Tyson Provides Road Map For Use Of Sampling In FCA Cases Law360, New York (March 31, 2016, 11:13 AM ET) -- The circumstances under which statistical sampling can be used in order to establish liability in False Claims Act cases remains a controversial issue. Defendants argue that sampling results in a “trial by formula” that can result in ruinous liability based on evidence relating to a relatively small number of claims, while the U.S. Department of Justice and relators’ counsel argue that statistical sampling is an accepted form of evidence, and that limiting its use in FCA cases would have the perverse effect of immunizing the largest frauds from the full scope of liability. District courts have recently taken different approaches to the use of statistical sampling in FCA cases to establish liability. In addition, a case pending before the Fourth Circuit may provide the first appellate-level guidance. See United States ex rel. Michaels v. Agape Senior Community Inc., No. 15-2145 (4th Cir.). And while it did not involve or address FCA claims specifically, the U.S. Supreme Court’s March 22 opinion in Tyson Foods v. Bouphakeo offers support to both proponents and detractors of the use of statistical sampling to prove liability in FCA cases. In Tyson, a class of workers at a meat processing plant sued their employer under the Fair Labor Standards Act, alleging that the company failed to pay them overtime compensation for time spent donning and doffing their protective wear. Tyson did not maintain records of how long each employee spent donning and doffing each day. To determine Tyson’s liability, the plaintiffs retained an expert to estimate the average time employees in different departments spent donning and doffing their applicable protective gear. Another expert compared these averages to each employee’s documented work time to identify workers with respect to whom Tyson had allegedly violated the FLSA. The central dispute in the case was whether individual variances in time spent donning and doffing made this approach too speculative to sustain classwide liability. The court, affirming the lower court’s judgment, refused to create a “categorical exclusion” of statistical sampling to establish liability, instead explaining that the use of statistical sampling would hinge, like other evidence, on its reliability. Though Tyson is not an FCA case, the court’s rejection of broad-based arguments against the use of sampling seems to suggest that arguments in favor of a blanket rule prohibiting sampling to establish FCA liability are unlikely to find a welcome reception going forward in the lower courts. However, Tyson does offer ammunition to those seeking to limit the use of sampling, at least in particular types of FCA cases. While eschewing a blanket prohibition on sampling, the court noted that the “fairness and utility of statistical methods in contexts other than those presented here will depend on facts and circumstances particular to those cases.” Factors the court raised as impacting whether sampling is appropriate including: the extent to which the sample is meaningfully representative of the universe of claims; the extent to which the statistical sampling methodology is otherwise reliable; the elements of the underlying cause of action; the 2 purpose for which the evidence is introduced; the harm to a defendant’s ability to assert individual defenses; and the availability of substitute evidence, including direct proof. Within this range of considerations, the opinion preserves certain arguments weighing in favor of a limited role for representative evidence of liability in FCA suits. Thus far, plaintiffs have primarily sought to use statistical sampling to establish liability in FCA cases based on lack of medical necessity, such as those involving hospice eligibility. Defendants have argued that representative evidence is inappropriate in such cases given the highly fact-intensive nature of each claim. Under this view, statistical sampling robs defendants of the opportunity to present patient-specific evidence in defense. The court in Tyson made clear that application of the expert’s average donning and doffing times to establish classwide liability “did not deprive [Tyson] of its ability to litigate individual defenses.” This was so because, given the absence of individualized donning and doffing records, “there were no alternative means for the employees to establish their hours worked.” Therefore regardless of whether each employee litigated alone, Tyson’s defense would be premised on attacking the reliability of the study, i.e., this was a defense common to each employee’s claim. In contrast, defendants in FCA suits will generally have individualized evidence and they can continue to maintain their right to present claim-specific defenses. The role of statistical sampling in cases in which there is no missing individual evidence — indeed where defendants are demanding an opportunity to present additional evidence — remains an open question even after Tyson. The court implied that sampling is more likely to be appropriate if the defendant’s own failure adequately to maintain and preserve records caused the evidentiary gap that would otherwise hinder the plaintiffs. Even the case on which the court relied for its holding involved statistical sampling “to fill an evidentiary gap,” and Supreme Court Justice Clarence Thomas interpreted the majority’s evidentiary rule as “limited to cases where the employer breaches its obligation to keep records of employees’ compensable work.” This suggests that sampling could be available in a case involving inadequate recordkeeping, but not, for example, in a case challenging medical necessity in which the underlying medical records exist. Indeed the district court in United States ex rel. Michaels v. Agape Senior Community Inc. adopted just such a position, constraining statistical sampling to cases in which direct proof of liability is not available. The Tyson opinion preserves, and arguably reinforces, the position that direct proof is preferable where available. On this point, defendants can further leverage arguments Justice Thomas raised in his dissent. According to Justice Thomas, the class should not have been certified because the considerable variation in each employee’s donning and doffing times meant that the expert’s average estimates were not reliable common proof. The majority seemingly resolved this concern by relying on the fact that individual times could not be recreated and therefore the average estimates were the same evidence Tyson would need to defend against in either a class or individual action. However, the conclusion that statistical sampling is appropriate in these narrow circumstances does not necessarily extend to FCA cases in which there may be not only a rich record of individualized evidence, but also significant variability among the claims at issue. 