. 30, 2011).  Accord Carlin v. Superior Court, 920 P.2d 1347, 1355 (Cal. 1996); Craft v. Peebles, 893 P.2d 138,154-155 (Haw. 1995); Pittman v. Upjohn Co., 890 S.W.2d 425, 430 (Tenn. 1994); Shanks v. Upjohn Co., 835 P.2d 1189, 1195 (Alaska 1992); King v. Danek Medical, Inc., 37 S.W.3d 429, 445 (Tenn. App. 2000); Valentine v. Baxter Healthcare Corp., 81 Cal. Rptr.2d 252, 263 (Cal. App. 1999); Rosci v. Acromed Inc., 669 A.2d 959, 969 (Pa. Super. 1995); Makripodis v. Merrell-Dow Pharmaceuticals, Inc., 523 A.2d 374, 378 (Pa. Super. 1987); Peterson v. Parke Davis & Co., 705 P.2d 1001, 1003 (Colo. App. 1985); Stilwell v. Smith & Nephew, Inc., 482 F.3d 1187, 1194 (9th Cir. 2007); Phelps v. Sherwood Medical Industries, 836 F.2d 296, 302-303 (7th Cir. 1987); In re Zyprexa Products Liability Litigation v. Eli Lilly & Co., 489 F. Supp.2d 230, 267 (E.D.N.Y. 2007); Soufflas v. Zimmer, Inc., 474 F. Supp.2d 737, 751 (E.D. Pa. 2007); Chamian v. Sharplan Lasers, Inc., 2004 WL 2341569, at *6 n.6 (Mass. Super. Sept. 24, 2004). Bartlett did neither.  Rather, the court reached the singular conclusion that “ordinary consumers” have expectations about prescription drugs (which require a prescription precisely because they have risks) beyond those stated in the labeling.  That’s a unique and dangerous holding.

But none of the First Circuit’s mangling of New Hampshire law matters much any more – at least in Bartlett.  Those are all state-law issues, and if the defendant decides to “take it up” to the Supreme Court (as our prior post urged), there must be a federal question.

That question is preemption.

The First Circuit’s treatment of the preemption question is, if anything, even more peculiar than its flights of fancy under the guise of New Hampshire law.

Essentially, the Bartlett court thumbed its collective noses at the Mensing decision.  Instead the decision purports to follow Levine despite:  (1) Levine having nothing to do with generic drugs and (2) Levine having even less to do with design defect claims.  Indeed, in Levine the Supreme Court went through contortions specifically to avoid having to decide a “failure to contraindicate” claim:

[Plaintiff] also offered evidence that the [FDA-approved use] should be contraindicated and that [the drug] should never be administered [that way].  Perhaps for this reason, the dissent incorrectly assumes that the state-law duty at issue is the duty to contraindicate. . . .  But, as the Vermont Supreme Court explained, the jury verdict . . . did not mandate a particular replacement warning, nor did it require contraindicating [the use]. . . .  We therefore need not decide whether a state rule proscribing [the FDA-approved use] would be pre-empted.

555 U.S. at 564-65.

Pretty obviously, there was not a majority in Levine to allow state tort law to ban FDA-approved uses.  Bartlett is even more extreme – rather than simply banning one FDA-approved use of a drug, it allowed state tort claims predicated on banning an FDA-approved drug altogether.  We're not being hyperbolic defense lawyers here.  Take a look:

[Defendant] argues . . . that the generic maker also cannot alter the composition of the drug and so [Mensing]'s policy of encouraging generics by preempting state tort claims should extend to design defect as well as claims based on inadequate warning.  But although [defendant] cannot legally make [the drug] in another composition . . . it certainly can choose not to make the drug at all . . . . This is second-guessing the FDA (unless new information emerged known to the maker but not the FDA), but . . . while the generic maker has no choice as to label – the decision to make the drug and market it in New Hampshire is wholly its own.

2012 WL 1522004, at *4 (citations omitted) (emphasis added).

This is a startling result.  The claim Bartlett permits, that state tort law can penalize purportedly “unsafe” generic drugs simply for being on the market in the first place, has been raised on numerous occasions and universally rejected.  Indeed, it was rejected in Mensing itself.  The Eighth Circuit held, in the opinion that the Supreme Court later reversed:

The generic defendants were not compelled to market [their drug]. If they realized their label was insufficient but did not believe they could even propose a label change, they could have simply stopped selling the product.

