DOJ Defends Use of Statistical Sampling To Prove FCA Liability

In a case in which it chose not to intervene, DOJ has stepped in to defend the relator’s attempt to use statistical sampling to prove FCA liability.  The relators allege that Select Medical Corporation, a related entity in Evansville, Indiana, and a physician violated the FCA by making medically unnecessary admissions to long term care facilities and increasing reimbursements by manipulating patients’ lengths of stay and falsifying diagnoses.  The relators contend that their claims encompass Select’s facilities nationwide, and advised the Court that they intended to propose a plan to use statistical sampling to establish FCA liability based on evaluation of a subset of medical records from facilities throughout the country.  The defendants contend that the claims should be limited to the named facility in Evansville.

After the parties raised the dispute with the court, the magistrate judge issued an order rejecting the relator’s proposal and limiting the scope of discovery to the one facility.  As a preliminary matter, the magistrate judge rejected the relators’ attempts to characterize their allegations as encompassing “a nationwide Medicare fraud case,” finding that the allegations were most appropriately read to be limited to the one named facility in Indiana.  But in addition, the court held, sampling would be inappropriate because “it ignores what the plaintiffs would ultimately need to prove to prevail in this case,” noting that relators had presented “no authority for the proposition that proving that a particular Medicare reimbursement claim was fraudulent based on a theory of lack of medical necessity can be done by a random-sampling method that does not evaluate whether each particular claim for which the plaintiffs seek relief was actually knowingly false within the meaning of the FCA.”  The court concluded that “fraud will have to be proved on a claim-by-claim basis based on the patient’s actual medical condition and actual medical care,” on an individualized basis.

The relators subsequently filed with the district judge objections to the magistrate judge’s order.  On May 11, DOJ submitted a Statement of Interest in support of the relators’ objections, in which it argues that the magistrate judge’s order should be set aside as clearly erroneous “because it is contrary to long-established precedent recognizing statistical sampling as an admissible and valid method of proof in complex cases involving large numbers of claims, including cases brought under the FCA.”  DOJ argues that if the government cannot utilize sampling in FCA cases, “then defendants would be incentivized to commit fraud on a large scale, knowing that the government could not present all of the defendant’s false claims individually to the jury.”

As we have previously reported, contrary to DOJ’s contention, the appropriateness of statistical sampling to establish liability in FCA cases (as distinct from damages once liability has already been established), has been a hotly contested issue among lower courts, with little guidance from courts of appeal or the Supreme Court.  How this issue is resolved has significant implications on the scope of FCA claims going forward, particularly those based on lack of medical necessity.  The district court has not yet ruled on the relator’s objections, but we will continue to monitor and report on these developments.

The case is U.S. ex rel. Conroy v. Select Medical Corporation, No. 3:12-cv-00051 (S.D. Ind.).