November 23, 2009
Guest Blogger Ben Zipursky on "The Differential Treatment of Medical Devices and Drugs in Preemption Doctrine: A Justified Distinction?"
One year ago, many tort scholars and securities analysts believed that a major shift was about to occur in products liability law: our nation’s highest court was about to declare that claims against pharmaceutical manufacturers were preempted. We all know that this did not occur, that in the much-awaited decision of Wyeth v. Levine, six members of the Court decided to leave state tort law for prescription drugs alone, at least for the moment. Although I believe that the case was correctly decided, I am happy to admit that the decision came as something of a surprise. Part of the reason for that surprise was the Court’s lopsided decision in favor of preemption just one year earlier, in Riegel v. Medtronic. There, eight members of the Court voted in favor of a broad preemption doctrine in medical device cases. Five of the Justices who voted for preemption in Riegel – Stevens, Kennedy, Thomas, Souter, and Breyer – voted against preemption in Wyeth.
The combination of preemption in Riegel and no-preemption in Wyeth does not present any truly sharp inconsistencies or paradoxes, but it nevertheless merits exploration. There is no actual contradiction between the cases for many reasons. Most notably, the two cases presented questions of statutory interpretation, but did not involve the same statute. To get right to the point, Riegel applied a statute that actually contained an express preemption clause, while Wyeth applied a statute that lacked such a clause. From this point of view, one could have known everything one needed to know about Wyeth by looking at Justice Stevens’ concurring opinion in Riegel, which essentially announced: “Make no mistake about it; I am voting for preemption here only because there is a preemption clause in the text of this statute.”
Nevertheless, it is quite unsatisfying to explain away the distinction between these cases simply by reference to the textual difference. To begin with, if we are candid we must admit that there was virtually no federal preemption in any cases involving FDA approved products a couple of decades ago; there has been a strong, industry-based push to create preemption doctrine, and it has been quite successful. Where it matters most, however – the Supreme Court – there is a striking discrepancy. There is a clear acceptance for (a large class of) medical devices and a clear rejection (so far) for drugs. Yet the policy motivation is exactly the same – it makes sense to defer to federal agency expertise, and to forestall the costly state-by-state variations, especially given the relatively low probativeness of jury decisions on safety.
Second, notwithstanding the eight votes for Medtronic in Riegel, the statutory interpretation argument in that case is far weaker than the Court admitted. The Court has been fooling around with the concept of a “requirement” in express preemption clauses. There is no evidence that Congress was conceiving of state tort liability as a form of legal requirement and, if there is any straightforward reading of the text, it does not cut in favor of preemption. Obviously, there was already good precedent for Justice Scalia to build upon (Cippollone and Geier) to make this argument. But those arguments were far from unassailable, both because each came with a fractured Court whose members wrote highly nuanced analyses, and because each of those products (cigarettes, airbags) was plainly unique in a way that mattered to the significance of agency decisionmaking. More importantly, the precedents themselves reflected that many of the Justices were engaged by pro-preemption motivations.
The other side of the puzzle is that, while nothing in the text of the statute applicable in Wyeth permitted an easy, across-the-board express preemption argument there, plenty of opportunities to rule for Wyeth presented themselves. Quite apart from the opportunities for deference to the FDA, which had taken a strong pro-preemption position, there were quite powerful implied preemption arguments made. And the analogy to Geier, although imperfect, was quite strong. Moreover, as Justice Alito’s quite fact-based and careful dissent illustrates, there was a fairly persuasive argument for Wyeth on the facts.
And now, let’s admit that part of the point of preemption doctrine is that the Court gets to play the role of rationalizing the big picture scheme of the law. If two aspects of the law – state and federal – are too badly out of whack with one another, then the Court will exercise its power to create a smoothing-out of the law. It is odd, then, that the Court seems happily to countenance a stark difference in treatment between device law and drug law. The discussion above is merely meant to remind us that there was nothing inevitable about this. Finally, even if one were to take the view that statutory differences fully warranted the discrepancy, one would want to ask why medical device makers were given so much more protection against state tort liability.
