MEMORANDUM AND ORDER
FULLAM, Chief Judge.
Plaintiff is a Pennsylvania non-profit corporation which operates Temple University Hospital. It has brought this action against various officials of the Commonwealth of Pennsylvania, invoking 42 U.S.C. § 1983, asserting that the defendants have deprived Temple of rights secured by Title XIX of the Social Security Act, 42 U.S.C. § 1396a
et seq.
in their administration of the payment system for in-patient hospital care under the Pennsylvania Medical Assistance Program.
At issue are such matters as whether defendants are meeting the statutory requirement that the State’s Medical Assistance program must provide payments to hospitals that are reasonable and adequate to meet the costs that must be incurred by efficiently and economically operated hospitals; whether Pennsylvania’s Medical Assistance program is based upon the statutorily required findings and certification; and whether the Program adequately takes into account the special problems of hospitals which, like Temple, serve a disproportionate number of low-income patients. Before addressing the merits, however, it is necessary to refer briefly to defendants’ threshold argument that plaintiff cannot maintain this action under 42 U.S.C. § 1983 for violation of alleged federal statutory rights, because Title XIX gives rise to no rights which can be asserted by hospitals. This precise argument has been persuasively rejected by the Third Circuit Court of Appeals in
West Virginia Univ. Hospitals, Inc. v. Casey, et al.,
885 F.2d 11 (3d Cir., 1989), and I am bound by that decision.
I. FACTUAL BACKGROUND
For several years, hospitals received reimbursement for costs actually expended in the care of Medicaid patients, on the basis of cost figures submitted to and audited by the appropriate state authorities. In the belief that this reimbursement system provided inadequate incentives to hospitals to operate efficiently, and in order to cope with rapidly escalating Medicaid hospital costs, Congress, as part of the ’81 Omnibus Reconciliation Act (OBRA), P.L. 97-35, established a new standard of hospital reim
bursement. Whereas, previously, hospitals were to be reimbursed “the reasonable cost” of rendering in-patient services; the OBRA replaced that standard with the current standard requiring payments to hospitals at rates which are “reasonable and adequate to meet the costs which must be incurred by efficiently and economically operated facilities”. 42 U.S.C.A. § 1396a(a)(13)(A) (West Supp.1989). The statute and accompanying regulations afford considerable leeway to the States to determine the precise methods by which payments to hospitals will be calculated in order to meet the statutory standard, but require States to make findings and give assurances that the plan adopted does in fact comply with federal requirements.
Pennsylvania’s Medical Assistance Program established a complex system for classifying hospitals into groups thought to face similar external constraints affecting their costs; classifying medical procedures performed in hospitals into groups thought to involve essentially similar costs; calculating the average cost per procedure, experienced by hospitals in each group; and, after various adjustments including a so-called “budget neutrality factor” and an upward adjustment for hospitals burdened disproportionately with indigent patients, resulting in a formula by which hospitals are paid prospectively, without regard to their actual expenditures. The changeover from the old system to the new system was phased in over a three-year period.
The Third Circuit Court of Appeals, in
West Virginia Univ. Hospitals, supra,
held that this precise medical assistance plan was invalid, for non-conformity with the requirements of Title XIX, insofar as it affected out-of-state hospitals. That holding is not, of course, directly controlling in the present case, because we are dealing with in-state hospitals, and with a different reimbursement formula. But the thorough discussion and analysis of the statute, the regulations, and the legislative history provided by Judge Rosenn’s opinion in that case makes it unnecessary to undertake a similar exposition here. It is sufficient to note that Judge Rosenn identifies three criteria which medical assistance plans must meet in order to conform to Title XIX: (1) the plan must take into account the situations of those hospitals which serve a disproportionate number of low-income patients; (2) the plan must be based upon a finding and certification that the rates are reasonable and adequate to meet the necessary costs of an efficiently operated hospital; and (3) the plan must be based upon a finding and certification that the rates will assure Medicaid patients reasonable access to in-patient hospital care. Judge Rosenn refers to these as the “disproportionate share” requirement, the “reasonable and adequate” requirement, and the “reasonable access” requirement. As to the first and third, judicial review is plenary, but as to the “reasonable and adequate” standard, judicial review is limited to inquiring whether the State’s determination is arbitrary and capricious (slip op. pp. 24, 25). Notwithstanding this deference to the State’s determination of what constitutes a reasonable and adequate rate of reimbursement, the process by which that determination was reached — the adequacy of its factual investigation and findings— must also conform to the statutory requirements.
