Pomona Valley Hospital Medical Center v. Azar

Opinion

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA

____________________________________

)

POMONA VALLEY HOSPITAL ) MEDICAL CENTER, )

)

Plaintiff, )

)

v. ) Civil Action No. 18-2763 (ABJ)

)

ALEX M. AZAR II, ) Secretary, United States Department ) of Health and Human Services, )

)

Defendant. )

____________________________________)

MEMORANDUM OPINION

In this lawsuit against the Secretary of the U.S. Department of Health and Human Services, plaintiff Pomona Valley Hospital Medical Center challenges certain payments it received for Fiscal Years 2006 through 2008 under the Medicare statute. Specifically, it asserts that the Secretary improperly calculated payments owed to it under the disproportionate share hospital (“DSH”) adjustment, which provides an additional payment to hospitals that serve a disproportionately large number of low-income patients. Plaintiff filed an administrative appeal of the calculation to the Provider Reimbursement Review Board, which upheld the calculation. The Secretary adopted the Board’s decision, and plaintiff has filed this lawsuit, arguing that since the calculation was not based on the best available data, the decision to uphold the calculation did not comport with the applicable statute and regulations. Because the Board’s decision is not supported by substantial evidence, the Court will grant plaintiff’s motion for summary judgment in part and remand the matter to the agency for further proceedings consistent with this decision.

2

BACKGROUND

I. Legal Framework A. The Medicare Statue The Medicare Act, 42 U.S.C. § 1395 et seq., provides health insurance to elderly and disabled individuals. The Secretary of the Department of Health and Human Services administers the Medicare program through the Centers for Medicare and Medicaid Services (“CMS”), a component of the department, and CMS contracts with Medicare Administrative Contractors (“MACs”), 1 typically private insurance companies, to determine amounts to be paid to Medicare providers, including hospitals such as plaintiff. 42 U.S.C. § 1395kk; id. § 1395h(a); 42 C.F.R. § 413.24(f).

Medicare is divided into five parts, Parts A through E. Ne. Hosp. Corp. v. Sebelius, 657 F.3d 1, 2 (D.C. Cir. 2011), citing 42 U.S.C. §§ 1395c–1395i–5. Among other things, Medicare Part A provides payments to hospitals for inpatient services provided to Medicare beneficiaries. 42 U.S.C. § 1395c et seq. Hospitals are reimbursed for these services based on their operating costs using standardized rates subject to certain adjustments, such as the DSH adjustment at issue here. 42 U.S.C. § 1395ww(d); Baystate Med. Ctr. v. Leavitt, 545 F. Supp. 2d 20, 22 (D.D.C. 2008).

B. The DSH Adjustment The DSH adjustment provides additional payments to hospitals that serve a disproportionately large number of low-income patients. 42 U.S.C. § 1395ww(d)(5)(F); Adena Reg’l Med. Ctr. v. Leavitt, 527 F.3d 176, 177–78 (D.C. Cir. 2008) (explaining that Congress

1 MACs were formerly referred to as “fiscal intermediaries.” 42 U.S.C. § 1395h(a), 42 C.F.R. § 413.24(f).

3

determined any hospital that serves a disproportionately large percentage of low-income patients should be reimbursed at a higher rate “because the more low-income patients a hospital treats, the more it costs on average to care for Medicare patients”). The Medicare statute provides that a hospital’s DSH adjustment is established using the “disproportionate patient percentage” (“DPP”), 42 U.S.C. § 1395ww(d)(5)(F)(v) and (vi), which is a “proxy” calculation of how many low-income patients a hospital serves. Ne. Hosp. Corp., 657 F.3d at 3. The higher the DPP proxy, the larger the DSH adjustment and the higher the DSH payment a hospital receives. See Cath. Health Initiatives Iowa Corp. v. Sebelius, 718 F.3d 914, 916 (D.C. Cir. 2013).

1. The Disproportionate Patient Percentage DPP is the sum of two fractions. Cath. Health, 718 F.3d at 916. The first fraction seeks to capture those patients served by a hospital who are eligible for Medicare and Supplemental Security Income (“SSI”), which is income provided by the federal Social Security Administration (“SSA”) to financially needy individuals who are aged, blind, or disabled. Smith v. Berryhill, 139 S. Ct. 1765, 1772 (2019); see 42 U.S.C. § 1381 et seq. This fraction is referred to as the Medicare/SSI fraction or simply the SSI fraction. See Cath. Health, 718 F.3d at 916. The second fraction seeks to account for patients who are not eligible for Medicare, but who receive Medicaid, which is a state-administered program for low-income individuals and families. See id. The two fractions provide separate indicators of low income that, when added together, serve as “an indirect, proxy measure for low income.” Id.

This lawsuit concerns the SSI fraction, specifically, the numerator of this fraction.

4

2. The SSI Fraction and Its Numerator The Medicare statute defines the SSI fraction as follows:

[T]he numerator . . . is the number of such hospital’s patient days for such period which were made up of patients who (for such days) were entitled to benefits under part A of this subchapter and were entitled to supplementary security income benefits (excluding any State supplementation) under subchapter XVI of this chapter, and the denominator . . . is the number of such hospital’s patient days for such fiscal year which were made up of patients who (for such days) were entitled to benefits under part A of this subchapter . . . .