This is all by way of a lengthy lead up to a tentative suggestion, one that is truly meant to invite discussion. Riegel – like most medical device cases (I believe … but would be happy to be corrected) – is a design defect case. Wyeth – like most prescription drug cases – is a failure-to-warn case. That means that, at a gestalt level (and, to some extent, at a practical level), the Court was deciding in Riegel whether juries would be allowed to displace FDA decisions on medical device design, while the Court in Wyeth was deciding whether juries would be allowed to displace FDA decisions on what information drug companies must convey to prescribing healthcare providers. Is it possible that the difference between displacing design decisions and displacing informational decisions is part of what led us to where we are now? If so, would this be a rational line to draw? Does the difference between the two types of questions shed any light on why medical devices and drugs might be drawn differently? (Is this distinction overdrawn?) If so why?
When I put this question to myself, my mind tracked over to the somewhat analogous field of medical malpractice. As torts professors and medical malpractice lawyers know, courts and legislatures do not look kindly on juries second-guessing experts on technical and scientific questions regarding the standard-of-care in medical malpractice cases. It is black-letter law that juries cannot second-guess the standard-of-care accepted by experts in the field; T.J. Hooper does not apply to medical malpractice. However, in a large number of jurisdictions today, informed consent cases work differently: jurors may second-guess what experts have decided is the right level of information disclosure. That is, in part, because courts are willing to empower juries to second-guess experts on normative decisions about how much safety information should be disclosed. This is the so-called “patient rule” rather than the “physician rule.”
The question raised is thus: does the Riegel/Wyeth dichotomy derive from a distinction similar to the medical malpractice/informed consent distinction? Are the five justices who voted for Medtronic and against Wyeth preserving juror power to second-guess experts on informational decisions? If so, is this a justifiable position to take?
N.B. This contribution is written in the spirit of conversation, not scholarship. Many of the suggestions thrown out are off the cuff – some, perhaps, are demonstrably false. Feel free to point this out, if so.
- Professor Benjamin C. Zipursky
Fordham University School of Law
TrackBack URL for this entry:
Listed below are links to weblogs that reference Guest Blogger Ben Zipursky on "The Differential Treatment of Medical Devices and Drugs in Preemption Doctrine: A Justified Distinction?":
I find myself in agreement with you all the way up to your distinction based on type of defect. Riegel resulted in an overwhelming decision for preemption, such that it looked like rather than Congressional intent, it was based on several Justice's views about policy. Justice Ginsburg's dissent reveals how weak the textual argument was. I believed the Court would continue in the same vein in Levine.
But I depart when it comes to design cases justifying deference while warnings do not. The simple reason is that the approval process for medical devices does not include review of the design to determine whether it reasonably safe (in the Third Restatement sense). Rather, all it requires is that the manufacturer demonstrate that the device works (efficacy) and that its benefits outweigh its risks. So, in one sense, designs for medical devices deserve no deference by common law courts and less than warnings, which the FDA explicitly does approve. The difficulty in deferring to warnings is the dynamic process of identifying risks once the drug is approved and out in the larger market, a point that the Supreme Court got in Levine.
Posted by: Mike Green | Nov 23, 2009 8:59:08 AM
Could the fact that one therapy is " Life Support" and the other "symptom supression" be of any consideration? Hardware, material, function, design vs Software, compounding. One stays in the body the other absorbed..not sure if they, the products, should be mentioned on the same page. Only the FDA and the human body is held in common. Just a thought. P.
Posted by: Peter | Nov 23, 2009 11:04:48 AM
I agree with Michael Green. The FDA's standard for permission to market does not even approximate the product liability standard. The approval standard is simply superiority to placebo, with no comparative element.
After permission to market comes the mass uncontrolled human experiment stage where there is no affirmative duty of product stewardship (which would require makers of new drugs to actively monitor how the experimental drug performs in its subjects' bodies). (See my `Punctuated Equilibrium', Review of Litigation 2007)
Although product stewardship has not been embraced, courts have favored warning theories of drug product liability because informed patient decision-making is a high value - especially among the highly-educated.