A.
Temple’s Experience
Temple University Hospital serves a North Philadelphia community which is principally black, hispanic and indigent. The majority of Temple’s patients are blacks and hispanics who live in poverty. Fifty-percent of Temple’s patients have Medicaid insurance coverage; 20% are covered by Medicare; and 5% have no coverage.
Approximately 2,100 children are born at Temple Hospital each year. The neo-natal mortality rate in that community is more than twice the national average. Approximately 20% of the mothers enter Temple Hospital without having had any prenatal care. Many are addicted to drugs; about 20% of the babies born at Temple Hospital show evidence that their mothers consumed cocaine during pregnancy. Six percent of the babies born at Temple suffer from low
birth weight or other problems, including sexually transmitted diseases.
Nursing costs represent approximately 22% of Temple’s total operating costs. Although Temple does not have the highest pay scale in the area, the rate of pay for nurses at Temple increased more than 37% between 1980 and 1988. The average compensation of nurse practitioners increased 44% between 1983 and 1988. Because of its pay scale, Temple has difficulty retaining nurses; its ratio of nurses to occupied beds is among the lowest of the teaching hospitals. Temple’s professional liability insurance costs and similar obligations beyond its control have increased dramatically in recent years. For example, its premium for Pennsylvania's Catastrophic Loss Fund increased approximately 400%. Utility costs have increased at a rate of approximately 10% to 12% per year; the cost of disposable supplies and pharmaceuticals has increased at the rate of approximately 9.5% each of the last several years.
On average, Philadelphia medical school hospitals employ 6.1 full-time employees per occupied bed; Temple’s rate is 5.3. Temple’s housekeepers are required to clean between 30% and 50% more square footage than the industry norm.
Temple’s current occupancy rate is approximately 84% to 85%, an increase of 21% since 1984.
Under the Department of Public Welfare’s ranking system, Temple is one of seven hospitals with the highest score; it has the least costs of these seven hospitals; its costs are below the weighted average cost of all hospitals classified in the same group under the Medical Assistance Program (Class 1).
Under the Pennsylvania Medical Assistance Program, Temple receives approximately 81% of its actual costs for medical assistance patients; during fiscal year 1988-89, Temple will lose approximately $7.8 million on in-patient care to medical assistance patients.
During each of the years since the inception of the prospective payment system, Temple’s medical assistance payments have been inadequate to cover its medical assistance reimbursable costs. The shortfall between its costs for medical assistance in-patient care and its reimbursement have been as follows:
FY ’1985 - $2,504,617
FY ’1986 - $2,525,899
FY ’1987 - $5,086,863
FY ’1988 - $6,218,785
FY ’1989 - (estimated) $7,790,427
To some extent, Temple has been able to achieve cross-subsidization from other payors, but it will nevertheless experience a loss of $3,256,155 on in-patient care during the current fiscal year. Its loss for fiscal year 1989-90 is expected to rise to somewhere between $4.5 and $5.2 million on in-patient care.
Plaintiff presented a mass of evidence, which stands unrebutted, to the effect that it has cut costs in every conceivable way, and that, as a practical matter, no further “efficiency” or “economy” is possible.