42 U.S.C. § 1395ww(d)(5)(F)(vi)(I). This means that the numerator seeks to count the hospital’s number of patient days – meaning, overnight stays – of patients who were entitled to benefits under both Medicare Part A and SSI at the time they were receiving inpatient services at the hospital, and the denominator is the total number of the hospital’s overnight stays for all patients, who for such days, were entitled to Medicare Part A benefits. Id. The fraction “effectively asks, out of all patient days from Medicare beneficiaries, what percentage of those days came from Medicare beneficiaries who also received SSI benefits?” Cath. Health, 718 F.3d at 917 (emphasis in original).

The Secretary, through his delegate the Centers for Medicare and Medicaid Services, is responsible for computing each hospital’s SSI fraction. See 51 Fed. Reg. 31,454, 31,459 (Sept. 3, 1986) (making CMS responsible for this task because hospitals would have difficulty identifying their Medicare patients who are also SSI recipients).

Pursuant to a regulation issued in 2010, CMS computes the SSI fraction by matching data from the Social Security Administration with Medicare inpatient data in CMS’s own files by looking for one of three codes appearing in SSA’s files – C01, M01, and M02 – to identify a patient’s entitlement to SSI benefits. See Medicare Program, Final Rule, 75 Fed. Reg. 50,041,

5

50,281 (Aug. 16, 2010) (stating that using SSI codes “C01, M01, and M02 accurately captures all SSI-entitled individuals during the month(s) that they are entitled to receive SSI benefits”). 2 CMS matches individuals appearing in the SSA data denoted with these three codes with individuals appearing in its own Medicare Provider Analysis and Review (“MedPAR”) file, which contains information for all Medicare beneficiaries using hospital inpatient services. See Baystate, 545 F. Supp. 2d at 23–24; see also 75 Fed. Reg. at 50,276; 51 Fed. Reg. 16,772, 16,777. CMS identifies the individuals appearing in both two data sets to determine the number of patients, and the inpatient days for those patients at each hospital, for the applicable fiscal year to calculate the hospital’s SSI numerator. See Cath. Health, 718 F.3d at 916.

The Medicare Administrative Contractor then uses the SSI fraction calculated by CMS to determine what a hospital will receive under the DSH adjustment, which is a component of the total Medicare payment to a given hospital. See 42 C.F.R. § 412.106(b)(2)–(5).

C. Providers’ Access to SSA Data The Medicare statue requires the Secretary to “arrange to furnish . . . hospitals . . . the data necessary for such hospitals to compute the number of patient days used in computing the disproportionate patient percentage . . . for that hospital for the current cost reporting year.”

2 The 2010 Final Rule was the product of the 2008 ruling in Baystate Medical Center v. Leavitt, amended by 587 F. Supp. 2d 37 (D.D.C. 2008), in which the court held that CMS’s process for matching Medicare and SSA data failed to use “the best available data” to determine the number of patients entitled to SSI benefits. 545 F. Supp. 2d at 58–59 (finding CMS failed to use superior data readily available to it, including updated SSA data available before the end of the settlement period that would have reflected retroactive SSI eligibility determinations, forced pay SSI records, and inactive or “stale” SSI records omitted from the SSI fractions for two fiscal years). After that ruling, the Secretary issued a new regulation that revised CMS’s matching process. Final Rule, 75 Fed. Reg. at 50,277 (explaining that Final Rule was based on the “revised match process used to implement the Baystate decision [which] addressed all of the concerns found by the court”).

6

Medicare Modernization Act, Pub. L. No. 108-173, § 951, 117 Stat. 2066, 2427 (2003) (codified at 42 U.S.C. § 1395ww Note); see 70 Fed. Reg. 47,278, 47,439 (explaining that a hospital will be provided this data “to calculate and verify its Medicare fraction, and to decide whether it prefers to have the fraction determined on the basis of its fiscal year rather than a Federal fiscal year” and that it “will be the same data set CMS uses to calculate the Medicare fractions for the Federal fiscal year”). To accomplish this, CMS gives hospitals data from its MedPAR Limited Data Set 3 “contain[ing] the matched patient-specific Medicare Part A inpatient days/SSI eligibility data on a month-to-month basis.” 70 Fed. Reg. at 47,440.

But given the confidentiality of information retained by the Social Security Administration, CMS does not give the hospital the complete SSI eligibility file that it receives from SSA. See id. (rejecting proposal that CMS release the data file of SSI eligibility information that the Social Security Administration gives CMS because CMS is prohibited from disclosing SSI eligibility information).

D. Administrative Review A hospital may obtain administrative review of a MAC’s payment determination by requesting a hearing before the Provider Reimbursement Review Board (“PRRB” or the “Board”). See 42 U.S.C. § 1395oo(a). A decision of the Board must be

3 The MedPAR Limited Data Set or MedPAR LDS “contains a summary of all services furnished to a Medicare beneficiary, from the time of admission through discharge, for a stay in an inpatient hospital or skilled nursing facility, or both; SSI eligibility information; and enrollment data on Medicare beneficiaries.” 70 Fed. Reg. at 47,439.

The MedPAR LDS is protected under the Privacy Act, 5 U.S.C. § 552a, and the Health Insurance Portability and Accountability Act of 1996, Pub. L. No. 104-191, 110 Stat. 1936 (1996), but these disclosures are permissible under an applicable routine use. See id. at 47,439.

7

based upon the record made at such hearing, which shall include the evidence considered by the intermediary and such other evidence as may be obtained or received by the Board, and shall be supported by substantial evidence when the record is viewed as a whole.

42 U.S.C. § 1395oo(d). A hospital bears the “burden of production of evidence and burden of proof by establishing, by a preponderance of the evidence, that the provider is entitled to relief on the merits of the matter at issue.” 42 C.F.R. § 405.1871(a)(3).

