Baystate Medical Center v. Leavitt, et al

MOTION for Summary Judgment , Memorandum of Points and Authorities in Support, Statement of Material Facts

D.D.C.February 1, 2007

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA BAYSTATE MEDICAL CENTER, ) ) Plaintiff, ) ) v. ) C.A. No. 06-1263 (JDB) ) MICHAEL O. LEAVITT, et al., ) ) Defendants. ) PLAINTIFF’S MOTION FOR SUMMARY JUDGMENT Pursuant to Fed. R. Civ. P. 56, Plaintiff respectfully moves for summary judgment in its favor on Count I on the grounds that there are no material facts in dispute and that Plaintiff is entitled to judgment as a matter of law. In support of this motion, the Court is respectfully referred to the accompanying Plaintiff’s Memorandum of Points and Authorities in Support of Its Motion for Summary Judgment, and Plaintiff’s Statement of Material Facts As To Which There Is No Genuine Dispute. A proposed order setting the relief requested by this motion is also attached. Respectfully submitted, /s/ Christopher L. Keough Christopher L. Keough DC Bar No. 436567 John M. Faust DC Bar No. 433553 Stephanie A. Webster DC Bar No. 479524 VINSON & ELKINS L.L.P. 1455 Pennsylvania Avenue, N.W. Suite 600 Washington, D.C. 20004-1008 (202) 639-6500 (phone) (202) 639-6604 (fax) Counsel for Plaintiff February 1, 2007 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 1 of 145 UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA BAYSTATE MEDICAL CENTER, ) ) Plaintiff, ) ) v. ) C.A. No. 06-1263 (JDB) ) MICHAEL O. LEAVITT, et al., ) ) Defendants. ) PLAINTIFF’S MEMORANDUM OF POINTS AND AUTHORITIES IN SUPPORT OF ITS MOTION FOR SUMMARY JUDGMENT Respectfully submitted, /s/ Christopher L. Keough Christopher L. Keough DC Bar No. 436567 John M. Faust DC Bar No. 433553 Stephanie A. Webster DC Bar No. 479524 VINSON & ELKINS L.L.P. 1455 Pennsylvania Avenue, N.W. Suite 600 Washington, D.C. 20004-1008 (202) 639-6500 (phone) (202) 639-6604 (fax) Counsel for Plaintiff February 1, 2007 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 2 of 145 ii TABLE OF CONTENTS I. INTRODUCTION .............................................................................................................. 1 II. STATUTORY AND REGULATORY BACKGROUND.................................................. 5 A. Medicare Payment, Appeals, and Payment Corrections......................................... 5 B. Medicare DSH Adjustment .................................................................................... 6 III. SUMMARY OF THE FACTS.......................................................................................... 10 IV. SUMMARY OF ARGUMENT ........................................................................................ 16 V. ARGUMENT.................................................................................................................... 18 A. Standard of Review............................................................................................... 18 B. The Secretary Wrongfully Denied the Hospital Relief Based On the Hospital’s Alleged Failure to Prove the Precise Financial Impact Of the Secretary’s Errors With Information That Only the Government Has and That the Secretary Refused To Produce....................... 20 1. The Hospital Does Not Have the Burden to Show the Impact Of the Secretary’s Errors. ............................................................. 21 2. The Secretary’s Errors Resulted in Substantial Underpayments ......................................................................................... 22 C. The Secretary Incorrectly Alleges that the Law Provides For No Right To Relief From the Secretary’s Errors .................................................. 23 1. Right to Relief........................................................................................... 24 2. Unsound Management, Lack of Quality Control, and Concealment ............................................................................................. 26 3. Lack of Notice and Meaningful Opportunity to Comment....................... 30 D. The Secretary Did Not Use the “Best Data Available” to Calculate the SSI Fractions at Issue...................................................................................... 31 E. The Secretary’s Decision to Exclude Non-Cash SSI Benefits is Contrary to Law .................................................................................................... 32 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 3 of 145 iii F. The Secretary’s Final Decision Applies the Wrong Standard for Determining the Proper Count of Medicare Days in the Denominator of the SSI Fraction .......................................................................... 34 1. Exhausted Benefit and MSP Days Should Not be Included..................................................................................................... 35 2. Medicare HMO Days Should Not be Included......................................... 40 VI. CONCLUSION................................................................................................................. 45 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 4 of 145 iv TABLE OF AUTHORITIES Cases Alvarado Community Hosp. v. Shalala, 1998 U.S. App. LEXIS 33973 (9th Cir. 1998) ................................................................. 25 AT&T Corp. v. F. C. C., 86 F.3d 242 (D.C. Cir. 1996) ............................................................................................ 19 * Atlanta Coll. of Med. & Dental Careers, Inc. v. Riley, 987 F.2d 821 (D.C. Cir. 1993) ................................................................................ 3, 16, 21 Biloxi Reg’l Med. Ctr. v. Bowen, 835 F.2d 345 (D.C. Cir. 1987) .......................................................................................... 19 Burlington Truck Lines, Inc. v. United States, 371 U.S. 156 (1962).......................................................................................................... 19 Cabell Huntington Hosp. v. Shalala, 101 F.3d 984 (4th Cir. 1996) ........................................................................................ 7, 38 Califano v. Boles, 443 U.S. 282 (1979) ........................................................................ 23, 26, 27 Centra Health, Inc. v. Shalala, 102 F. Supp. 2d 654 (W.D. Va. 2000) .............................................................................. 25 Cheshire Hosp. v. New Hampshire-Vermont Hospitalization Serv., Inc., 689 F.2d 1112 (1st Cir. 1982)......................................................................... 43 Clark Reg’l Med. Ctr. v. United States Dep’t of Health and Human Servs., 314 F.3d 241 (6th Cir. 2002) ............................................................... 7 * County of Los Angeles v. Shalala, 192 F.3d 1005 (D.C. Cir. 1999) ........................................................................ 6, 17, 23, 25 Deaconness Health Serv. Corp. v. Shalala, 83 F.3d 1041 (8th Cir. 1996) .............................................................................................. 7 Department of Treasury v. FLRA, 837 F.2d 1163 (D.C. Cir. 1988) .................................................................................. 18, 34 Georgetown Univ. Hosp. v. Bowen, 862 F.2d 323 (D.C. Cir. 1988) .......................................................................................... 25 * Denotes authorities principally relied upon. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 5 of 145 v Gutierrez-Rogue v. INS, 954 F.2d 769 (D.C. Cir. 1992) .......................................................................................... 22 * In re Medicare Reimbursement Litig., 309 F. Supp. 2d 89 (D.D.C. 2004), aff’d, 414 F.3d 7 (D.C. Cir. 2005), cert. denied, 126 S. Ct. 1672 (2006).............................................. passim Int’l Union v. NLRB, 459 F.2d 1329 (D.C. Cir. 1972).......................................................... 3, 21 Jewish Hosp. Inc. v. Secretary of Health and Human Servs., 19 F.3d 270 (6th Cir. 1996)............................................................................... 8, 37 Kaczmarczyk v. INS, 933 F.2d 588 (7th Cir.), cert. denied, 112 S. Ct. 583 (1991) ................................................................................... 22 Legacy Emanuel Hosp. & Health Ctr. v. Shalala, 97 F.3d 1261 (9th Cir. 1996) ........................................................................................ 7, 37 Loma Linda Cmty. Hosp. v. Shalala, 907 F. Supp. 1399 (C.D. Cal. 1995) ................................................................................. 21 Methodist Hosp. of Sacramento v. Shalala, 38 F.3d 1225 (D.C. Cir. 1994) .......................................................................................... 24 Michigan Public Power Agency v. FERC, 405 F.3d 8 (D.C. Cir. 2005) .................................................................................. 18, 19, 38 Monmouth Med. Ctr. v. Thompson, 257 F.3d 807 (D.C. Cir. 2001) .......................................................................................... 38 Morall v. DEA, 412 F.3d 165 (D.C. Cir. 2005) ................................................................. passim Motor Vehicle Mfrs. Ass’n of the United States, Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29 (1983)........................................................ 19 National Treasury Employees Union v. FLRA, 399 F.3d 334 (D.C. Cir. 2005) ............................................................................. 18, 19, 38 Orloski v. FEC, 795 F.2d 156 (D.C. Cir. 1986)........................................................................ 42 Portland Adventist Med. Ctr. v. Thompson, 399 F.3d 1091 (9th Cir. 2005) ............................................................................................ 7 Riverside Methodist Hosp. v. Thompson, No. C2-02-94, 2003 U.S. Dist. LEXIS 15163 (S.D. Ohio, July 31, 2003) ................................................................................................ 25 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 6 of 145 vi Samaritan Inns, Inc. v. District of Columbia, 114 F.3d 1227 (D.C. Cir. 1997) ........................................................................................ 21 Sarasota Mem. Hosp. v. Shalala, 60 F.3d 1507 (11th Cir. 1995) .......................................................................................... 25 Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555 (1931) ....................................................................................... 21 Thomas Jefferson Univ. v. Shalala, 512 U.S. 504 (1994).................................................................................................... 19, 38 Transactive Corp. v. United States, 91 F.3d 232 (D.C. Cir. 1996) ...................................................................................... 41, 43 United Scenic Artists v. NLRB, 762 F.2d 1027 (D.C. Cir. 1985) ........................................................................................ 22 United States v. Mead Corp., 533 U.S. 218 (2001).......................................................................................................... 19 United States v. Wilson, 290 F.3d 347 (D.C. Cir. 2002), cert. denied, 537 U.S. 1028 (2002)................................................................................... 42 Statutes 5 U.S.C. § 706........................................................................................................................... 18 42 U.S.C. § 426(c)(1)................................................................................................................ 36 42 U.S.C. § 1381......................................................................................................................... 8 42 U.S.C. § 1381a ....................................................................................................................... 8 42 U.S.C. § 1382h(b) .......................................................................................................... 33, 34 42 U.S.C. §§ 1395c–1395i-4....................................................................................................... 5 42 U.S.C. § 1395d(a)(1)........................................................................................................ 5, 36 42 U.S.C. §§ 1395mm(a)(1) ..................................................................................................... 43 42 U.S.C. §§ 1395mm(a)(3) ..................................................................................................... 43 42 U.S.C. § 1395oo................................................................................................................... 10 42 U.S.C. § 1395oo(a) ................................................................................................................ 5 42 U.S.C. § 1395oo(d) ................................................................................................................ 5 42 U.S.C. § 1395oo(f)........................................................................................................... 5, 18 42 U.S.C. § 1395oo(h) ................................................................................................................ 5 42 U.S.C. § 1395ww(d)(5)(C)(i)................................................................................................. 6 * 42 U.S.C. § 1395ww(d)(5)(F)............................................................................................ passim 42 U.S.C. § 1395ww(d)(5)(G) .................................................................................................. 39 42 U.S.C. § 1395ww(h)(3)(C) .................................................................................................. 41 42 U.S.C. § 1395y(b)(2)(A)...................................................................................................... 37 Social Security Amendments of 1983, Pub. L. No. 98-21 § 601(e) ........................................... 6 Balanced Budget Act of 1997, Pub. L. No. 105-33 § 4625 ...................................................... 42 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 7 of 145 vii Regulations 20 C.F.R. § 416.264 (1995) .......................................................................................... 18, 33, 34 41 C.F.R. §§ 201-20.303(b)(2)(i) ............................................................................................. 27 42 C.F.R. § 405.1835.................................................................................................................. 5 42 C.F.R. § 405.1875.................................................................................................................. 5 42 C.F.R. § 412.106 .............................................................................................................. 6, 36 * 42 C.F.R. § 412.106(b)(2)................................................................................................. 8, 9, 10 42 C.F.R. § 412.106(b)(3)......................................................................................................... 10 42 C.F.R. § 412.106(b)(4)........................................................................................................... 7 42 C.F.R. § 412.106(c)(1)........................................................................................................... 6 42 C.F.R. § 412.106(d) ............................................................................................................... 6 42 C.F.R. § 416.101 .................................................................................................................... 8 42 C.F.R. § 416.105 .................................................................................................................... 8 42 C.F.R. § 417.524 .................................................................................................................. 43 42 C.F.R. § 417.584 .................................................................................................................. 43 51 Fed. Reg. 31,454 (Sept. 3, 1986) ........................................................................... 7, 9, 10, 37 54 Fed. Reg. 40,286 (Sept. 29, 1989) ................................................................................. 41, 43 55 Fed. Reg. 35,990 (Sept. 4, 1990) ................................................................................... 39, 43 59 Fed. Reg. 37,906 (July 25, 1994)......................................................................................... 28 60 Fed. Reg. 45,778 (Sept. 1, 1995) ......................................................................................... 30 61 Fed. Reg. 6,428 (Feb. 20, 1996) .......................................................................................... 28 68 Fed. Reg. 27,154 (May 19, 2003) ........................................................................................ 36 Other Authorities Medicare Claims Processing Manual (CMS Pub. 100-04), Ch. 3, § 40 ................................... 39 Medicare Intermediary Manual (CMS Pub. 13-3) § 3685........................................................ 40 Medicare Secondary Payer Manual (CMS Pub. 100-05), Ch. 1, § 40...................................... 40 Office of Management and Budget Circular No. A-130........................................................... 27 S. Rep. No. 99-146 (1985) ................................................................................................... 22-23 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 8 of 145 UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA BAYSTATE MEDICAL CENTER, ) ) Plaintiff, ) ) v. ) C.A. No. 06-1263 (JDB) ) MICHAEL O. LEAVITT, et al., ) ) Defendants. ) PLAINTIFF’S MEMORANDUM OF POINTS AND AUTHORITIES IN SUPPORT OF ITS MOTION FOR SUMMARY JUDGMENT I. INTRODUCTION This is an action for judicial review of the final decision of the Secretary of the Department of Health and Human Services (the “Secretary”), concerning the Secretary’s determinations of the Medicare “disproportionate share hospital” (or “DSH”) payments due to plaintiff hospital, Baystate Medical Center (the “Hospital”), for fiscal years 1993-1996. Based on an extensive record, including uncontroverted testimony from the government’s own current and former employees, the Secretary’s Provider Reimbursement Review Board (the “PRRB” or “Board”) issued a meticulous decision conclusively establishing that the Secretary’s calculations of the DSH payments are flat wrong. The Board found that the Secretary’s many errors are systemic and understate the payments due to this hospital and all other hospitals that shoulder the financial burden of providing care to a disproportionate share of low-income Americans. PRRB Dec. No. 2006-D20, Mar. 17, 2006 (“PRRB Dec.”).1 1 The pertinent facts and citations to evidence in the Administrative Record (“AR”) are set forth in Plaintiff’s Statement of Material Facts As To Which There Is No Genuine Issue (“Plaintiff’s Statement” or “Pl. Stmt.”). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 9 of 145 2 The Board’s decision was improperly reversed by the Administrator of the Centers for Medicare and Medicaid Services (“CMS”), in a May 17, 2006 decision (“Adm’r Dec.”), from which this appeal is taken. Although the record below is voluminous, the Secretary’s inability to raise any genuine dispute as to the Board’s identification of serious errors in CMS’ calculations makes this case simple. The Secretary does not, and cannot, dispute the Board’s findings and the largely uncontroverted record evidence showing that CMS’ calculations are wrong. The record of the agency’s errors, omissions, and concealments stands uncontested – a secret, years-long, and knowing failure to use the data and processes readily available to CMS to accurately produce the straightforward DSH calculations that the Secretary undertook to perform on the hospitals’ behalf. The Administrator has chosen to defend that record with legal defenses that are contrary to D.C. Circuit precedents and allegations that are contrary to the undisputed record evidence and Board findings. For example, while the Administrator does not refute the Board’s findings regarding the several systemic errors in CMS’ calculations, all of which improperly understate the DSH payments, CMS has the temerity to argue that the Hospital is entitled to no relief because it did not prove the financial impact of CMS’ errors with sufficient precision. In fact, the record shows that the likely dollar impact of the agency’s miscalculations is in the range of hundreds of millions of dollars, possibly more, for all hospitals nationally. If correction of the Secretary’s errors results in only a ten percent increase in the Hospital’s SSI fractions for the years at issue, the DSH payments for all four years would increase by more than $1.75 million. AR 4820. But, more importantly, even if that proof had been lacking below, the Court of Appeals has made clear that, under circumstances like these, the government must bear the consequences of the uncertainty it has created by having refused to produce the evidence that would have shown Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 10 of 145 3 precisely what was and was not counted in CMS’ calculations, as well as the precise impact of the errors and omissions in those calculations. The evidence of errors and omissions in CMS’ calculations, which is almost entirely undisputed, shifted the burden to the Secretary “to produce countervailing evidence or a reason, not based on the insufficiency of the [Hospital’s] evidence, that explains why [the Hospital’s] allegations [should not be] accepted,” and this burden cannot be avoided merely by asserting generalized scenarios that could affect the impact of CMS’ errors and omissions and that “could only be eliminated with information unavailable to the [Hospital].” Atlanta Coll. of Med. & Dental Careers, Inc. v. Riley, 987 F.2d 821, 830-31 (D.C. Cir. 1993) (emphasis added). Further, it is well-established “that when a party has relevant evidence within his control which he fails to produce, that failure gives rise to an inference that the evidence is unfavorable to him.” Int’l Union v. NLRB, 459 F.2d 1329 (D.C. Cir. 1972). An adverse inference is required here. As explained below, the Secretary took it upon himself to calculate the DSH adjustment factor at issue here, without input or oversight from anyone, thereby assuming a fiduciary obligation to get the calculation right. His refusal to show his agency’s work raises a now- irrebuttable presumption that his errors materially disfavor the Hospital. Next, in the interest of “finality,” the Administrator claims the right to apply an admittedly unwritten “policy” (Answer ¶ 102) requiring that the agency’s calculations – however wrong and however easily they could be corrected by later and better data available to CMS – cannot be revisited. This brand of “finality” can be achieved only at the expense of the rule of law. Following the decisions of this Court and the Court of Appeals in In re Medicare Reimbursement Litigation, 309 F. Supp. 2d 89 (D.D.C. 2004), aff’d, 414 F.3d 7 (D.C. Cir. 2005), cert. denied, 126 S. Ct. 1672 (2006), the Secretary cannot immunize his erroneous DSH Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 11 of 145 4 calculations from judicial review and retrospective correction. In that case, the Court of Appeals affirmed this Court’s 2004 decision compelling the Secretary to revise erroneous and unlawful DSH payment determinations that were issued as early as 1994, for fiscal years that ended even earlier. If hospitals have a right to appeal the Secretary’s DSH calculations (and even the Administrator does not contest that right), they have a right to a remedy. Finally, the Administrator alleges that CMS, as a matter of policy, uses the best data available to it and thus achieved an “acceptable” rate of error here. Those assertions simply cannot be squared with the uncontroverted record evidence. The testimony of current and former government employees reflects that the agency’s calculation process was slapdash, that it was inconsistent with established federal standards for data management and verification, and that better and more current data was always available to the agency at the time it performed its calculations. See Morall v. DEA, 412 F.3d 165, 167 (D.C. Cir. 2005) (an agency’s decision cannot stand when it fails to consider contradictory record evidence). Moreover, there is simply no authority for the proposition that, so long as the resulting error is “acceptable” to an agency, the agency is entitled to produce a result its personnel know to be wrong. There are no material facts in genuine dispute here, and Plaintiff is entitled to judgment as a matter of law. The relief sought in Count I of the Complaint, which is the subject of this motion, is straightforward and closely tailored to remedy what occurred here. The Hospital seeks an order retaining jurisdiction over this case and remanding to the Secretary with instructions that he correct his errors and omissions in a manner that is transparent to and can be verified by this Court and the plaintiff hospital, recompute the DSH payments due the Hospital for the years at issue as a result of those corrections, and pay the Hospital the additional sums to which it is entitled, plus interest. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 12 of 145 5 II. STATUTORY AND REGULATORY BACKGROUND A. Medicare Payment, Appeals, and Payment Corrections This case concerns Part A of the Medicare Act, 42 U.S.C. §§ 1395c–1395i-4, which provides for payments from the Federal Hospital Insurance Trust Fund, of which the Secretary is a trustee, for institutional health care services furnished to Medicare beneficiaries by participating “providers of services” like the Hospital. See 42 U.S.C. § 1395d(a)(1). Medicare payments to hospitals are determined by fiscal intermediaries, private firms (usually insurance companies) that contract with CMS. Two to three years after a provider’s fiscal year, the intermediary analyzes a cost report prepared by the provider and issues a Notice of Program Reimbursement, or “NPR,” that informs the provider of the intermediary’s final determination of the provider’s Medicare reimbursement for the cost reporting period at issue. PRRB Dec. at 32 & n.183, AR 162. A provider, like the Hospital, is entitled to a hearing before the PRRB2 if it is dissatisfied with an intermediary’s determination in an NPR as to the amount of Medicare payment due the provider for a cost reporting period. 42 U.S.C. § 1395oo(a); 42 C.F.R. § 405.1835. The Board may reverse, affirm or modify the final determination of the intermediary. 42 U.S.C. § 1395oo(d). The final decision of the PRRB is subject to review by the Administrator of CMS pursuant to delegation of authority by the Secretary to the Administrator. See 42 U.S.C. § 1395oo(f); 42 C.F.R. § 405.1875. The Secretary’s final decision may be reviewed in a civil action before this Court. 42 U.S.C. § 1395oo(f). 2 The PRRB is an administrative tribunal appointed by the Secretary. 42 U.S.C. § 1395oo(h). The Medicare Act prescribes that the appointed members of the PRRB must be “knowledgeable in the field of payment of providers of services.” Id. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 13 of 145 6 B. Medicare DSH Adjustment Medicare pays most hospitals under a prospective payment system (“PPS”) for services furnished to inpatients. See County of Los Angeles v. Shalala, 192 F.3d 1005, 1008 (D.C. Cir. 1999). Under PPS, Medicare pays predetermined, standardized rates for each discharge. Id. When Congress enacted the PPS in 1983, it authorized the Secretary to provide an adjustment to the standardized PPS payment rates per discharge for hospitals that serve a disproportionate share of low-income patients. See Social Security Amendments of 1983, Pub. L. No. 98-21 § 601(e), codified at 42 U.S.C. § 1395ww(d)(5)(C)(i) (1983). This is the disproportionate share hospital, or “DSH,” payment adjustment at issue in this case.3 Under the statutory criteria in effect during the periods at issue here, a hospital that serves a disproportionate share of low-income patients – a DSH – is entitled to an upward percentage adjustment to the standardized PPS rates. See 42 U.S.C. § 1395ww(d)(5)(F); see also 42 C.F.R. § 412.106. The DSH payment adjustment has always been determined for each hospital retrospectively, after the end of the fiscal year, through the cost report settlement process described above, which typically occurs years later. PRRB Dec. at 8, AR 138; see also 42 U.S.C. § 1395ww(d)(5)(F); 42 C.F.R. § 412.106. In all but a small handful of cases, the DSH payment adjustment is based on a hospital’s “disproportionate patient percentage.” See 42 U.S.C. § 1395ww(d)(5)(F)(i)(I) and (d)(5)(F)(v); 42 C.F.R. § 412.106(c)(1). That percentage determines a hospital’s qualification for any DSH payment at all as well as the amount of the DSH payment for most qualifying hospitals (including the Plaintiff hospital). See 42 U.S.C. § 1395ww(d)(5)(F)(iv) and (vii)-(xiii); 42 C.F.R. § 412.106(d). The percentage is defined in 42 U.S.C. § 1395ww(d)(5)(F)(vi) as the sum 3 From the beginning, CMS has been hostile to the very concept of the DSH payment. See Compl. ¶¶ 34-35, 42; n.4, infra. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 14 of 145 7 of two fractions, commonly called the “Medicaid fraction” and the “Medicare/SSI fraction” (sometimes referred to interchangeably as the “Medicare fraction” or the “SSI fraction”). The Medicaid Fraction. The Medicaid fraction is based on a hospital’s percentage of total inpatient days for a cost reporting period attributable to patients who were “eligible for medical assistance” under a State Medicaid plan. 42 U.S.C. § 1395ww(d)(5)(F)(vi)(II). In general, the more Medicaid inpatient days, the larger the DSH payment. The Secretary does not permit estimates for the Medicaid fraction. Pl. Stmt. ¶¶ 5, 189. A hospital’s Medicaid fraction for a hospital cost reporting period is determined retrospectively by the fiscal intermediary, 42 C.F.R. § 412.106(b)(4), in an NPR that is generally issued two to three years after the close of the hospital’s cost reporting period. PRRB Dec. at 32, AR 162. Until 1997, the Secretary required that the numerator of the Medicaid fraction could include only those inpatient days that were actually paid by Medicaid, even though the statute required only that the patient be eligible for Medicaid. 51 Fed. Reg. 31,454, 31,460-61 (Sept. 3, 1986) (final rule). Ensuing litigation culminated in four appellate courts holding uniformly that the Secretary’s exclusion of Medicaid eligible-but-unpaid days contravened the plain meaning of the DSH statute.4 After the Secretary finally acquiesced in the adverse judicial decisions, 4 See Cabell Huntington Hosp. v. Shalala, 101 F.3d 984 (4th Cir. 1996); Legacy Emanuel Hosp. & Health Ctr. v. Shalala, 97 F.3d 1261 (9th Cir. 1996); Deaconness Health Serv. Corp. v. Shalala, 83 F.3d 1041 (8th Cir. 1996); Jewish Hosp. Inc. v. Secretary of Health and Human Servs., 19 F.3d 270 (6th Cir. 1996). These and subsequent decisions recognize the Secretary’s historic resistance to the legislative will as expressed in the DSH statute. See Portland Adventist Med. Ctr. v. Thompson, 399 F.3d 1091, 1099 (9th Cir. 2005) (“This appears to be the latest in a series of cases in which the Secretary has refused to implement the DSH provision in conformity with the intent behind the statute.”); Clark Reg’l Med. Ctr. v. United States Dep’t of Health and Human Servs., 314 F.3d 241, 249 (6th Cir. 2002) (“the Department cannot simply interpret the [DSH] regulation so as to always disadvantage the subject hospital”); Cabell Huntington Hosp., 101 F.3d at 990 (characterizing the Secretary’s interpretation as an attempt to “rewrite the will of Congress” by an agency that has been “hostile from the start to the very idea of making the Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 15 of 145 8 through an administrative ruling issued in February 1997, this Court held in 2004 that the Secretary’s ruling to acquiesce only prospectively was a “notice of inconsistency with law” that triggered a mandatory duty on the Secretary’s part to retrospectively reopen past DSH payment determinations in NPRs that were issued as far back as 1994 for fiscal years as early as 1991. In re Medicare Reimbursement Litig., 309 F. Supp. 2d 89 (D.D.C. 2004), aff’d, 414 F.3d 7 (D.C. Cir. 2005), cert. denied, 126 S. Ct. 1672 (2006). The SSI Fraction. The other fraction in the DSH payment calculation – the SSI fraction – is the subject of the dispute in this case. The DSH statute defines the SSI fraction as: the fraction (expressed as a percentage), the numerator of which is the number of such hospital’s patient days for such [cost reporting] period which were made up of patients who (for such days) were entitled to benefits under Part A of this title [Medicare] and were entitled to supplemental security income benefits (excluding any State supplementation) under title XVI of this Act, and the denominator of which 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 title. . . . 42 U.S.C. § 1395ww(d)(5)(F)(vi)(I). The supplemental security income (“SSI”) program referred to in the DSH statute, quoted above, is a federal assistance program for low-income individuals who are aged, disabled or blind. 42 U.S.C. § 1381 et seq. The SSI program is administered by the Social Security Administration (“SSA”). See 42 U.S.C. § 1381a; 42 C.F.R. §§ 416.101, 416.105. The Secretary computes the SSI fraction annually for every hospital, undertaking the sole and exclusive responsibility to perform the necessary calculation. See 42 C.F.R. § 412.106(b)(2); see also PRRB Dec. at 3, AR 133. As was stipulated before the Board in this case, apart from the transmission of claims data for Medicare inpatient hospital stays, hospitals payments at issue”); Jewish Hosp., 19 F.3d at 276 (finding “credible and compelling” evidence of the agency’s “hostility to the concept of the disproportionate share adjustment”). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 16 of 145 9 and fiscal intermediaries have no input into, and are not parties to, the Secretary’s calculation of the SSI fractions. See Pl. Stmt. ¶¶ 6-7; 42 C.F.R. § 412.106(b)(2); 51 Fed. Reg. 31,454, 31,459- 60 (Sept. 3, 1986) (final rule). The Secretary’s preamble to the September 1986 final rule instituting DSH stated: “Recognizing the difficulty hospitals would have identifying their Medicare patients who are also SSI recipients, we have undertaken the task (and expense) of determining for each hospital the number of patient days of those dually entitled to Medicare Part A and SSI and have removed this burden from hospitals.” 