Mensing v. Wyeth, Inc., 588 F.3d 603, 611 (8th Cir. 2009), rev’d, 131 S. Ct. 2567 (2011).  The Supreme Court, of course, reversed Mensing in toto – finding preempted every claim that the plaintiffs raised.  Then, as other courts have described, plaintiffs highlighted the duty-to-withdraw theory in seeking reargument in Mensing and lost again:

The Supreme Court rejected this reasoning on appeal and denied a rehearing despite the contentions . . . “that the Petitioners could have ‘independently’ complied with both state and federal law simply by suspending sales. . . .”

Gross v. Pfizer, Inc., ___ F. Supp.2d ___, 2011 WL 5865267, at *3 (D. Md. Nov. 22, 2011) (quoting plaintiffs’ reconsideration brief in Mensing).  See also Mensing v. Wyeth, Inc., 658 F.3d 867 (8th Cir. 2012) (vacating portion of opinion that permitted failure-to-withdraw argument).

Moreover, as a regulatory matter, the FDCA requires generic drugs to be bioequivalent to the branded ("reference listed") drug.  As the Court noted in Mensing, a “generic drug” must be “identical in active ingredients, safety, and efficacy.  131 S.Ct. at 2574 n.2 (emphasis added).  The statute requires “that the active ingredients of the new drug are of the same pharmacological or therapeutic class as those of the listed drug.”  21 U.S.C. § 355(j)(2)(A)(iv).  The FDA’s regulations are similar, mandating rejection of an abbreviated new drug application (what an application for generic approval is called) if the ANDA “is insufficient to show that the active ingredient is the same as that of the reference listed drug. 21 C.F.R. §314.127(a)(3)(i); see 21 C.F.R. §314.127(a)(3)(ii) (identical language for products with multiple active ingredients)

There’s that word again – “same.”  Seems like we’ve seen that somewhere before.  Where does Bartlett get the idea that the “sameness” requirement for design is less exacting than the “sameness” requirement for warnings held preemptive in Mensing?  Nowhere.  The opinion offers nothing but rank speculation:

[T]he FDCA might permit states to tell [defendant] it ought not be doing so [selling a generic drug] if risk-benefit analysis weights against the drug, despite what the Supreme Court made of similar arguments in the labeling context.

Bartlett, 2012 WL 1522004, at *4.  The mind boggles.  In Mensing the plaintiffs also offered speculation about what the FDA might have done.  The Court flat out rejected such hypotheses – derisively calling them a “mouse trap game.”  Preemption is simply not affected by speculation that the FDA “might” do something different:

Accepting [plaintiffs’] argument would render conflict pre-emption largely meaningless because it would make most conflicts between state and federal law illusory.  We can often imagine that a third party or the Federal Government might do something that makes it lawful for a private party to accomplish under federal law. . . .  If these conjectures suffice to prevent federal and state law from conflicting for Supremacy Clause purposes, it is unclear when, outside of express pre-emption, the Supremacy Clause would have any force.  We do not read the Supremacy Clause to permit an approach to pre-emption that renders conflict pre-emption all but meaningless.

Mensing, 131 S. Ct. 2578-79 (emphasis original).  Bartlett’s speculation that the identical word, “same,” might for some unknown and unarticulated reason be interpreted (presumably by the FDA) differently in the design context than it has been in the warning context is precisely the kind of speculation that Mensing rejected.  Anybody could always avoid liability by not selling any products at all – but that would make preemption "illusory," and also totally defeat the purpose of the Hatch-Waxman Amendments to encourage production of generic drugs.  Sooner or later, one plaintiff or another will argue that every generic drug ever approved should be removed from the market.

Not surprisingly, before Bartlett the post-Mensing precedents had universally rejected arguments that supposed state-law duties (no state high court has ever recognized such a duty) to remove generic drugs from the market altogether could survive preemption.

[A] state law duty that would compel generic manufacturers to stop production of a drug that under federal law they have the authority to produce . . . would directly conflict with the federal statutory scheme in which Congress vested sole authority with the FDA to determine whether a drug may be marketed in interstate commerce.

