the Secretary managed to reply adequately in two scant columns of Federal Register prose to the 600-plus comments that occupy five thick volumes of the administrative record. Unfortunately, no such miracle has been witnessed by this court, which instead finds the Secretary's response woefully inadequate.
Particularly disturbing is the complete lack of response to comments that challenged the statistical validity of the Westat Report, the sole empirical study on which the Malpractice Rule was based. Several commenters questioned the adequacy of the data base. For example, the American Hospital Association noted that providers who self-insure were excluded from the study even though self-insurers constituted 20% of all community hospitals in 1976 and 53% of community hospitals over 500 beds. Administrative Record, Vol. 5, Tab A, at p. 9. The Hospital Corporation of America argued that because the data were collected from only nine insurance companies -- out of the 50 which then offered hospital malpractice coverage -- and because only claims closed during a four-month period (July to October 1976) were examined, the study offered a quantitatively inadequate sample from which to extrapolate. Vol. 6, Tab H, at 6. The preamble to the final rule contains not a word acknowledging that any commenter criticized the Westat Report; far less did it make an adequate response to those criticisms.
Nor did the Secretary address other points raised by commenters save in a highly conclusory fashion. For example, the American Hospital Association noted that 30% to 40% of the cost of insurance is attributable to administrative expenses such as risk management, and that claims handling expenses (such as incident investigations and legal fees) do not necessarily parallel ultimate payments. Vol. 5, Tab A, at 8. The Association cited a study by the Rand Corporation which concluded that "only about 35 cents of every dollar in malpractice premiums is paid to successful claimants." Id. at 5-6. In light of these shortcomings, the Secretary's basis and purpose statement cannot be considered adequate.
Plaintiffs also contend that adoption of the Malpractice Rule was arbitrary, capricious, and an abuse of discretion, thus violating § 706(2)(A) of the APA. The Supreme Court recently reiterated the proper role for a court reviewing agency action challenged under that provision: to determine "whether the decision was based on a consideration of the relevant factors and whether there had been a clear error of judgment." Motor Vehicle Manufacturers Association v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 103 S. Ct. 2856, 2865, 2867, 77 L. Ed. 2d 443 (1983) (citations omitted). Plaintiffs argue that the Secretary has committed such an error by basing the Malpractice Rule on inadequate empirical information. Reliance on faulty data has been held to be a clear error of judgment in analogous circumstances. See Almay, Inc. v. Califano, 187 U.S. App. D.C. 19, 569 F.2d 674 (D.C. Cir. 1977). In Almay, the D.C. Circuit held that the Food and Drug Administration had acted arbitrarily in adopting a regulation on the basis of a survey, where the director of the bureau that sponsored and designed the survey had seriously questioned the statistical integrity of the survey results. Id. at 682.
Here, the authors of the Westat Report noted that "conclusions must be drawn very cautiously because of the possibility of biased data." Westat Report, Administrative Record, Vol. 3, at 2-2. One source of bias arose because many hospitals self-insure, and claims against self-insurers obviously could not show up in a survey of claims closed by insurance companies. Id. More telling is a memorandum to the HEW project officer in charge of the Medical Malpractice Study from senior analysts at Westat, who were authors of several chapters of the Final Report. The memorandum's authors noted that the design of the 1976 study made it
"virtually impossible to generalize beyond the sample . . . to annual totals or to national totals, to cite the most elementary limitation. Beyond that, there was no assurance that the sample was in any way representative of hospitals[,] of physicians, of injuries, of geographic regions, of claimants."
Record, Vol. 4, Tab 3.
Of course, an agency need not await perfect data before taking regulatory action. There are limits, however, to the degree of imperfection that is permissible, which the Secretary has surpassed here.
Plaintiffs further challenge the rule as arbitrary and capricious, on the grounds that it produces "bizarre and unfair reimbursement consequences" and is inconsistent with other Medicare reimbursement practices.
Because the issues raised by this contention are virtually identical to those presented by plaintiffs' challenge to the rule under the Medicare Act, these issues are addressed in Part IV.
Finally, plaintiffs contend that the Malpractice Rule is substantively invalid under the Medicare Act because it fails to reimburse hospitals for insurance costs reasonably incurred in treating Medicare patients.
The Act requires that hospitals be reimbursed for "reasonable costs or customary charges," whichever is less. 42 U.S.C. § 1395f(b). The parties to this litigation agree that it is the "reasonable costs" standard which is involved here. The Act's definition of reasonable costs is lengthy but leaves the Secretary considerable discretion to adopt regulations specifying methods for determining reasonable cost. Id. § 1395x(v)(1)(A). The statute does specify, however, that "the reasonable cost of any services shall be the cost actually incurred, excluding therefrom any part of incurred cost found to be unnecessary in the efficient delivery of needed health services." Id. Furthermore, the Act requires that reimbursement standards "take into account both direct and indirect costs . . . in order that . . . the necessary costs of efficiently delivering covered services to [Medicare patients] will not be borne by [non-Medicare patients]." Id.
