In 1979 the Secretary of Health, Education, and Welfare1 issued a new regulation for reimbursing Medicare health providers for the portion of malpractice insurance premiums which is attributable to Medicare patients (codified at 42 C.F.R. Sec. 405.452(b)(1)(ii)(1982)). This regulation, commonly referred to as the Malpractice Rule, reimburses malpractice premiums based on the ratio of Medicare malpractice claims paid to total malpractice claims paid during the year for which reimbursement is sought and the preceding four years. The District Court2 concluded that the Malpractice Rule is invalid because it fails to give the Medicare provider reasonable reimbursement for the premium costs incurred by Medicare patients, contrary to the Medicare Act, 42 U.S.C. Sec. 1395 et seq. (1982), because it is arbitrary and capricious, see 5 U.S.C. Sec. 706 (1982), and because the Secretary failed to supply an adequate basis-and-purpose statement for the rule as required by 5 U.S.C. Sec. 553(c)(1982).3 We affirm.I.

On March 15, 1979, the Secretary published a notice of proposed rulemaking that expressed her desire to promulgate a new rule for reimbursing malpractice premiums. This rule was intended to prevent Medicare from reimbursing a "disproportionate" share of the premiums. 44 Fed.Reg. 15,744 (1979) (to be codified at 42 C.F.R. pt. 405) (proposed March 5, 1979). Under the regulation then in effect, Medicare reimbursed a hospital for malpractice premiums in proportion to the utilization that Medicare patients made of its services during the year in question. Malpractice premiums were pooled together with other general and administrative (G & A) costs, and then reimbursed according to the ratio of Medicare patient utilization. Thus, if Medicare patients accounted for 30% of the patient-days at a hospital for a given year, Medicare would reimburse the hospital for 30% of G & A costs, including malpractice premiums. This method assumed that costs for which Medicare patients incur a disproportionate share, such as the costs of processing Medicare paperwork, would be balanced out by costs for which non-Medicare patients incur a disproportionate share, such as the G & A costs associated with younger patients.

The Secretary proposed the new regulation because she believed that the existing "method result[ed] in Medicare paying a disproportionate amount of malpractice costs." 44 Fed.Reg. 15,745 (1979). She based this belief on the following observations:

A study conducted by an HEW consultant4 indicates that malpractice awards for Medicare and Medicaid patients are significantly lower in amount than losses for other patient population. The lower awards for these patients result because their income potential and life expectancy are less than the remainder of the patient population. Thus, the use of overall Medicare utilization to allocate malpractice costs results in Medicare paying for a disproportionate amount of malpractice costs.

44 Fed.Reg. 15,745 (1979).

The proposed rule removed malpractice insurance premiums from the G & A pool and placed them in a separate fund where they would be reimbursed using a new ratio instead of the utilization ratio. The new ratio is determined by dividing the malpractice losses paid to Medicare patients by the total malpractice losses paid to all patients over the current year and the preceding four years. 44 Fed.Reg. 15,745 (1979). If a hospital did not have malpractice losses over the five-year period, its reimbursement ratio was to be determined using an actuarial estimate of Medicare's share of the malpractice costs of that particular hospital.

The Secretary received nearly 600 comments on the proposed rule from health-care institutions, consumers, accounting firms, insurance companies, actuaries, health-care consultants, physicians and nurses, and Medicare beneficiaries. All of the comments opposed the rule. The Secretary nevertheless issued the final Malpractice Rule, which differs from the proposed rule only in its treatment of providers with no malpractice losses during the five-year period. Instead of compensating these providers using a ratio arrived at through an actuarial estimate, the final rule compensates them using the national ratio of malpractice awards paid to Medicare patients to malpractice awards paid to all patients.5

The plaintiffs make three separate challenges to the Malpractice Rule. They argue that the Malpractice Rule is arbitrary and capricious in violation of the Administrative Procedure Act (APA), 5 U.S.C. Sec. 706(2)(A) (1982); that the rule fails to contain an adequate basis-and-purpose statement as required by 5 U.S.C. Sec. 553(c)(1982); and that the Rule violates the Medicare Act, 42 U.S.C. Sec. 1395f(b)(1)(1982), in that it fails to reimburse providers for the "reasonable cost" of their services.