51 Fed. Reg. at 31,459-60. When the Secretary assumed responsibility in 1986 for the calculation of the SSI fractions, and imposed a regulatory scheme whereby hospitals were to play no role in that process but instead were required to rely upon the agency’s calculation of the SSI fraction, the Secretary assumed a fiduciary obligation toward the hospitals to perform the calculation of the SSI fraction accurately. The “Match” Process Used to Calculate the SSI Fraction. The Secretary computes the SSI fractions retrospectively for each Federal fiscal year ending on September 30. The SSI fractions are computed solely by CMS based upon SSI data that CMS receives annually from SSA and CMS’ match of that data against CMS’ own inpatient hospital records. PRRB Dec. at 3, AR 133. In general, CMS’ calculation of the SSI fraction comprises two basic data sources: (i) SSI entitlement data provided to CMS by SSA about six months after the September 30 year- end for the subject Federal fiscal year (i.e., the following March); and (ii) Medicare inpatient hospital stay records accumulated in a CMS file by about nine months after the end for the subject Federal fiscal year (i.e., by the end of the following June). Id. CMS performs a “match” of these two data sets (in July or August) in order to determine a hospital’s number of patient days in the prior Federal fiscal year attributable to individuals who were entitled to both Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 17 of 145 10 Medicare Part A benefits and federal SSI benefits during their inpatient hospital stay. The more Medicare patient days that are matched with the SSI eligibility records, the larger the numerator of the SSI fraction and the greater the DSH payment to the hospital. For each fiscal year at issue in this case, between the time when CMS calculated the SSI fraction (about 10-11 months after the Federal fiscal year) and the issuance of the DSH payment determination in an NPR for each fiscal year (two to three years later), CMS performed a series of six additional matches of its inpatient hospital stay records against updated and corrected SSI data files subsequently received (for the same fiscal year) from SSA. Pl. Stmt. ¶¶ 11, 27, 38-40, 166, 169, 172. The results of those later matches that were performed with updated and corrected data were not used to determine the DSH payments made to the Hospital for the years at issue. Pl. Stmt. ¶¶ 11, 13, 15, 17, 19, 135. Medicare fiscal intermediaries are required to apply the SSI fraction computed by CMS when they determine a hospital’s DSH payment for a hospital fiscal year. 42 C.F.R. § 412.106(b)(2)-(3). The Secretary has always acknowledged, however, that hospitals have the right to contest CMS’ calculations of the SSI fractions through the usual administrative appeals process established under 42 U.S.C. § 1395oo. See 51 Fed. Reg. at 31,458 (Sept. 3, 1986) (“all hospitals have the right to appeal . . . the denial of eligibility for the disproportionate share adjustment or the amount of the adjustment they receive” after settlement of the hospital’s cost report and receipt of an NPR for a fiscal year). III. SUMMARY OF THE FACTS The Hospital’s fiscal years 1993 through 1996 are at issue here. The Board found that each of those years required reopening and correction in light of pervasive errors in the data and Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 18 of 145 11 match process used by CMS to calculate the SSI Fraction (and the resulting DSH payment) for all hospitals, including Plaintiff. This is not the first time that the Medicare DSH payment and the Secretary’s unlawful calculation of DSH have been before this Court. Most recently, in the lead case on the “core issue” in the hundreds of Medicare DSH payment actions consolidated as In re Medicare Reimbursement Litigation, Misc. No. 03-0090 (PLF), this Court exercised its mandamus jurisdiction to order the Secretary to correct his erroneous and unlawful calculations of the Medicaid fraction and retrospectively recalculate hospitals’ DSH payments for past years. Among other things, the Court concluded that the public interest in the Secretary’s following the rule of law eclipsed the Secretary’s purported concerns over the administrative burden of correcting his past errors. In re Medicare Reimbursement Litig., 309 F. Supp. 2d 89, 99 (D.D.C. 2004), aff’d, 414 F.3d 7 (D.C. Cir. 2005), cert. denied, 126 S. Ct. 1672 (2006). This case concerns a similar need for retroactive correction of the agency’s DSH calculation, but this time, the other fraction – the “SSI fraction” – is involved. This case, too, has been before this Court before. In 2000, the Hospital sued in this Court to compel the Secretary to comply with a PRRB subpoena requiring production of records reflecting the patients and patient days that were counted in CMS’ calculation of the Hospital’s SSI fraction for fiscal year 1995, one of the four fiscal years at issue now. Baystate Medical Center v. Thompson, Civ. No. 00-02940 (PLF). The Secretary eventually moved to dismiss the earlier enforcement action for lack of subject matter jurisdiction, submitting the supporting declaration of an agency employee (the “Thomas Declaration”) who stated, under penalty of perjury, that the data files that the agency had used to calculate the Hospital’s SSI fraction for 1995 did not exist. Pl. Stmt. ¶¶ 221-23. This Court later dismissed that action. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 19 of 145 12 Apparently, the government’s hope was that the sworn declaration submitted to this Court was “literally true” (CMS counsel’s words), even if grossly misleading. See Pl. Stmt. ¶ 227. But it was not true. Two years later, in a discovery hearing before the PRRB, the Hospital proved that the data files that the Secretary used to compute the SSI fraction for the Hospital’s fiscal year 1995 did exist. Through counsel, the Secretary immediately responded to that revelation with threats, issuing an extraordinary warning to the appointed members of the PRRB that they should “proceed very cautiously with respect to any further discovery orders in this case.” Pl. Stmt. ¶ 231. Nevertheless, the Hospital uncovered still more evidence, including evidence of numerous long-standing, systemic errors and omissions in the Secretary’s calculations of the SSI fraction, including some errors that had been known to the agency for many years but had never before been disclosed by the agency to the Hospital or, upon information and belief, to any other Medicare-participating hospital. See generally, PRRB Dec., AR 131-73; Pl. Stmt. ¶¶ 91-95, 123- 42. In the end, the Secretary’s Board found, and the CMS Administrator does not dispute, that the Secretary’s calculations of the SSI fraction for DSH are wrong in several respects, that almost all of the errors result in a systematic understatement of the SSI fractions and, therefore, the DSH payments that are properly owed to the predominantly non-profit hospitals, like the plaintiff Hospital, that serve low-income patients. The Board’s findings and the largely uncontroverted record evidence establish several errors in the Secretary’s calculations, all of which are systemic and recurring, tend to deflate the Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 20 of 145 13 SSI fractions and thus reduce the DSH payments to hospitals, easily could have been avoided in the first place, and are not burdensome to fix:5 • For all years prior to 1995, CMS did not count Medicare/SSI patient days that were required to be counted because the patients were receiving SSI benefits during their hospital stays but that were not counted only because the individual later died or his SSI record otherwise fell off the active SSI rolls by the time CMS did its calculation. See Pl. Stmt. ¶¶ 123, 125-50. • CMS did not count SSI patient days attributable to Medicare patients who received manual or forced payments of SSI benefits. See Pl. Stmt. ¶¶ 123, 151-60. • CMS omitted all days attributable to Medicare beneficiaries who were receiving non- cash SSI benefits during their inpatient hospital stays. See Pl. Stmt. ¶¶ 123, 174-76. • CMS did not count SSI patient days for individuals who were entitled to SSI benefits during their hospital stays but whose SSI benefits were awarded or reinstated by SSA after CMS performed one of the earliest of its several data matches, the one CMS used to calculate the SSI fraction. See Pl. Stmt. ¶¶ 41, 123, 161-73. • CMS did not use individuals’ own Social Security numbers to match records, as the Secretary said he would in the preamble to the original, 1986 DSH rule; and, until 1996, the Secretary had no means of capturing all SSI patient days when individuals received SSI or Medicare benefits under different numbers, which was possible given the types of identifiers that CMS was improperly using in its match process. See Pl. Stmt. ¶¶ 48-60, 62-71, 91-92, 97-104. The Board’s findings and undisputed record evidence also establish the lack of quality control that produced these problems: • CMS did not provide SSA with appropriate specifications or requirements for the SSI data needed to calculate the SSI fractions, did not undertake any meaningful efforts to test or validate the data files and programs that it used to compute the SSI fractions, and did not maintain required documentation for its calculations or the systems it used to perform those calculations. Pl. Stmt. ¶¶ 86, 178-83. • The Secretary cannot now say, and does not now have records to show, specifically which patient days were counted in his calculation of the SSI fraction for two of the 5 Each of these errors results in an undercounting of SSI patient days in the numerator of the SSI fraction. The Board also found a number of discrepancies in the count of Medicare inpatient days that were included in the denominator of the SSI fractions, which the Secretary was unable to explain. Pl. Stmt. ¶¶ 191-217. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 21 of 145 14 Hospital’s four fiscal years at issue (1993 and 1994), because CMS did not maintain the supporting records for its calculations for these years. Pl. Stmt. ¶¶ 42, 61, 83, 86, 139, 172; Answer ¶ 7. • The data that the Secretary claims to have used to calculate the fractions for the other two years at issue (1995 and 1996) do not agree with the SSI fractions that CMS actually applied to determine the Hospital’s DSH payments for those years, and the agency has admitted that it cannot explain the differences or why they occurred. Pl. Stmt. ¶¶ 86, 229-30. • The Secretary cannot now say how CMS’ match programs ran at the time when the agency computed the SSI fractions for the fiscal years at issue because CMS failed to keep systems documentation for its calculation process. Pl. Stmt. ¶¶ 42, 61, 83, 85, 86, 137-40, 178-83. The Board’s findings and the record evidence also show that CMS failed to use the best data available to the agency and that the agency faced no significant administrative burden in correcting the resulting errors: • CMS was unable to explain why it did not correct the errors that it discovered even before it performed the erroneous calculations at issue. Pl. Stmt. ¶¶ 91-95, 124, 140, 148. • Several errors would have been substantially reduced or eliminated altogether (including errors relating to inactive records and later reinstatement of SSI benefits) if only CMS used its later data matches that were performed before, but not considered in, the final DSH payment determinations that were issued for each year. CMS performs 10 quarterly data matches with updated SSI records for every federal fiscal year, and the latest of these matches is not run until more than two years after the end of each federal fiscal year, but before the final DSH payment determination issued for each year. With these later matches already in hand, government witnesses testified that the “fix” for this problem with CMS’ use of an early cut-off would have been simple: CMS would need only to have divided A (SSI days), by B (Medicare days) to compute C – the most accurate and up-to-date SSI percentage for the fiscal year in question. Pl. Stmt. ¶¶ 40, 43, 168. • The SSI data needed for CMS’s match process took only one hour each year for SSA to produce and it would not have been burdensome for SSA to have furnished CMS with updated and corrected SSI data more frequently than just once per year. Pl. Stmt. ¶¶ 29, 114. • CMS cannot identify any factor considered in establishing the early cut-off date for the match results used in the final DSH payment determinations, or why CMS never Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 22 of 145 15 used the later matches it performed for each year in the final DSH payment determinations that were later issued for each year. Pl. Stmt. ¶¶ 42, 61, 86. The Board’s decision and the underlying record evidence also establish the likely economic impact of the agency’s error. The undisputed evidence reflects that seemingly slight percentage changes in the Medicare/SSI fraction substantially impact the resulting DSH payments to hospitals. Pl. Stmt. ¶¶ 119-22, 135, 146-49, 186. Finally, the Board’s findings and the undisputed record evidence establish CMS’ knowledge and concealment of its errors and their magnitude: • CMS knew about the “inactive” records problem as early as 1993, before the Secretary used that incorrect SSI data to calculate the SSI fractions at issue. Pl. Stmt. ¶¶ 124, 136. A “quick and dirty” fix of that problem was implemented only later (in 1996), along with attempts to correct some other problems with the match process, but only prospectively and without disclosure to hospitals that prior years were contaminated by error. Pl. Stmt. ¶¶ 124, 128-29, 140. • CMS disseminated false and misleading statements, including the Thomas Declaration mentioned above (which falsely asserted that records of the agency’s calculations did not exist); various Federal Register notices regarding the SSI calculations and the data used to determine the SSI days in the numerator of the fraction; the agency’s federal court briefs regarding the count of Medicare days in the denominator of the fraction; sworn deposition testimony given by the agency’s designated witnesses; and what the Board charitably referred to as “disingenuous” testimony at the hearing by the agency’s lead employee on DSH policy matters. Pl. Stmt. ¶¶ 72-74, 124, 137-39, 172, 209, 222-23, 225-33. • CMS consistently resisted efforts by hospitals and the Board, in this case and others, to discover evidence in the government’s sole possession that would shed light on the totality and financial impact of the pervasive errors and omissions in the Secretary’s calculations. In fact, the CMS Administrator declined the Board’s express request that the agency enforce the Board’s subpoenas. Pl. Stmt. ¶¶ 23, 144, 157, 171, 218- 56. • CMS destroyed records that would have showed the precise impact, to the day, of the Secretary’s omission of all inactive SSI entitlement records in the calculations of the SSI fractions for at least 1993 and 1994. In early 1996, the Secretary obtained corrected records from SSA to replace CMS’ files that were “contaminated” (CMS’ word) by the omission of all inactive records of SSI recipients. CMS requested this corrected data from SSA for the purpose of rerunning matches for prior years and correcting SSI fractions for those years if differences were material. But, after it Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 23 of 145 16 reran the new matches, CMS staff received instructions to go in a “different direction” and never corrected fractions for prior years. The corrected tapes existed as late as 1997, but then were lost or destroyed or allowed to “disappear” after the SSI entitlement data previously had been kept under one staffer’s desk. See Pl. Stmt. ¶¶ 124, 128-29, 135, 139-40, 144, 148. The Secretary, in his final decision reversing the Board, does not dispute the Board’s factual findings or the record evidence on the systemic problems with the Secretary’s calculations of the SSI fractions. Pl. Stmt. ¶¶ 22, 30, 35, 52, 60, 65, 71, 88, 115, 143-44, 157, 171, 176, 200. Instead, the CMS Administrator rests his decision to reverse the Board on unsound legal argument and the unsupported assertion that his agency uses the “best data available” in its Medicare payment calculations and therefore, as a matter of “policy,” declines to revisit them even if wrong. The Administrator thus assumes the conclusion of his own argument in declaring that CMS achieved an “acceptable” rate of error for the SSI fraction, even while no rate of error is accepted by the Secretary for the other fraction – the Medicaid fraction – that is used to compute the DSH payment. Pl. Stmt. ¶¶ 5, 24, 116, 140, 149, 159, 172-73, 177-90, 201-03. IV. SUMMARY OF ARGUMENT The CMS Administrator concluded that the Hospital is not entitled to relief for three essential reasons: (1) the Hospital failed to prove the precise financial impact of the Secretary’s errors with data only the Government has and which the Secretary’s agency refused to produce; (2) the Secretary’s calculations are fixed when performed and can never be revised, no matter how egregiously wrong they are, to reflect corrected data; and (3) the Secretary’s calculations were performed with the “best data available.” As shown below, the Court may easily dispose of the Secretary’s first two arguments because they are contrary to controlling Circuit precedents. First, under the Court of Appeals’ decision in Atlanta Coll. of Med. & Dental Careers, 987 F.2d at 830-31, the demonstrated errors Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 24 of 145 17 and omissions in the Secretary’s calculation of the SSI fractions shifts the burden to the Secretary to show why the Hospital’s allegations should be disregarded. Second, the Court of Appeals’ affirmance of this Court’s decision compelling the Secretary to reopen and retrospectively revise DSH payment determinations issued as many as 10 years earlier forecloses the Secretary’s argument that the Hospital is not entitled to retrospective relief from the undisputed errors and omissions in the SSI fractions used to compute its DSH payments. In re Medicare Reimbursement Litig., 414 F.3d 7. Third, the Secretary’s assertion that CMS used the “best data available” to compute the SSI fractions simply cannot withstand the overwhelming weight of uncontroverted evidence in the record. Current and former government employees testified that CMS knew that its data was incorrect when it performed the calculations at issue and that CMS never took any meaningful steps to implement quality control systems for this calculation. Morall, 412 F.3d at 167 (an agency’s decision cannot stand when it fails to consider contradictory record evidence and when it ignores the credibility findings of the administrative tribunal that heard the evidence directly). Indeed, because the Secretary never even documented the systems and data used to perform these calculations for what is now a $6 billion dollar per year DSH payment adjustment nationwide, the Secretary cannot say with certainty which patients and patient days were counted in the agency’s calculations of the Hospital’s SSI fractions or how the programs that were used compute the fractions worked at the time the calculations were performed. The Secretary’s failure to use the best data available underscores the propriety and necessity of retrospective correction. See County of Los Angeles v. Shalala, 192 F.3d 1005, 1020-23 (D.C. Cir. 1999). Additionally, the Administrator’s 2005 decision in this case that the SSI fractions should not include non-cash SSI benefits is invalid, and must be reversed, because it is inconsistent with Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 25 of 145 18 the plain language of the SSA regulation that the Administrator purports to rely upon. 20 C.F.R. § 416.264 (1995). Moreover, this Court owes no deference to CMS’ interpretation of SSA’s regulations concerning the SSI program that SSA administers. Department of Treasury v. FLRA, 837 F.2d 1163, 1167 (D.C. Cir. 1988). Finally, the Administrator wrongly concluded that the denominator of the SSI fractions should include Medicare “covered or utilized days,” even when days are not paid by Medicare Part A, and days for Medicare beneficiaries enrolled in a health maintenance organization. That standard is inconsistent with the plain language and manifest intent of the DSH statute and the agency’s original interpretation of the statute and regulations that were in effect during the periods at issue. Moreover, the agency’s erratic and inconsistent interpretation of the statute and regulation is invalid. National Treasury Employees Union v. FLRA, 399 F.3d 334, 337 (D.C. Cir. 2005); Michigan Public Power Agency v. FERC, 405 F.3d 8, 12-14 (D.C. Cir. 2005). V. ARGUMENT A. Standard of Review Jurisdiction over this action lies under 42 U.S.C. § 1395oo(f), which provides that the case “shall be tried pursuant to the applicable provisions under” the Administrative Procedure Act (“APA”). The applicable provisions of the APA require the Administrator’s decision to be set aside if it is arbitrary, capricious, an abuse of discretion, unsupported by substantial evidence, or otherwise not in accordance with law. 5 U.S.C. § 706. An agency’s interpretation of a statute or regulation that it is entrusted to administer generally is afforded deference. But, an agency’s interpretation is not entitled to deference when the agency interprets a statute or regulation that it has not been entrusted to administer, Department of Treasury, 837 F.2d at 1167, when the agency’s interpretation is inconsistent with Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 26 of 145 19 the plain language of the statute or regulation, United States v. Mead Corp., 533 U.S. 218, 227- 31 (2001); Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512 (1994), when an alternative interpretation of a regulation is indicated by the agency’s intent when it adopted the regulation, Thomas Jefferson Univ., 512 U.S. at 512, or when the agency has erratically and inconsistently construed its charter. National Treasury Employees Union, 399 F.3d at 337. An agency’s unexplained departure from its established precedent is also arbitrary and capricious. See, e.g., Michigan Public Power Agency, 405 F.3d at 12-14. The scope of review under arbitrary and capricious standard, however, also entails a careful, searching inquiry as to whether: the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise. Motor Vehicle Mfrs. Ass’n of the United States, Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). Substantial evidence “means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion,” taking into account “whatever in the record fairly detracts from its weight.” AT&T Corp. v. F. C. C., 86 F.3d 242, 247 (D.C. Cir. 1996) (internal quotes and citations omitted). An agency’s final decision cannot withstand review when it fails to consider contradictory record evidence and when it ignores the credibility findings of the administrative tribunal that heard the evidence directly. Morall, 412 F.3d at 167. Finally, a reviewing court may uphold agency action only on the basis articulated by the agency in its decision, not on post-hoc rationalizations offered by the agency or its counsel in litigation. See, e.g., Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168-69 (1962); Biloxi Reg’l Med. Ctr. v. Bowen, 835 F.2d 345, 348 n.12, 351 n.18 (D.C. Cir. 1987). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 27 of 145 20 B. The Secretary Wrongfully Denied the Hospital Relief Based On the Hospital’s Alleged Failure to Prove the Precise Financial Impact Of the Secretary’s Errors With Information That Only the Government Has and That the Secretary Refused To Produce. The Secretary cannot lawfully refuse to correct undisputed errors in the calculations of the Hospital’s SSI fractions on the premise that the Hospital failed to prove the precise impact of those errors. Only the Secretary, not the Hospital, has access to the information necessary to make that showing. Answer ¶ 105. Moreover, while the Administrator’s decision (AR 37) admits only to a “lengthy and contentious disagreement about the release of data,” CMS took every conceivable step to keep the Hospital and the Board from obtaining that information. 6 See PRRB Dec. at 25, AR 155; Pl. Stmt. ¶¶ 23, 218-56. For example, CMS, through counsel, refused to cooperate with an alternative proposal that was acceptable to SSA and that would have yielded the correct number of SSI days for all of the Hospital’s Medicare patients, even though the Hospital offered to bear the full cost of that alternative process. Pl. Stmt. ¶ 244. 6 As noted above, the Hospital’s earlier action to compel CMS to produce records was dismissed after CMS submitted a sworn statement – later proven false before the Board – that the records did not exist. Pl. Stmt. ¶ 222-26. The agency’s mindset about production of relevant documents in this case is reflected in a later email from agency counsel who sought to confirm with a CMS staffer that the false declaration that had been submitted to this Court was “literally true, I hope.” Pl. Stmt. ¶ 227. Later, despite CMS counsel’s warning that the Board should “proceed cautiously,” the Board requested that the Secretary’s counsel initiate legal action to enforce subpoenas requesting SSI entitlement records from SSA, but the Secretary’s counsel refused and invoked the aid of the Administrator in denying the Board’s request. See Pl. Stmt. ¶¶ 231-36. Moreover, the evidence in this case later revealed that SSI entitlement records, which had been furnished to CMS by SSA, which had at one time been kept under a CMS’ employee’s desk, and which were confirmed by agency counsel to be in the hands of CMS staff as late as 1997, were later lost, destroyed or allowed to disappear. Those records would have shown the impact of CMS’ past omission of “stale” SSI records to the day and even as thousands of other appeals challenging the Secretary’s calculations were pending before the PRRB. See Pl. Stmt. ¶ 139. Despite all that the agency has successfully hidden from view, however, there is no question that Medicare/SSI days were omitted from the numerator of the SSI fractions at issue, and thus necessarily deflated the resulting percentages and DSH payments incorrectly. The only remaining question is how many days were omitted and how many dollars the Hospital lost as a result. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 28 of 145 21 The Secretary’s own Board graciously called the Secretary’s position “disingenuous.”7 This Court should pronounce it arbitrary and capricious. 1. The Hospital Does Not Have the Burden to Show the Impact Of the Secretary’s Errors. Given the Secretary’s withholding and destruction of relevant evidence in the Government’s sole possession, the decision to deny relief because the Hospital did not prove the precise impact of the agency’s errors is, to put it mildly, arbitrary and capricious. The undisputed evidence supporting the Board’s findings of errors and omissions in CMS’ calculations of the SSI fraction shifted the burden to the Secretary “to produce countervailing evidence or a reason, not based on the insufficiency of the [Hospital’s] showing, that explains why [the Hospital’s] allegations [should not be] accepted.” Atlanta Coll. of Med. & Dental Careers, 987 F.2d at 830-31 (emphasis added).8 The Secretary has not met this burden.9 Moreover, it is well-established that “when a party has relevant evidence within his control which he fails to produce, that failure gives rise to an inference that the evidence is unfavorable to him.” Int’l Union v. NLRB, 459 F.2d 1329, 1336 (D.C. Cir. 1972). Thus, even if the Secretary were correct that the record evidence fails to show a substantial adverse financial 7 PRRB Dec. at 25 & n.135, 32, AR 130, 137. 8 While a plaintiff ordinarily must prove the fact of injury with reasonable certainty, the plaintiff does not bear the same burden as to the amount of damages – particularly when the defendant is the cause of plaintiff’s inability to prove damages with mathematical certainty. See Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555, 563 (1931); Samaritan Inns, Inc. v. District of Columbia, 114 F.3d 1227, 1235, 1237 (D.C. Cir. 1997). 9 See also Loma Linda Cmty. Hosp. v. Shalala, 907 F. Supp. 1399, 1404 (C.D. Cal. 1995) (finding that “the Secretary acted arbitrarily, not in accordance with the law, and abused her discretion in ruling that [the agency’s SSI calculation] was presumed correct and that [the hospital] had not met its burden of disproving it, even though [the fiscal intermediary] provided no supporting data for its calculation and claimed that the Privacy Act precluded its disclosure of such data to [the hospital]”). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 29 of 145 22 impact (and he is not), see Pl. Stmt. ¶ 186, the inference to be drawn from the Secretary’s refusal to produce relevant evidence on that point is that there is a substantial impact adverse to the Hospital. For this reason alone, the Secretary’s decision must not stand. Agencies cannot presume the propriety of their own actions in the face of adjudicated facts to the contrary.10 CMS cannot create evidentiary gaps in a hearing record and then fill those gaps with its Administrator’s self-serving, after-the-fact speculation. That would effectively “sanction the creation of an unregulated back door through which unrebuttable, non- record evidence could be introduced against [the Hospital] outside of the statutorily-mandated hearing context.” Kaczmarczyk v. INS, 933 F.2d 588, 596 (7th Cir.), cert. denied, 112 S. Ct. 583 (1991). 2. The Secretary’s Errors Resulted in Substantial Underpayments. While the Administrator claims that the agency achieved an “acceptable rate of error” in the calculations at issue, Adm’r Dec. at 26-27, AR 27-28, the DSH statute does not allow for the Secretary’s guess as to a hospital’s SSI fraction for a year.11 Congress recognized that the Secretary would have to use the best historical data available to estimate preliminary, interim DSH payments to a hospital, but Congress intended the Secretary “to develop accurate data” for the SSI fraction used in the “final settlement” for each fiscal year. S. Rep. No. 99-146 at 291 10 See United Scenic Artists v. NLRB, 762 F.2d 1027, 1034-35 (D.C. Cir. 1985) (overruling an agency finding “presumed as matter of law, whether or not this squares with the actual facts”); Gutierrez-Rogue v. INS, 954 F.2d 769, 773 (D.C. Cir. 1992) (the due process right to a meaningful hearing requires an agency to provide a party with the opportunity to rebut an officially noticed fact). 11 It is important to consider the source when the CMS Administrator declares that the agency achieved an “acceptable rate of error.” The CMS analyst in charge of DSH testified that a DSH payment issue “involving ‘a few hundred million dollars’ would ‘[n]ot necessarily’ have left a lasting impression on her, ‘given the amount of money involved,’” Pl. Stmt. ¶ 141. In his Answer to the Complaint, the Secretary himself denied that an error of $100 million on DSH payments would be significant. Answer ¶ 203. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 30 of 145 23 (1985), reprinted in 1986 U.S.C.C.A.N. 42, 258 (emphasis added). Consistent with Congressional intent, the DSH statute directs the Secretary to determine the SSI fraction based upon “the number” of days (not an estimate) attributable to patients who were entitled to SSI and Medicare Part A benefits and “the number” of days (again, not an estimate) attributable to patients who were entitled to Medicare Part A. 42 U.S.C. § 1395ww(d)(5)(F)(vi)(I) (emphasis added). In any event, despite the agency’s efforts to “hide the ball,” and despite the fact that the Hospital was under no obligation to prove the magnitude of CMS’s errors, the record nonetheless clearly establishes the substantial impact of those errors. It runs to the hundreds of millions of dollars for all hospitals nationwide during the years in question. See Pl. Stmt. ¶ 186. As the Court of Appeals has previously observed, even “seemingly modest percentage differences [can] represent substantial sums of money” in the Medicare reimbursement context. County of Los Angeles, 192 F.3d at 1010. And that is certainly true in this case. As the PRRB observed in its decision in this case, while the financial impact of an error in the SSI calculation “may be relatively small when averaged over all PPS hospitals, it can have a significant impact on those PPS hospitals that serve a large low-income population.” Pl. Stmt. ¶ 132. C. The Secretary Incorrectly Alleges that the Law Provides For No Right To Relief From the Secretary’s Errors. The Administrator claims that the right to relief is merely a “trapping[] of the adversary process” that can be dispensed with when fairness is assured by “sound management [sic] techniques and quality control designed to achieve [sic] acceptable rates [sic] of error.” Adm’r Dec. at 26, AR 27 (misquoting Califano v. Boles, 443 U.S. 282 (1979)). Tellingly, the Administrator does not claim that CMS actually employed such measures for the SSI calculations, but simply declares that the agency’s errors were acceptable despite the proven Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 31 of 145 24 absence of any meaningful quality control and the agency’s efforts to conceal known errors. Id. Additionally, the Administrator alleges that the Secretary gave hospitals notice and an opportunity to comment on the fact that CMS was calculating the SSI fractions retrospectively, based on the next June update to the agency’s MEDPAR file, for the prior federal fiscal year and that hospitals would be precluded from “using any other data” for the SSI calculations. Id. at 27 & n.41, AR 28. As shown below, each of these arguments lacks merit. 1. Right to Relief The Administrator is correct in noting that our Court of Appeals has held that hospitals are not necessarily entitled to retrospective relief from errors in the Secretary’s determination of Medicare PPS payment rates that are established prospectively and are based on the best data available at the time the rates were set. See Methodist Hosp. of Sacramento v. Shalala, 38 F.3d 1225 (D.C. Cir. 1994). But, this principle has no application with respect to errors in payment determinations, like the SSI fractions, that are not prospectively established or are not based on the best data available. The reason why courts have allowed the Secretary to implement prospective-only corrections of errors in prospectively established PPS calculations is to promote certainty and predictability with respect to the truly prospective PPS rates, as well as minimizing administrative burdens. Id. at 1232. But, these concerns are absent in appeals from Medicare payment determinations, like the DSH payment calculation in general and the SSI fraction in particular, that are performed retrospectively, after the fiscal year to which they apply. Thus, the Court of Appeals gave no credit to the Secretary’s claim of undue burden in 2005, when it affirmed this Court’s order compelling the Secretary to reopen and retrospectively revise DSH payment determinations that had been issued up to 11 years earlier in a manner that was contrary Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 32 of 145 25 to law. In re Medicare Reimbursement Litig., 414 F.3d at 13 (“Having to pay a sum one owes can hardly amount to an equitable reason for not requiring payment.”) And, indeed, no court has ever denied retrospective relief in an appeal, like this one, contesting errors in CMS’ retrospective payment determinations, even when they involve a hospital-specific payment adjustment under the PPS.12 On the contrary, the courts in this Circuit and others have not hesitated to compel retrospective relief even from errors in PPS calculations that truly are prospectively determined when those calculations were not based on the best data available.13 Even the Secretary has acknowledged elsewhere that retrospective relief clearly is warranted when his calculations of prospectively determined rates are shown not to have been calculated on the basis of the best data available. See Sarasota Mem. Hosp. v. Shalala, 60 F.3d 1507, 1513-14 (11th Cir. 1995) (noting the Secretary’s stipulation that retrospective relief is available to the plaintiff hospitals upon a showing that the Secretary’s calculation of the PPS wage index was incorrect). 12 See Georgetown Univ. Hosp. v. Bowen, 862 F.2d 323 (D.C. Cir. 1988) (invalidating a CMS rule providing for prospective-only correction of a hospital-specific PPS payment rate); see also Riverside Methodist Hosp. v. Thompson, No. C2-02-94, 2003 U.S. Dist. LEXIS 15163 (S.D. Ohio, July 31, 2003) (requiring the Secretary to retrospectively correct his final determination regarding the PPS payment adjustment for a hospital’s indirect medical education costs). 13 County of Los Angeles v. Shalala, 192 F.3d at 1020-23 (ordering the Secretary to provide a rational explanation for the use of 1981 data in setting prospectively-determined thresholds for PPS “outlier” payments or to recompute the thresholds using later data and pay hospitals the additional sums due prior years); Alvarado Community Hosp. v. Shalala, 1998 U.S. App. LEXIS 33973, *28 (9th Cir. 1998) (ordering the Secretary to recompute PPS outlier thresholds because the Secretary did not use the best data available in calculating those thresholds); Centra Health, Inc. v. Shalala, 102 F. Supp. 2d 654, 659-61 (W.D. Va. 2000) (granting retrospective relief from the calculation of the PPS wage index that was not based on the best data available). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 33 of 145 26 2. Unsound Management, Lack of Quality Control, and Concealment The Administrator’s reliance (Dec. at 26, AR 27) on the Supreme Court’s decision in Califano v. Boles, 443 U.S. 282 (1979), is woefully misplaced. There the Court noted that the SSA may sometimes adopt procedures (e.g., suspension or termination of benefits without a hearing in advance) to accommodate adjudication of vast numbers of social security benefits claims because “fairness can best be assured by Congress and the [SSA] through sound managerial techniques and quality control designed to achieve an acceptable rate of error.” Id. at 285. While there is no substantial evidence in the record supporting the Administrator’s pronouncement that CMS one way or another achieved an acceptable rate of error here,14 the complete absence of sound management or anything resembling quality control in connection with the calculations at issue, coupled with the agency’s active concealment of known errors, defies the Administrator’s reliance on Califano. The record shows that CMS performed these calculations for the DSH payment adjustment, which currently amounts to more than $6 billion per year, “on the fly,” without any meaningful efforts to test or validate any aspect of the SSI calculations – neither the SSI data, nor the match process, nor the count of Medicare days – beyond simply “eyeball[ing]” the numbers. Pl. Stmt. ¶¶ 180-83. Despite public representations that the Secretary would perform the data match using Social Security Numbers, his agency never did that, and undercounted the SSI 14 For example, the Administrator’s decision grossly understates his alleged rate of error in the count of Medicare days by comparing the numbers of Medicare-paid days for fiscal years 1993 and 1994 with numbers of Medicare days reflected in later, corrected data for those years that was not used to calculate the SSI fractions at issue for those years. Pl. Stmt. ¶¶ 201-03. In short, because CMS did not maintain the detailed support for its calculations for any hospital’s SSI fraction for fiscal years 1993 and 1994, Pl. Stmt. ¶ 86, the Administrator computed a rate of error by comparing to later, corrected data that the agency has but never used in the DSH payment calculations at issue. Pl. Stmt. ¶ 203. If the Administrator had used the actual numbers reflected in the SSI fractions at issue for fiscal years 1993 and 1994, the Administrator’s alleged rate of error would be approximately 23 times greater than the rate suggested in his decision. Id. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 34 of 145 27 Fraction as a result. Id. at ¶¶ 48-60, 62-71, 75-88, 117-22. No one ever tested the match process or validated its results, even after problems were discovered, and CMS representatives admit to knowing very little about how the agency’s own process actually worked.15 Id. at ¶¶ 42, 61, 83, 86, 93. CMS’ failure to maintain reasonable documentation for its calculations and programs, and its failure to have undertaken any reasonable measures to test the accuracy and validity of its calculations, not only belie the Administrator’s reliance on Califano but also render the calculations performed for the years at issue arbitrary and capricious, especially given that those failures violated applicable Federal Information Processing Standards. Pl. Stmt. ¶¶ 136, 177-85; see Shafer, Hearing Tr. at 332-34, 409-13, 348-49, 632-33, 651. During the years at issue, Federal Information Resources Management Regulations, 41 C.F.R. § 201-20.303(b)(2)(i), and Office of Management and Budget (“OMB”) Circular No. A- 130 required all Federal agencies to comply with Federal Information Processing Standards (“FIPS”). FIPS 101 established Federal standards for software validation, verification and testing.16 CMS complied with none of those standards here. 15 The CMS staffer who has been principally responsible for the match process since 1995, Dean, gave wholly contradictory testimony as to how his match process actually worked during the period at issue and admitted that CMS has no way to know now how the match process worked then because the agency did not maintain systems documentation for the process. See Pl. Stmt. ¶¶ 75-88. The Board thus found that “CMS’ employees charged with the responsibility for implementing DSH demonstrated a lack of comprehensive knowledge of the match process, i.e., what data was included in or excluded from the SSA tapes, as well as how, when, why and to what extent the shortcoming of that data might lead to a failure to match CMS’ MEDPAR file.” PRRB Dec. at 21-22, AR 151-52. In contrast, the Board found that “the SSA witnesses were credible. All had over 20 years of experience with SSA and were extremely knowledgeable about SSI programs and SSA’s operations, including recordkeeping.” PRRB Dec. 21; AR 151. The Administrator’s decision must be reversed for failure to address the Board’s findings concerning the credibility of the present and former SSA and CMS employees who testified before the Board. Morall, 412 F.3d at 167. 16 Validation must examine the correctness of the final version of the software. FIPS 101 at 4, AR 3840. Verification must examine the software’s integrity and internal consistency both within each phase of the software’s development and from one evolutionary phase of the Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 35 of 145 28 In addition, OMB Circular A-130 stands on the fundamental principle that: Systematic attention to the management of government records is an essential component of sound public resources management which ensures public accountability. Together with records preservation, it protects the government’s historical record and guards the legal and financial rights of the government and the public. 59 Fed. Reg. 37,906, 37,910 (July 25, 1994). More specifically, the OMB circular requires Federal agencies to “[r]ecord, preserve, and make accessible sufficient information to ensure the management and accountability of agency programs, and to protect the legal and financial rights of the Federal Government.” 59 Fed. Reg. 37,910. All federal agencies,17 therefore, are required to “[i]ncorporate records management and archival functions into the design, development, and implementation of information systems.” Id. In addition, agencies must develop “adequate and proper documentation of agency activities.” Id. at 37,911. As the Director of OMB stated, “Federal agencies cannot carry out their mission in a responsible and responsive manner without adequate recordkeeping.” Id. at 37,921. In this case, however, the Secretary cannot even say with certainty which particular patient days were counted in the SSI fractions at issue,18 or how the match process worked during the periods at issue. Pl. Stmt. ¶¶ 42, 61, 83, 86, 93. software’s development to another. Id. Testing must examine the software’s behavior using sample data sets. Id. Validation, verification and testing were required in all phases of software development: requirements, design, programming, testing, installation, operations, and maintenance. Id. 17 The obligation to comply with OMB Circular A-130 extends to all Federal agencies. 61 Fed. Reg. 6,428, 6,436 (Feb. 20, 1996). 18 The Secretary has no records available reflecting which days were counted in his calculations of the SSI fraction for any hospital for fiscal years 1993 and 1994. Pl. Stmt. ¶ 139; Answer ¶ 7. For fiscal years 1995 and 1996, CMS produced the data that was supposed to have been used to calculate the Hospital’s fractions, but those records do not agree with the fractions that were applied to determine the Hospital’s DSH payments and CMS has admitted that it cannot account for the differences. Pl. Stmt. ¶¶ 236-37. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 36 of 145 29 Worse than all of these complete lapses in quality control, when CMS found out that it had a problem with the SSI data, not later than 1993, the agency embarked on a course of delay, disinformation and concealment. Pl. Stmt. ¶¶ 124, 140. In violation of the Secretary’s fiduciary duties to disproportionate share hospitals, the agency quietly resolved discrepancies with some hospitals that had gotten too close to discovering errors in the process, while CMS staff testified falsely in proceedings initiated by other hospitals that the SSI data used in the calculation of the SSI fraction were verified and correct. Pl. Stmt. ¶ 140. Meanwhile, the agency published a rule change in the Federal Register in 1995 stating falsely that recalculations of the SSI fraction invariably resulted in lower SSI ratios due to problems, not with CMS’ data but with the requesting hospitals’ data, and claiming falsely that SSA, not CMS, computes the SSI fractions based on data that CMS is not given. Pl. Stmt. ¶ 140. The evidence before the Board proved those representations to be not only false, but a grossly misleading deflection of comments from hospitals asking CMS to release the underlying data to them so that they could verify the accuracy of the calculations. See id. Then, after having obtained corrected SSI records from SSA and having created new match records for prior years, which would have shown the precise impact of its errors to the day, the agency later lost or destroyed those records. Id. CMS never fixed its prior errors for affected hospitals and never disclosed the errors to a single hospital until 2003, when they first began to be uncovered by the Plaintiff through its discovery efforts in this case. Pl. Stmt. ¶¶ 91-95, 124, 136, 139-40. Measured against the duties and obligations of the Secretary as a fiduciary with respect to the calculation of the SSI fractions, this course of conduct is tantamount to fraud or similarly wrongful conduct. In any event, it is enough that the Secretary’s calculations are wrong, and that CMS’ concealment of those errors precludes the Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 37 of 145 30 Administrator’s assertion that CMS did the best it could or that the Secretary ought to be relieved of the obligation to fix CMS’ mistakes. 3. Lack of Notice and Meaningful Opportunity to Comment In the face of all this carelessness and obfuscation the Secretary tries to blame the hospitals, suggesting that his agency’s admittedly unwritten policy against retrospective correction of errors (Answer ¶ 102) and the early cut-off for the match process were vetted by public notice and comment. Not at all. The Administrator’s decision is flat wrong when it alleges that the Secretary notified hospitals that the SSI fractions would be calculated based on a data match using the June MEDPAR update, nine months after the end of the fiscal year to which the calculations apply, and that hospitals would be precluded from “using any other data” for the SSI calculations. Adm’r Dec. at 27 & n.41, AR 28. According to the Administrator’s decision, this notice was given in the appendices to the final rules establishing the prospective PPS rates published in advance of federal fiscal years 1992, 1993, 1994 and 1995. Id. But, what those notices actually say is that the prospective PPS rates established for each of those federal fiscal years were calculated using data in CMS’ MEDPAR file as of the prior month of June. None of the notices cited in the Administrator’s decision (n.41) address when CMS was calculating, or what MEDPAR data run CMS was using to calculate, the SSI fractions for the prior federal fiscal year that was already completed when each notice was published. Indeed, one of the very notices the Administrator relies upon in support of this argument stated falsely that SSA calculated the fractions and that CMS (then called “HCFA”) did not have the underlying data that SSA allegedly used in those calculations.19 60 Fed. Reg. 45,778, 45,812 (Sept. 1, 1995). And, having 19 Indeed, regarding CMS’ use of the June matches to compute the SSI fractions for the prior federal fiscal year, the PRRB found that CMS’ own policy staff was not aware of what exactly Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 38 of 145 31 given no notice that CMS was using the June MEDAR data matches to compute the SSI fractions, hospitals had no opportunity to comment on that approach. Finally, the Administrator’s assertion that the agency’s calculations of the SSI fraction are fixed when performed, and can never be revised to reflect updated or corrected data, is belied by the fact that the agency itself has recalculated the SSI fractions for at least two fiscal years. In October 1999, CMS recalculated the SSI fractions for all hospitals for fiscal year 1998, and that revision reduced the Provider’s DSH payment for 1998 by nearly $35,000.20 Pl. Stmt. ¶ 172. D. The Secretary Did Not Use the “Best Data Available” to Calculate the SSI Fractions at Issue. The Secretary invokes Medicare data “policy” and broad principles of administrative law such as “finality” and “administrative burden” to reach the conclusion that the calculations at issue were based, as a matter of fact, on the “best data available.” Adm’r Dec. at 37-39, 42-43, 53-54, AR 38-40, 42-44, 54-55. The Administrator asserts that “CMS has historically used best available data in computing [PPS] payments,” and thus concludes that the agency must have followed that practice here, too. Adm’r Dec. at 23, 28, AR 24, 29. This conclusion cannot be squared with the evidence in the record, however, and should be summarily rejected. the agency was doing, noting that CMS’ own witness, “Ms. [Anne] Tayloe [formerly Anne Rudolph], the individual responsible for [CMS’] DSH policy during the periods in issue, apparently believed that the latest available data was being used to compute the Medicare fraction.” PRRB Dec. at 33, AR 163. Similarly, the CMS employee who had served as the director of the agency division responsible for DSH payment policy, Phillips, testified to his misimpression that the SSI fractions for each federal fiscal year were calculated in advance of the year. AR 5358. He asserted this mistaken belief as the reason why he felt that hospitals should not be entitled to retrospective correction of errors in the agency’s calculations. Id. at 5345, 5388. 20 More recently, the SSI fractions for fiscal year 2005 were initially posted on CMS’ website on August 2006, but were later taken off the website, while CMS considered whether to change them, and were not posted again on the website until October 2006. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 39 of 145 32 See Morall, 412 F.3d at 167. The Board’s findings and the uncontroverted record evidence show that the calculations at issue were not based on the best data available: • CMS never bothered to take any meaningful step to ensure that the processes it used, and that data it requested from SSA, to perform the calculations of the SSI fractions were accurate and adequately documented. Pl. Stmt. ¶¶ 179-84. • The SSI data used in the calculations at issue was incomplete and incorrect in several respects. Pl. Stmt. ¶¶ 123-76. • CMS agency failed to request or use an accurate data match process that captured the SSI entitlement information for all Medicare/SSI patients, and the agency never used individual’s own Social Security numbers to match records, as the Secretary’s 1986 DSH rule said it would.21 Pl. Stmt. ¶¶ 21-122. • The Medicare inpatient stay data was inaccurate, and the Secretary cannot explain the discrepancies. Pl. Stmt. ¶¶ 211-17. The Secretary’s “best data available” argument, therefore, is entitled to the same consideration the Secretary gave to ensuring the accuracy of his calculations – none. E. The Secretary’s Decision to Exclude Non-Cash SSI Benefits is Contrary to Law. The Administrator belated conjures an ineffective legal defense for one of the SSI data omissions that resulted from CMS’ mismanagement of the SSI calculations. The Administrator’s decision asserts now that the agency’s inadvertent omission of SSI records for individuals who received non-cash SSI benefits is legally correct. AR 49-50. The PRRB found that the Secretary’s calculations of the SSI fractions at issue are systemically understated in part because the calculations omitted all days attributable to 21 If the Secretary matched records based on Social Security numbers, as he represented in the preamble to the 1986 DSH rule, the problems stemming from the identifiers used in his match process likely would have been substantially mitigated if not avoided altogether. Notably, SSA has for years performed a monthly match to CMS records for patients entering or leaving nursing homes in order to adjust SSI payments to those individuals, and SSA runs those matches on Social Security numbers first, and then a secondary match using names and date of birth. Pl. Stmt. ¶¶ 51, 108, 112, 114. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 40 of 145 33 Medicare beneficiaries who were receiving non-cash SSI benefits under title XVI of the Social Security Act during their inpatient hospital stays. PRRB Dec. at 29-30, AR 159-60. The Board reasoned that “Section 1619(b) of the Social Security Act, as amended, 42 U.S.C. § 1382h(b), establishes a work incentive for disabled individuals entitled to SSI benefits by allowing for the continuation of Medicaid benefits even though the individual’s income is too high to qualify for an SSI cash payment during periods of work.” PRRB Dec. at 29, AR 159. Quoting from SSA regulations codified during the period at issue at 20 C.F.R. § 416.264 (1995), the PRRB noted: [t]he special SSI eligibility status applies for the purposes of establishing or maintaining your eligibility for Medicaid. For these purposes, we [SSA] continue to consider you to be a blind or disabled individual receiving benefits even though you are in fact no longer receiving regular SSI benefits or special SSI cash benefits. . . . Special SSI eligibility status also applies for purposes of reacquiring status as eligible for regular SSI benefits or special SSI cash benefits. PRRB Dec. at 29 (emphasis in italics added by the PRRB), AR 159. Based on this provision, the PRRB concluded that “[a]s stated in the regulation, certain qualifying disabled recipients are considered to be entitled to Federal SSI benefits even though no SSI cash payment is made, both for purposes of allowing the recipient to continue receiving Medicaid coverage and to resume receiving the SSI cash benefit when the individual is unable to work, without having to reapply.” Id. The Administrator does not controvert the PRRB’s finding that CMS’ calculation of the SSI fractions for periods at issue categorically omitted all SSI patient days attributable to individuals who received non-cash SSI benefits. See Adm’r Dec. at 37-41, AR 38-42. Instead, the Administrator reversed the Board’s ruling on the legally incorrect ground that “all of the regulations relied upon by the Board in this case expressly state that they are applicable only for purposes of Medicaid eligibility.” Adm’r Dec. at 49, AR 50. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 41 of 145 34 The Administrator’s decision on this point is contrary to law, and must be reversed, because it ignores the last sentence of the SSA regulation that the Board relied upon, as quoted above, and omits that last sentence from the portion of the regulation that is quoted in the Administrator’s decision (at 48). AR 49. The sentence in the regulation that is omitted in the Administrator’s decision states that the “Special SSI eligibility status” afforded under Section 1619(b) of the Social Security Act applies not only for Medicaid purposes, as the Administrator’s decision asserts, but “also applies for purposes of reacquiring status as eligible for regular SSI benefits or special SSI cash benefits” under the SSI program. Both types of non-cash SSI benefits mentioned in the SSA regulation that the PRRB relied upon, 20 C.F.R. § 416.264 (1995), are afforded to individuals under the provisions of Title XVI of the Act governing the SSI program. See 42 U.S.C. § 1382h(b). Thus, as the PRRB correctly concluded, individuals who receive these special non-cash benefits are in fact “entitled to supplemental security income benefits . . . under title XVI” as described in the pertinent provision of the Medicare DSH statute, 42 U.S.C. § 1395ww(d)(5)(F)(vi)(I).22 The Secretary’s current position on non-cash benefits should be seen for what it is – a post-hoc rationalization for yet another mistake. The omission of these days is not the product of reasoned decision-making by the agency when it performed the calculations at issue. F. The Secretary’s Final Decision Applies the Wrong Standard for Determining the Proper Count of Medicare Days in the Denominator of the SSI Fraction. Perhaps not surprisingly, the Secretary’s complete inability to keep track of his own policies, never mind his own data processes, has also led to a legal error regarding the count of Medicare days in the denominator of the SSI fraction. The Administrator affirmed the Board’s 22 The CMS Administrator’s interpretation of the SSI statute and regulations, which are administered by SSA, not CMS, is not entitled to deference. Department of Treasury, 837 F.2d at 1167. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 42 of 145 35 incorrect decision that the SSI fraction should include Medicare “covered or utilized days even though the hospital may not have received payment” under Medicare Part A. PRRB Dec. at 37- 38, AR 167-68; Adm’r Dec. at 28, AR 29. This aspect of the Secretary’s decision should be reversed insofar it would include either: (i) days that are not paid by Medicare Part A because a beneficiary had exhausted Medicare Part A coverage for inpatient hospital services (hereinafter “Exhausted Benefit” days); (ii) days that are not paid by Medicare Part A because Medicare’s liability was secondary to some other payer’s (hereinafter “Medicare Secondary Payer” or “MSP” days), or (iii) days of a Medicare beneficiary who was enrolled in a Medicare health maintenance organization that was responsible for payment for the inpatient hospital services rendered (hereinafter “Medicare HMO” days). 1. Exhausted Benefit and MSP Days Should Not be Included As shown below, the Secretary’s definition of Medicare days in this case is invalid, and should be set aside for two reasons. First, it is inconsistent with the Medicare DSH statute. Second, the Secretary’s adoption of the “covered or utilized days” standard in this case is an unexplained departure, without notice and comment rulemaking, from the Secretary’s original and long-standing interpretation of the DSH statute and regulation to include only Medicare paid days in the SSI fraction. Indeed, it even conflicts with the Secretary’s discovery responses in this case. Moreover, even if “covered or utilized days” were the correct standard, the Secretary’s own program rules and guidance reflect that Exhausted Benefit days and MSP days are not considered to be covered or utilized under the Part A inpatient hospital services benefit, and cannot lawfully be considered so here. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 43 of 145 36 a. The Correct Standard is Medicare Paid Days During the years at issue, the SSI fraction included only Medicare paid days.23 Pl. Stmt. ¶¶ 218-19. This was consistent with the plain language of the statute, the Secretary’s original interpretation of that statute in the DSH regulation, 42 C.F.R. § 412.106, and the Secretary’s representations to the federal courts regarding the agency’s interpretation of the statute and regulation. To the extent that the Secretary’s decision would now include days not paid by Medicare Part A in the SSI fraction, his decision is invalid and should be set aside. First, the Secretary’s current view of Medicare “entitlement” is inconsistent with the plain language and design of the DSH statute and the Medicare Act as a whole. The DSH statute defines the SSI fraction as consisting of patient days attributable to individuals who were “entitled to benefits” under Medicare Part A. The “benefit” provided under Medicare Part A is the right to have payment made on one’s behalf for covered inpatient hospital services or other post-hospital skilled nursing facility or home health services.24 See 42 U.S.C. § 426(c)(1) Thus, when payment is not made under Part A for hospital services, either because the patient has exhausted Part A benefits25 or because Medicare’s payment liability is secondary to some other 23 It is clear that the Secretary never counted unpaid Medicare beneficiary days in the SSI fractions for periods prior to the 2004 amendments to the DSH regulation. To be sure, CMS said otherwise in a notice of proposed rulemaking published in 2003. 68 Fed. Reg. 27,154, 27,207 (May 19, 2003). But, CMS later admitted in a 2004 website posting that Exhausted Benefit days were never counted in the SSI fraction. See AR 4375. Consistent with the 1986 rules implementing the DSH statute, the 2004 website posting acknowledged that prior CMS policy excluded Exhausted Benefit days. Id. 24 By definition, a Medicare beneficiary is never entitled to have payment made for skilled nursing or home health services on a day that is counted as an inpatient hospital. 25 The Part A benefit for inpatient hospital services covers 90 days per spell of illness with a lifetime reserve of 60 days. 42 U.S.C. § 1395d(a)(1). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 44 of 145 37 payer,26 the beneficiary is not “entitled” to benefits under Part A with respect to those days, and those days should not be counted in the SSI fraction.27 The Secretary’s current spin on Medicare “entitlement” also flies in the face of his past representations to hospitals and the federal courts. When the Secretary adopted the DSH regulation in 1986, the Secretary stated that the Medicaid fraction includes only Medicaid paid days and the SSI fraction includes only Medicare paid days. 51 Fed. Reg. 31,454, 31,460-61 (Sept. 3, 1986). Thus, an inpatient day for an individual enrolled in both Medicaid and Medicare could only be counted as one type of day – a Medicaid day or a Medicare day. Id. at 31,460. In subsequent litigation challenging and ultimately overturning the Secretary’s policy of counting only Medicaid paid days in the Medicaid fraction,28 the Secretary defended that policy on the ground that it was consistent with his policy of counting only Medicare paid days in the SSI fraction. In all of those cases, CMS consistently represented to the federal circuit courts that the SSI fraction includes only Medicare paid days. The Ninth Circuit’s decision in Legacy Emanuel, for example, expressly noted the Secretary’s position that “the Medicare proxy only counts patient days paid by Medicare.” 97 F.3d at 1265 (emphasis added). Likewise, the 26 A Medicare beneficiary is not entitled to have Medicare Part A payment made on his or her behalf when another third-party payer is responsible for the inpatient hospital care. See 42 U.S.C. § 1395y(b)(2)(A). 27 As the Sixth Circuit held in Jewish Hospital, while the Medicaid fraction considers eligibility for Medicaid, regardless of whether the patient was entitled to Medicaid payment for inpatient hospital services, the SSI fraction “fixes the calculation upon the absolute right to receive an independent and readily defined payment.” 19 F.3d at 275 (emphasis in original). Similarly, as the Ninth Circuit concluded in Legacy Emanuel Hospital, “entitlement” to Medicare Part A “means ‘the absolute right to . . . payment.’” 97 F.3d at 1265 (quoting Jewish Hospital, 19 F.3d at 275). 28 Four federal circuits invalidated the Secretary’s policy of counting only Medicaid paid days in the Medicaid fraction because the Secretary had impermissibly conflated the statute’s use of the broader term “eligible” (for Medicaid) with its use of the narrower term “entitled” (to Medicare Part A benefits). See supra n.4. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 45 of 145 38 Secretary’s brief to the Fourth Circuit in Cabell Huntington asserted (at 23) that the Medicare fraction “counts only those ‘patient days’ that are actually paid by the Medicare program.” AR 4345, 4346. The Secretary’s brief also went on to explain (at 23) that this construction of the SSI fraction is consistent with the intent of Congress, which “concededly did not intend for the Secretary to count unpaid days of hospital care provided to Medicare patients.” Id. (underscoring in original).29 The Secretary cannot do an about-face now. The Secretary’s own discovery responses in this case admitted that Exhausted Benefit days should not be counted in the SSI fractions. AR 4610. Insofar as the Secretary’s final decision in this case would include all “covered or utilized days even though the hospital may not have received payment,” that standard is invalid and should be set aside. See Michigan Public Power Agency, 405 F.3d at 12-14 (an agency’s unexplained departure from its established precedent is arbitrary and capricious). See also Thomas Jefferson Univ., 512 U.S. at 512, 515 (no deference is owed to an agency’s interpretation of a regulation that is inconsistent with indications of the agency’s intent at the time of its adoption); National Treasury Employees Union, 399 F.3d at 337 (no deference is owed to an agency’s erratic and inconsistent interpretations). Moreover, once an agency has established its interpretation of a regulation, it cannot change that interpretation without invoking the notice and comment rulemaking procedure prescribed by the Administrative Procedure Act. Monmouth Med. Ctr. v. Thompson, 257 F.3d 807, 814 (D.C. Cir. 2001). 29 Likewise, CMS’ lead DSH analyst, Ann Rudolph (now Tayloe) testified in a 1995 deposition that the SSI fraction includes only Medicare “paid” days. AR 876-77. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 46 of 145 39 b. Exhausted Benefit and MSP Days Are Not Covered or Utilized Days. Even if “covered or utilized days” were the correct standard (and it is not), the Secretary’s own rules and program guidance state that Part A Exhausted Benefit days and MSP days are not covered or utilized days under the Medicare Part A benefit covering inpatient hospital services. The Secretary’s treatment of Exhausted Benefit days and MSP days for purposes of the SSI fraction is inconsistent with its policy on those days for other Medicare purposes, and cannot be sustained. For example, in the preamble to a 1990 Federal Register notice, the Secretary articulated that “[e]ntitlement to payment under part A ceases after the beneficiary has used 90 days in a benefit period and has either exhausted the lifetime reserve days or elected not to use available lifetime reserve days.” 55 Fed. Reg. 35,990, 35,996 (Sept. 4, 1990) (emphasis added).30 Further, the Secretary’s Medicare Claims Processing Manual clearly states in § 40 that “Days after benefits are exhausted are not charged against the beneficiary’s utilization even though the hospital may receive the full DRG payment.”31 Medicare Claims Proc. Man. (CMS Pub. 100-04), Ch. 3, § 40, available at http://www.cms.hhs.gov/manuals/downloads/clm104c03.pdf (emphasis added). Likewise, the Secretary’s Medicare Intermediary Manual states in § 3685.B that no utilization of benefits is charged to a Medicare beneficiary for days that are not paid for by Medicare Part A under the Medicare Secondary Payer provisions of the Medicare Act. See 30 Significantly, the Secretary articulated this policy when interpreting a statutory provision that uses Medicare “entitlement” language identical to the language used in the DSH statute. Specifically, the Medicare statute defines a Medicare-dependent hospital as one in which “not less than 60 percent of its inpatient days or discharges . . . were attributable to inpatients entitled to benefits under part A.” 55 Fed. Reg. at 35,996 (quoting 42 U.S.C. § 1395ww(d)(5)(G)(iii)(IV)) (emphasis added). 31 A hospital may receive the full Medicare payment for a hospital stay if a patient’s Medicare benefits are exhausted during the patient’s stay. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 47 of 145 40 Medicare Intermed. Man. (CMS Pub. 13-3) § 3685, AR 4950-56. (“If payment by the primary payer for Medicare covered services . . . equals or exceeds the hospital’s full charges for those services or the Medicare payment rate . . . no payment is due from Medicare and no utilization is charged to the beneficiary”) (emphasis added). The Medicare Secondary Payer Manual further explains: “Where no Medicare secondary payer benefit is payable, no utilization is charged the beneficiary.” Medicare Secondary Payer Man. (CMS Pub. 100-05), Ch. 1, § 40, available at http://www.cms.hhs.gov/manuals/downloads/msp105c01.pdf. Thus, to the extent that the Secretary’s decision in this case interprets “entitled to benefits under Part A”, as that term is used the DSH statute, to mean “covered or utilized days”, it is inconsistent with the Secretary’s longstanding understanding of Medicare “entitlement,” and is invalid. 2. Medicare HMO Days Should Not be Included. The utter inconsistency in and between Medicare policy and practice also dooms the Secretary’s position on Medicare HMO days in the DSH context. The Board’s decision correctly notes (at 40, AR 170) that, prior to 1998, Medicare HMO days were “recorded as non-Medicare days” for all purposes on the Medicare cost report that is used to calculate payments to a hospital for services rendered to Medicare Part A beneficiaries, and, consequently, that these days were not counted as Medicare days in the Medicare payment for graduate medical education (“GME”), “even though the statutory definitions of the two programs [GME and DSH] are very similar.” The Board also concluded, and the Administrator affirmed, however, that the Secretary had an established “policy” of counting HMO days as Medicare days for DSH purposes and that this policy was valid notwithstanding the unexplained departure from the Secretary’s policy of treating Medicare HMO days as non-Medicare days for every other Medicare payment purpose. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 48 of 145 41 This aspect of the Secretary’s decision below also is invalid and should be set aside. As shown below, the inclusion of Medicare HMO days in the SSI fraction conflicts with Congress’ manifest intent that Medicare HMO days not be regarded as entitled to benefits under Medicare Part A. Moreover, the decision in this case to include Medicare HMO days in the SSI fraction is arbitrary and capricious because the Secretary has not provided a rational explanation for treating these days as Medicare Part A days solely for purposes of the Medicare DSH calculation and not for any other Medicare payment purpose. Transactive Corp. v. United States, 91 F.3d 232, 237 (D.C. Cir. 1996). Probably as a result of this policy aberration, the Secretary never implemented his professed policy to include Medicare HMO days in the SSI fraction. a. Congress Did Not Intend To Include Medicare HMO Days. During the periods at issue, the Secretary did not count Medicare HMO days in the computation of the Medicare payment to a teaching hospital for graduate medical education (“GME”). The GME statute specified then, as it does now, that the GME payment, like the DSH payment, is based in part on a hospital’s number of patient days “attributable to patients with respect to whom payment may be made under part A.” See 42 U.S.C. § 1395ww(h)(3)(C) (defining the “Medicare patient load” fraction that is used to compute the GME payment to teaching hospitals); cf. 42 U.S.C. § 1395ww(d)(5)(F)(vi)(I) (defining the SSI fraction as consisting of days attributable to patients “entitled to benefits under part A”). In the preamble to the 1989 implementing rule, the Secretary construed the GME statute provision to exclude Medicare HMO patient days. The Secretary explained that Medicare HMO days “are recorded as non-Medicare days” for all Medicare payment purposes. See 54 Fed. Reg. 40,286, 40,294-95 (Sept. 29, 1989). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 49 of 145 42 In 1997, Congress, presumably aware of the Secretary’s existing policy of recording Medicare HMO days as non-Medicare days,32 enacted legislation providing for a separate, additional GME payment specifically for hospitals that treat Medicare HMO enrollees. Balanced Budget Act of 1997, Pub. L. No. 105-33 § 4625, 1997 U.S.C.C.A.N. (111 Stat.) 251, 479 (adding 42 U.S.C. § 1395ww(h)(3)(D)(i)). If Congress had intended for Medicare HMO patients to be counted as patients who are entitled to benefits under Medicare Part A, then Medicare HMO patient days would have needed to be included in the calculation of the standard GME payment under the pre-existing statutory and regulatory scheme. And, if that were the intent, then Congress would have simply directed the Secretary to count Medicare HMO days in the GME payment calculation specified under the pre-existing law. But, Congress did not do that. Instead, by enacting a provision for a separate GME payment for Medicare HMO patient days, Congress manifested its intent that Medicare HMO enrollees not be regarded as patients “with respect to whom payment may be made under part A.” Those patient days, therefore, should not be counted in the SSI fraction either. Like the GME statute, which fixes the GME payment on a hospital’s number of patient days attributable to patients who are entitled to payment under Medicare Part A, the DSH statute defines the SSI fraction to consist of a hospital’s number of days attributable to patients who are “entitled to benefits” under Medicare Part A. The Secretary’s decision in this case that Medicare HMO days should be included the SSI fraction unlawfully conflicts with the will of Congress. 32 See, e.g., United States v. Wilson, 290 F.3d 347, 357 (D.C. Cir. 2002) (stating that Congress is presumed to have legislated with awareness of the Executive Branch’s interpretation and implementation of existing law), cert. denied, 537 U.S. 1028 (2002); Orloski v. FEC, 795 F.2d 156, 166 (D.C. Cir. 1986) (stating that Congress’ failure to modify an agency’s interpretation of a law when it amended other related terms of the statute signifies Congress’ acquiescence in the agency’s interpretation). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 50 of 145 43 b. The Unexplained Inconsistency With the Treatment of HMO Days For All Other Medicare Payment Purposes is Arbitrary and Capricious. The Secretary’s decision regarding Medicare HMO days is also invalid, and should be reversed, because the Secretary has never provided a rational explanation for counting Medicare HMO days as Medicare days for DSH purposes but not for any other Medicare payment purpose, including GME payment. During the periods at issue, Medicare HMO days were not counted as Medicare days for any purpose on the Medicare cost reports that were used to determine Medicare payments to hospitals for services that are covered under Part A and paid by the intermediaries.33 See PRRB Dec. at 39-40, AR 169-70. Pl. Stmt. ¶ 215. As the PRRB put it, “CMS has never explained the reason for this inconsistent treatment of Medicare HMO days, see 54 Fed. Reg. 40,294-95; 55 Fed. Reg. 35,990, 35,994 (Sept. 4, 1990).” PRRB Dec. at 40, AR 170. This inconsistency is arbitrary and capricious. See Transactive Corp., 91 F.3d at 237; see also Cheshire Hosp. v. New Hampshire-Vermont Hospitalization Serv., Inc., 689 F.2d 1112, 1126 (1st Cir. 1982) (ruling that it is arbitrary and capricious for the Secretary to apply one rule in some cases and another rule in other cases without providing a rational explanation for the inconsistency). Moreover, the Secretary’s decision to include Medicare HMO days now is arbitrary and capricious because the Secretary never implemented a genuine policy to include these days in the SSI fraction. While the preamble to a 1990 rule said that the agency had been counting Medicare HMO days in the SSI fraction since 1987, 55 Fed. Reg. 35,990, 35,994 (Sept. 4, 1990), the 33 The Medicare Act authorized the Secretary to enter into contracts with HMOs and other risk- bearing entities under which the Secretary paid the HMOs capitation rates per enrollee in lieu of amounts that would otherwise be paid to hospitals under Medicare Part A through the Medicare fiscal intermediaries. See 42 U.S.C. §§ 1395mm(a)(1) and (a)(3); 42 C.F.R. §§ 417.524 and 417.584. Hospitals then received contractual payments from the HMOs. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 51 of 145 44 evidence in the record, including the Government’s own reports, shows that the agency never came close to capturing all, or even most, Medicare HMO days in the SSI calculation and it knew that was the case even as it told the hospitals otherwise in the preamble to the 1990 rule. Thus, despite what the agency said in 1990, the evidence reflects that it did not actually adopt or follow policy to count Medicare HMO days in the Medicare/SSI calculation. The evidence in this case shows that the Hospital’s Medicare HMO days were counted in the SSI fraction only when the Hospital had mistakenly billed a Medicare HMO patient stay to the fiscal intermediary for payment under the Part A fee-for-service program, and payment was denied.34 See PRRB Decision at 37, 39, AR 167-69. Moreover, two contemporaneous government reports and the testimony of a CMS witness show that as far back as 1987, and continuing thereafter, CMS knew that few HMO days were counted in the SSI fraction because hospitals and HMOs had little incentive and no contractual obligation to submit data to CMS for Medicare HMO patient stays. Pl. Stmt. ¶¶ 212-13. Indeed, CMS’ principal MEDPAR programmer, Dean, testified that the HMO field on the MEDPAR file had not been used since at least 1995, when he started working with MEDPAR (which period would have included the time when CMS computed the SSI fractions for at least two or three of the fiscal years at issue). Pl. Stmt. ¶ 214. Before the Secretary amended the DSH regulation in 1995, CMS instructions required a hospital seeking a recalculation of the SSI fraction for the hospital’s own cost reporting period – instead of the federal fiscal year in which that period began – to obtain the fiscal intermediary’s certification that the hospital’s submission to CMS of the hospital’s 34 Thus, while CMS’ calculations of the Medicare/SSI fractions reflect that the Hospital had 250,494 Medicare patient days for the four fiscal years at issue, Pl. Stmt. ¶ 119, CMS’ calculations of the Hospital’s SSI fractions included only 22 Medicare HMO patient stays accounting for only 359 days (slightly more than one one-thousandth of the total Medicare days). Pl. Stmt. ¶ 211. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 52 of 145 45 Medicare Part A days for its cost reporting period agreed with the fiscal intermediary’s report that identifies only Medicare-paid days and excludes all Medicare HMO days. See AR 4789-90, 5430. In summary, the Secretary’s decision to include days not paid by Medicare Part A and Medicare HMO days in the denominator of the SSI fraction is contrary to law and should be reversed. VI. CONCLUSION For these reasons, Plaintiff respectfully requests that this Court grant its motion for summary judgment as to Count I of the Complaint and enter the accompanying proposed order remanding to the agency with specific instructions to correct the deficiencies at issue and retaining jurisdiction to assure compliance with this Court’s order. Respectfully submitted, /s/ Christopher L. Keough Christopher L. Keough DC Bar No. 436567 John M. Faust DC Bar No. 433553 Stephanie A. Webster DC Bar No. 479524 VINSON & ELKINS L.L.P. 1455 Pennsylvania Avenue, N.W. Suite 600 Washington, D.C. 20004-1008 (202) 639-6500 (phone) (202) 639-6604 (fax) Counsel for Plaintiff February 1, 2007 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 53 of 145 UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA BAYSTATE MEDICAL CENTER, ) ) Plaintiff, ) ) v. ) ) C.A. No. 06-1263 (JDB) MICHAEL O. LEAVITT, et al., ) ) Defendants. ) PLAINTIFF’S STATEMENT OF MATERIAL FACTS AS TO WHICH THERE IS NO GENUINE DISPUTE Pursuant to Federal Rule of Civil Procedure 56 and Local Civil Rule 7(h), the Plaintiff hereby submits its statement of material facts as to which there is no genuine issue (“Plaintiff’s Statement”). To facilitate the Court's review of the decisions below, in light of Defendant Leavitt's failure even to acknowledge many factual findings of the Provider Reimbursement Review Board and the record evidence supporting those findings, and in consideration of the voluminous record, Plaintiff is submitting this detailed statement of material facts with detailed citations to the evidence in the Administrative Record underlying those findings. The statement is organized as follows: A. Parties (¶¶ 1-3) B. Calculation of the Medicare DSH Payment (¶¶ 4-12) C. The DSH Payment Determination at Issue (¶¶ 13-20) 1. Fiscal Year 1993 (¶¶ 13-14) 2. Fiscal Year 1994 (¶¶ 15-16) 3. Fiscal Year 1995 (¶¶ 17-18) 4. Fiscal Year 1996 (¶¶ 19-20) Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 54 of 145 2 D. CMS’ Use of a Systemically Flawed Match Process (¶¶ 21-122) 1. SSA’s Annual SSI Data Tapes (¶¶ 25-30) 2. CMS’ Inpatient Hospital Stay Records (MEDPAR File) (¶¶ 31-35) 3. CMS’ Data Match (¶¶ 36-122) a. CMS’ Early Cut-Off (¶¶ 38-47) b. CMS’ Failure to Use Social Security Numbers (¶¶ 48-52) c. SSA’s Identification Numbers (¶¶ 53-65) d. CMS’ Identification Number (HICAN) (¶¶ 66-74) e. Testimony of CMS’ Witness, Dean, Regarding the Identification Numbers CMS Uses in the Match Process (¶¶ 75-90) f. CMS’ Admission of A Systemic Flaw In the Match Process (¶¶ 91-96) g. Additional Flaws in the Match Process (¶¶ 97-115) h. Impact of Flaws in the Match Process (¶¶ 116-122) E. SSI Data Used (¶¶ 123-176) 1. Omission of Inactive Records (¶¶ 125-150) 2. Omission of Forced or Manual SSI Records (¶¶ 151-160) 3. Omission of Retroactive Determinations and Hold and Suspense Cases (¶¶ 161-173) 4. Omission of Non-Cash SSI Benefits (¶¶ 174-176) F. CMS’ Allegedly Acceptable Rate of Error (¶¶ 177-190) G. CMS’ Count of Medicare Days in MEDPAR (¶¶ 192-217) 1. Unexplained Discrepancies in CMS’ Count (¶¶ 192-203) 2. Days That Should Not Be Counted in the Denominator of the SSI Fraction (¶¶ 204-217) a. Exhausted Benefit Days (¶¶ 204-209) b. Medicare HMO Days (¶¶ 210-217) H. CMS’ Resistance to Discovery Efforts (¶¶ 218-256) Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 55 of 145 3 A. Parties 1. Plaintiff, Baystate Medical Center (“Baystate” or “Hospital”), is a Medicare participating, general acute care hospital located in Springfield, Massachusetts. Provider Reimbursement Review Board Decision 2006-D20 (“PRRB Dec.”) at 3, Administrative Record (“A.R.”) 133. 2. Defendant, Michael O. Leavitt (“the Secretary”), is the Secretary of the Department of Health and Human Services (“HHS”), the federal agency that administers the Medicare program and the disproportionate share hospital (“DSH”) payment at issue in this case. See Defendant’s Answer to Complaint (“Answer”) ¶¶ 15, 16. 3. The Centers for Medicare and Medicaid Services (“CMS”) is a component of HHS and is responsible for day-to-day operation and administration of the Medicare program. CMS was formerly known as the Health Care Financing Administration (“HCFA”) until the name of the agency was changed in July 2001. Answer ¶ 17; PRRB Dec. at 2, AR 132. B. Calculation of the Medicare DSH Payment 4. A hospital’s Medicaid fraction for DSH is determined retrospectively by the fiscal intermediary, 42 C.F.R. § 412.106(b)(4), in and through the Medicare cost report settlement process that is generally completed two to three years after the close of the hospital’s cost reporting period. PRRB Dec. at 32, AR 162; Gouger, Hearing Transcript (“Hearing Tr.”) at 1821-22, AR at 5504-05. 5. The Secretary does not allow a hospital to estimate the number of Medicaid patient days included in the Medicaid fraction in the DSH calculation for a hospital fiscal year. PRRB Dec. at 8, AR 138; Gouger, Hearing Tr. at 1816, AR 5503; see also 42 C.F.R. § 412.106(b)(4). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 56 of 145 4 6. The SSI fraction is computed solely by CMS, based upon SSI data that CMS receives annually from SSA and CMS’ match of that data against CMS’ own inpatient hospital utilization records. PRRB Dec. at 3, AR 133; Parties’ Stipulations at ¶¶ 2.8, 3.2, 3.3, AR 5024. 7. As was stipulated before the Board in this case, apart from the transmission of claims data for Medicare inpatient hospital stays, hospitals and fiscal intermediaries have no input into, and are not parties to, the Secretary’s calculation of the SSI fractions. Parties’ Stipulations at ¶¶ 3.2, 3.3, AR 5024; see 42 C.F.R. § 412.106(b)(2); 51 Fed. Reg. at 31459-60. 8. In general, CMS’ calculation of the SSI fraction comprises two basic data sources: (i) inpatient hospital stay records reflecting Medicare patient days in each hospital in the country (which forms the denominator of the SSI fraction); and (ii) SSI entitlement data provided to CMS by SSA. PRRB Dec. at 3, AR 133; 51 Fed. Reg. 31454, 31459 (Sept. 3, 1986) (final rule). 9. CMS attempts to “match” these two data sets annually in order to determine a hospital’s number of patient days in a year attributable to individuals who were entitled to both Medicare Part A benefits and federal SSI benefits during their inpatient hospital stays. PRRB Dec. at 3, AR 133. The more Medicare patient days that are matched with the SSI eligibility records, the larger the numerator of the SSI fraction and the greater the DSH payment to the hospital. Answer ¶ 50. 10. Since the inception of the DSH payment adjustment in 1986, the Secretary has computed the SSI fractions retrospectively for every Medicare-participating hospital based on discharges occurring in the Federal fiscal year ending on September 30. PRRB Dec. at 2 & n.1, 8, AR 132, 138. The SSI fraction computed by the Secretary for a Federal fiscal year applies to Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 57 of 145 5 a hospital fiscal year (or cost reporting period) that begins in that Federal fiscal year, unless the hospital later requests a recalculation of the SSI fraction based on discharges occurring in the hospital’s own fiscal year (if different than the Federal fiscal year, e.g., a fiscal year ending June 30). PRRB Dec. at 2, 12, AR 132, 142; 42 C.F.R. § 412.106(b)(2)-(3). 11. For the periods at issue, the Secretary calculated the SSI fractions that were applied retrospectively to determine Medicare DSH payments based on a match of (i) SSI data that CMS received from SSA in the following March, about six months after the September 30 year-end for subject Federal fiscal year, against (ii) Medicare inpatient hospital stay records accumulated in a CMS file by the following June, about nine months after the September 30 year-end for the subject Federal fiscal year. See infra ¶¶ 27, 38-40, 46. The CMS match for a given fiscal year thus occurs in June of the following year. Answer ¶ 52. 12. Medicare fiscal intermediaries are required to apply the SSI fraction computed by CMS when they determine a hospital’s DSH payment for a hospital fiscal year. See infra ¶ 48. C. The DSH Payment Determination at Issue 1. Fiscal Year 1993 13. The fiscal intermediary issued a DSH payment determination for the Hospital’s fiscal year ending September 30, 1993 in a notice of program reimbursement (“NPR”) dated October 2, 1996. AR 6445. This NPR for the Hospital’s fiscal year 1993 reflected an SSI fraction computed by CMS in June 1994. See Supra ¶ 11. 14. The parties stipulated below that the Secretary’s determination of the SSI fraction for the Hospital’s fiscal year 1993 included the following: Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 58 of 145 6 SSI Days Medicare Days SSI Ratio 4746 65780 0.07215 Parties’ Stipulations at ¶ 1.5, AR 5023. 2. Fiscal Year 1994 15. The fiscal intermediary issued a DSH payment determination for the Hospital’s fiscal year ending September 30, 1994 in an NPR dated March 11, 1997. AR 6448. This NPR for the Hospital’s fiscal year 1994 reflected an SSI fraction computed by CMS in June 1995. See supra ¶ 11. 16. The parties stipulated below that the Secretary’s determination of the SSI fraction for the Hospital’s fiscal year 1994 included the following: SSI Days Medicare Days SSI Ratio 5274 64251 0.08208 Parties’ Stipulations at ¶ 1.5, AR 5023. 3. Fiscal Year 1995 17. The fiscal intermediary issued a DSH payment determination for the Hospital’s fiscal year ending September 30, 1995 in an NPR dated March 23, 1998. AR 6453. This NPR for the Hospital’s fiscal year 1995 reflected an SSI fraction computed by CMS in June 1996. See supra ¶ 11. 18. The parties stipulated below that the Secretary’s determination of the SSI fraction for the Hospital’s fiscal year 1995 included the following: SSI Days Medicare Days SSI Ratio 5569 63024 0.08836 Parties’ Stipulations at ¶ 1.5, AR 5023. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 59 of 145 7 4. Fiscal Year 1996 19. The fiscal intermediary issued a DSH payment determination for the Hospital’s fiscal year ending September 30, 1996 in an NPR dated September 29, 1998. AR 6472. This NPR for the Hospital’s fiscal year 1996 reflected an SSI fraction computed by CMS in June 1997. Supra ¶ 11. 20. The parties stipulated below that the Secretary’s determination of the SSI fraction for the Hospital’s fiscal year 1996 included the following: SSI Days Medicare Days SSI Ratio 5324 57439 0.09269 Parties’ Stipulations at ¶ 1.5, AR 5023. D. CMS’ Use of a Systemically Flawed Match Process 21. In its decision below, the PRRB found that CMS’ match process is flawed in ways that understate the numerator of the resulting SSI fractions, that additional or alternative match criteria, such as patients’ own unique Social Security numbers, must be used to correct the flaws in CMS’ process, and that using such additional criteria would impose no significant administrative burden on the agency. PRRB Dec. at 10-21, AR 140-151. 22. The Administrator’s decision does not dispute the Board’s finding that CMS’ match process failed to identify some Medicare inpatients of the Hospital as federal SSI recipients for the periods at issue and continues to do so. See generally Adm’r Dec., AR 2-55. 23. The Administrator’s decision asserts that the Hospital failed to quantify the financial impact of the flaws in CMS’ match process. Adm’r Dec. at 36, AR 37. It is undisputed, however, that any such quantification would require access to records that only CMS Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 60 of 145 8 and SSA possess, Answer ¶ 105, and that CMS resisted efforts by the Hospital and the Board in this case to obtain access to the information contained in those records. See PRRB Dec. at 25 & n.135, AR at 155; Adm’r Dec. at 36, AR 37. 24. The Administrator concluded that CMS achieved an “acceptable rate of error” in calculations of the Hospital’s SSI fractions for the years at issue. Adm’r Dec. at 25-26, 42-43, AR 26-27, 43-44. The Administrator’s decision, however, does not address what “rate of error,” if any, is “acceptable” for Medicare payment purposes generally or for DSH payment purposes in particular or what overall error rate the Administrator believes CMS achieved in its calculations of the Hospital’s SSI fractions for the years at issue. Adm’r Dec., AR 2-55. 1. SSA’s Annual SSI Data Tapes 25. The Board made detailed findings of fact concerning the data elements included in the SSI data tapes that the Secretary uses to compute the SSI fractions for DSH. PRRB Dec. at 10-11, AR 140-141. 26. The Board’s findings concerning the data elements in SSA’s annual SSI tapes are based in part on the testimony of Pat Cribbs. PRRB Dec. at 10-12, AR 140-142. Cribbs worked for SSA for 40 years, including 24 years in the SSI department. See Cribbs, Hearing Tr. at 110- 14, AR 5054-55; PRRB Dec. at 10, AR 140; Answer ¶ 108. She was an SSA team leader for SSI database analysis and, in 1988 or 1989, took over responsibility for preparing the annual SSI tapes that CMS used to calculate the SSI fractions for DSH. See Cribbs, Hearing Tr. at 127, 131, AR 5058, 5059; PRRB Dec. at 10, AR 140; Answer ¶ 108. CMS did not controvert Ms. Cribbs’ testimony before the Board. The Administrator did not controvert her testimony in his decision. See generally Adm’r Dec., A.R.2-55. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 61 of 145 9 27. Based on the record evidence, the Board made the following specific findings regarding the SSI eligibility data: (a) “SSA makes SSI eligibility determinations and maintains the SSI eligibility files.” PRRB Dec. at 10, AR 140; see Cribbs, Evidentiary Hearing Tr. at 262-66, AR 5679-80. (b) “Before April of each year, generally in March, SSA prepares and sends CMS a cartridge or tape containing SSI eligibility information covering a 42-month period.” PRRB Dec. at 10, AR 140; see Cribbs, Evidentiary Hearing Tr. at 176- 77, 124-25, AR 5070-71, 5057-58. (c) “The 42-month period covers the 36 months in the three prior calendar years and the first six months of the calendar year in which the tape is prepared. Thus, the data included in the tape projects forward three months through the end of June in the calendar year in which the tape is created.” PRRB Dec. at 10, AR 140; see Cribbs April 29, 2003 Evidentiary Hearing Tr. at 284-85, AR 5097-98. (d) “The SSA file contains select information prepared solely for CMS’ use; SSA does not use the data file for any purpose.” PRRB Dec. at 10, AR 140; see Cribbs, Hearing Tr. at 126, AR 5058. 28. Regarding the data elements on the special-purpose SSI tapes furnished to CMS by SSA, the Board found that the tapes for the periods at issue include the following data elements for each SSI recipient who was included on each tape: (a) truncated last name and first initial; (b) Social Security number, which is referred to as the personal account number, or “PAN;” (c) date of birth; (d) gender; (e) the Social Security or railroad retirement program identification number (called a “Title II number” or “CAN”), if the SSI recipient received monthly social security or railroad retirement benefits; and, (f) 42 monthly indicators (ones and zeros) denoting the payment or non-payment of Federal SSI cash benefits during the period covered by the SSA tape. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 62 of 145 10 PRRB Dec. at 10-11, AR 140-141; see also Cribbs, Hearing Tr. at 143-46, AR 5062-63; Cribbs, Evidentiary Hearing Tr. at 265-71, 344-51, AR 5736-37, 5680-81; Dean, Evidentiary Hearing Tr. at 112-14, AR 1743. 29. Cribbs testified that she spent about one hour per year on the annual SSI tapes for CMS, which were produced through computer runs performed in about two hours over a weekend. PRRB Hearing Tr. at 125-26, AR 5058 (“Since it never changes, all you do is just plunk the program in when the time comes. You change dates. That’s it.”); see also PRRB Hearing Tr. at 175-76, AR 5070; Cribbs, Hearing Tr. at 125-26, 175-76, AR 5645, 5657. Considering the little time required for SSA to produce the special-purpose SSI tapes, it would not have been burdensome for SSA to run and furnish CMS with updated SSI tapes at least once every quarter instead of just once a year. See Answer ¶ 111. 30. The CMS Administrator’s decision does not controvert the Board findings, or Cribbs’ testimony, described in paragraphs 25-29 above. See, e.g., Adm’r Dec. at 15, AR 16. 2. CMS’ Inpatient Hospital Stay Records (MEDPAR File) 31. The Board also made detailed findings of fact concerning the data elements included in the CMS files, called “MEDPAR” files, containing the inpatient hospital stay records for Medicare beneficiaries that CMS uses to compute the SSI fractions for DSH. PRRB Dec. at 11-12, AR 141-42. 32. The Board’s findings concerning the inpatient hospital stay records in the MEDPAR files are based on stipulations entered into below and uncontroverted record evidence, including the testimony of CMS’ witness, William Anthony Dean, who “had been CMS’ principal MEDPAR programmer since 1995 and was responsible for producing the MEDPAR Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 63 of 145 11 and maintaining the database.” PRRB Dec. at 11, AR 141; see Dean, Evidentiary Hearing Tr. at 53-55, AR 5722. 33. Based on the stipulations of the parties below, the Board found that CMS’ MEDPAR files contain the following data fields for each inpatient hospital stay that is included: (a) the hospital’s Medicare provider number; (b) the patient’s health insurance claim account number (“HIC” or “HICAN”); (c) the dates of admission to, and discharge from, the hospital; (d) the total length of the inpatient hospital stay; (e) the number of days in the stay that were covered under Medicare Part A; and (f) the number of days in the stay for which the patient was determined, through the match process described below, to be entitled to Federal SSI benefits. PRRB Dec. at 11-12, AR 141-42; see also Parties’ Stipulations at ¶ 6.1, AR 5024; Dean, Evidentiary Hearing Tr. at 52, AR 1775. 34. Regarding the HIC numbers (or HICANs) that appear in CMS’ MEDPAR files, the Board found: “The HIC number assigned to each Medicare beneficiary is either a Social Security number or railroad retirement number (but not necessarily the individual’s own number) followed by an alpha-numeric beneficiary identification code.” PRRB Dec. at 12 n.39, AR 142; see CMS On-Line Manual, Medicare Pub. 100-1, Ch. 2, §§ 50.2 – 50.4, AR 3696-99. 35. The CMS Administrator’s decision does not controvert any of the Board findings described in paragraphs 31-34 above. See Adm’r Dec. at 16, AR 146. 3. CMS’ Data Match 36. The Board made detailed findings concerning the processes that CMS and SSA use to match data records, not only for purposes of CMS’ calculations of the SSI fractions for Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 64 of 145 12 DSH, but also for purposes of SSA’s administration and payment of SSI cash assistance. PRRB Dec. 12-13, AR 142-43. 37. The Board’s findings concerning the match processes are based on uncontroverted record evidence, principally the testimony of Cribbs, the long-time SSA employee who prepared SSA’s annual SSI tapes for CMS, and of two CMS witnesses: Dean, the CMS MEDPAR programmer, and Daryl Rosenberg, another CMS employee who also was involved in CMS’ calculations of the SSI fractions for DSH. PRRB Dec. at 12-13, AR 142-43; see generally Cribbs, Hearing Tr. 109-284, AR 5054-97; Cribbs, Evidentiary Hearing Tr. 260-377, AR 5678- 5708; Dean, Hearing Tr. 1334-1453, AR 5383-5412; Dean, Evidentiary Hearing Tr. 51-230, AR 5710-66; Rosenberg, Hearing Tr. 2038-2154, AR 5559-88; Rosenberg, Evidentiary Hearing Tr. 246-314, AR 5770-87. a. CMS’ Early Cut-Off 38. Based on Dean’s testimony, the Board found that CMS creates multiple MEDPAR files for every federal fiscal year: “Each quarter, CMS creates several separate MEDPAR files that collectively cover multiple calendar and fiscal years.” PRRB Dec. at 12, AR 142; see Dean, Evidentiary Hearing Tr. at 64-71, A.R.1758-60; AR 1553. The Board further found that CMS creates multiple MEDPAR files for every Federal fiscal year over a two year period after the end of the subject Federal fiscal year: Over [a period of] two and a quarter years, CMS performs 10 MEDPAR runs and creates 10 MEDPAR files for each Federal fiscal year, the latest of which is run in the third December after the end of the subject fiscal year. For example, the 10th and last MEDPAR file for Federal fiscal year ended September 30, 1994 was produced in a December 1996 MEDPAR run. PRRB Dec. at 12, AR 142; see Dean, Evidentiary Hearing Tr. at 64-71, AR 1758; AR 1555. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 65 of 145 13 39. The Board found, again based on Dean’s uncontroverted testimony, that: “CMS matches its inpatient hospital stay records against the most recent annual SSA tape every time it creates one of these quarterly MEDPAR runs.” PRRB Dec. at 12, AR 142; see Dean, Evidentiary Hearing Tr. at 67-69, 87, AR 1759-60, 1764. 40. The Board found, based on undisputed evidence, that, (i) although CMS creates 10 separate MEDPAR files for every Federal fiscal year and matches the inpatient hospital stay records in each file to the latest SSI data that CMS has received from SSA, (ii) although the latest of these MEDPAR files for a given Federal fiscal year is created more than two years after the end of the Federal fiscal year, and (iii) although final DSH payment determinations typically are not made until 2-3 years after the end of a hospital fiscal year, CMS nonetheless uses one of the earliest of the 10 MEDPAR files to compute the SSI fractions that are applied to determine hospitals’ DSH payment: Despite having MEDPAR runs that are updated to the time of the final cost report settlement [which is usually 2-3 years after the end of hospital’s fiscal year], it is undisputed that CMS used the MEDPAR file based on the month of June following the end of the Federal fiscal year to compute the SSI fractions at issue for each year under appeal. For example, CMS used the June 1996 MEDPAR run to calculate the SSI fraction for Federal fiscal year ended September 30, 1995. PRRB Dec. at 32, AR 162; see CMS Discovery Response to Provider Interrogatory 15, AR 3631-32. 41. The Board further found, based on uncontroverted evidence, that the SSI fractions that CMS applies to determine hospitals’ DSH payments for a given year do not reflect SSI benefits that may be granted or reinstated after March of a given calendar year, retroactively for a period within the preceding Federal fiscal year. PRRB Dec. at 32, AR 162. This has the effect of reducing the number of Medicare/SSI patient days counted in the numerator of the SSI Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 66 of 145 14 fractions, thereby understating the SSI fraction and the resulting DSH adjustment for that period. Shafer, Evidentiary Hearing Tr. at 376-78, AR 5120-21. 42. Regarding CMS’ use of June MEDPAR files, the Board found that CMS’ own policy staff was not aware of what exactly the agency was doing, noting that CMS’ own witness, “Ms. [Anne] Tayloe [formerly Anne Rudolph], the individual responsible for [CMS’] DSH policy during the periods in issue, apparently believed that the latest available data was being used to compute the Medicare fraction.” PRRB Dec. at 33, AR 163; see Tayloe, Hearing Tr. at 2005-06, AR 5551. CMS conceded in written discovery responses below that the agency has no recollection or documentation concerning any factor or factors ever considered by the agency in deciding to use the June matches to determine the DSH payments. See CMS Discovery Response to Provider Interrogatory 16, AR 3632. 43. Moreover, according to another CMS staffer, Daryl Rosenberg, it would have been “simple” for the Secretary to have distributed the results of CMS’ later data matches for each fiscal year to the Secretary’s fiscal intermediaries and hospitals for use in the later determinations of the DSH payments due for each year. Hearing Tr. at 2056, 2073-74, AR 5563, 5567-68. 44. CMS’ written discovery responses and the testimony of another CMS employee, Steve Phillips, reflect the agency’s concession that the latest data available for a year is the best, most accurate data for that year. CMS Responses to Provider Interrogatories 18, 19 & 20, AR 3632-34; see Hearing Tr. at 1163-67, AR 5340-41. 45. The Secretary did not disclose CMS’ later data matches for the fiscal years at issue to the Medicare intermediaries or hospitals, and its latest calculations for those years were Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 67 of 145 15 not applied to determine the Hospital’s DSH payments for the periods at issue. See supra ¶¶ 13, 15, 17, 19. Even though the final DSH payment determination for each period at issue was issued after the Secretary had performed the latest data match for each year at issue, the Hospital’s NPR reflected the results of an SSI fraction calculation performed with much earlier, and concededly less accurate, data. Hearing Tr. at 1819-25, AR 5504-05; CMS Response to Interrogatories, AR 3631-33; see PRRB Dec. at 33, AR 163. 46. Concerning CMS’ publication of SSI fractions, the Board found: “Following its computation of the SSI fractions for a particular Federal fiscal year, generally in July or August, CMS transmits the results of the computations for each hospital (i.e., the hospital’s total number of Medicare Part A covered days, its total number of Medicare/SSI days, and the resulting percentage) to the Medicare fiscal intermediaries.” PRRB Dec. at 12, AR 142; see Rosenberg, Evidentiary Hearing Tr. at 256-60, AR 5772-73; SSI Ratios Published by CMS, AR 1547-50. 47. The Secretary’s audit and payment agents, the fiscal intermediaries, are bound to apply CMS’ calculation of the SSI fraction for a hospital: “Each intermediary is required by regulation to apply the SSI fraction computed by CMS in determining a hospital’s entitlement to the DSH payment.” PRRB Dec. at 12, AR 142 (referring to 42 C.F.R. §§ 412.106(b)(2) and (b)(5)); see also 51 Fed. Reg. 16772, 16777 (May 6, 1986) (interim final rule); 51 Fed. Reg. 31454, 31457-61 (Sept. 3, 1986) (final rule); Nancy Edwards Letter, AR 1017-18. b. CMS’ Failure to Use Social Security Numbers 48. Regarding the mechanics of CMS’ match process, the Board found: “The fundamental requirement in determining how many SSI recipients are among a hospital’s inpatient population is that the patient identifier in the SSA file must match the patient identifier Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 68 of 145 16 in the MEDPAR file.” PRRB Dec. at 12, AR 142; see Dean, Hearing Tr. at 116-18, 121-22, AR 5737-39; Declaration of Jack Gallagher (SSA), AR 1512-17. 49. The Board further found that “CMS said in the preamble to the 1986 rule [implementing the DSH statute] that it would use the Social Security number in the match[,] . . . stating ‘we believe that matching Social Security numbers on a Federal fiscal year basis is the most feasible approach.’” PRRB Dec. at 12 (quoting 51 Fed. Reg. 16772, 16777 (May 6, 1986)), AR 142; see also AR 4597. CMS had access to Social Security numbers, which appear on SSA’s annual SSI tapes and in CMS’ own Medicare beneficiary records. Id. Before the Board, the parties did not dispute that “the Social Security number is the best unique identifier.” PRRB Dec. at 12, AR 142; Dean, Hearing Tr. at 1393-95, AR 5397-98. 50. The Board found that “despite CMS’ stated intent to use Social Security numbers as the patient match criteria, the evidence showed that CMS has never matched Social Security numbers on the SSI data tape with Social Security numbers on the MEDPAR in the calculation of the SSI ratios.” PRRB Dec. at 13, AR 143; see Dean, Evidentiary Hearing Tr. at 106, 113-14, 121-22, AR 1769, 1771, 1773. 51. The Board further found that, because the SSI fraction match process does not use Social Security Number as the patient identifier, this match process is different from the match process employed by SSA and the Secretary for a different government program: In a similar program, SSA and CMS use Social Security numbers (or PANs), as well as alternative identifiers (name and gender), to match SSA records with CMS’ data in connection with SSA’s monitoring of SSI recipients’ residence in nursing homes. When SSA set up the nursing home match process with CMS, SSA and CMS worked to identify the functional requirements for the match, and the agencies tested and validated the match program. In contrast, the only criteria established for the matching process of SSI data files with MEDPAR data for hospitals were memorialized in two letters. Cribbs’ understanding was that a Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 69 of 145 17 programmer called her predecessor, who no longer works for SSA, and they decided over the phone what criteria to use. The program was already written when Cribbs took over and nothing had changed. PRRB Dec. at 13, AR 143; see Cribbs Hearing Tr. at 117-23, AR 5056-57. Significantly, a failure to properly match records in connection with the nursing home match process would produce greater government SSI payments to individuals than they are entitled to receive when residing in an institution for care that is paid by Medicaid, whereas in the DSH context, a failure to match would produce a lower SSI fraction and thus a lower DSH payment to hospitals. PRRB Dec. at 20-21, AR 150-51. 52. The CMS Administrator’s decision does not controvert any of the Board findings regarding the match processes employed by CMS as described in paragraphs 6, 9, and 25-51 above. See, e.g., Adm’r Dec. at 15-17, AR 16-17. c. SSA’s Identification Numbers 53. The Board made detailed findings of fact concerning the different types of identification numbers that are used by SSA in its administration of the Social Security retirement program (established under title II of the Social Security Act) and in its administration of the SSI program (established under title XVI of the Act) as well as the identification number that is used by CMS in its administration of the Medicare program (established under title XVIII of the Act). PRRB Dec. at 13-15, AR 143-45. The Board also made detailed findings concerning the identification numbers that the Secretary used, or now alleges to have used, in calculating the SSI fractions for DSH. PRRB Dec. at 13-15, 16-20, AR 143-45, 146-50. 54. The Board’s findings concerning the various identification numbers mentioned in paragraph 53 above are based principally on the testimony of Cribbs, the long-time SSA employee who ran the programs that produced SSA’s special-purpose SSI tapes for CMS, Alan Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 70 of 145 18 Shafer, another former, 34-year SSA employee who has appeared numerous times on behalf of SSA as an expert on management of information systems and data and who served as Associate Commissioner for Management’s representative to the SSI systems work group that set up the SSI system, and Dean, CMS’ own MEDPAR programmer. PRRB Dec. at 10-11, 13-14, 16-18, 20-21, 26-28, AR 140-41, 143-44, 146-48, 150-51, 156-68; see Cribbs, Hearing Tr. 109-284, AR 5054-97; Cribbs, Evidentiary Hearing Tr. 260-377, AR 5678-5708; Shafer, Hearing Tr. 286- 88, AR 5098; Dean, Hearing Tr. 1334-1453, AR 5383-5412; Dean, Evidentiary Hearing Tr. 51- 230, AR 5710-66. 55. The Board found: “SSA uses claim account numbers, called Title II numbers or CANs, to track individuals’ benefits under the Social Security Retirement program established under Title II of the Social Security Act.” PRRB Dec. at 13, AR 143; see Shafer, Hearing Tr. at 292-93, AR 5099-5100. 56. The Board found that SSA’s Title II number “consists of two elements: the Social Security number of the Title II record holder (the person whose work history qualifies for benefits) and an alpha-numeric suffix, called a beneficiary identification code or ‘BIC,’ that denotes the relationship between the beneficiary and the Title II record holder.” PRRB Dec. at 13, AR 143; see Shafer, Hearing Tr. at 293-94, AR 5100. 