A Board decision is final “unless the Secretary, on his own motion, and within 60 days after the provider . . . is notified of the Board’s decision, reverses, affirms, or modifies the Board’s decision,” after which time a hospital may obtain judicial review of the decision by filing a civil action with a district court. See 42 U.S.C. § 1395oo(f)(1). II. Factual and Procedural History A. CMS’s Calculation of Plaintiff’s DSH Adjustment Plaintiff Pomona Valley Hospital Medical Center is a nonprofit general acute-care hospital in Pomona, California that furnished hospital services to patients, including Medicare beneficiaries, during the fiscal years 2006 through 2008, the years at issue in this case. Compl. [Dkt. # 1] ¶ 6.

CMS calculated plaintiff’s SSI fractions for those fiscal years as follows:

FY 2006 14.74%

FY 2007 14.73%

FY 2008 14.40%

See Pl.’s Post-Hearing Brief, Exs. P-40, P-41, and P-42, AR 00119, 00121, 00123; see also AR 00135, 00152, 00170. 4

4 Citations to the Administrative Record will use “AR” and the Bates numbers appearing at the bottom right of each page of the record.

8

B. Plaintiff’s Efforts to Verify the Calculation Plaintiff sought to verify CMS’s calculation of the SSI fraction that was used in determining its total DSH adjustment because it thought that the fraction was lower than it should be. See PRRB Hr’g Tr., AR 00339–476 (“Tr.”) at 143–44, AR 00374; see also id. at 145–46, AR 00375 (explaining that over the years the hospital sought to validate the fraction because the fraction fluctuated while its patient population remained consistent). Plaintiff requested and obtained CMS’s MedPAR data for federal FYs 2006 and 2007 – the data CMS used to calculate the FY 2006 SSI fraction. See Pl.’s Final Position Paper, AR 01830; see also 70 Fed. Reg. at 47,440 (CMS gives hospitals data “contain[ing] the matched patient-specific Medicare Part A inpatient days/SSI eligibility data”).

Plaintiff also sought the underlying SSA data from CMS, so it could review both the matched and unmatched data against its own patient files. See PRRB Decision No. 2018-D50 (Oct. 1, 2018), AR 0006–15 (“Board Decision”) at 6, AR 00011 (plaintiff “made numerous efforts to obtain the source SSA data”); see Compl. ¶ 34. CMS declined the request. Answer [Dkt. # 7] ¶ 19 (admitting that “CMS generally does not share the underlying SSA data that it uses in the revised matching process and refused Plaintiff’s request related to such data in connection with the Medicare/SSI fractions at issue”).

Plaintiff then sought to have either CMS or the Social Security Administration review a sample of thirty to fifty “unmatched” patients and days and compare that sample against SSA’s data to ascertain whether they had been correctly excluded. See Letter from Candice Le-Tran to Tzvi Hefter (Jan. 21, 2016), AR 01058–59 (seeking assistance to communicate with CMS and SSA about carrying out the review); see also Tr. at 97–113, AR 00363–67 (testifying about plaintiff’s efforts to obtain the review of the sample records). It offered to abide by the results of

9

the review if it confirmed the accuracy of the CMS calculation. See Tr. at 97–113, AR 00363–67 (testifying about plaintiff’s willingness to abide by the result of the review); Pl.’s Post-Hr’g Brief at 17, AR 00059 (asserting that “if the review showed an overwhelming result of no matches, Pomona would have dropped the challenge.”). Plaintiff also enlisted the help of a Member of Congress to try to persuade the agencies to cooperate with this effort. See Letter from U.S. Senator Dianne Feinstein to Christina Walters, SSA (May 20, 2016), Ex. P-25, AR 01360–62; Letter from U.S. Senator Dianne Feinstein to Carolyn W. Colvin, SSA (Oct. 26, 2016), Ex. P-26, AR 01365– 66. But this request was declined as well. Tr. at 238–239, AR 00398; Letter from Carol L. Blackford, CMS, to U.S. Senator Dianne Feinstein (June 23, 2017), Ex. P-32, AR 01452–53.

C. Plaintiff’s Recalculation of the SSI Numerator Unsuccessful in obtaining either the underlying SSA data or a review of sample data, plaintiff set out to recalculate its SSI fraction numerator using data to which it did have access from the State of California Department of Health Care Services Medi-Cal program. 5 See Board Decision at 6, AR 00011. Plaintiff matched individuals appearing in its own patient files and CMS’s MedPAR data files with individuals appearing in Medi-Cal’s files.

Plaintiff identified individuals in Medi-Cal’s files designated with “aide codes” 10, 20, and 60 – codes that indicate that an individual is eligible for federal SSI. See Tr. at 309, AR 00416 (explaining that an aide code indicates how an individual became eligible for Medi-Cal, the source of eligibility, the source of the benefits, and how much money the state can claim from the federal

5 Medi-Cal is California’s Medicaid program. Asante v. Azar, 436 F. Supp. 3d 215, 220 (D.D.C. 2020), citing Cal. Welf. & Inst. Code § 14000, et seq. Medicaid is a cooperative federalstate program authorized by Title XIX of the Social Security Act that finances medical care for people who cannot afford medical services. Id.

