Photo of Bexis

Judge Pratter’s recent decision in Wawrzynek v. Statprobe, Inc., No. 05-1342, 2007 WL 3146792 (E.D. Pa. Oct. 25, 2007), recently caught our eye. The decision raises a host of interesting questions.
First, how the heck do you pronounce that plaintiff’s name?
But wait, there’s more!
In a nutshell, Statprobe is a contract research organization that helps drug and device companies monitor clinical trials and analyze statistical data from the trials. Gliatech hired Statprobe to provide services in connection with its clinical trials of ADCON-L, a Class III medical device (requiring approval under the PMA process) intended to reduce the development of scar tissue after lower back surgery. Mr. and Mrs. Unpronounceable claim that Statprobe is liable to them for fraud, negligence, and loss of consortium.
Judge Pratter’s opinion reports these facts: Under the ADCON-L study protocol, a single, blinded neuroradiologist was to examine MRIs and record each patient’s “scar score” in pen. When interim data did not show ADCON-L’s efficacy, Statprobe released the randomization code to Gliatech, and the neuroradiologist then re-read the MRIs in the presence of two Gliatech employees (who knew which patients were receiving active treatment and which were in the control group). The Gliatech employees recorded the new results in pencil; a later FDA investigation found many erasures within the data. The re-read significantly improved the performance results for ADCON-L.
Statprobe submitted a final clinical study report that did not mention the substituted data or the multiple violations of the study protocol. On May 27, 1998, the FDA gave Gliatech conditional approval to distribute ADCON-L. The Department of Justice later prosecuted Gliatech, which entered a guilty plea in March 2002 for failure to submit adverse event reports, failure to maintain accurate files, and (most importantly, to our eye) submitting a false or misleading report to the FDA.
In February 1999 — after the FDA had granted conditional approval to ADCON-L, but before Gliatech had submitted the final report to the FDA — Mrs. Wawrzynek had back surgery in which ADCON-L was used. Plaintiffs’ expert says that the the use of ADCON-L caused a dural leak requiring a second surgery; the defense expert says it was not ADCON-L, but an infection, that necessitated the second surgery. Either way, Mrs. Wawrzynek suffered a bad outcome and two additional back operations.
The Wawrzyneks sued the health care providers for medical malpractice; the doctors won. The Wawrzyneks sued Gliatech; the parties settled. And, in 2005, the Wawrzyneks sued Statprobe. Statprobe moved for summary judgment.
The Court rejected Statprobe’s statute of limitations defense. Those facts, and that analysis, are largely case-specific, so they don’t interest us too much.
But the Court then went on to consider Statprobe’s preemption argument. Mention the word “preemption,” and you know you’ve caught our eye.
Judge Pratter discussed Medtronic v. Lohr, 518 U.S. 470 (1996), in which the Supreme Court largely rejected the preemption defense in the context of medical devices that enter the market through the 510(k) process. And she discussed Buckman v. Plaintiffs’ Legal Committee, 531 U.S. 341 (2001), in which the Court found that claims of fraud on the FDA are preempted. (Judge Pratter did, however, butcher her statement of the facts in Buckman. In Buckman, AcroMed Corp. was the manufacturer of the medical device; Buckman Company was the consultant that was accused of defrauding the FDA. Your humble scribes devoted most of the 1990s to that litigation; trust us.)
Judge Pratter found the the Wawrzyneks could sue Statprobe for two fraud claims. First, they could sue for fraud on the FDA. Since Gliatech had pleaded guilty to defrauding the FDA, the state court jury would not be asked to find a fraud that the FDA did not believe had occurred. A claim of fraud on the FDA is not preempted by Buckman when the drug company is convicted of that very fraud.
Second, the Wawrzyneks could also pursue a claim of fraud not on the FDA, but on the general public and medical community (because Statprobe may have been required to disclose the concealed data to the FDA, and all physicians who use a device rely on the FDA’s approval). Again, Judge Pratter held that the FDA’s finding of fraud (and Gliatech’s guilty plea) permitted these claims to proceed.
Finally, Judge Pratter rejected, for summary judgment purposes, Statprobe’s arguments that it did not owe a duty to the Wawrzyneks (needed to support a negligence claim) and could not have proximately caused Mrs. Wawrzynek’s injuries.
What do we think?
Actually, this case made us think a lot. What do we think about preemption of fraud on the FDA cases where the FDA has actually gone out and successfully prosecuted a company for fraud? Everybody who reads this blog knows that, when a drug or device company is at peril, no one heeds the call to arms more quickly than the two of us. But what about this? Ultimately, at least on the facts of Wawrzynek, we’ve concluded that we disagree with Judge Pratter on preemption. Here’s why.
First, the basic reason that the Supreme Court in Buckman decided that fraud on the FDA claims should be preempted had to do with FDA prosecutorial power and discretion. The Court held, first and foremost:

The conflict stems from the fact that the federal statutory scheme amply empowers the FDA to punish and deter fraud against the Administration, and that this authority is used by the Administration to achieve a somewhat delicate balance of statutory objectives.