Defendants’ response, in essence, is that, since the Medical Assistance Plan is designed to provide reasonable and adequate reimbursement of the costs which would necessarily be incurred by an efficient institution, the shortfall between Temple’s costs and the reimbursement provided by the Plan is proof that Temple is not an efficient hospital. Although this seems, at first blush, to be a rabbit-in-the-hat type of argument, it is one which follows inevitably from the design of the Medical Assistance Plan established by the defendants, as will be discussed in the following section.
II. REASONABLE AND ADEQUATE COSTS INCURRED BY EFFICIENT AND ECONOMICAL HOSPITALS
Opponents and proponents of the Pennsylvania Medical Assistance Plan (MAP) agree on certain basic principles. What in-patient care should reasonably cost depends upon the nature of the care provided, and the circumstances of the provider. The first step in the analysis, therefore, is to identify and classify the diseases or conditions which require in-patient hospital care. Thus, Pennsylvania’s MAP identifies some 477 categories of Diagnostic-Related Groups (DRGs). For each DRG, a flat fee
is established, regardless of whether the patient actually received more or less than the standard treatment. In Pennsylvania, the relative value for each DRG is computed by:
(a) determining the total standardized cost for all approved claims in the data base (i.e., previous experience);
(b) determining the total number of medical assistance hospital cases in the data base;
(c) dividing the total standardized cost by the total number of cases to establish a statewide average cost per case for all eases;
(d) determining the total standardized costs and total number of cases for each DRG;
(e) dividing the total costs for each DRG by the corresponding number of cases for that DRG to establish an average cost per case for each DRG; and
(f) dividing the average cost per case for each DRG by the statewide average cost per case for all cases to establish the relative value for each DRG [Stipulation of Facts 35]. All cases with a cost more than three standard deviations from the arithmetic or logarithmic mean cost for that DRG are excluded [Stipulation 38].
Thus, historical costs for similar kinds of hospital admissions are the starting point for calculating reimbursements under the plan.
The next step under the Plan was to classify hospitals into groups, in an effort to treat similarly situated institutions in similar fashion. The classification process involved ranking each institution on a series of 13 variables in 4 categories, teaching, medical assistance volume, environment and cost. The variables included such matters as number of full-time employed physicians/residents/interns per bed; number of full-time equivalent physicians, residents and interns; and number of residency programs (the teaching concept variables); medical assistance reimbursable in-patient costs, separately and in relation to total in-patient costs; acute care in-patient medical assistance in-patients days, separately and in relation to total acute care in-patient days (medical assistance volume concept); percentage of persons below the poverty level in that county, median family income in that county, percentage of unemployment in that county (environmental characteristics); and Medicare area wage index, total in-patient expenses adjusted for direct medical education and capital/total in-patient admissions; and total in-patient expenses adjusted for direct medical education and capital/total in-patient days (hospital costs concept). [Stipulation 41.] For each of these 13 variables, a hospital would receive a numerical score which resulted in a separate ranking within each concept area.
On the basis of these numerical rankings, hospitals were placed in any one of eight groups. Three children’s hospitals were placed in a separate group, and the remaining hospitals divided into seven groups. Group 1 hospitals have the highest cost-factors, group 7 the lowest. A separate group payment rate was established for each group.
The defendants pegged the group rate for each group at somewhat more than the costs incurred by the lowest-cost hospital within that group, but considerably less than the costs incurred by the average hospital within that group. Presumably, this was done on the theory that similarly situated hospitals should be encouraged to emulate the example of the lower-cost hospitals within that group.
There are several problems with this approach, both theoretical and practical. Assuming that the hospitals are grouped appropriately, and that the hospitals within a particular group are indeed similarly situated, and assuming that the lowest-cost hospitals within that group are the most efficient, the incentive is double-edged: whereas some fairly low-cost institutions might be encouraged to reduce their costs still further, those at the higher end of the spectrum within a particular group would do better to increase their costs and become less “efficient”, in order to qualify for membership in the next-higher group. In short, grouping hospitals according to their actual cost experience (rather than in
accordance with the similarity of their circumstances affecting costs), produces a result having little or nothing to do with whether the hospitals are being run efficiently and economically.