The government argues that the Secretary should be accorded substantial deference in establishing the method or methods for determining reasonable costs. Of course, this court cannot overrule the Secretary's interpretation merely because it might have interpreted the term "reasonable cost" in another fashion. See Batterton v. Francis, 432 U.S. 416, 425, 97 S. Ct. 2399, 2405, 53 L. Ed. 2d 448 (1977). Nonetheless, as the Supreme Court recently noted:
While reviewing courts should uphold reasonable and defensible constructions of an agency's enabling Act . . . they must not "rubberstamp . . . administrative decisions that they deem inconsistent with a statutory mandate or that frustrate the congressional policy underlying a statute."
Bureau of Alcohol, Tobacco and Firearms v. Federal Labor Relations Authority, 464 U.S. 89, 97, 104 S. Ct. 439, 78 L. Ed. 2d 195 (1983) (citations omitted.) See also Pacific Coast Medical Enterprises v. Harris, 633 F.2d 123, 131 (9th Cir. 1980) ("Even though the Medicare reimbursement area is complex, and to a great degree left to the Secretary to structure, his interpretations are nonetheless subject to our examination.") Bethlehem Steel v. EPA, 651 F.2d 861, 867 (3d Cir. 1981).
In determining the validity of the plaintiffs' challenge under the Medicare Act and under § 706 of the APA,
we must examine how the rule operates in practice. Under the rule, a hospital is reimbursed for that percentage of its malpractice insurance costs equal to the dollar percentage of malpractice awards it has paid in the past five years to Medicare (as opposed to non-Medicare) patients. Thus, if a hospital had three malpractice claims within that time, one of which, for $100,000, was paid to a Medicare patient, and the remaining claims, for $200,000, involved non-Medicare claimants, the hospital would be reimbursed for one-third of its malpractice premiums for the current reporting period. If all three claims had been paid to Medicare patients, the hospital would be reimbursed for 100% of its malpractice insurance expenditures; but if all three claims had been paid to non-Medicare patients, then the hospital would receive not one cent in reimbursement. The Malpractice Rule takes no account of the hospital's rate of utilization by Medicare patients, so the figures in the above example would apply regardless of whether 1% or 99% of the hospital's patients were Medicare recipients. If a hospital has paid no malpractice claims at all during the past five years, reimbursement is figured according to the national ratio currently set at 5.1%. Such hospitals would be reimbursed for only 5.1% of their malpractice insurance costs, again without regard to their Medicare utilization levels.
The government argues that the Secretary has acted within permissible discretion in establishing a rule that reimburses hospitals under the Medicare program according to the portion of malpractice loss that Medicare patients have historically created. The superficially appealing logic of this formulation, however, entirely neglects two key facts: first, a hospital's malpractice insurance premiums are not established solely with reference to its paid claims history,
and second, malpractice insurance does not function solely to ensure that assets will be available to cover a successful claimant's damages. This latter fact is itself a function both of the nature of insurance and of Pennsylvania law. Under 40 Pa. Cons. Stat. Ann. § 1301.701, all hospitals must carry malpractice insurance. Such insurance is thus a necessary, though indirect, cost of providing health care to Medicare patients, because a Pennsylvania hospital cannot provide any services absent such insurance.
Furthermore, malpractice insurance protects an insured's assets against losses that could otherwise threaten its financial stability.
Continuity of service benefits all patients, Medicare and other. To the extent that malpractice service costs represent the cost of protecting hospital assets, they are "reimbursable to the extent that this indirect cost benefitted Medicare patients." Presbyterian Hospital v. Harris, 638 F.2d 1381, 1386 (5th Cir.), cert. denied, 454 U.S. 940, 102 S. Ct. 476, 70 L. Ed. 2d 248 (1981). Clearly, the extent of benefits to Medicare patients is not simply a function of how many Medicare versus non-Medicare claims a hospital has paid during the past five years, and the Malpractice Rule consequently fails to reimburse actual costs incurred by hospitals.
The statute plainly requires that indirect costs be allocated so as to ensure that non-Medicare patients do not bear costs attributable to Medicare patients, and vice-versa. Since we are dealing with indirect costs, there is in any such allocation, almost by definition, an element of uncertainty. Reasonable estimates and extrapolations are certainly permissible; mathematical precision is not to be expected. But this does not mean that the Secretary is free to choose an allocation formula which is least likely to reflect reality -- indeed, as in this case, a formula which seemingly guarantees that the statutory mandate will inevitably be violated in most cases.