Gross, 2011 WL 5865267, at *3.  “Finding that state law prohibits [defendant] from doing what federal law explicitly requires [it] to do would be tantamount to conferring supremacy upon the state law.”  Coney v. Mylan Pharmaceuticals, Inc., 2012 WL 170143, at *5 (S.D. Ga. Jan. 19, 2012) (rejecting failure-to-withdraw claim).  See also Eckhardt v. Qualitest Pharmaceuticals, Inc., ___ F. Supp.2d ___, 2012 WL 1511817, at *6 (S.D. Tex. April 30, 2012) (“a state law requirement that the drug be completely withdrawn from the market . . . would also impermissibly conflict with federal law and be preempted”); Fulgenzi v. PLIVA, Inc., 2012 WL 1110009, at *7 n.5 (N.D. Ohio March 31, 2012) (“reject[ing] Plaintiff's argument that . . . [defendant] could have complied with both state and federal law by choosing to stop selling [the drug]”); Metz v. Wyeth LLC, ___ F. Supp.2d ___, 2012 WL 1058870, at *4 (M.D. Fla. March 28, 2012) (“[t]o the extent Plaintiffs contend that [defendant] should have pulled the generic version . . . from the market, such claim is also preempted”); Cooper v. Wyeth, Inc., 2012 WL 733846, at *6 (M.D. La. March 6, 2012) (“[i]f state law could require a generic drug manufacturer to wholly withdraw from the market . . . it necessarily must repudiate the label approved by the FDA”); In re Darvocet, Darvon & Propoxyphene Products Liability Litigation, 2012 WL 718618, at *3 (E.D. Ky. March 5, 2012) (“the idea that [defendants] should have simply stopped selling [the drug] is an oversimplified solution that could apply anytime the issue of impossibility preemption arises: avoid a conflict between state and federal law by withdrawing from the regulated conduct altogether”); Bowman v. Wyeth, LLC, 2012 WL 684116, at *6 (D. Minn. March 2, 2012) (“reject[ing] th[e] argument” that “manufacturers could have complied with state and federal law by suspending sales”); Moretti v. Mutual Pharmaceutical Co., ___ F. Supp.2d ___, 2012 WL 465867, at *5 (D. Minn. Feb. 13, 2012) (failure-to-withdraw “reject[ed]” as an “additional attempt to recast [plaintiff’s] failure to warn claims”); Lyman v. Pfizer, Inc., 2012 WL 368675, at *4 (D. Vt. Feb. 3, 2012) (the “’federal duty of sameness,’ . . . applies to the design or composition of the drug as well as to its labeling”); Fullington v. PLIVA, Inc., 2011 WL 6153608, at *6 (E.D. Ark. Dec. 12, 2011) (argument that “generic manufacturers could have complied with both state and federal law by simply pulling the drug off the market entirely” has been “overruled”).

Basically, the court in Bartlett decided it liked the outcome in Levine better than it liked the result in Mensing, even though Mensing dealt specifically with generic drugs, the product involved in Bartlett.  Thus the court tried to minimize Mensing as merely “carv[ing] out an exception" to LevineBartlett, 2012 WL 1522004, at *4.  That’s just result-oriented sophistry.  The First Circuit’s claim that, in Levine, “the Court adopted a general no-preemption rule, 2012 WL 1522004, at *5, is simply wrong – an attempt to resurrect a presumption against preemption that the Mensing court refused to apply to generic drugs, and that four justices would abolish altogether in all cases.  131 S. Ct. at 2579-80.

Actually, both Levine and Mensing, being implied preemption cases, set forth generally applicable rules.  The rule in Levine, as much as we strongly believe otherwise, was that “the manufacturer bears responsibility for the content of its label at all times.”  555 U.S. at 570-71.  Note the word "label."  Levine was conspicuously silent about drug design (having deliberately avoided failure to contraindicate), let alone generic drug design.  The general principle established by Mensing was “when a party cannot satisfy its state duties without the Federal Government's special permission and assistance, which is dependent on the exercise of judgment by a federal agency, that party cannot independently satisfy those state duties for pre-emption purposes.”  131 S. Ct. at 2581.  That test is met for all drug designs, generic or branded.  The notion that Levine was based on some general principle was rejected in Mensing, which recognized that Levine turned on what the FDA allowed branded manufacturers (unlike generics) to do:

[Levine] is not to the contrary.  In that case, as here, the plaintiff contended that a drug manufacturer had breached a state tort-law duty to provide an adequate warning label.  The Court held that the lawsuit was not pre-empted because it was possible for . . . a brand-name drug manufacturer, to comply with both state and federal law.  Thus, the federal regulations applicable [in Levine] allowed the company, of its own volition, to strengthen its label in compliance with its state tort duty.

Mensing, 131 S. Ct. at 2581.  The FDA's regulations governing generic drugs do not confer this ability.  Id. at 2575-76.

Thus, having now thoroughly examined the substantive holdings in Bartlett (its evidentiary aspects, a rather tautological discussion of adverse event reports, must await another day), we remain convinced that the First Circuit decided that it would affirm no matter what, and in the process gave the back (or possibly some other part) of the hand to the Supreme Court’s Mensing decision.  Not only does that make Bartlett an excellent candidate for review and reversal by the High Court, but we would not be at all surprised if a such a reversal were ordered summarily.