This difficulty is exacerbated by the fact that hospitals are assigned to groups, not because their scores on the ranking test were comparable, but simply in order to achieve seven groups of near-equal size. Thus, there is greater variation among the hospitals in group 1 (the scores in that group range from 913 to 724) than there is between the highest-scoring hospital in group 3 (608) and the lowest-scoring hospital in group 7 (439).
Group 1 hospitals rank highest in teaching, medical assistance volume, environmental factors, and costs. But the hospitals assigned to that group are far from homogeneous. The group includes the six Philadelphia teaching hospitals (Temple, Penn, Jefferson, Medical College of Pennsylvania, Hahnemann, and Osteopathic); Germantown Hospital, a primary-care hospital in Philadelphia; and Springfield Hospital in Delaware County. Within that group, the ratio of full-time residents per bed varies from .688 (Medical College of Pennsylvania) to .124 (Springfield); the ratio of acute-care medical assistance costs to total costs varies from .361 (Temple) to .042 (Springfield).
In order to receive full medical assistance reimbursement, a hospital must rank relatively low within its group. It is simply impossible for a hospital such as Temple, which ranks high in the group, to receive payments equal to its costs. Plaintiffs’ statistical expert, Dr. Siskin, has convincingly demonstrated that whether a hospital will or will not receive full-cost reimbursement is, under the MAP, essentially unrelated to efficiency; it depends upon where it ranks, within that particular group, on the variables; the higher the ranking within the group, the greater the likelihood, even certainty, of a shortfall. But the very purpose of rank-scoring on the variables was to identify the hospitals entitled to higher rates of reimbursement.
A further, major, difficulty with Pennsylvania’s MAP is that, after determining the group rate for each group, application of a “budget neutrality” factor results in an across-the-board 14% reduction. That is, hospitals within each group receive only about 86% of the costs incurred by the hospitals at the lower end in that group.
The defendants’ attempts to justify this reduction are not entirely consistent with each other, or with the evidence. Initially, the concept of a budget-neutrality adjustment was introduced as a part of the transition to the system of prospective payment; it was designed to make sure that total payments under the prospective payment system would not exceed what they would have been for the same level of services, under the old system. Another justification advanced by the defendants is that, under the old system, there was usually a shrinkage between the cost figures submitted by hospitals in claiming reimbursement, and the final audited figures. Since the prospective payment formula was based in part upon the hospitals’ (unaudited) costs, it was reasonable to suppose that there should be a similar shrinkage between the defendants’ initial projections and the correct prospective payment estimate.
These theoretical justifications do not hold water. The discrepancy between audited and unaudited cost figures under the old system seldom, if ever, exceeded 2%; and no attempt has been made to determine, in recent years, how the total cost under the new system compares with the costs which would have been incurred had the old system remained in effect. The defendants’ own calculations at the time projected little or no difference. Plaintiff’s expert, Dr. Coelen, performed a study suggesting that the maximum budget neutrality adjustment should be only about 2.4%.
The evidence makes very clear that the “budget neutrality” adjustment, like other features of the MAP, is entirely budget-driven. It is simply a mechanism for keeping total medical assistance costs within the Welfare Department budget. Moreover, the across-the-board approach — apply
ing the adjustment equally to all hospitals and all groups without regard to their relative level of efficiency or other pertinent circumstances — is utterly inconsistent with the notion of rewarding efficiency.
In short, Pennsylvania’s reimbursement rates are simply arbitrary.
III. DISPROPORTIONATE SHARE ISSUES
In order to comply with Title XIX, the MAP must use rates established according to standards “which, in the case of hospitals, take into account the situation of hospitals which serve a disproportionate number of low-income patients with special needs ...”. There is no dispute about the fact that Temple does serve a disproportionate number of such indigent patients. Pennsylvania’s MAP provides for additional payments to such hospitals, ranging from a low of 0.5% to a high of 2.5%. Temple qualifies for the 2.5% add-on. Unfortunately, it is also undisputed that this represents only a small fraction of the total increase in costs attributable to Temple’s status as a disproportionate-share hospital. As the defendants’ own calculations demonstrate, full recognition of Temple’s status as a disproportionate-share hospital would result in an add-on of at least 16%. The parties disagree as to whether this is in conformity with Title XIX.