57. To illustrate the relationship between a Title II number and an individual’s Social Security number, the Board wrote: [A]n individual, ‘A,’ may receive Title II benefits only through his or her spouse, ‘B,’ who has sufficient work history to be entitled to benefits on his or her own account. In this circumstance, ‘A’ would be assigned a Title II number that includes ‘B’s’ nine-digit Social Security number with an alpha-numeric beneficiary identification code (such as B2) at the end. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 71 of 145 19 PRRB Dec. at 13, AR 143; see Shafer, Hearing Tr. at 293-94, 392-93, 400-01, AR 5100, 5124- 25, 5126-27. 58. Contrasting SSA’s Title II number with an individual’s own unique Social Security number, the Board found, based on the uncontroverted record evidence, that the Title II CAN is not an individual-unique identification number: Unlike the Social Security number, a Title II number (or CAN) is not an individual unique identifier. An individual may have more than one Title II number at the same time because, as CMS acknowledges, an individual may receive Title II benefits simultaneously under or on account of more than one Title II number holder’s account. In addition, an individual’s Title II number may change over time, for example, with marriage, divorce, death of a spouse or changes in work status. PRRB Dec. at 13, AR 143; see CMS Discovery Response to Provider Interrogatory 29, AR 3640; Shafer, Hearing Tr. at 298-99, 392-96, AR 5101, 5124-25; Cribbs Hearing Tr. at 150- 51,156, 158-60, AR 5064-66. 59. SSA’s annual tapes do not include a Title II number for every individual who was entitled to SSI benefits. Answer ¶ 139. This occurs because an “individual who receives SSI but does not receive Title II Social Security benefits, for example, may not have a Title II number displayed on his or her SSI record, and a Title II number for that individual would not appear on SSA’s annual tape.” PRRB Dec. at 13-14, AR 143-44; see Cribbs, Hearing Tr. at 173, AR 5070; Dean, Evidentiary Hearing Tr. at 106, 113-16, AR 1769, 1771; Cribbs, Evidentiary Hearing Tr. at 267-68, 346-47, AR 1713, 1733. 60. The Administrator’s decision does not controvert any of the Board’s findings regarding SSA’s Title II numbers, or CANs, as set forth in paragraphs 53-59 above. See, e.g., Adm’r Dec. at 15-17, AR 16-18. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 72 of 145 20 61. The Administrator’s decision alleges that CMS’ employee, Dean, runs a program that generates assumed Title II numbers based on the Social Security numbers reflected in SSA’s annual SSI tapes for purposes of calculating the SSI fractions for DSH. Adm’r Dec. at 42, AR 43. That assertion is based solely on Dean’s testimony before the Board in the 2004 hearing in this case, and, as the Board observed, was squarely contradicted by his earlier testimony in a 2003 evidentiary hearing conducted by the Board in this same case. PRRB Dec. at 16-18, A.R.146-48; see infra ¶¶ 75-90. Dean also testified, however, that it is impossible for him or CMS at this point to know how CMS’ programs worked before 1995, when Dean assumed his current position, because CMS kept no records documenting how its programs worked. PRRB Dec. at 16-18, AR 146-48; Dean Hearing Tr. at 1349, AR 5387. Dean’s admission that it is impossible for CMS to know now how its programs ran before 1995 is not acknowledged in the Administrator’s decision, and it contradicts the Administrator’s assertion that CMS ran a program to generate assumed Title II numbers for purposes of calculating the Hospital’s SSI fractions for the 1993-96 fiscal years at issue. Adm’r Dec. at 16-17, AR 17-18. 62. SSA’s annual tape includes a field for only one Title II number. Answer ¶ 142. SSA’s annual tapes never include more than one Title II number for individuals who receive Title II Social Security benefits under more than one Title II number, or for individuals whose Title II number changes over the course of the 42-month date range covered by an annual tape. Answer ¶ 142. 63. The Board credited Cribbs’ uncontroverted testimony that “in preparing SSA’s annual tape, SSA’s program takes the Title II number from the ‘unearned income’ or ‘UMI’ field on the SSI master record.” PRRB Dec. at 14, AR 144; Cribbs, Hearing Tr. at 148, 227, AR Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 73 of 145 21 5063, 5083. The Board further found, again based on Cribbs’ uncontroverted testimony: “The SSA program takes the first Title II number that appears within the unearned income field within the 42-month date range covered by the tape, and it is ‘more than likely’ that there will be more than one Title II number appearing in the unearned income field within the 42-month date range.” PRRB Dec. at 14, AR 144; Cribbs, Hearing Tr. at 149-50, AR 5064. The Board also found that “because the Title II numbers are usually sorted from oldest to newest in the unearned income field, the first number to appear in that field, and the number that would be put on SSA’s annual tape, ordinarily would be the oldest Title II number within the 42-month date range.” PRRB Dec. at 14, AR 144; Cribbs, Hearing Tr. at 150-51, AR 5064. 64. Quoting Cribbs’ uncontroverted testimony, the Board further found: [O]n the [SSI record] they have what they call unearned income as they get Title II Social Security benefits and that’s where I have to look – I look at a certain part and I have to select to see if they happen to have a type “A” [Social Security benefits] within the 42-month period and whatever their CAN [i.e., Title II number] is I just pick it up because they only have room for one so the first one wins. PRRB Dec. at 14, AR 144; Cribbs, Evidentiary Hearing Tr. at 346-47, AR 5700 (emphasis added); see Answer ¶ 144. 65. The Administrator’s decision does not controvert any of the Board’s findings regarding the Title II numbers that appear on SSA’s special-purpose SSI tapes or the process by which SSA selects the Title II numbers that appear on those tapes, as set forth in paragraphs 62- 64 above. See Adm’r Dec. at 15-17, AR 16-18. d. CMS’ Identification Number (HICAN) 66. The Board found that the Health Insurance Claim Account Number, sometimes referred to as a “HIC number” or “HICAN,” for each Medicare beneficiary “is a Social Security Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 74 of 145 22 number or railroad retirement number (but, in either case, not necessarily the individual’s own number) followed by an alpha-numeric beneficiary identification code.” PRRB Dec. at 14, AR 144; see CMS On-Line Manual, Medicare Pub. 100-1, Ch. 2, §§ 50.2–50.4, AR 3696-99. 67. Regarding the relationship and differences between CMS’ HIC numbers and SSA’s Title II numbers, the Board found: “Often, an individual’s HIC number is the same as his or her Title II number or CAN, but while an individual can have more than one Title II number simultaneously, an individual cannot receive Medicare Part A benefits under more than one HIC number at one time.” PRRB Dec. at 14, AR 144; see Shafer, Hearing Tr. at 296-98, 300-01, 303, AR 5100-02. 68. The Board found that “because the beneficiary identification codes are ‘equated’ in CMS’ match process, the equated HIC number that is used in that process is not an individual unique identifier.” PRRB Dec. at 14-15, AR 144-45; see Dean, Evidentiary Hearing Tr. at 121- 22, AR 5739; Shafer, Hearing Tr. at 529, AR 5178. In other words, after the HIC numbers are “equated” by the programs CMS employs in its match process, two or more different Medicare beneficiaries could be shown as having the same HIC number. 69. Regarding the components of CMS’ HIC number, the Board found: “The first part of a HIC number, the Social Security or railroad retirement number, may or may not be the Medicare beneficiary’s own Social Security or railroad retirement number.” PRRB Dec. at 15, AR 145; see CMS On-Line Manual, Medicare Pub. 100-1, Ch. 2 § 50.2, AR 3696. 70. To illustrate the function and use of an individual’s Social Security number in the HICAN, the Board gave the following example: [A] married woman (“W”) who is over the age of 65 may receive Medicare Part A benefits on the account of her husband (“H”) if H has sufficient quarters of Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 75 of 145 23 work to qualify for Social Security benefits. In this example, W’s HIC number would include H’s nine-digit Social Security number followed by an alpha- numeric beneficiary identification code indicating W’s spousal relationship with H, on whose account W may receive Social Security benefits (if any) under Title II of the Act. PRRB Dec. at 15, AR 145; see generally Shafer, Hearing Tr. at 527-30, AR 5178. 71. The Administrator’s decision does not controvert the Board findings regarding CMS’ HIC numbers that are used in the match process CMS employs to calculate the SSI fractions for DSH, except as noted in paragraphs 72 and 74 below. See, e.g., Adm’r Dec. at 42- 43, AR 43-44. 72. Quoting from his own notice published in the Federal Register in 2005, after the Board conducted its hearing in this case, the Administrator’s decision contradicts the Board’s findings by claiming that SSA’s “social security numbers are used on a ‘wage earner’ basis that is not necessarily specific to an individual Medicare beneficiary (or hospital patient).” Adm’r Dec. at 42; AR 43; see 70 Fed. Reg. 47278, 47440 (August, 12, 2005). 73. Nothing in the Administrator’s decision in this case, nothing in the 2005 Federal Register notice that is quoted in his decision, and nothing in the record in this case explains what is meant by the Administrator’s use of the term “‘wage earner’ basis” or the factual basis for the allegation that SSA’s Social Security numbers are “not necessarily specific to an individual.” 74. Again quoting from his own notice published in the Federal Register in 2005, the Administrator’s decision contradicts the Board’s findings by claiming that “HICAN are [sic] unique to each beneficiary.” Adm’r Dec. at 42-43, AR 43-44 (quoting 70 Fed. Reg. at 47440- 41). Nothing in the Administrator’s decision, nothing in the 2005 Federal Register notice, and nothing in this case explains or identifies any factual basis for that allegation. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 76 of 145 24 e. Testimony of CMS’ Witness, Dean, Regarding the Identification Numbers CMS Uses in the Match Process 75. The Board found that CMS’ principal MEDPAR programmer, Dean, “became involved in the MEDPAR process in 1995 when he inherited the programs used to produce the MEDPAR. He knew little about his predecessors’ work on the program.” PRRB Dec. at 16, AR 146; see Dean Evidentiary Hearing Tr. at 108-111, AR 5735-36. 76. The Board found that “Dean testified at the 2003 Evidentiary Hearing that CMS runs a program, called SSISORT, that reformats the SSI data received from SSA, drops all SSI records that do not include a Title II number, a term that he used ‘loosely’ to refer to entitlement to Medicare, and he puts the remaining SSI records in a new output file.” PRRB Dec. at 16, AR 146; see Dean, Evidentiary Hearing Tr. at 106, 113-16, AR 5735-37. The Board further found that “[a]t the 2003 Evidentiary Hearing, Dean testified that CMS’ programs have used just one identifier to match the inpatient hospital stay records in the MEDPAR file with the SSI entitlement information derived from tapes obtained from SSA.” PRRB Dec. at 16, AR 146; see Dean, Evidentiary Hearing Tr. at 106, 113-14, 121-22, AR 5735, 5737, 5739. 77. Regarding Dean’s 2003 testimony concerning the use of just one identifier in the match process CMS used to calculate the SSI fractions for DSH, the Board quoted the following testimony of Dean at the 2003 hearing: Q. What are the criteria or identifiers on which the match is run? A. The health insurance claim account number. The match program will read a stay record and look at the HICAN, health insurance claim account number, and then will go to the – my version of the SSI file and search for the HICAN. And when there is a match, that is when the SSI calculation is performed. PRRB Dec. at 16, AR 146; Dean, Evidentiary Hearing Tr. at 121, AR 5739. Dean further explained at the 2003 hearing “that the Title II number taken from the SSA file is also called a Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 77 of 145 25 HICAN in the new file [Dean] creates after the SSISORT program is run.” PRRB Dec. at 16, AR 146; Dean, Evidentiary Hearing Tr. at 113-14, AR 5737. 78. Dean also testified in 2003 that CMS’ match program will drop any SSI records that do not display a Title II number from SSA, stating that “if . . . there is no Title II number, which is all we are, you know, on our end, that is all we are really concerned about . . . this process would not write that record out. So I wouldn’t keep that record.” PRRB Dec. at 16, AR 146; Dean, Evidentiary Hearing Tr. at 113-14, AR 5737. 79. Dean was clear in 2003 that the Social Security number is not used at all in CMS’ match process even though that number appears on SSA’s annual SSI tapes for every individual record included on those SSI data tapes. PRRB Dec. at 16, AR 146; Dean, Evidentiary Hearing Tr. at 122, AR 5739. According to Dean in 2003, “It is strictly HICAN” that CMS uses for the match. Id. Dean testified that “all other data elements on SSA’s annual tapes, including the Social Security numbers, or PANs, are ‘useless.’” PRRB Dec. at 16, AR 146; Dean, Evidentiary Hearing Tr. at 123-24, AR 5739. The Board found that, “[i]n response to a question about whether Social Security numbers are included in the data fields on the MEDPAR, [Dean] responded [in his 2003 testimony] that ‘[i]t is a field we would never use.’” Id. 80. After the 2003 evidentiary hearing in this case, but prior to the hearing the Board conducted in September 2004, the Hospital submitted a pre-hearing brief (entitled “Provider’s Position Paper”) that detailed several systemic flaws in CMS’ match process, including problems stemming from CMS’ failure to match against SSI records that do not display a Title II number on SSA’s annual SSI tape and are dropped from the file before it is matched against CMS’ MEDPAR file. The Provider’s Position Paper also highlighted problems stemming from CMS’ Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 78 of 145 26 failure to match records based on individuals’ own unique Social Security numbers or any other secondary match criteria. AR 441-48. 81. Contradicting his sworn testimony in 2003, Dean testified at the September 2004 hearing “that CMS runs matches against two numbers, the Title II CAN plus a HIC-like number that the CMS program creates from the Social Security number [that appears on SSA’s tape].” PRRB Dec. at 17, AR 147; Dean, Hearing Tr. at 1340-41, AR 5384. The Board found that Dean testified at the 2004 hearing “that CMS uses two numbers for matching because ‘our process doesn’t want to take any chances,’ and ‘[w]e want to make sure we’re matching everybody that they [SSA] send us.’” PRRB Dec. at 17, AR 147; Dean, Hearing Tr. at 1347-48, 1360, AR 5386, 5389. The Board further found that “[w]hen confronted with what appeared to be inconsistencies between the 2003 Evidentiary Hearing [and] the September 2004 Hearing, Dean asserted that the number CMS creates from the Social Security number is a HICAN, and that when he testified in 2003, he was using the term Title II number ‘in a generic sense to mean eligibility for Medicare.’” PRRB Dec. at 17, AR 147 (bold, underscored text emphasized in italics in PRRB decision); see Dean, Hearing Tr. at 1390-92, AR 5397. 82. Regarding additional inconsistencies in Dean’s testimony, the Board found: In further contrast to his prior [2003] testimony, Dean stated [at the 2004 hearing] that the SSISORT program does not discard all records in SSA’s annual tape that do not contain a Title II number. On the contrary, he said [in 2004] that “. . . we create a record from every record Social Security sends us by generating a HIC off of the Social Security number” and “. . . we’ll match against numbers that are provided in the Title 2 field and we’ll match against numbers provided in the Social Security field because we generate a high CAN [sic] from that.” Dean elaborated that his program runs the SSA data file through the SSISORT program and “. . . we come out with a version of the SSA tape, and this version is then applied against the claims data . . . but we’re not dropping claims if they’re not matched and keeping claims if they are matched. It’s our entire version of the SSA tape.” Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 79 of 145 27 PRRB Dec. at 17, AR 147; see Dean Hearing Tr. at 1391-93, 1449, AR 5397, 5424. 83. The Board found that Dean’s testimony in 2004 was inconsistent with deposition testimony that he gave under oath in 1995, when he admitted that “he did not know how the match program worked at that time . . . ; that he had not been given written guidelines or instructions concerning the match program . . . ; that he did not understand the reference in the match program to Title II numbers . . . ; and that he did not know what the match program did with duplicate records from SSA.” PRRB Dec. at 17 n.79, AR 147; Deposition Tr. at 59, 68-69, 79, 109-110, AR 1337, 1339-40, 1342, 1350. 84. The Board additionally found that Dean’s 2004 testimony is inconsistent with a written discovery response that CMS furnished to the Hospital in this case. PRRB Dec. at 19, AR 149; see CMS Discovery Response to Provider Interrogatory 46, AR 3650-51. 85. Regarding the inconsistency between Dean’s 2004 testimony and CMS’ written discovery response referred to in the preceding paragraph, the Board found: With regard to Dean’s testimony in 2004 that CMS matches on two numbers – the HICAN as well as a HIC-like number derived from the Social Security number - CMS’ discovery response did not indicate that CMS ever matches against a number derived from the Social Security numbers included in SSA’s annual tapes. CMS stated that the agency ‘ . . . has no recollection, and no documentation concerning, whether CMS ever saw a need . . . to employ secondary or alternative matching criteria’ in addition to, or in lieu of, the HICAN. PRRB Dec. at 19, AR 149; see CMS Discovery Response to Provider Interrogatory 46, AR 3650-51. 86. The Board explained its decision not to credit Dean’s inconsistent testimony in 2004 in part as follows: “Dean admitted that CMS cannot explain how the match ran, or what criteria were used in the match program that was run for the calculations of the SSI fractions for fiscal years 1993 and 1994, nor can CMS prove how the match program ran, or what criteria Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 80 of 145 28 were used in the match program that was run, for the calculations of the SSI fractions for fiscal years 1995 and 1996, because CMS did not have an archiving mechanism for the output.” PRRB Dec. at 18, AR 148; see Dean, Hearing Tr. at 1349-50, AR 5386-87. 87. Additionally, the Board found that “[a]t the 2004 hearing, Dean confirmed that he was still the person primarily responsible for the matching process and for maintaining the program, but no one [from CMS] had yet asked him to look at the data and see if there appeared to be a problem.” PRRB Dec. at 18, AR 148; see Dean, Hearing Tr. at 1447, AR 5411. 88. The Administrator’s decision does not controvert the Board findings regarding the internal inconsistencies and contradictions in Dean’s testimony, as set forth in paragraphs 75-87 above. See Adm’r Dec. at 15-17, 42, AR 16-18, 43. 89. The Administrator credited most of Dean’s testimony at the 2004 hearing, without addressing or even acknowledging the Board’s findings, set forth above, regarding the inconsistencies between Dean’s 2004 testimony and the sworn testimony that Dean gave in the 2003 hearing in this case and in his 1995 deposition testimony. Adm’r Dec. at 15-17, 42, AR 16- 18, 43. 90. Further, the Administrator’s decision fully credits Dean’s 2004 testimony regarding CMS’ alleged use of a secondary match identifier generated from the Social Security number appearing in SSA’s annual SSI tapes, without addressing CMS’ admissions in discovery that CMS had no documentation reflecting that the agency ever even saw a need to use secondary or alternative matching criteria. Adm’r Dec. at 16-17; AR 17-18; see PRRB Dec. at 19, AR 149; see CMS Discovery Response to Provider Interrogatory 46, AR 3650-57. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 81 of 145 29 f. CMS’ Admission of A Systemic Flaw In the Match Process 91. As the Board found, CMS admitted in discovery that, in 1996, a CMS employee, Joseph Del Pilar, changed the match program that CMS ran, called the SSISORT program, “so that the program would write out separate records in cases where an individual had Medicare entitlement on one HIC number and later became entitled to Medicare on another HIC number (the ‘Del Pilar change’).” PRRB Dec. at 18, AR 148; CMS Response to Provider Interrogatory 36, AR 3643-45. Based on record evidence, the Board further found that, in CMS’ discovery response regarding the Del Pilar change, CMS admitted the following: Had such change to the SSISORT program been in effect to 1996, such change would have tended to increase the numerator of the hospital’s Medicare [SSI] fraction for such period if an individual had a qualifying inpatient stay at such hospital, and was entitled to SSI during the month in which the discharge from such stay occurred, and appeared as entitled to SSI for such month on the tape sent by SSA to CMS for purposes of calculating the Medicare fraction for such fiscal year, and was entitled to title II on more than one account number, and if SSA did not furnish all such title II numbers to CMS. PRRB Dec. at 19, AR 149; CMS Response to Provider Interrogatory 36 at 2980-81, AR 3644- 45. 92. Del Pilar and Dean were the sources CMS identified for the discovery response quoted in paragraph 91 above. Answer ¶ 172. 93. When Del Pilar testified in the 2004 hearing he stated “that he merely filled in at that time [in 1996] for Mr. Dean and had no knowledge or understanding of the substantive changes he [had been] asked to make by [another CMS employee,] Ms. O’Leary, nor did he recollect anything about the changes.” PRRB Dec. at 18, AR 148; Del Pilar, Hearing Tr. at 840- 43, 851, AR 5258-59, 5261. Janet O’Leary “had worked for CMS or its predecessor since 1977. In the mid 1990’s she was in the National Claims History Branch and in 1996 was its Branch Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 82 of 145 30 Chief. Tony Dean and Joseph Del Pilar, both programmers, worked for her.” PRRB Dec. at 18, AR 148; O’Leary, Hearing Tr. at 668-69, AR 5213. Regarding O’Leary’s testimony concerning the 1996 change to CMS’ match process, the Board found: When Ms. O’Leary was questioned about the alleged changes [made in 1996], she also recalled very little about the Del Pilar changes. She did recall, though, in some detail the ‘stale records’ issue that was eventually corrected. She identified Provider Exhibit 64 at page 1550 as her own notes from February, 1996. The notes reflect that, at that time, there was a question whether a person could have more than one SSI record in a year and O’Leary stated that they questioned whether SSA was sending all those records. But Ms. O’Leary testified at the hearing in 2004 that “Again, we have no answer to that question that I know of.” PRRB Dec. at 18-19, AR 148-49; see O’Leary, Hearing Tr. at 669-81, 687, AR 5213-16, 5218. 94. The Board found that O’Leary had little or no understanding of the programs her immediate subordinates had been running in connection with CMS’ computations of the SSI fractions: O’Leary believed that the SSA tape had the health insurance claim [HIC] number on it, stating ‘there would be no other way for us to identify’ [the beneficiary] individually. She understood the HIC number and the Title II number to be the same. She acknowledged that the SSN would also be on the SSA tape but that it would not identify an individual for Medicare because Medicare uses the HIC number ‘which may or may not be the individual Social Security number . . . included in our claim number.[’] She believed that CMS would also have the SSN in the enrollment database in most instances, but possibly not all. PRRB Dec. at 19, AR 149; see O’Leary Hearing Tr. 688-91, AR 5218-19. 95. The Board found that “CMS does not allege that it ever corrected the SSI fractions that were computed prior to the 1996 change allegedly made by Del Pilar.” PRRB Dec. at 19, AR 149. 96. The Administrator’s decision does not address or acknowledge the Board’s findings, or CMS’ own admissions, as described in paragraphs 91-95 above, regarding the 1996 Del Pilar change to the match program. Nor does the Administrator disclaim CMS’ admission Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 83 of 145 31 that Del Pilar’s change, if implemented, “would have tended to increase the numerator of the hospital’s Medicare [SSI] fraction” for DSH. See Adm’r Dec., AR 2-56. g. Additional Flaws in the Match Process 97. The Board found that, regardless of whether CMS’ match process worked in the manner described by Dean in his 2003 testimony or whether it worked in the manner that Dean now says he thinks it might have worked, the process is systemically flawed in multiple respects that produce unreliable numbers and tend to deflate substantially the resulting SSI fractions. PRRB Dec. at 20-22, AR 150-52. 98. The Board found: If the evidence is correct that CMS only matches the HICAN in the MEDPAR files and the single Title II number on SSA’s annual tapes, as [Dean’s] 2003 testimony and CMS’ discovery responses indicate, it would follow that CMS’ inpatient hospital stay records would not have matched with the SSI records on SSA’s annual tape for some individuals who were entitled to SSI, because SSA’s annual tape never includes more than one Title II number, or CAN, for individuals who received Title II benefits under more than one Title II number. PRRB Dec. at 20, AR 150. 99. The Board further found: If the SSISORT program reformats the SSI data received from SSA to drop all SSI records that do not include a Title II number, as Dean testified in 2003, this creates a significant potential for error because, according to Shafer [the long-time SSA official], roughly half of the records on SSA’s annual tapes would not have a Title II number. PRRB Dec. at 20, AR 150; see Dean, Evidentiary Hearing Tr. at 113-16, AR 5737; Shafer, Hearing Tr. at 407-08, AR 5128. 100. Additionally, the Board found: The match process described in Dean’s testimony at the 2003 Evidentiary Hearing is also prone to error due to changes in Medicare beneficiaries’ HIC numbers over Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 84 of 145 32 the course of a year and CMS’ failure to use the only individual unique identifier, the Social Security number, in the match process. PRRB Dec. at 20, AR 150. 101. Regarding Dean’s inconsistent testimony at the 2004 hearing, the Board found: “Even if the SSI match program has always been run against both the Title II numbers and the HIC-like numbers derived from the Social Security numbers in SSA’s annual tapes, as Dean now hypothesizes, the evidence shows that CMS’ match process would likely still produce false negatives.” PRRB Dec. at 20, AR 150. 102. In the first of two situations noted by the Board in which CMS’ match process would produce false negatives even if the match program always ran in the manner that Dean hypothesized in his 2004 testimony, the Board found: One example of a situation in which the SSI match program would fail to match is when an individual, “Jane,” initially receives Title II and Medicare on her own account using her own Social Security number and an “A” beneficiary identification code (e.g., 000-00-0000A). “Jane” subsequently qualifies for Title II and Medicare on her spouse’s account, using her spouse’s Social Security number followed by a “B” beneficiary identification code (i.e., 111-11-1111B). After the change in the HICAN number, Jane is hospitalized. The HICAN for the period of the hospitalization would be Jane’s later Title II number (111-11- 1111B), which would not match the first, oldest, Title II number that SSA’s program would pick up from the unearned income field on the SSI master record. That number, in this example, would be 000-00-0000A. PRRB Dec. at 20, AR 150; see Dean Hearing Tr. at 1366-74, AR 5391-93. The Board found that, in his testimony at the 2004 hearing, “Dean initially agreed there would be no match in this first example, even assuming the SSI match program does what he now thinks it does.” PRRB Dec. at 20, AR 150; see Dean, Hearing Tr. at 1366-69, AR 5391. The Board noted, however, that Dean “later recanted his answer, based on the premise that SSA’s annual tape would have the spouse’s Social Security number and a “B” beneficiary identification code (i.e., 111-11- Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 85 of 145 33 1111B) in the Title II field.” PRRB Dec. at 20, AR 150, see Hearing Tr. at 1368-74, AR 5391- 93. The Board did not credit Dean’s second answer about how the match process might have worked properly, finding that “Dean’s assumption is contradicted by Cribbs’ testimony as to how her program actually works and how that program scans the unearned income field to pick up only the first and (oldest) Title II number. Further, Dean admits that he had never attempted to verify with SSA his supposition as to how SSA’s program works.” PRRB Dec. at 20, AR 150. 103. In the Board’s second example of a situation in which CMS’ match process would produce false negatives again (even if it always worked in the manner Dean hypothesized in his revised testimony on this point in 2004), the Board found: A second example of a situation in which the SSI match program would fail to match is when an individual, “Jane,” whose Social Security number is 000-00- 0000, initially receives Title II and Medicare under a deceased spouse’s account using the decedent’s Social Security number followed by a “B6” beneficiary identification code (i.e., 111-11-1111B6). Jane subsequently remarries and receives Title II and Medicare on her second spouse’s account, using the second spouse’s Social Security number and a “B1” beneficiary identification code (i.e., 222-22-2222B1). If Jane is hospitalized after the change in her HICAN number, then the HICAN for the period of the hospitalization would be the second number (222-22-2222B1), which would not match either the first, oldest, Title II number that SSA’s program would pick up and include on its annual tape (111-11- 1111B6) or her own Social Security number with an “A” beneficiary identification code tagged on at the end (000-00-0000A). See Hearing Tr. at 1375-76. Dean admitted that there is no way that CMS’ SSI match program would produce a match in this example, even if it actually worked the way he now thinks it might have worked. PRRB Dec. at 21, AR 151 (emphasis added); see Dean, Hearing Tr. at 1442-44, AR 5409-10. 104. In view of the foregoing findings and citing uncontroverted evidence in the record, the Board concluded that CMS’ match process is systemically flawed: “Regardless of how the match process may have worked, there is compelling evidence that it did not, in fact, produce accurate results.” PRRB Dec. at 21, AR 151. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 86 of 145 34 105. The Board further found that “CMS’ witness from SSA, Cliff Walsh, confirmed that the limited SSI records that the Provider was able to obtain from SSA show several examples of SSI stays that would have been on the SSA data file and yet inexplicably were not counted as SSI days in CMS’ MEDPAR data.” Answer ¶ 185; PRRB Dec. at 21, AR 151; Walsh, Hearing Tr. at 2195-2208, AR 5598-5601; see Provider Exhibits 183, 184, 185 and 186, AR 4041-49, 4050-4055, 4056-57, 4058-59. Walsh “is Branch Chief of the Master Records Support Section of SSI System within the Social Security Administration. He has been Branch Chief for six years, but has worked in that division since 1986 as the team leader of the Master Records Support Section. His branch maintains the SSI master file and its associated database.” Answer ¶ 185. 106. In response to questions from CMS’ own counsel, SSA’s Walsh testified that the Hospital had identified five patient stays attributable to Medicare beneficiaries whose inpatient days were not counted by CMS as Medicare/SSI patient days for purposes of the DSH calculation, but who were in the Hospital during the period at issue, who were receiving SSI benefits during that time, and whose SSI benefits certainly would have been reflected and included on the SSI tapes that SSA furnished to CMS for purposes of CMS’ calculations of the SSI fractions for the periods at issue. PRRB Dec. at 21, AR 151; Walsh, Hearing Tr. at 2195- 2208, AR 5598-5601; see Provider Exhibits 183, 184, 185 and 186, AR 4041-49, 4050-55, 4056- 57, 4058-59. 107. The Board found: “The omission of these stays from the SSI days in the numerator of CMS’ calculation proves there are flaws in CMS’ match process.” PRRB Dec. at 21, AR 151. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 87 of 145 35 108. The Board, therefore, concluded that “the numbers used to calculate the DSH Medicare fraction are unreliable, regardless of the ultimate process that CMS developed.” PRRB Dec. at 21, AR 151. According to the Board: [T]he evidence is convincing that there is a great likelihood that some matches will not be picked up because of the failure to match on a unique identifier and/or multiple criteria. Even if Mr. Dean’s most recent testimony is accepted, some SSI beneficiaries who qualify under another person’s SSN or HIC number would not be identified. Dean admits that unless SSA gives CMS all an individual’s Title II numbers, CMS will never get its match right under the match process that CMS now says it uses. A more accurate match would be undertaken using beneficiaries’ Social Security numbers and alternative identifiers (name and gender) when comparing the information on the SSA data tapes, which is what occurs when SSA monitors SSI recipients’ residence in nursing homes. PRRB Dec. at 21, AR 151 (emphasis added); see Dean, Hearing Tr. at 1436-37, AR 5408; Cribbs, Hearing Tr. at 117-23, AR 5056. 109. Concerning the credibility of the witnesses who testified regarding the match process and the identifiers used in that process, the Board found that “the SSA witnesses were credible. All had over 20 years of experience with SSA and were extremely knowledgeable about SSI programs and SSA’s operations, including recordkeeping.” PRRB Dec. 21, AR 151. “In contrast,” the Board found, “CMS’ employees charged with the responsibility for implementing DSH demonstrated a lack of comprehensive knowledge of the match process, i.e., what data was included in or excluded from the SSA tapes, as well as how, when, why and to what extent the shortcoming of that data might lead to a failure to match CMS’ MEDPAR file.” PRRB Dec. at 21-22, AR 151-52. The Board had occasion to warn at least one key CMS witness that her answers on critical questions before the Board were “disingenuous.” PRRB Dec. at 34- 35 & n.210, AR 164-65. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 88 of 145 36 110. The Administrator’s decision does not address the Board’s findings regarding the credibility of the CMS witnesses and the present and former SSA employees, but suggests that the Board should not have credited what the Administrator characterizes as “opinion testimony” of two witness from the Hospital’s DSH consulting firm, as well as the former long-time SSA employee, Shafer. Adm’r Dec. at 50-53, AR 51-54. 111. The Administrator acknowledges that the Hospital “did not proffer either witness [from its consulting firm] as experts [sic],” Adm’r Dec. at 50, AR 51, but nonetheless faults the Board for crediting a few pieces of testimony that he characterizes as “opinion” testimony. The Secretary’s counsel, however, did not object either before or during the Board hearing to the Hospital’s disclosed intent to call these witnesses to testify. Answer ¶ 338. The Secretary’s counsel also did not object on this basis to any testimony they gave at the hearing. Id. 112. Having concluded that CMS’ match process is flawed for the reasons set forth above, the Board found that other match criteria must be used to calculate the SSI fractions for DSH and that doing so would impose no significant administrative burden on CMS. PRRB Dec. at 22-23, AR 152-53; see Shafer, Hearing Tr. at 643-46, AR 5207; Walsh, Hearing Tr. at 2215- 16, AR 5603; Dean, Evidentiary Hearing Tr. at 202-12, AR 5759-61. In particular, the Board found that individuals’ own unique Social Security numbers should be used to match SSI records on SSA’s annual tapes with CMS’ inpatient hospital stay records in MEDPAR because “CMS stated in the [1986] Federal Register when the DSH legislation was first implemented that using the Social Security number was the best identifier, and [the government] uses the Social Security number, among other matching criteria such as a name and gender, for matching in a virtually identical program for skilled nursing providers.” PRRB Dec. at 22, AR 152; see Cribbs, Hearing Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 89 of 145 37 Tr. at 117-23, AR 5056-57 (testifying that nursing home matches used Social Security numbers); 51 Fed. Reg. 16772, 16777 (May 6, 1986). 