10

government). The record shows that Medi-Cal aide code 10 denotes “aged,” aide code 20 denotes “blind,” and aide code 60 denotes “disabled,” and these codes are used to indicate patients in each category who are eligible for federal SSI or state supplementary payments (“SSP”) benefits. 6 Tr. at 73, AR 00357; id. at 311, AR 00416. Medi-Cal assigns these codes to individuals appearing in its files using information from the Social Security Administration – the source of the data used by CMS. Tr. at 313–314, AR 00417; see also Rosenstein Decl. ¶ 3, AR 00104; Pl.’s Post-Hearing Brief, Exs. P-47 and P-48, AR 00180–84.

Plaintiff’s expert witness testified that this method accurately identified SSI and SSP patients from Medi-Cal’s files, see Tr. at 320–321, AR 00418–19, and a former Director of the CMS Division of Acute Care testified that Medi-Cal received SSI/SSP entitlement data directly from the SSA and that the data was reliable. Tr. at 216–217, AR 00392–93 (testifying that the data “seems to be a reliable basis for determining whether or not those patients truly were getting SSI or not”). 7 Plaintiff compared the results of its matching using the Medi-Cal data and codes with the results of CMS’s match of MedPAR and SSA data and found that the patient days between the two data sets either a) matched, meaning CMS’s data and plaintiff’s data were in agreement; b) did

6 “State supplementary payments” refers to the payments that some states, including California, make to supplement the federal payment benefits of the SSI program. See 42 U.S.C. § 1382e; see also Soc. Sec. Admin., Understanding Supplemental Security Income SSI Benefits – 2020 Edition, https://www.ssa.gov/ssi/text-benefits-ussi.htm (last visited Sept. 30, 2020).

7 In comparing its patient records to the Medi-Cal data, plaintiff excluded patient days if it could not locate an aide code to substantiate that a particular patient had been receiving SSI and/or SSP benefits. See Tr. at 68, AR 00355. And it only counted patient days when an aide code appeared for only a portion of a patient’s admission to the hospital. See id. at 65–66, AR 00355; id. at 68, AR 00355.

11

not match, meaning plaintiff found aide code 10, 20, or 60 for all days of a patient’s stay, but CMS did not include the patient and days in its data; or c) partially matched, meaning some but not all of a patient’s overnight stays were in CMS’s data. See Tr. at 76–77, AR 00357–58; see also PVHMC Summary of SSI Days by Aid Code, Ex. P-27, AR 01369–406. The result of plaintiff’s analysis using publicly available data was that there were additional Medicare/SSI patients in the hospital, and a greater number of patient days, than CMS had included in its calculation of the SSI fraction. See PVHMC Summary of SSI Days by Aid Code, Ex. P-27, AR 01369–406.

D. Plaintiff’s Administrative Appeal of the DSH Adjustment Calculation With that analysis in hand, plaintiff filed a timely appeal of the MAC’s calculation of its DSH adjustment for fiscal years 2006, 2007, and 2008 with the Provider Reimbursement Review Board, and the parties each submitted briefs. MAC’s Final Position Paper, AR 01600–16; Pl.’s Final Position Paper, AR 01823–38.

The issue before the Board was “[w]hether the Medicare Administrative Contractor properly calculated Pomona Valley Hospital Medical Center’s disproportionate share hospital reimbursement with respect to the Provider’s Supplemental Security Income percentage.” Board Decision at 2, AR 00007.

1. The Board’s Hearing

On August 17, 2017, the PRRB held a hearing on the matter. See, e.g., Tr. At the hearing, plaintiff presented three witnesses: Candice Le-Tran, plaintiff’s Director of Reimbursement and Analytics; Tzvi Hefter, former Director of CMS’s Division of Acute Care; and Stan Rosenstein, former Director of Medi-Cal. Tr. at 3–4, AR 00339. The Board accepted Mr. Rosenstein as an expert witness on the Medi-Cal program in general and Medi-Cal eligibility. Tr. at 294, AR 00412.

12

The MAC presented no evidence or testimony at the hearing. See Tr. 3–4, AR 00339 (showing that the MAC presented no witnesses and submitted only its final position paper with its exhibits 8 at the hearing). In its position paper, it argued that plaintiff’s recalculation of the SSI numerator was unreliable because it did not document that it excluded patients who received only state supplemental income, MAC’s Final Position Paper at 9, AR 01612; and that CMS gave the hospital the SSI data it was required to provide by regulation. See MAC’s Final Position Paper at 10, AR 01613. At the hearing, the MAC argued that plaintiff improperly relied on Medi-Cal data to recalculate the SSI fraction when Medicare regulations require the numerator to be determined using data from the SSA and CMS, Tr. at 40–41, AR 00348–49, 9 and that the interest of administrative finality should bar any effort by plaintiff to recalculate the SSI numerator using later data. Tr. at 42–43, AR 00349. 10

8 Although the MAC’s exhibits do not appear to be included in the Administrative Record, its Final Position Paper states that the exhibits were its calculations of the DSH adjustment and the legal authorities it relied upon in its paper. See MAC Final Position Paper at 14 (listing exhibits I-1 and I-2 as its August 20, 2012 Notice of Program Reimbursement and applicable pages of its Audit Adjustment Report and exhibits I-6–I-10 as 42 C.F.R. § 412.106, 70 Fed. Reg. 47,438-47, 439 (August 12, 2005); Baystate Medical Center v. Leavitt, 545 F. Supp. 2d 20 (D.D.C. 2008), as amended by 587 F. Supp. 2d 37 (D.D.C. 2008); 75 Fed. Reg. 50,275–286 (Aug. 16, 2010), and CMS Ruling 1498-R (April 28, 2010). Exhibits I-3–I-5 were exhibits showing that issues not pertaining to plaintiff’s SSI fraction were transferred to other appeals pending before the Board. See id.; see also id. 2–3.

