A variety of cases have addressed the role of statistical sampling in healthcare matters, including litigation before courts, administrative hearings and other matters … Statistical Sampling Case Law

The following cases, while not exhaustive, represent notable case facts and rulings related to statistical sampling:Statistical Sampling Case Law

Illinois Physicians Union et al. v. Jeffrey Miller et al. (1982)[1]

The plaintiff argued that the use of statistical sampling violated the Fourteenth Amendment and Illinois State law when used for the purposes of a routine audit. It argued that due process required itemized proof of liability for each unit (in this case, patient records).  The court ruled, in favor of the defendant, that a claim-by-claim review would be impractical due to the large volume of data and that statistical sampling was reasonable.  The court viewed the issue as whether the state can place the burden of showing the inaccuracy of the defendant’s calculations on the plaintiff, and it decided the burden was, in fact, the plaintiff’s.[2]  The court cited Story Parchment, in which an inexact calculation of damages would prove sufficient “if the evidence show[ed] the extent of the damages as a matter of just and reasonable inference”.[3]  The federal appeals court also agreed with the district court that “use of a sampling and extrapolation auditing procedure is not arbitrary, capricious or invidiously discriminatory”.Statistical Sampling Case Law

Ratanasen v. State of California Department of Health Services (1993)[4]

This case involved claims under California’s Medicaid program (“Medi-Cal”). The federal appeals court approved the use of statistical sampling in connection with Medicare and other similar programs, as long as opposing parties were allowed to rebut such evidence.[5]  The court’s opinion specifically addressed the sampling methodology.  Although the plaintiff argued a sample size representing only 3.4 percent of the total population was inadequate, the court cited Michigan Dept. of Education v. United States Dept. of Education, noting “there is no case law that states how large a percentage of the entire universe must be sampled.”[6]

Chaves County Home Health Service, Inc. et al. v. Louis W. Sullivan (1991)[7]

In this post-payment audit of Medicare payments, the court noted that “absent an explicit provision in the statute that requires individualized claims adjudication … the private interest at stake is easily outweighed by the government interest in minimizing administrative burdens; in light of the fairly low risk of error so long as the extrapolation is made from a representative sample and is statistically significant, the government interest predominates.[8]

Pruchniewski v. Leavitt, (2006)[9]

In this post-payment audit of Medicare payments, plaintiff’s statistical expert contended that the defendant was obligated to sample a minimum of 320 beneficiaries, however, the expert did not introduce any empirical findings to support his opinion that the defendant’s sample size of 30 was too small to be valid, or that a sample of 320 would have produced an estimated overpayment that was below the lower limit of the 90 percent confidence level calculated by the defendant. While the court was troubled by the defendant’s “one size fits all” approach, since it always employed a sample size of 30 regardless of the size of the sampling universe, the court stated it “must conclude that the ALJ’s decision [to admit the sampling analysis] was not contrary to applicable standards,” as neither the CMS guidelines nor the expert’s generally accepted principles were binding on the carrier, nor did they establish minimum sample size.[10]

Foot and Ankle Associations of NC, PLLC (Appellant) (Beneficiary) AdvanceMed (PSC) (Contractor), (2011)[11]

In this MAC decision, the overall precision rate in the stratified statistical sample was over 22 percent. While acknowledging the poor degree of precision (i.e. a high precision percentage), the council found this to be relevant in relationship to the ‘confidence interval’ in which the estimated overpayment falls, not to the validity of the analysis.  Further, the council found the Appellant provided no argument or factual support for the idea that a lower precision rate would have resulted in a lower demanded overpayment amount. Statistical Sampling Case Law

United States ex rel. Martin v. Life Care Centers of America, Inc. (2014)[12]

In this FCA suit, the court addressed the use of statistical sampling in order to prove liability. The court admitted that “[i]n the context of the FCA … statistical sampling has been generally limited to determine damages, rather than liability” and that “[u]sing extrapolation to establish damages when liability has been proven is different than using extrapolation to establish liability.” The court explicitly acknowledged that using “statistical sampling to find liability for extrapolated claims could be in conflict with the government’s burden to establish the elements of a FCA claim.” Although the court ultimately rejected arguments that individualized review was necessary to prove falsity because it “would consume an unacceptable portion of the Court’s limited resources,” it explicitly stated the limitations of its ruling:[13]