113. CMS’ “Dean agreed with the Provider’s counsel that if [CMS] had the Social Security numbers of each hospital inpatient, then matching a SSN against an SSN would ‘of course’ be a better match ‘in an ideal world.’” PRRB Dec. at 22, AR 152; Dean, Hearing Tr. at 1393-95, AR 5397-98. Citing a sworn statement previously submitted to this Court by Robyn Thomas, the Division Director for CMS’ Division of Information Distribution, the Board further found that “CMS’ Medicare Enrollment Data Base (like SSA’s annual tape) contains Medicare beneficiaries’ own Social Security numbers, their names and dates of birth.” PRRB Dec. at 22, AR 152; Declaration of Robyn Thomas, AR 596, 598. The Board also found that “[i]n 2001, CMS, or its contractor, used a ‘HIC finder’ program to pull Social Security numbers from the enrollment database and add them to MEDPAR files that were produced to the Provider for fiscal years 1993, 1994 and 1995 in response to the subpoenas issued by the Board in 2000 (which required Social Security numbers).” PRRB Dec. at 22, AR 152; see also AR 2439. 114. In view of the foregoing facts, the Board also found that “[t]here is no evidence of a significant administrative burden to create a far more accurate SSI calculation” by using Social Security numbers as an additional match criterion. PRRB Dec. at 22, AR 152. More specifically, the Board found that: [T]he overwhelming weight of the evidence, much of it from CMS’ own employees, indicates that to change the program to capture more accurate data would be routine, simple and not time consuming. It would not require access to old data, as an accurate calculation could be made using current data files. The fact[] that CMS has developed a similar program to match SSI information for Medicare beneficiaries in nursing homes and has for years run multiple matches for the same period at different times, including a time that would coincide more Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 90 of 145 38 closely with the settlement of the cost report, illustrate that a more reliable program is administratively and financially feasible. PRRB Dec. at 22-23, AR 152-53 (emphasis added); see Shafer, Hearing Tr. 643, AR 5207; Walsh, Hearing Tr. at 2216, AR 5603; Dean, Evidentiary Hearing Tr. at 203-12, AR 5759-61. 115. Except as previously mentioned above (see supra ¶ 111), the Administrator’s decision does not controvert the Board’s findings or CMS’ admissions concerning the effect of the flaws in CMS’ match process, the necessity of using individuals’ own Social Security numbers as additional match criteria and a later match of the MEDPAR file and SSI tapes in order to achieve more accurate results, or the lack of any administrative burden on CMS in making these changes. See Adm’r Dec. at 43, AR 44. h. Impact of Flaws in the Match Process 116. Although the Administrator’s decision does not genuinely address or controvert any of the Board’s findings regarding the flaws in CMS’ match process, the Administrator’s decision concludes that, “to the extent that the Provider may have identified a scenario for which the Medicare beneficiaries’ SSI days would not be credited (e.g., remarried widow on second husband’s account) the Provider only identified one instance in the sample drawn to exaggerate any such errors, [sic] where this occurred, a percentage of .15 percent, significant [sic] less than one percent. Therefore, the Secretary’s policy determination not to use alternative identifiers is supported by the record in this case.” Adm’r Dec. at 43, AR 44. 117. There is no record evidence to support the Administrator’s claim that the Hospital identified just one case in which a Medicare beneficiary’s SSI days were not credited in the numerator of the SSI fractions that CMS computed for the fiscal years at issue. See supra ¶¶ 105-06. As set forth in paragraph 106 above, CMS’ own witness from SSA, Walsh, testified Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 91 of 145 39 on examination by the Secretary’s counsel that based only on the limited SSI records that the Hospital has been able to obtain in this case, the Hospital identified at least five separate patient stays attributable to Medicare beneficiaries who were in fact receiving Federal SSI benefits during their inpatient stays but which were not identified by CMS as SSI beneficiaries due to flaws in its match process. See PRRB Dec. at 21, AR 151; Walsh, Hearing Tr. at 2194-2200, AR 5598-99. 118. Nor is there any evidence supporting the Administrator’s claim that the Hospital drew “a sample . . . to exaggerate any such errors” in CMS’ match process. Adm’r Dec. at 43, AR 44. It is uncontroverted the Hospital and the Board sought, and CMS consistently resisted the Hospital’s and Board’s efforts to obtain, the SSI records for all of the Hospital’s Medicare patients during the periods at issue. PRRB Dec. at 4, AR 134. Not only did the Hospital submit a request to SSA under the Freedom of Information Act for the SSI records of all Medicare patients discharged from the Hospital in fiscal years 1993 through 1996, but the Hospital also requested that the Board issue a subpoena to SSA for the SSI records of all Medicare patients discharged from the Hospital during those years. AR 1102-10; AR 1948-53; AR 1984-87. 119. The SSI fractions that CMS computed for the four Hospital fiscal years at issue included a total of 20,913 Medicare/SSI days and 250,494 total Medicare days, leaving 229,581 Medicare days that were not counted as SSI days by CMS. See supra ¶¶ 14, 16, 18, 20; Answer ¶ 198 (admitting accuracy of calculations). The application of a .75 percent error rate (.15 percent times 5) to the 229,581 Medicare patient days that CMS failed to identify as SSI patient days would result in the addition of 1,722 Medicare/SSI days to the numerators of the Hospital’s SSI fractions for the fiscal years at issue. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 92 of 145 40 120. For the four fiscal years at issue, the average effect of adding an SSI day to the numerator of the Hospital’s SSI fraction is to increase the DSH payment by $830. See AR 4820. The result, therefore, of adding 1,722 SSI days to the numerators of the Hospital’s SSI fractions for 1993-1996 is an increase in the Hospital’s DSH payment of approximately $1.4 million (1,722 days x $830 per day) for all four years combined. If the same financial impact were projected to the approximately 2,000 hospitals that received DSH payments for the four years at issue, the overall financial impact of the flaws in CMS’ match process would be approximately $2.8 billion ($1.4 million x 2,000) for those four years combined, and $700 million for each year. 121. The average of the SSI fractions computed by CMS for the Hospital fiscal years at issue is .08334, which is computed by dividing 20,913 Medicare/SSI days for all four years divided by 250,944 Medicare days for all four years. See supra ¶¶ 14, 16, 18, 20; Answer ¶ 200 (admitting accuracy of calculations). Adding 1,722 Medicare/SSI days to the numerators of the SSI fractions that CMS computed for these years would increase the Hospital’s average SSI ratio by 8.2 percent, from .08334 to .09020. Answer ¶ 200. 122. The Board calculated a $200 million impact associated with a 4.03 percent increase in the national average SSI ratio for 1993 alone. See infra ¶ 131; PRRB Dec. at 24, AR 154. If the national average SSI ratio actually increased by 8.2 percent (.9020/.08334 – 1), see ¶ 121 above, instead of 4.03 percent, the aggregate dollar impact would be over $400 million. Id. E. SSI Data Used 123. The Board found that the SSI data that CMS used to calculate the numerator of Hospital’s SSI fractions for the periods at issue omitted: (a) all inactive or “stale” SSI records for all years prior to 1995; Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 93 of 145 41 (b) SSI records relating to forced or manual payments of SSI benefits for all years; (c) SSI entitlement records relating to individuals whose SSI benefits were temporarily on hold or in suspense for all years; (d) SSI entitlement records for individuals whose SSI benefits were granted or reinstated retroactively for all years; and (e) SSI records relating to individuals who received special non-cash SSI benefits under Title XVI of the Social Security Act. PRRB Dec. at 23-35, AR 153-55. 124. As set forth more fully below, the Board also found that: (a) CMS knew at least by 1993 that “there were significant inaccuracies with the data used for the DSH calculation and that those inaccuracies may have diminished the DSH payment to providers” but CMS did not correct any of the problems it knew about until 1996 and even then the fix was only partial and implemented only prospectively and without notice or disclosure to affected hospitals. PRRB Dec. at 23-26, 35, AR 153-56, 155; (b) Even once CMS purportedly fixed one of the problems in 1996 (for years 1995 forward only), CMS never corrected the calculations it had previously performed with incomplete data that were, in CMS’ word, “contaminated” by the omission of ‘stale’ SSI records for all prior periods, including periods at issue in this case, nor did it undertake any effort to investigate other possible problems with the SSI data. PRRB Dec. at 34, AR 164; (c) CMS did not disclose the problems it knew about to affected hospitals, but instead disseminated false statements and misinformation concerning the accuracy of the data used in its calculations of the SSI fractions for DSH. PRRB Dec. at 34, AR 164; see CMS Response to Provider Interrogatory 68, AR 3664; and (d) The various systemic flaws in the SSI data that CMS has used in its calculations of the SSI fractions could be easily fixed without imposing administrative burden on the agency. PRRB Dec. at 26, AR 166; see Shafer, Hearing Tr. at 332-34, 339-50, 364-65, AR 5109-10, 5111-14, 5117-18. 1. Omission of Inactive Records 125. The Board found: “It is undisputed that, until approximately February 1996, SSA’s annual tapes omitted all SSI records that had been terminated and were inactive prior to Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 94 of 145 42 the time when SSA transmitted the tape to CMS (the ‘stale records problem’).” PRRB Dec. at 23, AR 153. This problem is conceded. Stipulation on the Record, Hearing Tr. at 561-63, AR 5186-87; see also Memorandum From Anne Rudolph to SSA Dated February 29, 1996, AR 2108; Cribbs, Hearing Tr. at 129, 165-67, 202-04, 214-23, 246-47, 258, 266-67, 271, AR 5059, 5068, 5077, 5080-82, 5088, 5091, 5093, 5094; Shafer, Hearing Tr. at 315-34, 546, 573-74, 624- 30, AR 5105-10, 5182, 5189, 5202-03; Walsh, Hearing Tr. at 2168-76, 2185-90, 2203-04, 2217, AR 5591-5593, 5596-97, 5600, 5604; Cribbs, Evidentiary Hearing Tr. at 276-80, 356, AR 5682- 5683. 126. Regarding the cause of the “stale records problem,” the Board found that “an ‘accounting page’ of an individual’s [SSI] record could be terminated [by SSA], even though the individual [may be still living and] remained entitled to SSI benefits, and that record [i.e., the accounting page] would become inactive immediately.” PRRB Dec. at 23, AR 153.; see Cribbs, Hearing Tr. at 219-23, AR 5081-82; Shafer, Hearing Tr. at 335-36, AR 5110. An accounting page of an individual’s SSI record could be terminated and would become inactive immediately, “either due to space limitations or in forced pay cases.” PRRB Dec. at 23, AR 153; see Cribbs, Hearing Tr. at 219-23, AR 5081; Shafer, Hearing Tr. at 335-36, AR 5110. In addition, the Board found that “the records of deceased individuals and individuals in terminated status would usually be moved into inactive files after 12 months. Records may have been moved to inactive status sooner in earlier years because of space limitation.” PRRB Dec. at 23-24, AR 153-54; see Cribbs, Hearing Tr. at 203-04, 246-47, AR 5077, 5088; Shafer, Hearing Tr. at 320, 622-30, AR 5106, 5201-03. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 95 of 145 43 127. As a result, a Medicare patient who was entitled to SSI during his hospital stay would not be counted in the numerator of the hospital’s SSI fraction solely because an accounting page of his SSI record may have been terminated due to his death or some other intervening event occurring between the time of his stay and the time when SSA generated the SSI records used by CMS to perform its match process for calculating the hospital’s SSI fraction for that year. PRRB Dec. at 23, AR 153; see Shafter, Hearing Tr. at 335-37, AR 5110-11. 128. Based upon the uncontroverted testimony of CMS’ employee, Rosenberg, and other undisputed evidence in the record, the Board found that “it is undisputed that CMS had SSA prepare new tapes beginning with the 1995 fiscal year that corrected the problem by including the stale records; that is, the records that had previously been omitted because they had become inactive were included on the data file prepared for CMS beginning with fiscal year 1995.” PRRB Dec. at 24 & n.120, AR 154; see Rosenberg, Hearing Tr. at 2045-46, AR 5561 (testifying that SSA supplied new tapes in 1996 that included the stale records); Memorandum from CMS to SSA dated Feb. 29, 1996, AR 2108. 129. Citing the uncontroverted testimony of another CMS employee, O’Leary, and other undisputed documentary evidence in the record, the Board found that “[a]fter it obtained corrected SSI records from SSA in 1996, . . . CMS created special MEDPAR files that included both the original (‘old’) and corrected (‘new’) SSI days for inpatient hospital stays in prior years.” PRRB Dec. at 24, AR 154; see O’Leary, Hearing Tr. at 685-741, AR 5217-31, AR 2080- 82. Based on O’Leary’s undisputed testimony, the Board found that the “intent of this ‘special MEDPAR’ project was to determine whether the differences were material and, if so, to recalculate prior SSI ratios.” PRRB Dec. at 24, AR 154. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 96 of 145 44 130. The Board found: “[A]s reflected on CMS’ summary of its special MEDPAR files, the ratio of new SSI additions to old SSI deletions was greater than 17:1 for most years,” though the omission of inactive SSI records was probably a larger problem in earlier years in the period from 1986-1994. PRRB Dec. at 24 & n.124, AR 154; see Pfeil, Hearing Tr. at 1575-84, AR 5443-45. 131. The Board further found that “the aggregate impact of the SSI additions (i.e. all PPS hospitals nationwide), net of SSI deletions, to be approximately $4 billion for years 1989 through 1996, assuming a 7-day average length of stay (ALOS) and an average value of $500 per SSI day.” PRRB Dec. at 24, AR 154; see Pfeil, Hearing Tr. at 1570, AR 5442. The Board found that the aggregate impact of the omission of inactive SSI records for fiscal year 1993 alone “would be approximately $202 million, . . . [e]ven though the new SSI data would have increased the national average SSI ratio for 1993 by only 4.03%, from .083746 to .087779[.]” PRRB Dec. at 24, AR 154; see Pfeil, Hearing Tr. at 1584-86, AR 5445-46. Thus, the Board found, based on record evidence, that “this seemingly minor change to the SSI ratios would have a major impact on hospitals that treat a large indigent population.” PRRB Dec. at 24, AR 154. 132. The Board further found that: While the financial impact may be relatively small when averaged over all PPS hospitals, it can have a significant impact on those PPS hospitals that serve a large low-income population. Hospitals must meet a certain threshold percentage of low-income patients before they qualify for the DSH adjustment. Even a fraction of a point under that threshold means they receive no DSH payment whatsoever. Once the threshold is met, however, the volume of low-income patients directly correlates to the amount of the DSH payment. PRRB Dec. at 25-26, AR 155-56. 133. Based on record evidence, the Board further found that the stale records problem is illustrated by comparison of two of CMS’ calculations of the SSI fraction for the Hospital’s Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 97 of 145 45 fiscal year 1994, one calculation performed before CMS fixed the stale records problem and one performed by CMS after it fixed the stale records problem. PRRB Dec. at 25, AR 155; see Smith, Hearing Tr. at 874-77, 927-29, AR 5267-68, 5280-81. 134. In particular, the Board compared the numbers included in CMS’ June 1995 calculation of the Hospital’s SSI fraction for the 1994 (which was calculated by CMS before it fixed the problem concerning inactive SSI records in 1996 but was nevertheless applied to determine the Hospital’s DSH payment for 1994) with the numbers included in CMS’ December 1996 calculation of the Hospital’s SSI fraction for 1994 (which was calculated by CMS after it fixed the stale records problem but was not applied to determine the Hospital’s DSH payment for 1994). PRRB Dec. at 25, AR 155; see Smith, Hearing Tr. at 874-77, 927-29, AR 5267-68, 5280- 81. The Board found that this comparison “showed that the Provider had 400 more SSI days while only 2600 more [Medicare] covered days which could be due to retroactive SSI eligibility determinations and/or restored [SSI] records.” Id. In other words, while the total number of Medicare days counted in the denominator of CMS’ December 1996 calculation was 4.03% greater than the number of Medicare days counted in CMS’s June 1995 calculation, the total number of Medicare/SSI days that were counted in numerator of the SSI fraction numerator grew by an even greater percentage, 7.5%. Thus, the Board found that the “discrepancy between the earlier calculation and the 1996 run [for the Hospital’s 1994 fiscal year] alone illustrates that the original data was inaccurate.” PRRB Dec. at 25, AR 155. 135. Though it found that even a seemingly small percentage change in the SSI ratios can have a major impact on hospitals that serve a large proportion of low-income patients, the Board also found, based on the testimony of CMS’ employee, Rosenburg, that “the only way to Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 98 of 145 46 know what the impact would be of including the inactive records that were omitted from original calculations for fiscal years 1993 and 1994 would be to rerun the actual calculations.” PRRB Dec. at 25, AR 155; see Rosenberg, Hearing Tr. at 2145-46; AR 5586. After it re-ran the new matches, however, CMS staff received instructions to go in a “different direction” and never corrected fractions for prior years. O’Leary, Hearing Tr. at 823, AR 5254. 136. The problem with CMS’ omission of inactive SSI records could have been avoided from the outset through adequate, reasonable communication and coordination between CMS and SSA: CMS could have given proper guidance to SSA programmers and systems analysts, who then could have developed proper functional requirements for the program that SSA used to create the annual SSI tapes in accordance with applicable Federal Information Processing Standards. Shafer, Hearing Tr. at 332-34, 348-49, 409-13, 632-33, 651, AR 5109-10, 5113-14, 5129-30, 5204, 5209. The Board found that because the “uncontroverted evidence shows that CMS knew, at least by 1993, that there was a problem with the SSI data that CMS had been receiving from SSA[,] . . . it is clear that CMS did not use the best data available when it calculated the SSI fractions at issue for fiscal years 1993-1996. This was due, in part, to the omission of ‘stale records’ from the SSI data tape prepared for CMS.” PRRB Dec. at 25, AR 155; see Rosenberg, Hearing Tr. at 2131, 2148, AR 5582; Pfeil, Hearing Tr. at 1481-1520, AR 5419-29. 137. The Board found: “It is disingenuous . . . to claim that the omission of SSI days has no significant impact or that the Provider must quantify the financial impact to prevail while CMS resisted efforts to obtain the information critical to making such proof.” PRRB Dec. at 25 & n.135, AR 155. The Board found that CMS’ employee, O’Leary, had “confirmed that the Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 99 of 145 47 special MEDPAR files would have shown the impact of the omission of stale records (as well as other updated SSI records) by year, by hospital, and by stay.” PRRB Dec. at 25, AR 155; see O’Leary, Hearing Tr. at 712-18, AR 5224-25. 138. A CMS employee, Dean, testified before the PRRB that all of the “MEDPAR” files reflecting the Secretary’s calculations were supposed to be kept “permanently available.” Evidentiary Hearing Tr. at 64-71, AR 5724-26. 139. The Board found that CMS now claims that, sometime after 1996, it “lost or destroyed” the records that would have shown the impact of the omission of stale SSI records by year, by hospital, and by stay. PRRB Dec. at 34, AR 164; see Rosenberg, Hearing Tr. at 2042, 2085-87, 2095-97, 2137-40, AR 5560, 5571, 5573-74, 5584; Hearing Tr. at 719-20, 731-33, 740, 799-800, 804, 806, 810, AR 5226, 5229, 5231, 5248-50; AR 2087. Until 1996, SSA’s annual tapes were kept under Dean’s desk at CMS. Dean, Hearing Tr. at 1396-97, AR 5398; AR 2113. Dean later destroyed these tapes, he says, because “there was no direction to keep them.” Dean, Hearing Tr. at 1397-99, AR 5398-99. Consequently, the Secretary cannot now say and does not now have records to show specifically which patient days were counted in his calculation of the SSI fraction for two of the Hospital’s four fiscal years at issue (1993 and 1994). Answer ¶ 7. The Board concluded, therefore, that “it is disingenuous for CMS to take a position that the Provider must demonstrate what financial impact the flaws in the process caused while knowing that its own record retention failures make such a showing impossible.” PRRB Dec. at 32, AR 162. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 100 of 145 48 140. The Board made several findings that established for the first time that CMS actively and willfully concealed its omission of inactive SSI records from the calculations of the SSI fractions for periods prior to 1995: (a) “In June 1995, at just about the same time that CMS was settling with [a hospital in Rhode Island that had noted discrepancies in CMS’ SSI fraction calculations], [CMS representative Ann] Tayloe testified in a deposition [in a contested DSH matter before the Board involving Saint Louis University Hospital] that the SSI data used in the calculation of the SSI fraction were verified and correct,” even though Tayloe had been directly and personally involved in responding to and settling the other hospital’s complaints on behalf of CMS. PRRB Dec. at 33, AR 163; see St. Louis University Hospital, Rudolph Deposition Tr. at 72, 88, 100-01, 200, AR 848, 864, 876-877, 976. (b) “CMS did not produce any supporting documents, and Tayloe did not indicate in her testimony that there were any documents relating to the potential for error in CMS’ calculation of the SSI fraction, even though the Board subpoena in that case [involving the Saint Louis University Hospital] required production of such documents in connection with the deposition.” PRRB Dec. at 33, AR 163. (c) “Also in June 1995, CMS published [a] proposed rule to change the process for recalculation of the SSI ratio based on a hospital’s cost reporting period. CMS indicated both in the proposed and final rule that the recalculations invariably resulted in lower SSI ratios due to problems with the requesting hospital’s data. It also represented that SSA, not CMS, computes the SSI fractions, and that SSA does not release the underlying SSI data to CMS.” PRRB Dec. at 33-34, AR 163- 164 (emphasis in original); see 60 Fed. Reg. 45778, 45811-12 (Sept. 1, 1995) (final rule), AR 3744-45; 60 Fed. Reg. 29202, 29223-24 (June 2, 1995) (proposed rule), AR 3725-30. These representations were false. (d) “Just five months after publishing the final 1995 rule, CMS requested corrected SSI records from SSA because its MEDPAR records for prior years were ‘contaminated’ [CMS’ own words] by the omission of ‘stale’ SSI records.” PRRB Dec. at 34, AR 164 (citing Provider Exhibit 64 at 1546, AR 2108 (Memorandum from HCFA (Tayloe) to SSA stating: “It has come to our attention that (SSI) files that we receive periodically from (SSA) do not contain stale records, that is, a record of any beneficiary that has died since the receipt of a check in an earlier entitlement period.”)). (e) “It is undisputed, however, that CMS has no documentation concerning whether, prior to 2003, CMS notified hospitals that SSA did not include all ‘stale records’ on the SSI tapes that it sent CMS prior to 1996 and that CMS has no record of Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 101 of 145 49 making any adjustment to hospitals’ Medicare fractions or making any payment adjustments as a result of CMS’ discovery in 1996 that SSA did not include all ‘stale records’ on the tapes it sent to CMS prior to 1996.” PRRB Dec. at 34, AR 164; see CMS Response to Provider Interrogatory 68, AR 3664. (f) “CMS witnesses were not able to explain why the agency did not correct its prior calculations of the SSI fractions that omitted ‘stale’ SSA records. Every CMS witness involved in the process, except Tayloe, said that ‘Policy’ would have made a decision as to the need to correct prior calculations, and that ‘Policy’ included Tayloe; but Tayloe alleges that she remembers nothing about the matter. PRRB Dec. at 34-35, AR 164-65; see Tayloe, Hearing Tr. at 1931-34, AR 5532- 33; Rosenberg, Hearing Tr. at 2136, AR 5583; O’Leary, Hearing Tr. at 676, AR 5212; Phillips, Hearing Tr. at 1149, AR 5215. 141. CMS’ counsel led the agency’s principal policy official on DSH matters for the periods at issue, Ann Tayloe, to testify that, “although she considers herself to be ‘a detail- oriented person,’ a DSH payment issue involving ‘a few hundred million dollars’ would ‘[n]ot necessarily’ have left a lasting impression on her, ‘given the amount of money involved.’” PRRB Dec. at 35, AR 165 (quoting Evidentiary Hearing Tr. at 142-43, AR 5062). All told, in the course of her later testimony before the Board in the 2004 hearing in this case, Tayloe stated that she “did not recall,” or words to that effect, more than 50 times over the course of testimony before the PRRB on September 22, 2004. Tayloe, Hearing Tr., AR 5518-59. Assessing Ms. Tayloe’s credibility and finding it wanting, the Board found: At the [September] 2004 Hearing, Tayloe also testified that she did not recall providing the responses to written interrogatories, between May 1 and August 15, 2004, which stated that CMS did not recall the answers to several Provider interrogatories. Tayloe, Hearing Tr. at 1934-75, AR 5533. The witness was then admonished by a Board member for being ‘a bit disingenuous’ in her testimony. Tayloe, Hearing Tr. at 1961, AR 5540. Tayloe did answer some detailed questions as to the composition of the ‘denominator’ of the SSI fraction when she testified on direct examination by the Office of the General Counsel at the 2004 Hearing. But, on cross-examination [by the Hospital’s counsel], Tayloe professed to be unable to say what types of days are included in the denominator of the fraction. Tayloe, Hearing Tr. at 2007-15, AR 5551. In the end, Tayloe testified that she did not recall what ‘denominator’ she was referring to when she testified Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 102 of 145 50 on direct examination by [the Secretary’s counsel]. Tayloe, Hearing Tr. at 2015- 16, AR 5551-5552. PRRB Dec. at 35 n.210, AR 165 (emphasis added). 142. Responding to the claim by CMS “that it would be too burdensome to revise the program to include the omitted records and recalculate the ratio based on the corrected data,” the Board found that “the evidence does not bear out the complaint. On the contrary, the evidence, even from CMS’ own witnesses, overwhelmingly supports the simplicity of correcting the problem.” PRRB Dec. at 26, AR 156. 143. The Administrator’s decision does not controvert the Board’s findings or CMS’ admissions concerning CMS’ omission of inactive SSI records, except that the Administrator disputes the Board’s findings regarding the possible financial impact of CMS’ omission of inactive SSI records from its calculations of the SSI fractions for fiscal years before 1995 and the lack of administrative burden in requiring CMS to revise its calculations to include the omitted records. See Adm’r Dec. at 43-47, AR 44-48. 144. The Administrator’s decision does not dispute that CMS omitted the inactive SSI records from its calculations of the SSI fractions for years prior to 1995, that CMS recomputed the SSI fractions for prior years using corrected SSI records in 1996, but did not use those recalculations to revise DSH payments for the prior years, that the Medicare/SSI days associated with the omitted records were not counted in the SSI fractions that were calculated before this problem was corrected, that CMS did not disclose the stale records problem to hospitals whose DSH payments for prior years were adversely affected, that (as CMS’ own witnesses admitted) the only way to know the precise financial impact of this error is to perform new calculations with all the SSI records included, that the government has sole and exclusive possession of the Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 103 of 145 51 SSI records needed to perform such a recalculation, that the Hospital and the Board sought to obtain such records through the Board’s subpoenas and the Hospital’s FOIA requests, or that, as the Board put it, CMS “resisted efforts” by the Board and the Hospital to obtain those records. See Adm’r Dec., AR 2-55. 145. Employing a series of assumptions and extrapolations from other data, the Administrator hypothesizes that CMS’ omission of “stale records” may have resulted in the omission of an average of three to four stays per hospital, per year, for fiscal years 1993 and 1994, a discrepancy the Administrator deemed too insignificant to merit correction. Adm’r Dec. at 44, AR 45. CMS’ own summary reports of its investigation in 1996 into the stale records issue reflect that, while only 48,616 stays for Medicare/SSI patients were omitted from the SSI days in CMS’ calculations for 1993 and 1994, many more stays were omitted for earlier years; for example, CMS’ summary report indicates that more than 320,000 stays were omitted for 1990 alone, an omission on the order of 13 times greater for 1990 than the omissions the agency uncovered for 1993 or 1994. See AR 2080-82; O’Leary, Hearing Tr. at 685-741, AR 5217-31. 146. Even if an average of only 3.5 patient stays were omitted from the SSI days in the numerator of the SSI fractions for each of fiscal years 1993 and 1994, as the Administrator alleges in his decision below, the financial impact of that error is significant. In the 1993 and 1994, the national average length of stay for Medicare inpatients was 7 days per stay. Answer ¶ 223. The result, therefore, of the omission of 3.5 patient stays from the numerator of the Hospital’s SSI fractions for fiscal years 1993 and 1994 would be approximately $20,335 per year (3.5 stays x 7 days per stay x $830 per day), and $40,670 for the two years combined. Id.; see supra ¶ 120. If the same financial impact were projected to the approximately 2,000 hospitals Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 104 of 145 52 that received DSH payments for the two affected years at issue, the overall financial impact of CMS’ stale records problem would be approximately $81.3 million for the two years combined, and $40.6 million for each year. Answer ¶ 223 (admitting accuracy of calculations). 147. Measured another way, the addition of an average of even 3.5 stays for Medicare/SSI patients erroneously omitted from CMS’ calculation for 1993 and 1994, with a 7- day average length of stay during this period, would add 49 Medicare/SSI patient days (i.e., 3.5 stays per year, times 2 years, times 7 days per stay) to the number of Medicare/SSI days (10,020) that were included in the numerators of the fractions that CMS computed for those Hospital fiscal years. See supra ¶¶ 14, 16, 18, 20, 131. Those 49 additional days would increase the Hospital’s average SSI ratio for those years by approximately .5% percent, from .07706 to .07744 (i.e., 10,020 SSI days per CMS’ calculations plus 49 additional SSI days associated with an average of 3.5 omitted stays, divided by 250,944 Medicare days equals an SSI ratio of .07744). Extrapolating from the Board’s calculation of a $200 million impact associated with 4.03% increase in the national average SSI ratio for 1993 alone equates to $25 million impact that would be associated with .5% increase in the SSI fractions (.5/4.03 x $200 million = $25 million). See PRRB Dec. at 24, AR 154. 148. The Administrator’s decision nowhere says that the calculation of financial impact described in his decision, or any other calculation, served as the basis for CMS’ decision in 1996 not to recalculate the SSI fractions it had previously computed in error for years prior to 1995. See Adm’r Dec. at 25-26, 42-47, AR 44 26-27, 43-44. At the hearing before the PRRB, none of CMS’ witnesses was able to state why CMS did not correct its error for the years prior to 1995. PRRB Dec. at 34-35, AR 164-165; see supra ¶ 140(f). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 105 of 145 53 149. The Administrator’s decision proclaims in effect, and incorrectly, that an underpayment to all DSH hospitals on the order of approximately $50-$80 million for fiscal years 1993 and 1994 is an “acceptable rate of error,” see Adm’r Dec. at 26, AR 27, at least when the error is CMS’ error, and when the error runs in favor of CMS and against hospitals. 150. The Secretary has expended substantial audit and enforcement resources in identifying, correcting, and recovering DSH overpayments to hospitals, Answer ¶ 227, resulting from errors in hospitals’ favor of a much smaller magnitude than $25 million. See infra ¶ 190. 2. Omission of Forced or Manual SSI Records 151. The Board found that CMS’ calculations of the SSI fractions for DSH are also flawed, and understated, because CMS’ calculations of the numerator have always omitted some SSI days attributable to Medicare beneficiaries who were receiving a forced or manual payment of SSI benefits during their inpatient hospital stay. PRRB Dec. at 26-27, AR 156-157; see Shafer, Hearing Tr. 307-08, AR 5103; Dean, Hearing Tr. 1378-81, AR 5394. The Board identified this error as a “systemic,” “common,” and “recurring error” that has the effect of reducing the SSI ratios computed by CMS. Id. 152. The Board found that, while “SSA’s system is designed to generate automated SSI payments,” sometimes “[an SSA] field office needs to make a forced or manual payment on a temporary basis.” PRRB Dec. at 26, AR 156; see Shafer, Hearing Tr. at 334-37, 474-79, AR 5110-11, 5164-66; Walsh, Hearing Tr. at 2177, AR 5594. In that case, “SSA will terminate the recipient’s existing record, start a new record that includes $0 due in the Federal amount field (in order to stop the system from making a duplicate payment), then terminate the second record and start a third to resume system payments prospectively after the manual payment action is Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 106 of 145 54 resolved.” PRRB Dec. at 26, AR 156; see Shafer, Hearing Tr. at 334-37, 474-79; AR 5110-11, 5164-66; Walsh, Hearing Tr. at 2177, AR 5594. 153. The Board further found: “In these cases [involving forced or manual payment], the earlier records, which would reflect a C01 [continuing payment eligibility] or M01 [manual payment] in the CMPH [computation history] field and some amount due in the FAM [Federal amount] field, will always be terminated, and thus would never have been included in the annual tapes that SSA sent to CMS before 1996. The later records, which Cribbs’ program did access for the data file, would show $0 due in the FAM field for periods in which a forced payment was made by a field office.” PRRB Dec. at 26, AR 156; see Shafer, Hearing Tr. at 341-44, 474-79, AR 5112, 5164-66; Cribbs, Hearing Tr. at 165-66, 202-03, 220-23, 250-58, AR 5068, 5077, 5081-82, 5089-91; AR 4029-39. “As a result,” the Board found, “individuals who received a forced payment during the periods of their inpatient hospital stays were not shown as having been entitled to SSI for such periods on SSA annual tapes.” PRRB Dec. at 26, AR 156; see Shafer, Hearing Tr. at 474-79, AR 5164-66; Cribbs, Hearing Tr. at 165-66, 202-03, 220-23, 250- 58, AR 5068, 5077, 5081-82, 5089-91. 