9 This is not a serious objection because plaintiff was attempting to replicate or estimate what the numerator should be since it did not have the SSA data. It was not asserting that the MAC should use its calculation was but rather that its calculation revealed potential issues with CMS’s matching.

10 Similarly, this objection is of little moment given the purpose of plaintiff’s calculation and the fact that plaintiff was not given the data that CMS used, and at bottom, the administrative finality argument is inconsistent with a provider’s statutory right to appeal.

13

After the hearing and in response to questions from the Board, plaintiff submitted a post-hearing brief and fact declarations. Pl.’s Post-Hearing Brief, AR 00038–00101; Rosenstein Decl., AR 00103–08; Le-Tran Decl., AR 00109–14. In that brief, following questions from the Board about how plaintiff accounted for patients who received only SSP benefits, plaintiff revised its calculation to correct an error in how the hospital originally estimated SSP-only beneficiaries. Pl.’s Post-Hearing Brief at 32, AR 00074; Le-Tran Decl. ¶ 6, AR 00112–13 (responding to Board members’ request for “additional clarity regarding the derivation of the ‘14%’ SSP-only population in California and how that number was derived” and correcting an error that increased the estimate of SSP-only beneficiaries “to just over 16%”). 11 Even as corrected, the calculation revealed a substantial divergence between CMS’s calculation and plaintiff’s best efforts to derive a number without access to the actual Social Security Administration data.

Year CMS Days Pl.’s Days % Difference CMS Patients Pl. Patients $ Difference 2006 4,886 5,841 19.55% 748 1,129 $ 770,837 2007 4,153 5,553 25.00% 757 1,197 $1,291,520 2008 4,238 5,500 22.95% 729 1,148 $1,232,627

11 Patients in Medi-Cal files assigned aide codes 10, 20 or 60 could be entitled to SSI-only benefits, SSP-only benefits, or both SSI and SSP benefits simultaneously. Rosenstein Decl. ¶ 5, AR 00105–06; Le-Tran Decl. ¶ 6, AR 00112–113; Tr. at 92–94, AR 00361–62. The Medicare statute counts only hospital days of patients who are entitled to “supplementary security income benefits (excluding any State supplementation),” see 42 U.S.C. § 1395ww(d)(5)(F)(vi)(I), but Medi-Cal records do not identify a patient as SSP-only, so following questions from the Board about the effect of SSP-only patients in plaintiff’s analysis, plaintiff reduced the total of patients and days for each year by approximately16.5%, which it determined was the statewide average of SSI/SSP patients having SSP-only benefits during California’s fiscal years 2006 to 2009. See Pl.’s Post-Hearing Brief, Exs. P-51 and P-46, AR 00191–93, 00178–79); see also Rosenstein Decl. ¶5, AR 00105–06.

14

Pl.’s Post-Hearing Brief, Exs. P-40, P-41, and P-42, AR 00118–123; see also Pl.’s Ex. P-27 at 501, 518, 535, AR 01370, 01387, 01404; Pl.’s Post-Hearing Brief at 8, AR 00050.

2. The Board’s Decision On October 1, 2018, the Board issued its decision on plaintiff’s administrative appeal, affirming the SSI fraction and DSH adjustment. See Board Decision; Letter from Lisa Ogilvie- Barr to Laurence D. Getzoff and Wilson C. Leong (Oct. 1, 2018), AR 00004. The Board acknowledged “Pomona’s difficulty in proving that CMS significantly understated Pomona’s SSI fractions for the three fiscal years under appeal” without access to the underlying SSA data or without the ability to test a sample of its data against the SSA data. Board Decision at 7, AR 00012. But it found a number of flaws in plaintiff’s matching and recalculation, specifically, that:

• plaintiff assumed all individuals with a Medi-Cal aide code of 10, 20, or 60 would “map to an SSI code of C01, M01, or M02,” when plaintiff itself recognized that the Medi-Cal aide codes included individuals who receive SSP payments but not SSI payments, Board Decision at 7, AR 00012;

• individuals who are eligible for Medi-Cal and go into a nursing home remain Medi-

Cal eligible with an aide code of 10, 20, or 60, and so would be counted as SSI-

eligible by Pomona, when CMS excludes such individuals when determining SSI-

eligible days, id.;

• there are differences in timing for when someone becomes eligible for Medi-Cal benefits and when someone becomes eligible for SSI benefits, affecting when an individual would appear with an aide code in the Medi-Cal file and when it would appear as entitled to SSI in the SSA data, Board Decision at 8, AR 00013; and

• plaintiff “did not explain or identify the potential reasons for differences between the data” from its own files, the Medi-Cal system data, the MedPAR SSI patient file, and other data sources plaintiff used, id.

The Board concluded that because plaintiff did not provide a “crosswalk” that mapped the Medi-Cal aide codes to the SSI codes; estimate the impact of the two issues identified by the Board; or explain or identify the reasons for differences between the Medi-Cal aide codes and the SSI

15

codes, plaintiff “did not submit sufficient quantifiable data in the record to prove that the SSI percentages calculated by CMS . . . were flawed.” Board Decision at 8–9, AR 00013–14. It found then, that the “[SSI] percentages used by the Medicare Contractor for Pomona Valley Hospital Medical Center’s . . . [DSH] adjustment for its 12/31/2006, 12/31/2007 and 12/31/2008 cost reports were proper.” Id. at 2, AR 00007.