531 U.S. at 348. The Court went on to say a lot of other things, but this was its main reason.
Well, in Wawrzynek, we can’t find any indication that the FDA “threw the book” at the company – ordering the device off the market or something like that. Nor did the plea agreement that concluded the prosecution contain (as far as Judge Pratter explains, anyway) any requirement for payment of restitution to persons injured by the device. If either of those things had happened, even we wouldn’t consider arguing for preemption in this case.
But they didn’t. Apparently the FDA, in its prosecutorial discretion, thought that the product should still be marketed and did not find the fraud bad enough to require restitution. Since that’s not what the FDA determined should be the appropriate punishment for this particular fraud, that’s not something a state-law cause of action can demand under Buckman. We think Judge Pratter should have followed the seven-justices who refused to create any categorical exception for FDA-prosecuted fraud – rather than go along with only two concurring justices who would have recognized that exception
Sure, it’s a closer call when the company has pleaded guilty to submitting false or misleading reports to the FDA. None of the practical considerations that the Court in Buckman mentioned to bolster its preemption conclusions apply when the FDA has already decided it was defrauded. There’s no preempted second-guessing of the FDA’s answer to the basic question of whether there’s a violation.
But there is still second-guessing of the FDA’s prosecutorial discretion, which as Buckman recognized is quite broad. 531 U.S. at 349. Since there’s no presumption against preemption in a fraud on the FDA case, we think that’s enough. The FDA didn’t impose the death penalty on the product here. Nor did it order restitution. State plaintiffs don’t have the power to second-guess the FDA’s choice of criminal penalties.
And Statprobe, of course, didn’t even involve the manufacturer that had pleaded guilty. It involved the contract research organization that had not been found guilty. Rejecting the preemption defense on those facts strikes us as even a greater stretch than to allow a claim against a guilty manufacturer. Despite what Judge Pratter said, there were exactly zero Supreme Court votes in Buckman for any claim against any defendant not already adjudicated guilty of fraud. Once again, the issue is prosecutorial discretion – only here the FDA presumably examined the CRO’s conduct and chose not to prosecute it at all. A later jury is thus not being asked to follow the FDA’s lead, but rather to find fraud where the FDA did not (or, at a minimum, where the FDA exercised its discretion not to prosecute). Permitting juries to second-guess the FDA’s decision that Statprobe did not engage in a fraud worthy of prosecution poses many of the risks that Buckman tried to avoid. CRO’s will be tempted to include unnecessary data in FDA submissions. And woe to any CRO that’s in the neighborhood when a manufacturer defrauds the FDA; guilty or no, plaintiff’s counsel will stink up the courtroom at trial, and the CRO will not come away smelling pretty.
Second, we don’t like the holding that a claim of fraud on the FDA can be transmogrified into a claim of fraud on the world — all treating physicians and the general public. Statprobe never communicated with either the Wawrzyneks or their health care providers. We’re not big fans of limitless fraud, where an alleged misstatement to one entity is transformed into a cause of action for hundreds — or thousands, or millions — of patients. In addition to the many legal objections to those types of claims, there’s also the policy ground that one achieves appropriate levels of deterrence only by crafting penalties that are tailored to the offense. (Of course, if everyone and his brother can sue for the fraud on the FDA, we’re not quite sure why it matters if the fraud is instead characterized as one on the general public. The number of potential plaintiffs seems pretty much the same, no matter the name the claim is given.)
Whether one agrees with the result in Statprobe or not, however, it certainly raises the stakes for everyone associated with a company that is convicted of defrauding the FDA. That triggers our two recommendations:
1. Drug companies: Don’t defraud the FDA! It’s wrong! It’s illegal! Don’t do it.
2. Contract research organizations: Monitor Statprobe and its progeny closely. This issue is high-stakes poker for everyone in your field.