154. The Board found that this systemic problem with manual or forced pay situations “was not fixed in 1996 when SSA began including all active and inactive records on the annual SSI tapes. As [CMS witness] Dean testified, CMS’ matching program eliminates duplicate records on SSA’s annual tapes. Thus, even after SSA started sending all active and inactive records, beginning in 1996, two of the three records generated in a forced pay situation would be eliminated by CMS, leaving only about a one-in-three chance of a valid match with the correct Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 107 of 145 55 SSI record.” PRRB Dec. at 26-27, AR 156-57 (emphasis added); see Dean, Hearing Tr. at 1379- 81, AR 5394. 155. In rejecting CMS’ allegation that it would be administratively difficult to correct this problem, the Board found: “[T]his problem could easily be fixed.” PRRB Dec. at 27, AR 157; see Shafer, Hearing Tr. at 332-34, 339-50, AR 5109-10, 5110-14. Specifically, based in part on the uncontroverted testimony of longtime SSA employee Shafer, the Board found that “the signal for a forced pay situation is a C01 or M01 code in the CMPH field and $0 due in the FAM and State amount (‘SAM’) fields. This situation only occurs when an individual is receiving a forced payment or non-cash benefits under section 1619(b) of the Social Security Act (these latter benefits are discussed further below). SSA’s computer program, therefore, could easily be written to create a ‘loop’ to go back and check an individual’s earlier records whenever it comes across a C01 or M01 in the CMPH field and no amount due in the FAM and SAM fields.” PRRB Dec. at 27, AR 157; see Shafer, Hearing Tr. at 339-48, AR 5110-13. 156. Responding to the argument by CMS that its omission of forced or manual payment cases “is too small to warrant retroactive correction,” the Board found that “Shafer’s testimony is credible that forced pay situations are common and concludes that omission of forced pay records is a systemic error that would likely deflate the DSH Medicare percentage. For the same reasons as discussed in the section on stale records, the Intermediary’s arguments regarding minimal impact and administrative burden are without merit.” PRRB Dec. at 27, AR 157. 157. The Administrator’s decision does not controvert the Board’s findings that CMS’ calculations of the SSI fractions for DSH have not counted SSI days attributable to Medicare Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 108 of 145 56 patients who received manual or forced payments of SSI benefits, nor does it controvert the Board’s findings concerning the lack of administrative burden in fixing this problem. See Adm’r Dec. at 41, AR 42. The Administrator’s decision does not dispute that the only way to know the precise financial impact of this error is to perform new calculations with the all SSI records included, that CMS and SSA have sole and exclusive possession of the records needed to perform such a recalculation, that the Hospital and the Board sought to obtain such records through the Board’s subpoenas and the Hospital’s FOIA requests, or that, as the Board put it, CMS “resisted efforts” by the Board and the Hospital to obtain those records. See Adm’r Dec., AR 2-55. 158. Without proffering evidence of his own, the Administrator asserts that “the Provider has failed to demonstrate that the data used by CMS for the calculation of the Medicare fraction adversely impacted the Provider’s reimbursement due to the omission of SSI days because of manual or forced payments to [sic] that the CMS calculation was improper.” Adm’r Dec. at 41, AR 42. 159. Extrapolating from the one stay relating to forced payment case that the Hospital determined had been omitted from CMS calculations for periods at issue, based on the Hospital’s review of the limited SSI records it was able to obtain in this case despite CMS’ resistance, the Administrator determined an error rate of “.15 percent,” and concluded, without more, that this represented an “acceptable rate of error.” Adm’r Dec. at 25-26, 41, AR 26-27, 42. 160. Insofar as the financial impact of a .75 percent error rate with respect to errors in CMS’ match process would equate to an underpayment to disproportionate share hospitals of approximately $400-$700 million per year, see supra ¶¶ 120-22, a .15 percent error rate Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 109 of 145 57 stemming from CMS’ omission of manual or forced payment cases produces a further underpayment of at least $80-$140 million per year (or one-fifth of the impact of .75 error rate) for all DSH hospitals nationwide. See Answer ¶ 238. 3. Omission of Retroactive Determinations and Hold and Suspense Cases 161. The Board found that CMS’ calculations of the SSI fractions for DSH are also systemically flawed, and understated, because CMS’ calculations have always omitted from the numerator some days attributable to Medicare beneficiaries whose SSI benefits were either granted or restored retroactively after CMS calculated the SSI fraction for a given Federal fiscal year or were temporarily on hold or suspense when CMS calculated the SSI fraction for a given Federal fiscal year but were subsequently reinstated retroactively, back to the period of the patients’ inpatient hospital stays. PRRB Dec. at 27-29, 32-33, AR 157-59, 162-63. The Board further concluded that there is no administrative burden imposed upon CMS in requiring CMS to fix this additional problem with its calculations. PRRB Dec. at 28-29, 32-33, AR 158-59, 162- 63; see Shafer, Hearing Tr. at 307-08, 330-31, AR 5103, 5109; Pfeil, Hearing Tr. at 1368-94, AR 5441-48. 162. Concerning hold and suspense cases, the Board found that the “omission of hold and suspense records is a systemic error that would likely deflate the DSH Medicare percentage” and that the arguments by CMS and the Intermediary about the administrative burden of fixing this issue are “without merit.” PRRB Dec. at 28, AR 158; see Shafer, Hearing Tr. at 362-64, AR 5117. Regarding the cause of this systemic problem in CMS’ calculation of the SSI fractions, the Board found: As discussed above, if there is a positive amount due in the FAM field, but any code other than C01 or M01, such as a hold or suspense code, appears in the CMPH field, then SSA’s program will assign a zero to that month in the annual Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 110 of 145 58 tape. In other words, the program would assign a zero to a month if an individual’s SSI benefits are temporarily on hold or in suspense when SSA prepares its annual tape. Of the limited records the Provider was able to obtain, Master ID number 13111 shows days omitted from the SSI count because payments were in suspense when the tape was prepared, but those benefits were awarded retroactive to several months prior to the patients’ hospital stays. PRRB Dec. at 27, AR 157; see Cribbs Hearing Tr. at 154-55, 186, 266, AR 5065, 5073, 5093; Shafer, Hearing Tr. at 153, 351-64, 518-21, AR 50065, 5114-17, 5175-76; Cribbs, Evidentiary Hearing Tr. at 318-19, 345, AR 5106, 5113; Provider Exhibit 155 (SSA list of payment status codes), AR 3682-85; Provider Exhibit 191 (SSI record for individual whose SSI days were not included in CMS’ data because the individual’s SSI benefits for the period of hospitalization were in suspense when SSA created the annual tape but were reinstated retroactively one month later), AR 4086-90. 163. The Board further found that “SSA’s list of payment status codes, Provider Exhibit 155, includes several temporary hold or suspense codes.” PRRB Dec. at 28, AR 158; see Provider Exhibit 155, AR 3682-85. The Board gave the following examples of just some of several hold and suspense codes utilized by SSA: Shafer testified that one common example is suspense code S08, used when SSA is looking for a representative payee who is able and willing to accept checks on behalf of an SSI recipient. Other common examples are suspense codes S21 and H80. Code S21 is used when an individual is presumptively disabled and has received benefits on that basis for six months. After that initial six-month period, benefits are suspended until all necessary state determinations and paperwork are completed. In virtually all of these cases, benefits that are temporarily in suspense after the sixth month are eventually granted retroactively for the full period. Hold code H80 is used during the period in which an application for SSI benefits is pending a state eligibility determination. PRRB Dec. at 28, AR 158; see Shafer, Hearing Tr. at 352-56, 359-63, 518-21, AR 5114-15, 5116-17. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 111 of 145 59 164. The Board concluded that CMS’ “exclusion of SSI days associated with individuals whose benefits were temporarily on hold or in suspense can be easily fixed by using the updated SSI data in later versions that SSA provides annually for rolling 42-month periods.” PRRB Dec. at 28, AR 158; see Shafer, Hearing Tr. at 364-65, AR 5117-18. 165. The Board also found that CMS’ “omission of retroactive awards is a systemic error.” PRRB Dec. at 28-29, AR 158-59; see Shafer, Hearing Tr. at 370-76, AR 5119-20. 166. Regarding CMS’ omission of SSI days attributable to Medicare patients whose SSI benefits were granted or reinstated retroactively to include the periods of their inpatient hospital stays, the Board found that “a significant portion of all SSI benefit awards are paid on appeal.” PRRB Dec. at 28, AR 158; see Shafer, Hearing Tr. at 370-74, 488-91, AR 5119-20, 5168-69; Provider Exhibit 192 at 3382-84, AR 4093-95. “Generally,” the Board added, “appeals are pending for an average of about one year. Consequently, a Medicare hospital inpatient whose SSI benefits are awarded on appeal are not picked up in CMS’ calculation of the SSI fraction unless an appeal happens to be resolved before SSA prepares its annual tape in March following the end of the fiscal year.” PRRB Dec. at 28, AR 158; see Shafer, Hearing Tr. at 375- 76, AR 5120. 167. The Board rejected the argument by CMS that the “effect of the omission of retroactive actions is . . . likely to be a wash” when measured against possible retroactive SSI benefits terminations, because “terminations of disability benefits are seldom made retroactively and most other denials are not made retroactively. In addition, CMS’ 1996 summary report [regarding its omission of ‘stale’ SSI records] shows that the ratio of the number of stays with new SSI days added (which would have included stays by individuals whose SSI benefits were Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 112 of 145 60 granted or reinstated retroactively) to the number of stays with old SSI days deleted (which would have included stays of individuals whose SSI benefits were denied retroactively) was greater than 17:1 for most years.” PRRB Dec. at 28 AR 158; see Shafer, Hearing Tr. at 377-78, 494-98, 501-02, 513-17, AR 5121, 5169-70, 5171, 5174-75; Pfeil, Hearing Tr. at 1575-84, AR 5443-45. 168. Regarding the lack of burden to correct this problem concerning omission of patients whose SSI benefits were determined retroactively, the Board found: “The problem with CMS’ omission of retroactive awards is similar to the omission of hold and suspense cases, and the same solution applies in each case—for CMS to use the updated SSI data that SSA provides annually for a rolling 42-month period.” PRRB Dec. at 29 AR 159; see Shafer, Hearing Tr. at 378 AR 5121. 169. The Board made several specific findings, regarding CMS’ unexplained failure to use the best available, most current data for its match process: (a) “The Medicare fraction is calculated relatively early in the cost report settlement process, which generally is not completed until two or three years after the end of a hospital’s fiscal year.” PRRB Dec. at 32, AR 162; see Gouger, Hearing Tr. at 1821-22, AR 5504-05. (b) During the 2-3 year cost report settlement process, “the provider is, among other things, collecting information from the state concerning data for the other component of its DSH calculation, the Medicaid fraction. Billing for services rendered during the year continues – up to 30 months after the end of the provider‘s cost report year – and that patient data from the billing and payment process is captured in the MEDPAR file used to determine the denominator of the Medicare fraction. At the same time, the determination of Medicare patients’ eligibility for SSI may still be in process and not be reflected on the SSA tape until much later.” PRRB Dec. at 32, AR 162; see Gouger, Hearing Tr. at 1821- 22, AR 5504-05; Shafer, Hearing Tr. at 370-76, AR 5119-20; Cribbs, Hearing Tr. at 134-36, 261-62, AR 5060, 5092. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 113 of 145 61 (c) “Over two and a quarter years [after the end of Federal fiscal year], CMS performs 10 MEDPAR runs and creates 10 MEDPAR files for each Federal fiscal year, the latest of which is run in the third December after the end of the subject fiscal year (e.g., the 10th and last MEDPAR file for Federal fiscal year ended September 30, 1994 was produced in a December 1996 MEDPAR run).” PRRB Dec. at 32, AR 162; see Dean, Evidentiary Hearing Tr. at 64-71, AR 5724-5726; AR 5725-5726, 5730. (d) “CMS matches its inpatient hospital stay records against the most recent annual SSA tape every time it creates one of these quarterly MEDPAR runs.” PRRB Dec. at 32, AR 162 (citing Dean’s undisputed testimony at Dean, Evidentiary Hearing, Tr. 67-69, 87, AR 5725-5726, 5730). (e) “Despite having MEDPAR runs that are updated to the time of the final cost report settlement,” two–three years after the end of fiscal year, “it is undisputed that CMS used the MEDPAR file based on the month of June following the end of the Federal fiscal years to compute the SSI fractions at issue for each year under appeal. For example, CMS used the June 1996 MEDPAR run to calculate the SSI fraction for Federal fiscal year ended September 30, 1995.” PRRB Dec. at 32, AR 162; see CMS Response to Interrogatory 15, AR 3631-32. Therefore, the calculations performed for fiscal year 1995 in September 1996, December 1996, and March, June, September and December of 1997 were not used to determine the DSH payment made for fiscal year 1995, even though the NPR for 1995 was issued later. Id.; see also Notice of Program Reimbursement, dated Sept. 25, 1997, AR 502. (f) “Even though MEDPAR is run quarterly, the SSA tape is run only once annually, and CMS always uses the annual run from the year following the end of the federal fiscal year. It is also undisputed that, because CMS always receives the annual SSA tape before April 1st, CMS’ inpatient hospital stay records are never matched with SSI records relating to individuals whose SSI benefits are granted or restored after April 1 retroactively to a period within the prior Federal fiscal year.” PRRB Dec. at 32, AR 162 (emphasis added). 170. For the foregoing reasons, the Board concluded: “Given the process actually used which fixes the Medicare fraction in June following the end of the FFY despite there being ongoing processes that affect the numerator and the denominator, it is inevitable that the data used to calculate the Medicare fraction is incomplete and not the ‘best available data.’” PRRB Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 114 of 145 62 Dec. at 33, AR 163; see Phillips, Hearing Tr. at 1240-41, AR 5359; Gouger, Hearing Tr. at 1817- 22, AR 5503-05. 171. The Administrator’s decision does not controvert the Board’s findings concerning CMS’ omission of hold and suspense cases and retroactive determinations. See Adm’r Dec. at 37-41, AR 38-42. Significantly, the Administrator does not dispute that CMS’ calculations of the SSI fractions have always omitted hold and suspense cases and retroactive determinations, that the Medicare/SSI days associated with the omitted records were not counted in the SSI fractions, that the only way to know the precise financial impact of these errors is to perform new calculations with the all SSI records included, that CMS and SSA have sole and exclusive possession of the records needed to perform such a recalculation, that the Hospital and the Board sought to obtain such records through the Board’s subpoenas and the Hospital’s FOIA requests, or that, as the Board put it, CMS “resisted efforts” by the Board and the Hospital to obtain those records. 172. The Administrator’s decision asserts: “CMS’ policy recognizes that not all retroactive determinations made on SSI entitlement (whether granting or denying) would be included in the SSI record forwarded to CMS. However, CMS has determined that such actions would have a minimal impact on a provider’s Medicare DSH fraction which combined with the need for finality in the process, makes the use of later data impractical and no more accurate.” Adm’r Dec. at 37, AR 38. Though the Administrator’s decision implies an established “policy” not to account for “retroactive determinations” in CMS’ calculations of the SSI fractions for DSH, the Administrator’s decision is conspicuously silent as to who established this alleged policy, when it allegedly was determined, what information allegedly was considered in reaching Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 115 of 145 63 this determination, or, most fundamentally, what evidence in the record reflects this alleged “policy” determination. Rather, the Administrator’s decision on this point rests exclusively on statements published by CMS in the 2005 Federal Register, several months after the Board heard this case and several years after the periods at issue. Adm’r Dec. at 38-39, AR 39-40 (quoting from 70 Fed. Reg. 47278, 47439-40). In addition, the statements in the 2005 Federal Register notice upon which the Administrator’s decision relies (at 38-39, AR 39-40) are unsupported by the record evidence in this case, are clearly incorrect, or both: i) The Administrator’s decision, in quoting from the 2005 notice, asserts that “there needs to be administrative finality to the calculation of a hospital’s Medicare [SSI] fraction;” but, neither the Administrator’s decision nor the 2005 notice acknowledge or address the Board’s undisputed findings that CMS performs 10 MEDPAR/SSI file matches for every Federal fiscal year, including several that are performed more than one and one-half years after the calculation of the SSI fractions that are applied for DSH payment purposes. ii) The Administrator’s decision, in quoting from the 2005 notice, asserts that “CMS has previously stated that its goal is to obtain reasonably accurate but not perfect calculations (51 Fed. Reg. 16777),” Adm’r Dec. at 38, AR 39; but, neither the Administrator’s decision nor the 2005 notice upon which it relies upon acknowledges or addresses the fact that the 1986 Federal Register notice cited in this portion of the Administrator’s decision concerned only CMS’ determination to calculate the SSI fractions on a Federal fiscal year basis and made no mention whatsoever of a policy determination of any sort having to do with retroactive determinations of SSI benefits. iii) The Administrator’s decision, in quoting from the 2005 notice, asserts that “our data have shown that 98 to 99 percent of SSI eligibility determinations are made and remain unchanged 6 months after the end of the Federal fiscal year,” Adm’r Dec. at 38, AR 39; but, neither the Administrator’s decision nor the 2005 notice upon which it relies cites to any supporting evidence. iv) The Administrator’s decision, in quoting from the 2005 notice, asserts that “[t]here will be a minimum of 6 months between the end of the hospital’s cost reporting period and April 1 date [sic] that we have receive [sic] SSI eligibility information.” Adm’r Dec. at 38, AR 39. Six months, however, is the maximum, not the minimum, time lag between the end of a hospital cost reporting period and CMS’ early, April 1 cut-off date. The Administrator’s assertion that six months is Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 116 of 145 64 the minimum time lag stands on the unstated premise that the SSI fractions computed for a Federal fiscal year apply to hospital cost reporting periods ending within the federal fiscal year. But, that assumption is not true. The fractions computed by CMS for each Federal fiscal year apply to cost reporting periods beginning in, and not those ending in, the Federal fiscal year ending September 30. See 42 C.F.R. § 412.106(b)(2). Although CMS performs many additional matches after the April 1st cut-off date, none of this more current data is used or reflected in CMS’ calculation of SSI fractions applied to determine the DSH payments. PRRB Dec. at 32-35, AR 162-65. v) The Administrator’s decision, in quoting from the 2005 notice, asserts that “[t]he time lag between the close of a hospital’s cost reporting period and the April 1 date that we receive eligibility information could actually be much longer for many hospitals,” Adm’r Dec. at 38, AR 39, but, again, this allegation is false and misleading. For the reasons noted in the preceding sub-paragraph, six months is the maximum, not the minimum, time lag. Likewise, while the Administrator’s decision and the 2005 notice state that “we will be using SSI eligibility from 17 months after a hospital’s year ends with a November 1 to October 31 cost reporting period,” Adm’r Dec. at 38, AR 39, that statement is also false and misleading. For example, it implies that the SSI fraction that would be applied to a cost reporting period beginning on November 1, 1993 is the fraction computed by CMS for the Federal fiscal year that ended on the prior September 30, 1993. But, in fact, the fraction that would be applied to a cost reporting period beginning November 1, 1993 and ending October 31, 1994 is the fraction computed by CMS for the Federal fiscal year that began on the prior October 1, 1993 and ended on September 30, 1994. See 42 C.F.R. § 412.106(b)(2). CMS calculated the SSI fraction for that year in June 1995. PRRB Dec. at 32, AR 162. vi) The Administrator’s decision, in quoting from the 2005 notice, asserts: “Given the time between the end of hospital’s [sic] cost reporting periods and when we are furnished with SSI eligibility information for that period [sic], we believe it is highly unlikely that a subsequent data run will produce data that is significantly different than one completed 6 months after the end of the Federal fiscal year[].” Adm’r Dec. at 38, AR 39. This allegation is conjecture, premised on the false assumptions described in the preceding sub-paragraphs. It leads one to question why, in October 1999, CMS recalculated the SSI fraction for all hospitals for fiscal year 1998, a revision that reduced the Provider’s DSH payment for 1998 by nearly $35,000. See Smith, Hearing Tr. at 957–63, AR 5288–89; Provider Ex. 216, AR 413–16; Provider Ex. 232, AR 4383–85. 173. The Administrator’s decision alleges that “the Provider has failed to demonstrate by a preponderance of the evidence that such omissions [of hold and suspense cases] have had Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 117 of 145 65 any impact on its DSH reimbursement.” Adm’r Dec. at 39-41, AR 40-42. Extrapolating from the one stay relating to retroactive SSI determination that the Hospital determined was omitted from CMS calculations of the Hospital’s SSI fractions for periods at issue, based on the Hospital’s review of the limited SSI records that it was able to obtain in this case despite CMS’ resistance, the Administrator determined an error rate of “1/6 of one percent” (or .1667%) and again concluded, without more, that this represented an “acceptable rate of error.” Adm’r Dec. at 25-26, 39, AR 26-27, 40. As shown in paragraph 160 supra, a .1667 percent error rate stemming from CMS’ omission of hold and suspense cases and retroactive determinations would produce a further underpayment to disproportionate share hospitals nationally of approximately $89-$156 million per year. 4. Omission of Non-Cash SSI Benefits 174. The Board found that CMS’ calculations of the SSI fractions for DSH are also flawed, and understated, because CMS’ calculations of the numerator have always omitted days attributable to Medicare beneficiaries who were receiving non-cash SSI benefits under title XVI of the Social Security Act during their inpatient hospital stays. PRRB Dec. at 29-30, AR 159-60; see Shafer, Hearing Tr. at 341, 381-82, AR 5112, 5122. The Board specifically found that CMS’ categorical omission of these patient days “is a systemic error.” PRRB Dec. at 30, AR 160. The Board found that, “certain qualifying disabled recipients are considered to be entitled to Federal SSI benefits even though no SSI cash payment is made, both for purposes of allowing the recipient to continue receiving Medicaid coverage and to resume receiving the SSI cash benefit when the individual is unable to work, without having to reapply.” Id. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 118 of 145 66 175. The Board found that all days associated with individuals who received non-cash SSI benefits always have been categorically omitted from CMS’ calculations of the SSI fractions for all years and that this problem can be fixed without imposing an administrative burden: SSA’s program for the annual SSI tape would assign a zero to any month in which an individual was working and not receiving a cash payment, but was considered to be receiving SSI under the 1619(b) work incentive program. As in the case of a forced payment the signal for someone who is receiving non-cash benefits under section 1619(b) of the Social Security Act is the presence of a C01 payment status code in the CMPH field and $0 due in the FAM field on the SSI master record. Thus, this problem, like the forced pay problem, can be fixed by putting a loop in the SSA program to go back to earlier records and check for a 1619(b) indicator. PRRB Dec. at 29-30, AR 159-60; see Shafer, Hearing Tr. at 341, 381-82, AR 5112, 5122. 176. The Administrator’s decision does not controvert the Board’s findings that CMS’ calculation of the SSI fractions for periods at issue categorically omitted all SSI patient days attributable to individuals who received non-cash SSI benefits under Title XVI of the Social Security Act. See Adm’r Dec. at 37-41, AR 38-42. The Administrator reversed the Board’s ruling on the incorrect ground that “all of the regulations relied upon by the Board in this case expressly state that they are applicable only for purposes of Medicaid eligibility.” Adm’r Dec. at 49, AR 50. The Administrator’s decision ignores the last sentence of the SSA regulation upon which the Board relied, which states that the “Special SSI eligibility status” afforded under Section 1619(b) of the Social Security Act “also applies for purposes of reacquiring status as eligible for regular SSI benefits or special SSI cash benefits” under the SSI program. 20 C.F.R. § 416.264. F. CMS’ Allegedly Acceptable Rate of Error 177. The Board found that even if the Medicare DSH statute permitted an approximation of the SSI fraction based on the best data available to CMS when it performed its Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 119 of 145 67 calculations, CMS did not, in fact, use the best data available when it calculated the Hospital’s SSI fractions for the periods at issue. PRRB Dec. at 9, AR 139. 178. The PRRB issued a subpoena in 2002 requiring CMS to produce systems documentation for the programs and files that were used to produce the SSI fractions computed by CMS for the fiscal years at issue. The Secretary claimed that CMS did not maintain such documentation. Subpoena, AR 7104-7110; see PRRB Dec. at 2, AR 134. 179. The Board also ordered CMS to produce documentary evidence reflecting any data specifications or systems requirements it gave to SSA regarding the files prepared by SSA for CMS in connection with CMS’ calculations of the SSI fractions for the years at issue and any efforts by CMS to test and validate the files and programs that were used to calculate those SSI fractions. PRRB Dec. at 2, AR 134; see CMS Response to Provider Interrogatory 45, AR 3650; Dean, Evidentiary Hearing Tr. at 125-27, AR 5740. 180. The PRRB found that the Secretary produced no documentation of any meaningful specifications or requirements given to SSA or of any meaningful efforts to test and validate the files and programs that CMS used to calculate the SSI fractions. See PRRB Dec. at 32-34, AR 162-64. 181. Pat Cribbs, a former long-time SSA employee who prepared the SSI data tapes used in CMS’ calculations, testified that CMS gave SSA only “general” and “vague” instructions concerning the SSI data that CMS needed. Evidentiary Hearing Tr. at 264, 269, 350-51, AR 5679, 5681, 5701. Essentially, she said, the SSI data tapes used in CMS’ calculations were prepared “on the fly.” Hearing Tr. at 111-12, AR 5054; Hearing Tr. at 239-40, AR 5086. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 120 of 145 68 182. A CMS employee who was involved in the calculations of the SSI fractions for the fiscal years at issue, Rosenberg, testified that she “eyeballs” the SSI fractions calculated by CMS for each fiscal year, and she acknowledged the insufficiency of this effort, noting the “old saying” that, as with any computer program, “garbage in” will produce “garbage out.” Rosenberg, Hearing Tr. at 2104-06, 2146-47, AR 5575-76, 5586. 183. Despite the size and importance of the DSH payment adjustment – a multi-billion dollar program annually – the Board found that the Secretary did not comply with the basic federal data management directives and guidelines applicable to all federal agencies, including requirements for validation, verification, and testing of software applications. PRRB Dec. at 30- 32, AR 160-162. 184. The Board also found that “it is disingenuous for CMS to take a position that the Provider must demonstrate what financial impact the flaws in the process caused while knowing that its own record retention failures make such a showing impossible.” PRRB Dec. at 32, AR 162. 185. The Administrator’s decision concludes that “despite the[] alleged problems, the policies and methods used by CMS resulted in an acceptable rate of error, to the extent of being an almost nonexistent rate of error.” Adm’r Dec. at 26-27, AR 27-28. 186. The Administrator’s “acceptable rate of error” may amount to underpayments to hospitals serving a disproportionate share of low-income Americans on the order of hundreds of millions of dollars per year for the periods at issue. See supra 120-122, 146-47, 160, 173. To recap the financial impacts of CMS’ errors and omissions in the calculation of the numerator of Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 121 of 145 69 the SSI fraction alone, id., the system-wide underpayments per year with respect to the errors and omissions shown by the uncontroverted record evidence amount to the following: Error FY 1993 FY 1994 FY 1995 FY 1996 Flaws in the Match $400-$700 million $400-$700 million $400-$700 million $400-$700 million Omission of Stale Records $25 million $25 million 0 0 Omission of Manual Payments $80-$140 million $80-$140 million $80-$140 million $80-$140 million Omission of Hold, Suspense and Retroactive Determinations $89-$156 million $89-$156 million $89-$156 million $89-$156 million Omission of Non-Cash Benefits Not Determined Not Determined Not Determined Not Determined Total Per Year (Not Counting Omitted Non-Cash Benefits) $594 million- $1.036 billion $594 million- $1.036 billion $569-$996 million $569-$996 million Total for Years 1993-1996: $2.3-$4.1 billion 187. The economic impact of CMS’ errors would be substantially greater if the foregoing assessment included the financial impact on hospitals that nearly missed qualifying for DSH, based on CMS’ calculations of the SSI fractions for the periods at issue, and would qualify if CMS’ calculations of the SSI fractions were correct. See PRRB Dec. at 25-26, AR 155-56. 188. The Administrator’s decision does not elaborate on what, if any, rate of error is considered by the Secretary to be acceptable for Medicare payment purposes generally or for the Medicare DSH payment in particular. See Adm’r Dec. at 25-26, AR 26-27. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 122 of 145 70 189. Though the Administrator found that CMS achieved an “acceptable rate of error” in its calculations of the SSI fractions, agency policy concerning other aspects of the DSH calculation, particularly with respect to the Medicaid fraction, does not allow hospitals to use estimates for the calculation. See Gouger, Hearing Tr. at 1816, AR 5503. CMS’ witnesses in this case confirmed that CMS does not permit “estimates” for the numbers of days that are counted in the Medicaid fraction for DSH. See PRRB Dec. at 8, AR 138; see Gouger, Hearing Tr. at 1814-16, AR 5503; Phillips, Hearing Tr. at 1237-39, AR 5358-59. CMS’ witnesses also confirmed that CMS does not permit “estimates” for another proxy (the interns-and-residents-to- beds ratio) that is used to determine another prospective payment system add-on, the indirect medical education adjustment. Like the DSH payment, this additional payment is determined through the cost report settlement process long after the end of hospital’s fiscal year. PRRB Dec. at 8, AR 138; Phillips, Hearing Tr. at 1237-39, AR 5358-59; Gouger, Hearing Tr. at 1814-15, AR 5503. 190. The Secretary’s Office of Inspector General (OIG) has concluded in at least three audit reports released in the last three years that DSH claims or payment determinations should be recalculated and revised when the number of Medicaid days included in the Medicaid fraction was overstated, with respect to discrete errors, by as few as 7 days and with a financial impact of as little as $2,247, and when the SSI ratio was misstated in a hospital’s favor by as little as 1.8%. OIG, Review of Gwinnett Medical Center's Claim for Medicare Disproportionate Share Hospital Payments for Fiscal Year 1999 (2003), available at http://oig.hhs.gov/oas/reports/region4/40202021.pdf; OIG, Review of Medicare Disproportionate Share Hospital Payments for Methodist Hospital-Memphis for Fiscal Year 1999 (2003), Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 123 of 145 71 available at http://oig.hhs.gov/oas/reports/region4/40302023.pdf.; OIG, Review Of Medicare Disproportionate Share Hospital Payments Claimed By Saint Mary’s Regional Medical Center for Fiscal Year Ending December 31, 2002 (2004), available at http://oig.hhs.gov/oas/reports/region1/10400506.pdf. G. CMS’ Count of Medicare Days in MEDPAR 191. The PRRB found that the count of Medicare days in CMS’ determination of the denominator of the SSI fraction is unreliable. PRRB Dec. at 35-40, AR 165-70. The Administrator reversed this finding, though there is no dispute that CMS’ count of Medicare days is not correct and that CMS cannot explain the discrepancies. See id.; Adm’r Dec. at 28-35, AR 29-36. 1. Unexplained Discrepancies in CMS’ Count 192. The Board found “compelling evidence” in the record showing that CMS’ “count of total Medicare days for the denominator [of the SSI fraction] is unreliable.” PRRB Dec. at 38, AR 168. That evidence relates to a reconciliation of CMS’ MEDPAR data and the data in the Intermediary’s Part A Provider Statistical & Reimbursement Report (“PS&R”), which was performed in 2003 by agreement of the Hospital and the Intermediary for only two of the four fiscal years at issue – 1994 and 1995. PRRB Dec. at 35-38, AR 165-168. 193. The Board found that the PS&R report, which is generated by the Medicare fiscal intermediary, “contains the details of the days for which a provider is paid by Medicare.” PRRB Dec. at 35. AR 165. Providers, like the Hospital, “are required to reconcile their cost report submission to the paid claims data furnished in the PS&R.” PRRB Dec. at 35, AR 165. The Board further found that, “[i]n most instances MEDPAR data and PS&R data would be the same; Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 124 of 145 72 i.e., both would show the days Medicare covered and paid for inpatient services for each Medicare beneficiary.” PRRB Dec. at 35, AR 165. 194. The Board also found that the “PS&R is unlike the MEDPAR in that the PS&R does not purport to contain days for which Medicare provided coverage and charged the beneficiary with ‘utilization’ of his/her available Medicare covered days but, for some technical reason, Medicare did not pay the hospital for the care. This can occur, for example, because the provider failed to timely bill Medicare for the services.” PRRB Dec. at 35, AR 165; see, e.g., Smith, Hearing Tr. at 889-96, AR 5271-72. 