On November 21, 2018, the CMS Administrator, as the Secretary’s delegate, notified plaintiff that she had declined to review the Board’s decision, making the decision final for purposes of judicial review under 42 U.S.C. § 1395oo(f). Letter from Jacqueline R. Vaughn to Laurence D. Getzoff (Nov. 21, 2018), AR 00001.

E. Plaintiff’s Lawsuit On November 27, 2018, plaintiff filed this lawsuit. Compl. On May 29, 2019, plaintiff filed a motion for summary judgment. Pl.’s Mot. for Summary J. [Dkt. # 11]; Mem. of P. & A. in Supp. of Pl.’s Mot. for Summ. J. [Dkt. # 11-1] (“Pl.’s Mot.”). On August 9, 2019, defendant filed a cross-motion and opposition brief. Def.’s Cross-Mot. for Summ. J.; Mem. in Supp. of Def.’s Cross-Mot. and Opp. to Pl.’s Mot. for Summ. J. [Dkt. # 13] (“Def.’s Mot. and Opp.”). On September 27, 2019, plaintiff filed its opposition and reply brief, Mem. of P. & A. in Opp. to Def.’s Mot. and Reply [Dkt. # 16] (“Pl.’s Reply”), and on November 26, 2019, defendant filed his reply brief. Reply Mem. in Supp. of Def.’s Cross-Mot. for Summ. J. [Dkt. # 17]. The Administrative Record was docketed with the Court on December 10, 2019. Joint Appendix [Dkt. # 18].

STANDARD OF REVIEW

Summary judgment is appropriate when the pleadings and evidence show that “there is no genuine dispute as to any material fact and [that] the movant is entitled to judgment as a

16

matter of law.” Fed. R. Civ. P. 56(a). However, in cases involving review of agency action under the Administrative Procedure Act (“APA”), Rule 56 does not apply due to the limited role of a court in reviewing the administrative record. Select Specialty Hosp.-Akron, LLC v. Sebelius, 820 F. Supp. 2d 13, 21 (D.D.C. 2011). Under the APA, the agency’s role is to resolve factual issues and arrive at a decision that is supported by the administrative record, and the court’s role is to “determine whether or not as a matter of law the evidence in the administrative record permitted the agency to make the decision it did.” Occidental Eng’g Co. v. INS, 753 F.2d 766, 769–70 (9th Cir. 1985), citing Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 415 (1971); see also Richards v. INS, 554 F.2d 1173, 1177 & n.28 (D.C. Cir. 1977).

Under the APA, a court must “hold unlawful and set aside agency action, findings, and conclusions” that are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” 5 U.S.C. § 706(2)(A), in excess of statutory authority, id. § 706(2)(C), or “unsupported by substantial evidence in a case . . . reviewed on the record of an agency hearing provided by statute.” Id. § 706(2)(E). However, the scope of review is narrow. See Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). The agency’s decision is presumed to be valid, see Citizens to Preserve Overton Park, 401 U.S. at 415, and the court must not “substitute its judgment for that of the agency.” State Farm, 463 U.S. at 43. A court must be satisfied, though, that the agency has examined the relevant data and articulated a satisfactory explanation for its action, “including a rational connection between the facts found and the choice made.” Alpharma, Inc. v. Leavitt, 460 F.3d 1, 6 (D.C. Cir. 2006) (citations omitted) (internal quotation marks omitted).

17

ANALYSIS

Plaintiff contends that the Centers for Medicare and Medicaid Services made calculation and/or matching errors in determining its DSH adjustment, and that the Board decision upholding that determination, which was adopted by the Secretary, is not supported by substantial evidence and is arbitrary and capricious. Compl. ¶¶ 2, 58–68.

I. The Court must determine whether the Board’s decision is supported by substantial evidence.

Defendant argues that it should be able to rely on the matching process established by its 2010 Final Rule to calculate plaintiff’s SSI numerator, asserting that the process is not arbitrary and capricious. See Def.’s Mot. and Opp. at 20 (arguing that the process was described in the Federal Register, see 75 Fed. Reg. at 24,002–06, and subjected to public notice and comment). But this case challenges a decision of the PRRB, not the 2010 Final Rule. See Pl.’s Mot. at 25 (asking the Court to set aside the SSI fraction, not the rule); see also Pl.’s Reply at 7, n.8 (contending that plaintiff does not challenge the agency’s interpretation of a statute but “whether [its] matching methodology has been applied accurately and whether the Secretary’s conclusions were based on the best available data”).

Courts review PRRB decisions pursuant to the standard of review set forth in the Administrative Procedure Act, 5 U.S.C. § 706. See 42 U.S.C. § 1395oo(f)(1); Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512 (1994). Under that standard, decisions “reviewed on the record of an agency hearing provided by statute,” like the Board’s decision at issue in this case, must be set aside if the agency’s “action, findings, and conclusions [are] found to be . . . unsupported by substantial evidence.” 5 U.S.C. § 706(2)(E); Biloxi Reg’l Med. Ctr. v. Bowen, 835 F.2d 345, 348– 49 (D.C. Cir. 1987) (“Our review in this case, like that of the District Court, is limited to

18

determining whether, on the record as a whole, the PRRB’s decision is supported by substantial evidence.”). The provider bears the burden of showing that the decision violates the APA standard. See Diplomat Lakewood, Inc. v. Harris, 613 F.2d 1009, 1018 (D.C. Cir. 1979).