3 Indeed, many FCA cases will likely implicate diverse claims, more akin to the variability between class members in Wal-Mart Stores Inc. v. Dukes — a class of employees the court refused to certify because the experiences of a subset of the employees was not probative of the experiences of the class. While the majority distinguished the Tyson case from Dukes, many defendants in FCA suits may be able successfully to analogize their circumstances to the diverse plaintiffs in Dukes. Even in FCA cases in which courts are willing to apply statistical sampling because they view the claims as having sufficient commonalities, Tyson confirms that the particular sampling methods used must be tested and shown to withstand scrutiny under the Federal Rules of Evidence. Sample size will be a critical component of any method’s reliability. As the dissent notes, the court in Dukes refused to credit 120 employee anecdotes as representative evidence across the class, distinguishing this tiny sample from the more “significant” 1-to-8 ratio that the court had previously approved in a case involving representative evidence. Tyson involved a 1-to-60 ratio between the sample and the full universe. Even where variability among claims falls short of convincing a court to rely on direct proof, it may nonetheless support defense arguments as to the appropriate size and structure of sample universes. The court’s statement that whether sampling is appropriate depends on “the elements of the cause of action” provides an additional reason more broadly to be skeptical of the appropriateness of sampling to establish FCA liability. The court found that the factual question tried in Tyson regarding whether employees worked time for which they were not compensated was susceptible to proof through statistical sampling, based on an examination of a valid sample of employees actually donning and doffing equipment. But in the FCA context, scienter — proof that the defendant “knowingly” submitted a false claim — is a critical element that was not at issue in Tyson. The appropriate use of statistical sampling to prove liability for fraud remains an arena for debate. Even if statistical sampling was appropriate to identify a set of claims representative of the entire universe with respect to falsity, it does not necessarily follow that sampling could establish that that the defendant “knowingly” caused that same population of “false claims” to be submitted. As one district court explained in the course of ordering a trial bifurcated between the falsity of the sampled claims and the remainder of the government’s burden of proof, “evidence of general corporate practices ... unrelated to the 233-patient sample is not relevant to whether a specific claim is false.” The disconnect between falsity and scienter at the very least creates trial challenges that may warrant additional protections for defendants. Tyson promises to be a source of justification for both sides of the sampling debate, and the interwoven considerations the court sets forth underscore that defendants in FCA suits retain room to raise arguments attacking the suitability of statistical sampling to establish liability. —By Scott D. Stein and Brenna E. Jenny, Sidley Austin LLP Scott Stein is a partner in Sidley's Chicago office. His practice focuses on complex health care civil litigation, with a special emphasis on False Claims Act and antitrust 4 litigation. Brenna Jenny is an associate in Sidley's Washington, D.C., office. She is part of the firm's health care group where she advises FDA-regulated companies and large provider groups on a range of regulatory and compliance matters, with an emphasis on government enforcement actions. The opinions expressed are those of the author(s) and do not necessarily reflect the views of the firm, its clients, or Portfolio Media Inc., or any of its or their respective affiliates. This article is for general information purposes and is not intended to be and should not be taken as legal advice.  Compare, e.g., United States ex rel. Martin v. Life Care Centers of America Inc., No. 08- cv-251 (E.D. Tenn. Feb. 18, 2015) (“[A]s long as the statistical sample is a valid sample that is representative of the universe of claims, the natural disparity between the claims does not preclude using sampling and extrapolation as evidence of the total number of claims for noncovered services.”) and United States ex rel. Ruckh v. Genoa Healthcare LLC, No. 11- cv-01303 (M.D. Fla. April 28, 2015) (same) with United States v. AseraCare Inc., No. 12-cv- 00245 (N.D. Ala. May 20, 2015) (allowing relators to use statistical sampling to establish liability but bifurcating the trial so that the first phase was limited to establishing falsity of claims within the sample) with United States ex rel. Michaels v. Agape Senior Community Inc., No. 12-3466 (D.S.C. June 25, 2015) (refusing to allow relators to use statistical sampling to establish liability where evidence was available for review and claims were highly fact-intensive), appeal docketed No. 15-2145 (4th Cir.).  United States v. AseraCare Inc., No. 12-cv-00245 (N.D. Ala. May 20, 2015).