The statute permits states to calculate the required additional payments to disproportionate-share hospitals either by adopting the same formula used in the Medicare Program, or by adopting their own formulas. It is clear that Pennsylvania rejected the Medicare formula because it would be too expensive. The defendants then worked out a formula which would have produced add-ons in the range of 4% to 5%. This was never implemented, however, because of budgetary considerations. In the final analysis, the defendants simply allocated to disproportionate-share hospitals the funds available; this turned out to produce the maximum reimbursement of 2.5%.
The statute does not mandate any particular level of payments for disproportionate-share hospitals. It does, however, require payments which are “reasonable and adequate to meet the costs which must be incurred by efficiently and economically operated facilities in order to provide care and services in conformity with applicable state and federal laws”; and it requires that such rates are to be determined in accordance with standards which “take into account the situation of hospitals which serve a disproportionate number of low-income patients with special needs”.
Whether the Pennsylvania plan can properly be said to have taken the situation of disproportionate-share hospitals into account when it falls so far short of meeting the medical assistance costs associated with disproportionate-share activities is a close question. By specifying either the Medicare system or an alternative system devised by the States, Congress seems to have contemplated that the State’s plan would produce comparable results. The 2.5% override provided by the Pennsylvania Plan is only about Vio of the amount which would be payable under the Medicare analysis (2.5% versus 20.93%).
Recognizing that States are given a considerable amount of flexibility in this area, and that reimbursement rates are to be fixed by the State, not by this court, I am nevertheless constrained to hold that Pennsylvania’s adjustment for plaintiff’s disproportionate-share status misses the mark by so wide a margin as to be inconsistent with the intent of Congress.
IV. PROCEDURAL ISSUES
Finally, it seems clear that Pennsylvania’s Medical Assistance Plan was adopted without compliance with the procedural requirements of Title XIX and the applicable regulations. Section 447.253(b) of the applicable regulations provides:
“(b)
Findings
— Whenever the Medicaid agency makes a change in its methods and standards, but not less often than annually, the agency must make the following findings:
“(1)
Payment Rates
—
“(i) the Medicaid agency pays for inpatient hospital services and long-term
care facility services through the use of rates that are reasonable and adequate to meet the cost that must be incurred by efficiently and economically operative providers to provide services in conformity with applicable state and federal laws, regulations, and quality and safety standards.
“(ii) With respect to in-patient hospital services—
“(A) the methods and standards used to determine payment rates take into account the situation of hospitals which serve a disproportionate number of low-income patients with special needs ...” 42 C.F.R. § 4471253(b) (1988).
So far as the record discloses, Pennsylvania did not make findings based upon empirical studies — on such matters, for example, as the characteristics of an efficient and economical hospital operation, the impact of the proposed reimbursement rates upon hospitals’ ability to survive, etc. — but merely certified that its plan complied with the statutory requirements. But, as stated by the Tenth Circuit Court of Appeals,
“Mere recitation of the wording of the federal statute is not sufficient for procedural compliance. There is a presumption that the State will engage in a bona fide finding process before it makes assurances to HCFA that the required findings have been made. To rule otherwise would completely eviscerate the federal requirements so long as the magic words are submitted to HCFA.
Amisub, supra,
879 F.2d at 797.
Indeed, our own Court of Appeals has expressly held that Pennsylvania’s plan is procedurally defective, at least with respect to out-of-state hospitals. The court pointed out:
“In structuring its out-of-state reimbursement program, Pennsylvania admits to gathering no information with respect to these hospitals’ actual costs. No empirical analysis was conducted to measure the effects of the reimbursement program on out-of-state hospitals ____ Federal law is not satisfied if a State merely makes conceptual policy decisions. A policy predicated upon provincialism and self-interest, not upon findings of reasonableness and adequacy, is unacceptable. We hold that the federal regulations unambiguously require the State to make findings____ In failing to make these requisite findings, Pennsylvania violated federal law.”