A court’s “review in substantial-evidence cases calls for careful scrutiny of the entire record” before the agency. Brown v. Bowen, 794 F.2d 703, 705 (D.C. Cir. 1986). It may not uphold an agency decision based on post-hoc rationalizations offered by the agency or its counsel. See Dep’t. of Homeland Sec. v. Regents of the Univ. of Cal., 140 S. Ct. 1891, 1908–09 (2020); see also Indus. Union Dep’t, AFL-CIO v. Am. Petroleum Inst., 448 U.S. 607, 631 n.31 (1980). When a court reviews the final decision of an administrative tribunal like the PRRB that follows an evidentiary hearing, “[t]he reviewing court must take the [tribunal’s] findings into account as part of the record,” and “the significance to be ascribed to them depends largely on the importance of credibility in the particular case.” Baystate, 545 F. Supp. 2d at 35–36, quoting Morall v. DEA, 412 F.3d 165, 179 (D.C. Cir. 2005) (citations and internal quotation marks omitted). “The final decision must ‘consider relevant contradictory evidence, including evidence that led the [tribunal] to contrary findings of fact and credibility,’ and failure to do so may result in reversal.” Id., quoting Morall, 412 F.3d at 180.

II. Plaintiff presented evidence that CMS’s matching process excluded patient days that should have been included in its SSI numerator.

Plaintiff presented evidence to the Board that supplied grounds to question whether CMS’s matching of MedPAR and SSA data undercounted the patient days used in the challenged SSI numerator. It presented evidence that aide codes 10, 20, and 60 denote patients in Medi-Cal’s files who are eligible for California state supplemental payments and/or federal supplemental security income, Tr. at 73, AR 00357; id. at 311, AR 00416, and that its matching of its own patient records

19

against Medi-Cal’s data showed that CMS failed to include hundreds of patients denoted with aide codes 10, 20, and 60, and thousands of patient days, in calculating plaintiff’s SSI numerator. Pl.’s Post-Hearing Brief, Exs. P-40, P-41, and P-42, AR 00118–123; Pl.’s Ex. P-27 at 501, 518, 535, AR 01370, 01387, 01404. The record also shows that although plaintiff’s original estimate of the average number of patients who were only SSP-eligible was incorrect, which resulted in an overcount of SSI beneficiaries in plaintiff’s original analysis, plaintiff submitted revised figures after the hearing. Le-Tran Decl. ¶ 6, AR 00112–13. Finally, the record shows that the MAC presented no evidence of its own to the PRRB to counter plaintiff’s evidence, relying only on its final position paper and exhibits at the hearing. Tr. at 3–4, AR 00339; see also Tr. at 16, AR 00342 (admitting the final position paper and exhibits).

III. The Board upheld the SSI calculation even though the MAC presented no evidence to contradict plaintiff’s evidence.

Based on this record, the Board agreed with the MAC that the DSH adjustment was properly calculated based on the CMS’s determination of plaintiff’s SSI numerator. Although the Board acknowledged that plaintiff would have “difficulty” proving that CMS understated its SSI fractions without the underlying SSA data or a sample tested against that data, Board Decision at 7, AR 00012, it found plaintiff’s showing, which was based on Medi-Cal data, to be insufficient.

The Board found that plaintiff wrongly “assume[d] that all individuals with an ‘aide code’

of 10, 20, or 60, will map to an SSI code of C01, M01, or M02,” but that plaintiff did attempt to adjust for SSP-only beneficiaries in its calculation. Board Decision at 7, AR 00012. It also noted evidence of other variances between the aide codes and SSA’s codes, including that plaintiff’s witness acknowledged that someone who goes into a nursing home would still remain Medi-Cal eligible denoted with an aide code of 10, 20, or 60 and counted as SSI eligible by the plaintiff,

20

when CMS does not count such a person as SSI-eligible in its calculations. Id. And it cited testimony from plaintiff’s witness that “Medi-Cal eligibility is month-specific” and begins on the first day of the month of the application, whereas SSI eligibility begins on either the first day of the month after an application is filed or the first day of the month after the applicant is determined to be eligible, whichever is later. Id. at 8, AR 00013. The Board recognized that plaintiff considered the effect of both these variances to be minimal, but it found that it could not know the extent of their effect because plaintiff failed to quantify their impact. Id. Finally, it found that although plaintiff “performed a detail comparison of its internal data, the Medi-Cal system data, the MedPAR SSI patient file, and multiple other data sources, it did not explain or identify the potential reasons for differences” among those sources. Id.; see also id. at 8–9, AR 00013–14 (stating that plaintiff could have reviewed the definitions of the SSI codes and aide codes and “built a crosswalk or diagram” to identify if there were other situations in which an individual would be assigned aide code 10, 20, or 60 but would not be assigned SSA code C01, MO1, or M02); id. citing Tr. 334–38, AR 00422 (testimony by plaintiff’s witness that he did not know if there was a one-to-one correlation between the Medi-Cal and SSA codes). Given this, the Board concluded that plaintiff failed to “submit sufficient quantifiable data in the record to prove that the SSI percentages calculated by CMS . . . were flawed.” Id. IV. The Board’s decision is not supported by substantial evidence.

But the question before the Court is not whether plaintiff presented sufficient quantifiable data to prove that CMS’s calculation was flawed, or whether plaintiff had ascertained the reasons for the discrepancies. The question is whether upon review of the entire record, there was substantial evidence to support the Board’s decision that plaintiff’s SSI fraction had been properly determined by CMS. 5 U.S.C. § 706(2)(E); Biloxi Reg’l Med. Ctr., 835 F.2d at 348–49.