195. The Board found that in 2003, the Hospital and the Intermediary agreed to reconcile the differences between the Intermediary’s PS&R reports and CMS’ available MEDPAR files for two of the four fiscal years at issue, 1994 and 1995. PRRB Dec. at 36 & n.211, AR 166; see Smith, Hearing Tr. at 884-85, AR 5269-70; Provider’s Response to CMS’ First Set of Interrogatories and Requests for Production of Documents at 14R-15R, 18R, AR 2909-10; CMS Responses to Provider’s Interrogatories 5.A, 10.A, 10.D, AR 3599-3600, 3611- 12, 3613-14. The Board further found that “[a] detailed summary of the data reflected on the PS&R, the MEDPAR, and the CWF documents produced by the Intermediary for each of the above stays [for 1994 and 1995] is set forth in paragraphs 12.1 - 13.15 and the accompanying tables in Appendices I and II in the Provider’s Position Paper. There is no dispute as to accuracy of the information presented in the tables in Appendices I and II in the Provider’s Position Paper.” PRRB Dec. at 36, n.211, AR 166; see Provider’s Position Paper, Appendices I – II, AR 469-77; Gouger, Hearing Tr. at 1771, AR 5492. Reconciliation of the Intermediary’s PS&R Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 125 of 145 73 reports and CMS’ available MEDPAR files for Hospital fiscal years 1994 and 1995 identified the following discrepancies: • 61 inpatient hospital stays that had Medicare paid days on the PS&R that were not included in the MEDPAR data; and, • 47 stays that had Medicare days in the MEDPAR that were not included as paid Medicare days on PS&R. PRRB Dec. at 36, AR 166; see Smith, Hearing Tr. at 884 AR 5269; Provider’s Response to CMS’ First Set of Interrogatories and Requests for Production of Documents at 14R-15R, 18R, AR 2909-10; CMS Discovery Responses to Provider’s Interrogatories 5.A, 10.A & 10.D, AR 3599-3600, 3611-12, 3613-14. 196. With respect to the 61 patient stays that were reflected as Medicare-paid days on the Intermediary’s PS&R reports but were not included in CMS’ MEDPAR files for 1994 and 1995, the Board found: (a) “57 [of the 61] stays had been paid before CMS ran the June MEDPAR files that were used to calculate the SSI fractions at issue for fiscal years 1994 and 1995”; (b) “46 of these 61 stays had utilized days charged on the CWF” [“Common Working File”]; (c) “CMS cannot explain why the Medicare paid days for these [46] stays are not included in the MEDPAR file”; (d) “As to the remaining 15 of these 61 stays, CMS’ answers to interrogatories state that the days associated with these stays were properly excluded from the MEDPAR because ‘utilized’ days were not, or should not have been counted in the CWF from which the MEDPAR report is drawn” (emphasis added); (e) But, “while most of the [15] stays had payments that were cancelled, the CWF continued to reflect ‘utilized’ days for some stays after payments were cancelled and some of the stays simply were not found on the CWF”; (f) And, “[t]he Intermediary does not know from which CWF field MEDPAR pulls the data, or whether the MEDPAR even looks to utilized days at all.” Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 126 of 145 74 PRRB Dec. at 36, AR 166; see Smith, Hearing Tr. at 887-89, 895, AR 5270-71; Gouger, Hearing Tr. at 1721, 1773-77, 1796, AR 5479, 5492-93, 5498; CMS Response to Provider’s Interrogatory 10.A; AR 3611-12; Provider’s Position Paper, Appendix I, Tables 12.5, 12.8, 12.15, AR 469- 470, 472. 197. Based on the “discrepancies that the Intermediary was unable to explain,” as described in the preceding paragraph, the Board concluded that “the MEDPAR count of total Medicare days for the denominator is unreliable.” PRRB Dec. at 38, AR 168. 198. In addition to unexplained discrepancies in the count of Medicare days in CMS’ MEDPAR files, the Board also found that CMS contradicted itself even as to the question of what days even should be counted as Medicare days in MEDPAR. For example, the Board noted that CMS reversed its position multiple times concerning its policy on whether the denominator of the SSI fraction should include days that are not paid by Medicare pursuant to the Medicare secondary payer provisions of the Act: [while] 13 Medicare secondary payer (“MSP”) stays . . . were counted as Medicare days in the MEDPAR to the extent that Medicare actually made payment[,] . . . [d]uring the course of this case, CMS reversed its answers to interrogatories asking whether MSP days should be counted in the denominator of the SSI fraction, and at the hearing, CMS’ witness, Phillips, initially disagreed with CMS’ revised interrogatory responses but then recanted his testimony. PRRB Dec. at 39 & n.226, AR 169; Phillips, Hearing Tr. at 1089-92, 1158-60, AR 5321-39; CMS Response to Provider’s Interrogatory 1.F, AR 3593, 3596. 199. The Board found, based on record evidence, that CMS’ “[t]reatment of certain categories of days,” like the Medicare secondary days, “sheds some light on MEDPAR’s deficiencies.” PRRB Dec. at 38, AR 168. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 127 of 145 75 200. The Administrator’s decision does not controvert the Board’s findings regarding the unexplained discrepancies between the Intermediary’s PS&R report, and CMS’ MEDPAR file, and CMS’ own contradictory statements as to what days should be counted in the MEDPAR file. See Adm’r Dec. at 28-31, AR 29-32. 201. The Administrator’s conclusion that the “MEDPAR data . . . was the best data available” is based, in part, on the Administrator’s allegation, unsubstantiated by citation to any record evidence, that “[t]here was no difference between the MEDPAR and the PS&R for FYEs 1993 and 1996.” Adm’r Dec. at 31, AR 32. As set forth above, the undisputed record evidence shows that the Intermediary and the Hospital agreed to reconcile only two of the four fiscal years at issue and that the files for 1993 and 1996, in contrast, were simply not reviewed for variances. See supra ¶ 195. 202. The Administrator’s conclusion that the “MEDPAR data . . . was the best data available” is also based, in part, on the Administrator’s allegation that “[t]here were approximately 35,000 stays on the PS&R for all four years involved” including the two years that the Hospital and the Intermediary did not attempt to reconcile. Adm’r Dec. at 31, AR 32. 203. The Administrator’s conclusion that the “MEDPAR data . . . was the best data available” is also based, in part, on the Administrator’s allegation, also unsupported by citation to record evidence, that “approximately 99.82 percent of the stays were accepted by both parties as accurately represented.” Adm’r Dec. at 31, AR 32. Again, the Administrator is incorrect in suggesting that the parties had reconciled the MEDPAR and PS&R files for 1993 and 1996, as well as 1994 and 1995, and had found no variances for 1993 and 1996. By agreement, the parties did not attempt to reconcile 1993 or 1996. PRRB Dec. 35-36, AR 165-66; Smith, Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 128 of 145 76 Hearing Tr. at 883-85, AR 5269-70. Further, with respect to the variances for 1994, the only data available from CMS to be reconciled for that year was not the data that went into the SSI fraction that was actually applied to determine the Hospital’s DSH payment for 1994 but rather was from a recalculation of the SSI fraction for 1994 that CMS performed later. See Smith, Hearing Tr. at 874-79, AR 5267-68; PRRB Dec. at 35, AR 165. That later calculation for 1994 included more than 2,600 additional Medicare patient days, for a total of 66,887 Medicare days. AR 5267; Smith, Hearing Tr. at 875-76. If a reconciliation had been performed between the PS&R and the number of Medicare days counted in the SSI fraction that was applied to determine the Hospital’s DSH payment for 1994 (64,251 Medicare days), see supra ¶ 16, the error rate would be approximately 23 times greater than what the Administrator’s decision suggests (66,887 – 64,251 = 2,636 / 64,251 = 4.10% error rate). In short, the Administrator’s decision attempts to use later, “corrected” data, which he did not use and which he refuses to use to determine the Hospital’s DSH payment for 1994, to determine the error rate in the very different data he did use is “acceptable.” 2. Days That Should Not Be Counted in the Denominator of the SSI Fraction a. Exhausted Benefit Days 204. The Board ruled that the denominator of the SSI fraction should include all “covered or utilized” days regardless of whether the patient was entitled to have Medicare benefits paid on his or her behalf for such days, and despite the Board’s finding that “CMS’ past position” on this point is “confusing.” PRRB Dec. at 37, AR 167. 205. The reconciliation of the PS&R and MEDPAR for Hospital fiscal years 1994 and 1995 showed that inpatient days occurring after a patient had exhausted Medicare Part A benefits were not included in MEDPAR’s count of Medicare days for the SSI fraction. Smith, Hearing Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 129 of 145 77 Tr. at 884, AR 5269; see Provider’s Response to CMS’ First Set of Interrogatories and Requests for Production of Documents, at 14R-15R, 18R, AR 2909-10; CMS Discovery Responses to Provider’s Interrogatories 5.A, 10.A & 10.D, AR 3599-3600, 3611-12, 3613-14. 206. CMS’ prior policy clearly established that a patient is not entitled to Part A benefits after he or she exhausted benefits for inpatient hospital services, as stated by the Secretary in a 1990 Federal Register notice. 55 Fed. Reg. 35,990 35996 (Sept. 4, 1990). 207. Despite whatever else CMS may have said in the 1980s, when CMS switched from using the PATBILL (patient billing) file to using the MEDPAR, without advance notice or opportunity for public comment, the PATBILL file included a field (not found in MEDPAR) consisting of “Cost report days,” which were defined as consisting of “days credited to the provider’s PS&R report as Medicare Days,” as stated in § 3870 of the Medicare Intermediary Manual. Medicare Intermed. Man. (CMS Pub. 13-1) § 3870, available at http://www.cms.hhs.gov/Manuals/; then follow “Paper-Based Manuals” hyperlink. 208. When the Secretary promulgated the original DSH regulation in 1986, he said that the Medicaid fraction would include only Medicaid-paid days and that the agency was construing the SSI fraction consistently to include only Medicare-paid days. 51 Fed. Reg. at 16777 (May 6, 1986); 51 Fed. Reg. 31454, 31460-61 (Sep. 3, 1986). 209. The Secretary’s original interpretation of the DSH statute as including only Medicaid paid days in the Medicaid fraction and only Medicare paid days in the Medicare/SSI fraction was consistently advanced by the Secretary in later federal court litigation concerning the application of this paid-days only policy with respect to the Medicaid fraction. See, e.g., Legacy Emanuel Hosp. & Health Ctr. v. Shalala, 97 F.3d 1261, 1265 (9th Cir. 1996) (expressly Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 130 of 145 78 noting the Secretary’s position that “the Medicare proxy only counts patient days paid by Medicare”); AR 4137-38, 4210, 4214, 4231, 4236, 4340-43, 4345-49, 4351-54, 4356-61, 4363- 66, 4368-74. b. Medicare HMO Days 210. The Board ruled that the denominator of the SSI fraction for DSH should include days attributable to Medicare patients enrolled in a health maintenance organization (“HMO”). PRRB Dec. at 39-40, AR 169-70. 211. The 2003 reconciliation of the MEDPAR and PS&R files showed that the only Medicare HMO patient stays that were actually counted in CMS’ MEDPAR file for 1994 and 1995 – the only two years that were included in the reconciliation – were 22 Medicare HMO patient stays (accounting for only 359 days) that had been billed in error to the Intermediary for payment under regular Medicare (non-HMO) fee-for-service program and were, therefore, denied for payment by the Intermediary under that program. PRRB Dec. at 39-40, AR 169-70; see also PRRB Dec. at 36 & n.211, AR 166; Smith, Hearing Tr. at 893-95, AR 5272; Gouger, Hearing Tr. at 1796-97, AR 5498; see Provider’s Position Paper, Appendix II, Table 13.8, AR 475. 212. The undisputed record evidence, including two contemporaneous GAO reports, shows that as far back as 1987, and continuing thereafter, CMS knew that few HMO days were counted in the SSI fraction because hospitals and HMOs had little incentive and no contractual obligation to submit no-pay bills for HMO stays. PRRB Dec. at 39 and n.229, AR 169; see 1991 GAO Report, AR 4009-10; HCFA Fact Sheet, AR 4017; Phillips, Hearing Tr. at 1265-68, 1275- 80, AR 5365-66, 5368-69; 55 Fed. Reg. 35990, 35994 (September 4, 1990). Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 131 of 145 79 213. The undisputed record evidence shows that CMS acknowledged as far back as 1990 that Medicare fiscal intermediaries “do not necessarily know what to do with ‘no-pay’ bills for HMO days even when they are submitted.” PRRB Dec. at 39 (quoting from CMS’ notice in the 1990 Federal Register at 55 Fed. Reg. 35990, 35994 (September 4, 1990)). 214. The Board found that CMS’ own witness, Dean, testified “that the field on MEDPAR for HMO days ‘hasn’t been used since the time that I started running the MEDPAR [in 1995].’” PRRB Dec. at 40, AR 170; Dean, Hearing Tr. at 1387, AR 5396. “Therefore,” the Board found, “HMO days could not have been included in the SSI fraction in any case, even if a no-pay bill had been submitted.” PRRB Dec. at 40, AR 170. 215. It is undisputed that, and the Board found: “For periods [at issue] prior to 1998, HMO days were excluded from all aspects of the Medicare cost report [other than the SSI fraction for DSH] and were not counted as Medicare days in the calculation of the Medicare patient load for [graduate medical education (“GME”)] payment purposes even though the statutory definitions of the two programs are very similar.” PRRB Dec. at 40, AR 170; see also Johnson, Hearing Tr. at 977-78, AR 5293; Gouger, Hearing Tr. at 1798-99, AR 5499; Phillips, Hearing Tr. at 1293-96, AR 5372. 216. The Board found: “CMS has never explained the reason for this inconsistent treatment of Medicare HMO days.” PRRB Dec. at 40, AR 170; see 54 Fed. Reg. 40286, 40294- 95; 55 Fed. Reg. 35990, 35994 (September 4, 1990). 217. The Administrator concludes that Medicare HMO days should be included in the SSI fraction for the fiscal years at issue because “HMOs are required to cover inpatient Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 132 of 145 80 hospitalization services, and utilization is charged to the beneficiary for such days.” Adm’r Dec. at 34, AR 164. H. CMS’ Resistance to Discovery Efforts 218. The Hospital engaged in extensive discovery in an attempt to obtain from CMS and SSA the specific data elements for each patient and patient day used by CMS in calculating the DSH adjustment. See PRRB Dec. at 4, AR 134, AR 6913-15; AR 7239-54. 219. The PRRB issued subpoenas in 2000 requiring CMS to produce specific data elements relating to each patient and patient day that was included in the numerator and the denominator of the SSI fractions at issue for fiscal years 1993, 1994, and 1995. AR 5855-59, 5850-54, 5845-49; see PRRB Dec. at 4, AR 134. 220. When CMS failed to timely respond to the subpoenas, Plaintiff initiated a civil action in this Court (Civ. No. 00-02940(PLF)) to compel compliance with the subpoena issued for fiscal year 1995. PRRB Dec. at 4, AR 134; Complaint for Declaratory and Injunctive Relief for Enforcement of Administrative Subpoena, AR 519. 221. The Secretary moved to dismiss the prior civil action on June 6, 2001, asserting that “the Secretary . . . has, in fact, complied with the PRRB’s subpoena” and that the case should be dismissed for lack of subject matter jurisdiction. Memorandum of Points and Authorities in Support of Defendant’s Motion to Dismiss For Lack of Subject Matter Jurisdiction at 12, Civ. No. 00-02940 (PLF), AR 593. 222. In support of the motion to dismiss, the Secretary’s counsel submitted a declaration of CMS employee Robyn Thomas, who stated under penalty of perjury that the data that CMS had used to calculate the Hospital’s SSI percentage for fiscal year 1995 was no longer available: Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 133 of 145 81 Baystate Medical Center requested data that was included in the calculation of the numerator and denominator of its Medicare SSI percentage for fiscal year ending September 30, 1995. The files that HCFA originally used to calculate that per-centage [sic] are no longer available. Therefore, HCFA requested from the Social Security Administration (SSA) a re-creation of SSA’s 1995 SSI eligibility file, which is the basis for the SSI eligibility information contained in the HCFA MEDPAR file. Memorandum of Points and Authorities in Support of Defendant’s Motion to Dismiss For Lack of Subject Matter Jurisdiction, Exh. A, Declaration of Robyn Thomas ¶ 6, Civ. No. 00-02940 (PLF) (emphasis added), AR 596–97. The Thomas Declaration further attested that the Secretary “has answered the PRRB’s subpoena as completely and accurately as possible.” Thomas Declaration ¶ 12, AR 599. 223. The Secretary’s reply memorandum in support of the motion to dismiss reasserted that “the Secretary has answered the subpoena at issue as completely and as accurately as possible.” Defendant’s Reply to Plaintiff’s Response to Defendant’s Motion to Dismiss at 1, 4- 5, Civ. No. 00-02940 (PLF), AR 685, 688–89. The Secretary’s Reply further stated that CMS had conducted an exhaustive search for data that it had used to calculate the Hospital’s SSI fraction for fiscal year 1995: Indeed, an employee of the Secretary who is intimately familiar with the data has declared, under penalty of perjury, that the agency “has answered the PRRB subpoena as completely and accurately as possible.” Thomas Declaration, ¶ 12. Plaintiff’s refusal to accept that the agency has conducted an exhaustive search of its data banks in order to respond to the subpoena – a refusal based on nothing more than “skeptic[ism],” “suspicion,” and “belief,” – is a wholly insufficient basis upon which to look to recourse from this Court. Id. at 4-5, AR 688 (emphasis added). 224. In December 2002, this Court dismissed the suit on the ground that the Hospital must work through the Board to obtain judicial enforcement of the Board’s subpoena. Memorandum Opinion, AR 692–98. Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 134 of 145 82 225. Subsequently, during a discovery hearing conducted by the Board in 2003, the Hospital proved that the representations that CMS had made to this Court in the Thomas Declaration and in the memorandum in support of the Secretary’s motion to dismiss were false. PRRB Dec. at 18, AR 148. 226. During that hearing, CMS’ principal MEDPAR programmer, Dean (who reported directly to the declarant, Thomas), admitted upon examination before the Board that the SSI and Medicare days data that CMS had used to compute the SSI fraction for the Hospital’s 1995 fiscal year were still in CMS’ possession when Thomas’s sworn statement to the contrary was submitted, under penalty of perjury, to this Court. See PRRB Dec. at 18, 25 n. 135, AR 148, 155. 227. Dean’s admission before the Board in the 2003 discovery hearing came shortly after the Secretary’s counsel, Romano, sent Dean an April 23, 2004 email attaching a list of MEDPAR files and asking whether Thomas’ sworn assertion that the agency’s 1995 data no longer existed was “literally true (I hope) in light of the files listed below.” AR 1631 (emphasis added). 228. The Secretary’s counsel, Romano, ultimately produced that data to the Hospital under cover letter dated July 29, 2003 – the same data that CMS had used to compute the SSI fraction for the Hospital’s 1995 fiscal year, and the very data that the agency had sworn did not exist in a declaration submitted to this Court in support of dismissing the Hospital’s lawsuit. See AR 2685-88. 229. Specifically, the Secretary produced a CMS file purporting to reflect the patient- level detail for the patients and patient days that were counted in the number of SSI days and the Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 135 of 145 83 detail for the patients and patient days counted in the number of Medicare days reflected in the SSI fraction computed by CMS in June 1996 for the Hospital’s fiscal year 1995. However, the sums of SSI days and Medicare days reflected in the file that was produced to the Hospital in 2003 do not square with the numbers of days included in the Secretary’s determination of the SSI fraction for fiscal year 1995 as set forth in paragraph 18 above. AR 2685-88. 230. Likewise, the Secretary produced a CMS file purporting to reflect the same patient-level detail for patients and patient days reflected in the SSI fraction computed by CMS in June 1997 for the Hospital’s fiscal year 1996. Again, the sums of the SSI days and the Medicare days reflected in this file do not square with the numbers of days included in the Secretary’s determination of the SSI fraction for fiscal year 1996 as set forth in paragraph 20 above. AR 2685-88. 231. On May 1, 2003, the day after Dean had admitted to the existence of the 1995 data that CMS had used to calculate the Hospital’s SSI fraction for 1995, CMS counsel, Romano, issued a warning to the Board that it “should proceed very cautiously with respect to any further discovery orders in this case.” AR 5791 (emphasis added). 232. In response, the Chairman of the PRRB made the following remarks on the record: When I became Chairman in [2002] and started looking at some of the cases that we had, I saw that we had some very old cases that I couldn’t figure out why they were still sitting here. So I started asking some questions and I found out that this case and others, that were some apparently very intractable discovery disputes involved * * * And in digging into that a little bit more, I found that we had issued subpoenas and that there was at least a dispute over whether or not the responses had been complete. * * * From what I’ve seen, we do indeed [have] a genuine discovery dispute. * * * And I will also note that during the course of this proceeding the Board had requested the General Counsel’s office to enforce its subpoenas, which that office declined to do. With regard to whether or not Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 136 of 145 84 information has been submitted, * * * [m]y perception of the way that came out yesterday was it really didn’t come out very easily, and it had to be clarified through pretty extensive questioning, Now whether or not there was some intent to not be fully open with us or whether or not it some misunderstanding, I don’t know, but I know that it wasn’t – things didn’t flow concerning that revelation easily from my perspective. And it seems to me that just the fact of our having this evidentiary hearing seems to have produced some documents that otherwise had not been produced. . . . The Rudolph [Tayloe] memo [from February 1996] certainly appears to me to have been central to the issue that’s been pending all these years. * * * And there’s just one more. As I say, I’m not trying to form an opinion on whether or not anything was done by the client, CMS, willfully keep information out of the hands of the [hospital] and the Board or whether it was purely a lack of understanding of who was doing what, but I do know that the government has been very keen in at least the last decade with regard to the Federal False Claims Act . . . . And as I recall, one of the standards of culpability in the False Claims Act is willful ignorance. And I think that’s very reasonable and it should be imposed, but I also think that we to[o,] all of us that deal with federal health care funds should be held to the same standard. None of us should be excluded from that standard, so in analyzing your client’s [CMS’] caution that we tread very carefully here on other discovery orders from what I’ve seen it’s very clear that we just need to get to the bottom of this and get the facts out. * * * AR 5799-5801. 233. A few moments later, the Board Chair added the following additional comments in a further exchange with CMS’ counsel, Romano: I would just comment that just based upon the records availability yesterday that there did seem to be a semantics war going on there. It was very – it became very clear after a lengthy examination that even though Mr. Dean didn’t have the original physical tapes that SSA sent you had a problem. You had it all along. It was right there. AR 5803 (emphasis added). 234. On February 7, 2002, the PRRB issued a subpoena compelling CMS to produce all documents and programs relating to the policies and procedures used to calculate the SSI fraction. AR 7104-07. 235. CMS timely produced just three sets of documents, which included: (i) a copy of the transcript of Dean’s 1995 deposition in another case; (ii) a copy of the transcript of Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 137 of 145 85 Rosenberg’s 1995 deposition in that case; and (iii) printouts of three computer programs that purportedly are used to perform the match of MEDPAR and SSI records. See AR 1310. The computer programs that CMS produced were not the programs that the agency had used to compute the SSI fraction for fiscal years 1995 and earlier, however, as evidenced by the programs’ references to “Y2K” (i.e., year 2000 edit corrections). CMS produced additional documents relating to CMS’ policies and procedures for calculating the SSI fraction more than a year later, in connection with and as a follow up to the discovery hearing that the PRRB conducted in May 2003. AR 2116. But, even after the discovery hearing, CMS still did not produce any documented standards, procedures, or protocols for its calculation of the SSI fraction. See, PRRB Dec. at 16-18, AR 146-148. 236. The Board also issued a second subpoena directing SSA to produce its SSI entitlement records for a small sample of Medicare beneficiaries who were treated by the Hospital during periods at issue in here. AR 7108-10. SSA refused to produce those records on privacy grounds. By letter dated June 21, 2002, the Board requested that the Secretary’s counsel seek judicial enforcement of the subpoena “by the legal means addressed in the statute” at 42 U.S.C.§ 205(e). AR 7057–58. The Secretary’s counsel refused to enforce that subpoena. See infra ¶ 232. 237. The Hospital pursued other avenues in 2002 to obtain the SSI eligibility records for Medicare patients who were still living, whose whereabouts were still known to the Hospital, and who consented in writing to SSA’s release of their SSI records. This study indicated that some of these patients who had been entitled to SSI during their inpatient hospital stays had not Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 138 of 145 86 been counted as SSI patient days by CMS. AR 3267-68, 3286–88; Shafer Testimony, Hearing Tr. at 432-33, AR 5134–35. 238. Based on the results of the limited study, the Hospital submitted a request to SSA under the Freedom of Information Act for the SSI records of all Medicare patients discharged from the hospital in fiscal years 1993 through 1996. The Hospital also requested that the Board issue a subpoena to SSA for the SSI records of all Medicare patients discharged from the hospital in fiscal years 1993 through 1996. AR 1102-10, 1948-52, 1984-87. 239. On June 5, 2003, the PRRB issued a subpoena directing SSA to produce those SSI records for all of the Hospital’s Medicare patients during the years 1993 through 1996. AR 7229-7238; AR 7104-10; AR 6916-20. Those records would have allowed the Hospital to determine the exact number of its Medicare patient days attributable to individuals who were also entitled to Federal SSI benefits during their inpatient stays at the Hospital during the years at issue, and to measure precisely and quantify the financial impact of errors and omissions in CMS’ calculations of the Hospital’s SSI fractions for those years. See AR 7239-41. 240. In a partial response to the Hospital’s 2003 FOIA request, SSA furnished the Hospital with SSI eligibility records only for deceased individuals who had been treated by the Hospital in fiscal year 1994, contending that production of anything more would be too burdensome or prohibited by the Privacy Act. Shafer, Hearing Tr. at 425-26, AR 5133; AR 2705-06. 241. Having received no response from SSA concerning the Board’s subpoena for the SSI records for all of the Hospital’s Medicare inpatients during the years at issue, the Chairman of the PRRB proposed in a June 2003 pre-hearing conference that the parties attempt to work Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 139 of 145 87 through the political appointees at SSA and HHS for assistance in obtaining the SSI records required by the Board’s subpoena. AR 2166-67. The Hospital supported that proposal, but the Secretary’s counsel, Romano, opposed it, apparently preferring instead for the Secretary’s counsel to direct and control any discussions with SSA regarding its response to the PRRB’s subpoena. AR 2173-74, 2805–07. 242. In another pre-hearing conference held on July 28, 2003, faced with a continuing lack of progress, the Board restated its interest in pursuing through political channels the “fundamental data that’s needed to get the answers in this case.” AR 2617-18. 243. Soon after the July 2003 pre-hearing conference, the Board received a letter from SSA, dated July 28, 2003, stating that SSA would not produce the SSI records to the extent that the Board’s subpoena called for production of information regarding living individuals. AR 2642-43. Subsequently, by letter dated August 29, 2003, the Board requested the assistance of the Secretary in obtaining the subpoenaed SSI eligibility records from SSA. AR 2809-10. This request was consistent with the Medicare statute’s mandate that the Secretary shall “make available to the Board such secretarial, clerical, and other assistance as the Board may require to carry out its functions.” 42 U.S.C. § 1395oo(i) (emphasis added). 244. As all of these efforts were underway, counsel for the Hospital and the Secretary’s counsel, Romano, exchanged extensive correspondence regarding the Hospital’s proposal to have SSA release SSI data to a third party contractor of CMS, who would verify which of the hospital’s Medicare patients were entitled to SSI benefits during the periods at issue and then disclose the SSI data on those patients to the Hospital. AR 2604, 2708-12, 2786-88, 2790-92, 2802-03, 2805-07, 2830-36, 2838-41, 2843-47. Although the Hospital offered in writing to bear Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 140 of 145 88 all the cost of this alternative data match, CMS refused to cooperate, alleging (falsely, as it later turned out) that SSA would not agree to release the SSI data to CMS in connection with this proposal. AR 7497-7509. 245. On April 30, 2004, the Hospital filed a petition asking the Board to grant “expedited judicial review” (“EJR”) over the issues underlying the evidentiary impasse concerning the SSI entitlement records. AR 6326. 246. By letter dated May 28, 2004, the Board denied the Hospital’s petition on the ground that the Board had not yet formally requested that the Secretary’s counsel initiate legal action to enforce the Board’s June 5, 2003 subpoena for the SSI records for all Medicare patients who were discharged from the Hospital in fiscal years 1993 through 1996. The Board’s May 28, 2004, letter noted that the Board would make a formal request to HHS counsel for enforcement of its subpoenas, which it later did, and advised the parties that it would either grant EJR or initiate other appropriate action if its request for enforcement was not timely acted upon by the Secretary’s counsel. AR 6259-61. 247. Shortly after the Board issued its decision on the EJR petition, by a letter dated June 2, 2004, the Secretary’s counsel asked the Administrator of CMS to undertake an interlocutory review and issue an order reversing the Board’s determination to enforce its subpoena to SSA. Over the Hospital’s written objections, and in defiance of the Board, the Administrator eventually issued a decision concluding that the Board’s subpoenas to SSA should not be enforced. AR 3191-92, 3559-76. 248. By letter dated June 10, 2004, the Board Chairman wrote to counsel for the Hospital and counsel for CMS and suggested that the parties try to mediate the evidentiary Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 141 of 145 89 impasse concerning the SSI data, noting in reference to the Hospital’s alternative data match proposal that “there does appear to be a workable solution on the table, provided that the details could be worked out.” AR 3449. 249. By letter dated June 21, 2004, the Hospital’s counsel informed the Board and counsel for CMS that the Hospital was amenable to mediation of the evidentiary issue, but noting that the Hospital was being prejudiced by CMS’ continuing failure to provide answers to discovery requests. AR 7515. 250. The Secretary’s counsel wrote to the Board the same day, alleging that it was unclear to him whether SSA would agree to provide CMS with the SSI data needed to implement the Hospital’s alternative data match proposal and he did not know whether CMS “would be willing under any circumstance to participate in the [Hospital’s] data proposal.” AR 7513-14. 251. On that same date, by letter dated June 21, 2004, the Board requested that the Secretary’s counsel “seek enforcement of the June 5, 2003 subpoena issued to SSA [for the SSI records for all of the Hospital’s Medicare patients for the years at issue] and reiter[ated] its request for enforcement of the February 7, 2002 subpoena.” AR 6702. The Board’s four-page letter to CMS counsel noted: There is no authority in the statute or regulations for the CMS Administrator to block evidence from being sought or presented that the Board has determined is essential to a matter in dispute. These authorities, coupled with the inherent conflict of interest in OGC’s representation of both CMS and the Board, we believe should weigh heavily in favor of honoring the Board’s legitimate request. AR 6705. 252. On July 9, 2004, SSA counsel sent a letter to the Secretary’s counsel and the Hospital, with an additional copy sent under separate cover to the Attorney Advisor to the Administrator of CMS, concerning SSA’s willingness to supply CMS with the SSI data that Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 142 of 145 90 would be needed to facilitate the alternative data match that the Hospital had proposed as means of obtaining necessary SSI entitlement data for all of its Medicare patients during the fiscal years at issue. As noted in SSA’s transmittal letter to the CMS Administrator, SSA proposed its “business solution” as a means to “assist the Administrator in facilitating a resolution in [the PRRB] case.” AR 7492-94. 253. By letter dated July 12, 2004, the Hospital’s counsel furnished the Board with copies of SSA’s letters regarding its willingness to cooperate with the “business solution” to the evidentiary impasse in the pending PRRB appeals. In that same communication, the Hospital’s counsel also informed the Board that “[n]otwithstanding SSA’s agreement to cooperate with this proposal, Mr. Romano [the Secretary’s counsel], has notified us by e-mail ‘that OIS [the CMS Office of Information Services – the Office headed by Robyn Thomas] . . . is not willing to participate in the [Hospital’s] proposed data match.’” AR 7491. 254. On July 29, 2004, the Deputy Administrator of CMS signed an order concluding that the Board “does not have the authority to subpoena SSA under section 205 (d) and (e) of the Social Security Act. Thus, the Administrator will not request that the OGC [Office of General Counsel] seek enforcement of the Board subpoenas in Federal district court.” AR 6658. 255. The Deputy Administrator’s order of July 29, 2004, was effected through an extraordinary interlocutory administrative appeal process that the Administrator had previously ruled (in other cases) to be categorically unavailable in the Medicare administrative appeal process established under 42 U.S.C. § 1395oo. See Texas Disproportionate Share Group Appeal, Adm’r Dec. (July 22, 1996) Medicare & Medicaid Guide (CCH) ¶ 44,581 fn. 6; Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 143 of 145 91 Cardinal Cushing Hospital, Adm’r Dec. (Jan. 29, 2003) Medicare & Medicaid Guide (CCH) ¶ 80,975 fn. 4. 256. Approximately six weeks later, beginning on September 15, 2004, the PRRB commenced its six-day hearing on the merits of the Hospital’s appeals contesting CMS’ calculations of the hospital’s SSI fractions for fiscal years 1993-1996. AR 5028-5613. /s/ Christopher L. Keough Christopher L. Keough DC Bar No. 436567 John M. Faust DC Bar No. 433553 Stephanie A. Webster DC Bar No. 479524 VINSON & ELKINS L.L.P. 1455 Pennsylvania Avenue, N.W. Suite 600 Washington, D.C. 20004-1008 (202) 639-6500 (phone) (202) 639-6604 (fax) Counsel for Plaintiff Dated: February 1, 2007 Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 144 of 145 CERTIFICATE OF SERVICE I hereby certify that on this 1st day of February, 2007, I electronically filed the foregoing Motion for Summary Judgment, Memorandum of Points and Authorities, Statement of Material Facts As To Which There Is No Genuine Dispute, and Proposed Order with the Clerk of Court using the CM/ECF system, which will automatically send email notification of such filing to the following attorneys of record for the Defendants: LISA A. OLSON U.S. Department of Justice 20 Massachusetts Avenue, N.W. Room 7300 Washington, D.C. 20530 /s/ Christopher L. Keough Christopher L. Keough Case 1:06-cv-01263-JDB Document 15 Filed 02/01/2007 Page 145 of 145