947 F.3d 321
5th Cir.2020Background
- Palm Valley Health Care, a home health provider, was audited for claims from July 1, 2006 to January 31, 2009; 10,699 claims in the universe and a random sample of 54 were reviewed.
- The contractor found 29 sampled claims lacked coverage because beneficiaries were not "homebound" or did not need skilled care; sample overpayments totaled $81,681.03 and extrapolation produced a $12,589,185 repayment demand.
- Palm Valley pursued the full administrative appeals process (redetermination, reconsideration, ALJ, Medicare Appeals Council); final administrative outcome reduced the uncovered-sample findings to 25 claims and the Appeals Council affirmed the ALJ in large part.
- Palm Valley sued in district court challenging: (1) statutory deadline delays as a due process violation; (2) the legal standard for "homebound"; (3) the sufficiency of evidence that beneficiaries were not homebound; and (4) the statistical extrapolation/Daubert-style attack on methodology.
- The district court affirmed; the Fifth Circuit considered due process, exhaustion of the legal-standard challenge, substantial-evidence review of individual claim findings, and the propriety of extrapolation methodology.
Issues
| Issue | Plaintiff's Argument | Defendant's Argument | Held |
|---|---|---|---|
| Due-process claim based on administrative delays | Delays across appeal stages (multi-year) violated due process | Provider received hearings at each stage; statute violations don’t automatically equal constitutional violations; Palm Valley could have escalated to avoid delay | Denied — no due-process violation; no substantial prejudice shown and provider failed to use statutory escalation or seek injunction |
| Proper legal standard for "homebound" | "Should" language in 42 U.S.C. §1395f(a) means a normal inability to leave home is a useful but non-mandatory test | Agency and ALJ applied standard requiring a normal inability and that leaving home be "considerable and taxing" consistent with statute and manuals | Not considered on merits — claim not exhausted before Appeals Council; court declined to hear it |
| Substantial-evidence support that sampled beneficiaries were not homebound | Interviews were stale (years later) and therefore unreliable | Interviews, contemporaneous clinical records, and provider records corroborated non-homebound activity; factfinder can weigh witness credibility | Affirmed — substantial evidence supports non-homebound findings for 25 claims |
| Extrapolation/statistics and Daubert challenge | Sampling/extrapolation methodology is unreliable and insufficiently precise; seek Daubert exclusion | Daubert gatekeeping does not apply to agency proceedings; sample was a valid probability sample and agency used conservative estimate (lower 90% CI) | Affirmed — extrapolation methodology permissible and supported as substantial evidence |
Key Cases Cited
- Maxmed Healthcare, Inc. v. Price, 860 F.3d 335 (5th Cir. 2017) (upholding use of extrapolation in Medicare overpayment determinations)
- Caring Hearts Pers. Home Servs. v. Burwell, 824 F.3d 968 (10th Cir. 2016) (discussing the good-faith statutory defense under 42 U.S.C. §1395pp)
- Matthews v. Eldridge, 424 U.S. 319 (1976) (Mathews balancing test for due process)
- Cleveland Bd. of Educ. v. Loudermill, 470 U.S. 532 (1985) (statutory procedures vs. constitutional minimum due process)
- Family Rehab., Inc. v. Azar, 886 F.3d 496 (5th Cir. 2018) (addressing recoupment and timeliness in Medicare appeals)
- Donahue v. Barnhart, 279 F.3d 441 (7th Cir. 2002) (noting Daubert’s reliability concerns inform administrative substantial-evidence review)