While Defendant makes several compelling arguments regarding the inherent limitations associated with statistical sampling, these arguments are better considered by the fact finder rather than the Court. The Court’s ruling today simply holds that statistical sampling may be used to prove claims brought under the FCA involving Medicare overpayment, but it does not and cannot control the weight that the fact finder may accord to the extrapolated evidence. Rather, the burden of determining the weight of the evidence lies with the fact finder.  Read more…

United States ex rel. Wall v. Vista Hospice Care Inc. et al. (2016)[14]

The court granted summary judgment for portions of this False Claims Act suit. It found that an expert physician’s review of sampled hospice patients was not sufficient evidence as to the determination of whether or not a patient has six months to live, and therefore could not be extrapolated for the purposes of establishing liability.  The court said “Without any evidence about the nurses and doctors involved in treating or certifying the sampled patients for hospice, for relator to prevail at trial, jurors would have to take an impermissible inferential leap to conclude that those patients’ certifications were not based on the proper clinical judgment of physicians.”  Additionally, the government’s expert acknowledged errors in the sample’s selection, including having selected duplicate items and failing to appropriately stratify the population. Although the expert claimed the errors were ultimately corrected, insufficient documentation of those corrections was produced to allow sufficient scrutiny by the defendant, which was noted by the court as additional cause for its judgment.  Read more about extrapolating medical eligibility…

Statistical Sampling Case Law

  Explore our Statistical Sampling Services


This post does not, in any way, constitute legal advice, nor does it offer guidance or legal opinions about how courts or other bodies may interpret particular issues of statistical analysis. The cases cited are merely examples of relevant issues, and they are included to demonstrate how the issues were treated based on the facts and circumstances of each specific case.
[1] Illinois Physicians Union v. Miller, 675 F.2d 151, 154 n.4 (7th Cir. 1982).
[2] Weil, Lentz, Hoffman, Litigation Services Handbook, The Role of the Financial Expert, 5th Ed., Wiley, 2012, (Hereafter “Litigation Services Handbook”), p6.13.
[3] Story Parchment Co. v. Miller, v. Patterson Parchment Paper Co. 282 U.S. 555, 51 S. Ct. 248, 75 L.Ed. 544 (1931).
[4] Ratanasen v. State of California Department of Health Services, 11 F.3d 1467 (9th Cir. 1993).
[5] Litigation Services Handbook, p6.15.
[6] Michigan Department of Education v. United States Department of Education et al., 875 F.2d 1196 (1989).
[7] Chaves County Home Health Service, Inc. et al. v. Louis W. Sullivan, 931 F.2d 914, 915, 917-919 (D.C. Cir. 1991).
[8] Chaves County Home Health Service, Inc. et al. v. Louis W. Sullivan, 931 F.2d 914, 915, 917-919 (D.C. Cir. 1991) at 922.
[9] Pruchniewski v. Leavitt, 8:04-CV-2200-T-23TBM (M.D. Fla. 2006).
[10] Statistical Sampling – Evolving Legal Issues; Park R. and Perling L., American Health Lawyers Association.  Available at: https://www.healthlawyers.org/Events/Programs/Materials/Documents/MM12/papers/D_park_perling.pdf.
[11] Foot and Ankle Associations of NC, PLLC (Appellant) (Beneficiary) AdvanceMed (PSC) (Contractor), Claim for Part B Benefits, WL 7007022, Docket No. M-10-369 (2011).
[12] United States ex rel. Martin et al. v. Life Care Centers of America, Inc., 1:08-cv-251, Order, (E.D. Tenn. 2014).
[13] Statistical Sampling Debate: A Growing Web of FCA Cases, Doufekias D. and Kaufmann S., Law360 Expert Analysis, January 19, 2016.
[14] U.S. ex rel. Wall v. Vista Hospice Care Inc. et al., 3:07-cv-00604, (M.D. Tex. 2016).

© Forensus Group, LLC  |  2017 Statistical Sampling Case Law

Statistical Sampling Case Law

Statistical Sampling Case Law