West Virginia Hospitals, supra
at 30.
I therefore conclude that the Pennsylvania plan was adopted without adequate compliance with the procedural requirements of the statute.
V. SUMMARY OF LEGAL CONCLUSIONS
For the reasons set forth above, I have concluded that the Pennsylvania Medical Assistance Plan was adopted without compliance with the procedural requirements of Title XIX and the implementing regulations because of the absence of meaningful findings based upon reasonable investigation. I have further concluded that, as applied to the plaintiff, the Pennsylvania Medical Assistance plan violates the statute, 42 U.S.C.A. § 1396a(a)(13)(A) (West Supp.1989) both because it fails to provide payments at rates which are “reasonable and adequate to meet the costs which must be incurred by efficiently and economically operated facilities” and because the rates do not adequately “take into account the situation of hospitals [including plaintiff] which serve a disproportionate number of low-income patients with special needs;” the rates are properly characterized as arbitrary.
I accept defendants’ argument that plaintiff has made no showing, on this record, that Medicaid patients are being denied reasonable access to in-patient hospital care. But the fact that such denial of access has not yet occurred provides no assurance of reasonable access in the foreseeable future. Indeed, the evidence presented justifies the conclusion that, absent prompt corrective measures, denial of reasonable access to inpatient hospital care for Medicaid patients is almost inevitable.
VI. REMEDY
It is clear that the plaintiff is entitled to a declaratory judgment, and to an injunc
tion requiring the defendants to bring the Pennsylvania Medical Assistance Plan into conformity with federal requirements. That much relief will therefore be ordered at this time. What is less clear, however, is the extent to which retroactive relief can be ordered, given the limitations of the Eleventh Amendment and
Edelman v. Jordan,
415 U.S. 651, 94 S.Ct. 1347, 39 L.Ed.2d 662 (1974).
As explained in
Bennett v. White,
865 F.2d 1395, 1408 (3d Cir.1989),
“It is one thing for a state to insist that the Eleventh Amendment prevents retroactive relief which affects its fisc. It is quite another to insist that the State can confer an unwanted bonanza upon the United States by refusing to make an accounting and recover available funds from the federal treasury.”
The Eleventh Amendment does not, presumably, preclude prospective relief which directly affects the state
treasury
— e.g., an injunction requiring that, pending modification of the Medical Assistance Plan, the defendants make payments to plaintiff at a higher level, so as to preclude further damage from the statutory violations. It would also be appropriate, under
Bennett v. White, supra,
to require the defendants to take appropriate actions to recapture, so far as possible, additional sums from the federal treasury retroactively. But I believe it preferable to obtain additional clarification on these points before framing a final order addressing those issues.
ORDER
AND NOW, this 24th day of January, 1990, upon consideration of the arguments and evidence presented, it is hereby ADJUDGED that the Pennsylvania Medical Assistance Plan:
1. was adopted without compliance with the procedural requirements of the applicable federal laws and regulations;
2. applies payment rates which are arbitrary, and inadequate to meet the costs which must be incurred by efficiently and economically operated hospitals; and
3. applies rates which fail to adequately take into account the circumstances of hospitals which serve a disproportionate number of low-income patients with special needs.
IT IS THEREFORE ORDERED that the defendants shall promptly take all necessary steps to bring the Pennsylvania Medical Assistance Plan into compliance with federal requirements consistent with the accompanying Memorandum.
FINALLY, the parties are directed to clarify within ten (10) days their respective positions concerning the appropriate level of interim payments pending modification of the Plan, and possible retroactive recovery of additional federal or other funds, consistent with Eleventh Amendment restrictions.