It bears repeating that the following facts are, on this record, undisputed: (1) the ratio between malpractice claims paid on account of Medicare patients and malpractice claims paid on account of non-Medicare patients over the most recent five-year period has very little, if any, relationship to the hospital's current cost of maintaining malpractice protection. (2) The "national average" (currently estimated at 5.1%) of that ratio has even less bearing on the current malpractice coverage costs of any particular hospital. (3) The data relied upon in the Westat Report, including the data which produced the 5.1% figure, included both payments made to satisfy the liability of hospitals and claims paid to satisfy the liability of physicians. There is no evidence to support the conclusion that these combined payments provide reliable information about the malpractice payment experience of hospitals. Indeed, there is reason to believe that they do not: It would seem that Medicare patients stand a greater chance of being injured through the negligence of non-physician personnel than younger patients. (4) A very substantial portion -- certainly in excess of 50% -- of malpractice coverage expense is attributable to claim-handling, litigation, and other expenses unrelated to the amounts paid to claimants. Since the Westat data say nothing about the numbers of claims handled in each category, or the expenses attributable thereto, and since they include only payments in excess of applicable deductibles, even these limited data are open to serious challenge, not addressed in the administrative record.
It is not appropriate for this Court to attempt to specify what method of apportioning malpractice insurance costs should be adopted, notwithstanding plaintiffs' implicit argument to the contrary. And I certainly do not mean to suggest that the Secretary's perception that the exposure to malpractice insurance claims and costs attributable to non-Medicare patients is greater than that attributable to Medicare patients is necessarily invalid, or cannot be reflected in the allocation formula adopted. But, while the previous utilization rate method of allocation is not the only possible method, it seems self-evident that any method of allocation which totally ignores utilization rates fails to meet the requirements of the statute.
For the reasons discussed above, I have concluded that the Secretary violated the notice and comment provisions of the APA in adopting the malpractice rule, and that the rule itself is arbitrary and capricious, and fails to conform to the substantive provisions of the Medicare Act. I recognize that these conclusions run counter to other reported decisions on these questions. To date, four other courts have upheld the position of the Secretary. Athens Community Hospital v. Heckler, 565 F. Supp. 695 (E.D. Tenn. 1983); Cumberland Medical Center v. Heckler, 578 F. Supp. 39 (M.D. Tenn. 1983); Humana of Aurora, Inc. v. Heckler, No. 83-7-70 (D. Colo. Sept. 19, 1983); and Walter O. Boswell Memorial Hospital v. Heckler, 573 F. Supp. 884 (D.C.C. 1983). After careful consideration, I find myself unable to agree with those decisions.
The Cumberland and Humana courts merely followed the lead of the Athens Community Hospital court, without discussion. And the opinion of the Athens court reflects a rather limited analysis of the issues, as is evidenced by the fact that the administrative record appears not to have been part of the docket before the Athens court (see docket sheet, Appendix A to affidavit of H. Buckley Cole, Exhibit C of plaintiffs' memorandum in response to defendants' reply memorandum). I find it difficult to square this approach with the "searching and careful" review mandated by Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416, 28 L. Ed. 2d 136, 91 S. Ct. 814 (1971). Moreover, the Athens court did not address the anomalous results that the rule produces in practice.
A more complete exploration of the issues involved here is contained in the opinion of my esteemed colleague, Judge Bryant, in the Boswell case. Judge Bryant obviously was troubled by the "cloudy" aspects of the administrative record, and found the issue of substantive compliance with the Medicare Act the "thorniest" problem in the case, but the opinion reflects his ultimate conclusion that, in the long run, the imbalances created by the rule would probably balance out in the case of any particular hospital. If there is a distinction between Boswell and the present case, it lies in the fact that the Boswell plaintiff had a high percentage of Medicare patients and might therefore actually benefit from the new rule, in the long run. In the present case, I am not at liberty to indulge such an assumption. Be that as it may, with all due deference to the contrary opinions of other judges, I remain unpersuaded that the rule is valid.
[EDITOR'S NOTE: The following court-provided text does not appear at this cite in 576 F. Supp.]
AND NOW, this 8th day of December, 1983, it is ORDERED:
1. Defendants Motion for Summary Judgment is DENIED.
2. Plaintiffs' Motion for Summary Judgment is GRANTED.
3. It is ORDERED, ADJUDGED and DECLARED as follows:
a. that the Malpractice Rule challenged in this proceeding was promulgated in violation of the notice and comment procedures of 5 U.S.C. § 553;
b. that the Malpractice Rule is arbitrary, capricious, and an abuse of the Secretary's discretion, and is therefore substantively invalid under 5 U.S.C. § 706(2)(A);
c. that the Malpractice Rule is in conflict with § 1395(x)(v)(1)(A) of the Medicare Act.
4. This matter is remanded to the Secretary for further consideration, in conformity with the views expressed in the accompanying Opinion.