21

The Board did not make a serious effort to address this question. It concluded that because it could poke holes at what plaintiff provided, it did not need to examine the accuracy of what CMS did. Indeed, CMS, through the MAC, did not even bother to produce evidence at the hearing to justify its SSI fraction. Further, the Board’s decision fails to explain why the various potential flaws with plaintiff’s calculation undermined plaintiff’s conclusion so thoroughly that there was no reason to peek behind CMS’s methodology at all, even though it was CMS’s matching that was under review, not plaintiff’s.

The record shows that plaintiff used Medi-Cal data and aide codes 10, 20, and 60 to try to recalculate its SSI numerator and identify patient days that CMS may have missed in matching SSA data with Medicare files. It also shows that aide codes 10, 20, and 60 do not overlap precisely with SSA codes C01, M01, and M02, so the aide codes do not definitively indicate that a patient is SSI-eligible for purposes of determining the SSI fraction.

The Board highlighted these differences in reaching its conclusion, emphasizing that plaintiff either failed to explain the differences in the data sets and codes with sufficient detail or failed to estimate the impact of some of these differences. But even when plaintiff estimated differences in the codes and data – such as when it sought to quantify the effect of SSP-only patients in its calculations – rather than credit the estimate, the Board found that a minor error in the original estimate simply proved that plaintiff’s matching effort was flawed. See Board Decision at 7–8, AR 00012–13. So the Board made its decision not based on evidence presented by the agency but on its conclusion that plaintiff’s evidence was insufficient “to prove that the SSI percentages calculated by CMS . . . were flawed.” Board Decision at 8–9, AR 00013–14 (emphasis added).

22

While that may be a reasonable way to assess the data, there is not enough evidence in this record for the Court to conclude that “substantial evidence” supports the Board’s decision that the SSI numerators “were proper.” Board Decision at 2, AR 00007. “[W]here an agency is in sole possession of the records necessary to prove a party’s claim, the agency may not reject the aggrieved party’s allegations as insufficiently proven unless the agency comes forward with ‘countervailing evidence or a reason, not based on the insufficiency of the [movant’s] showing, that explains why the . . . allegations have not been accepted.’” Baystate, 545 F. Supp. 2d at 51 (edits, omissions, and emphasis in original), quoting Atlanta Coll. of Med. & Dental Careers, Inc. v. Riley, 987 F.2d 821, 830–31 (D.C. Cir. 1993). “[T]he burden of bringing forward evidence generally shifts when the defendant has greater access to information on a particular issue.” Id., quoting Atlanta Coll. of Med., 987 F.2d at 831.

The record here shows that the agency had the data that would have answered plaintiff’s allegations and that proving the allegations without it would be “difficult[ ],” Board Decision at 7, AR 00012, if not impossible. Nevertheless, the agency declined to provide plaintiff with any of the underlying SSI data or to conduct or facilitate a test of sample data, even though plaintiff agreed it would abide by the result of such a test. See Board Decision at 6, AR 00011. 12 And it declined to present any of that data to the Board on appeal. Without that evidence, the Board improperly rejected plaintiff’s allegations, Atlanta Coll. of Med., 987 F.2d at 830–31, and the Court finds that its decision is “unsupported by substantial evidence.” 5 U.S.C. § 706(2)(E).

12 The data CMS did provide to plaintiff pursuant to the statute, 42 U.S.C. § 1395ww Note, only included the “matched patient-specific Medicare Part A inpatient days/SSI eligibility data,” not the data for unmatched patients. 70 Fed. Reg. at 47,440. So that data would not allow plaintiff to determine whether CMS failed to match any patient or patient days it thought should have been included in its SSI numerator.

23

V. The Court will not impose an adverse inference against the defendant but will order remand of the matter.

Plaintiff asks the Court to impose an adverse inference against defendant given the agency’s failure to present evidence to the Board, Pl.’s Reply at 23, and argues that remand “is not necessary because a full administrative record has been provided to the Court.” Pl.’s Mot. at 44. It asks the Court to set aside the Board’s decision, issue a writ of mandamus, and order the Secretary to recalculate its SSI fraction using the correct SSI data, provide plaintiff the data and the programs CMS used to accomplish the recalculation, and pay the additional amounts due to it. Compl. ¶¶ 70, 72; id. at Request for Relief ¶ 1.

But that would be contrary to the law of this Circuit. In this circumstance, the burden is on the agency “to produce countervailing evidence or a reason, not based on the insufficiency of the [plaintiff’s] showing” that explains why the SSI numerator is accurate, but the ultimate burden of persuasion remains with plaintiff. See Atlanta Coll. of Med., 987 F.2d at 831 (emphasis in original), citing Texas Dep’t of Cmty. Affs. v. Burdine, 450 U.S. 248, 253 (1981). Mindful that it must not “substitute its judgment for that of the agency,” State Farm, 463 U.S. at 43, the Court finds that remand is the appropriate remedy at this stage of the proceedings. See PPG Indus., Inc. v. United States, 52 F.3d 363, 366 (D.C. Cir. 1995) (noting that an agency may reopen proceedings to take new evidence if a reviewing court finds the agency’s original findings invalid).

24

CONCLUSION

For the reasons set forth above, the Court will grant plaintiff’s motion for summary judgment in part and deny it in part [Dkt. # 11], deny defendant’s cross-motion [Dkt. # 13], and remand the matter to the agency for further proceedings consistent with this ruling.

AMY BERMAN JACKSON

United States District Judge

DATE: September 30, 2020