As Ben points out, courts are very comfortable with supporting informed consent in a robust way. Wyeth can be explained that way.
The strong Riegel majority's expansive use of the word `requirement' can best be explained by judicial skepticism about the "costs" of product liability. The way to that view is greased by judicial embrace of the myth of deference to the technical expertise of the FDA.
But the FDA does not compare the proposed device's design to its alternatives - except to exempt from vigorous review those for which there has been a "substantial equivalent" on the market.
When new devices without equivalents are proposed and more vigorous new device review is undertaken the FDA does not examine the omitted alternatives because it lacks the capability (and perhaps statutory authority) for that.
Such testing - for a reasonable alternative design - is, however, exactly what goes on in product liability design defect litigation.
So in device cases too the FDA's standard for approval is below that called for in product design defect litigation - the RAD test.
Riegel has barred such exploration. As an immediate result thousands of remedy-less defibrillator and pacemaker patients have implanted in their coronary arteries electrical lead wires which have been shown to be defective in design.
Posted by: George Conk | Nov 23, 2009 9:42:52 PM
Not all medical devices are hardware that remain in the human body until removed. For example, there is a group of Class III medical devices called anti-adhesion barriers that, like drugs, are absorbed into the human body after they are placed.
You make some excellent points, however, I believe that the discrepancy in the Court's decision boils down to one simple point, the one you made at the beginning of your article. In Riegel v Medtronic, the Court was encouraging Congress to clarify the meaning of the word "requirement" in the Medical Device Amendment of 1976. "Requirement" in the context of states not placing design demands upon medical devices is a reasonable mandate and was the original reasoning for the wording of the Medical Device Amendment. In 1976, the idea of removing tort liability had not been considered by Congress because it had never been done. The word "requirement" was used simply to refer to design specifications. And to go back and place our modern day definition of tort requirements into the Amendment is disingenuous and is something that the Supreme Court is encouraging Congress to clarify.
Posted by: Joe | Nov 24, 2009 4:50:19 AM
Mike Green and George Conk and "Joe" have suggested that Riegel was wrongly decided. I agree; my blog was not intended to suggest that I believe Riegel was correctly decided (although I obviously see why one might have thought so). I believe that statutory interpretation is the core of what the Court must do in a case like Riegel and that it was wrong on the statutory interpretation front. Cipollone’s assertion that a damages verdict under a state tort failure to warn claim should be understood as covered by the “requirement or prohibition” language of a preemption clause is questionable, to say the least; Justice Stevens’ analysis of that issue was meager, consisting largely of a citation to Prosser and one to Black’s Law Dictionary. However, tobacco and the legislation surrounding it, regarding warnings, were perhaps a special case. One can imagine a non-textual, intent-based motivation driving Justice Stevens; perhaps he thought that Congress was clearly cutting a deal with tobacco companies, and that liability in private law failure-to-warn claims was either being cut off or being assumed to be a failure as a matter of tort law. Moreover, the Court quite clearly struck a compromise in Cipollone, which could hardly be viewed as cutting off liability for tobacco companies.
When we move over to the context of medical devices in Riegel we simply see the entrenchment of a bad piece of legal and conceptual analysis (on “requirement”), taken to a statute where there is simply no reason to think either text or intent covered the common law action. I believed at the time that Justice Stevens concurred only because he saw that his vote made no difference, and he wanted to plant the text/no text distinction.
At a strategic level, Justice Stevens appears to have been correct; Kennedy, Thomas, Souter, and Breyer all moved with him. However, I do not believe this was a justifiable distinction. Moreover, even apart from Cipollone’s confusion, I do not believe there was a sufficient congealment of precedent on the “requirement” point, at least at the Supreme Court, to render the conceptual analysis of “requirement” something the Court needed to, or should have, taken as a given.
Posted by: Benjamin Zipursky | Nov 24, 2009 11:13:08 AM