These three issues have been thoroughly canvassed by the Seventh Circuit in St. James Hospital v. Heckler, 760 F.2d 1460 (1985); see also Walter O. Boswell Memorial Hospital v. Heckler, 749 F.2d 788 (D.C.Cir.1984). In St. James, the Seventh Circuit concluded that the Malpractice Rule was arbitrary and capricious because the Secretary based the rule on a faulty study and failed to consider critical aspects of the problem. 760 F.2d at 1465-69. It likewise found the basis-and-purpose statement inadequate because the Secretary failed to address significant criticisms made of the Rule. Id. at 1469-70. Finally, it held that the Rule violated the Medicare Act because the Secretary failed to show that the G & A pool had, before the Rule was adopted, impermissibly shifted costs to the Medicare program, and because the Rule fails to reimburse providers adequately for the actual costs of obtaining malpractice insurance for Medicare patients. Id. at 1471-72. We find the reasoning of the Seventh Circuit persuasive and see no need to add extensive discussion of our own. We make the following brief additional comments.

Under 5 U.S.C. Sec. 706, we are required to reject an agency rule which is arbitrary and capricious. The Supreme Court has recently reiterated the standards for this evaluation. Motor Vehicle Manufacturers Association of the United States, Inc. v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 103 S. Ct. 2856, 77 L. Ed. 2d 443 (1983). A rule is arbitrary and capricious if "the agency [fails to] articulate a satisfactory explanation for its action, including a 'rational connection between the facts found and the choice made.' " Id. 103 S.Ct. at 2866-67, quoting Burlington Truck Lines v. United States, 371 U.S. 156, 168, 83 S. Ct. 239, 246, 9 L. Ed. 2d 207 (1962). "Normally, an agency rule would be arbitrary and capricious if the agency ... entirely failed to consider an important aspect of the problem, [or] 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." 103 S. Ct. at 2867.

The Secretary did not offer a plausible explanation for the Malpractice Rule or consider several important aspects of the problem it was intended to address. Indeed, the Secretary failed to substantiate her conclusion that a problem ever existed, namely, that Medicare was paying a disproportionate amount of G & A costs. She arrived at this conclusion through several leaps in reasoning. She began with the premise, derived from the Westat Study, that "malpractice awards for Medicare ... patients are significantly lower in amount than losses for other patient population." 15 Fed.Reg. 15,745 (1979). From this she concluded that "Medicare pay[s] a disproportionate amount of malpractice [premium] costs." Id. And from this and the observation that malpractice costs are a significant part of the G & A pool, she finally concluded that Medicare was paying more than its share of the G & A pool. This logic crumbles upon analysis.

The Secretary's initial premise, that Medicare patients on average receive significantly lower malpractice awards than non-Medicare patients, came from the Westat Study. This study, however, came under significant criticism regarding the adequacy of its data base to generalize to national totals and the accuracy of the statistics relied on.6 Since these criticisms cast serious doubt on the premise grounding the Secretary's explanation, her failure to respond to them was arbitrary and capricious.7

Next, the Secretary's conclusion that Medicare was paying a disproportionate share of malpractice premiums rests on a leap of logic, quite apart from the questionable applicability of the Westat Study. Even if it is true that Medicare patients recover significantly lower awards, it does not necessarily follow that Medicare pays a significantly disproportionate amount of malpractice premiums. This would follow only if the Secretary showed that a reduced amount of malpractice recoveries for a specific patient population would result in significantly lower premiums for that population. Although the Secretary now points to some evidence in the record that loss experience makes up a significant part of premium levels, other factors, such as administrative costs, are also important ingredients in the setting of premiums. The agency itself made no attempt before promulgating the rule to examine the relationship between premiums and the various cost factors that insurance carriers may take into account. "The short--and sufficient--answer to [appellant's] submission is that the courts may not accept appellate counsel's post hoc rationalizations for agency action." Motor Vehicle, 103 S. Ct. at 2870. We also find it highly significant that insurers do not appear to charge higher premiums to hospitals with fewer Medicare patients.

The Secretary's final conclusion, that Medicare was paying a disproportionate amount of G & A costs, is likewise without rational support. Even assuming that the Secretary had justified her conclusion that Medicare was paying a disproportionate share of malpractice premiums, it does not follow that a disproportion existed in the overall allocation of G & A costs. To show this, the Secretary must show that other disproportions in her favor did not compensate for the malpractice disproportion. The Secretary made no such showing. Although she presented evidence that malpractice premiums increased dramatically during the 1970's, this still gives no indication whether other costs in her favor did not rise to a similar extent. Without this information, she was unable to draw any rational conclusions about the overall G & A pool which would warrant the issuance of a regulation with such substantial economic effects on health-care providers.

Under the Medicare Act, providers are entitled to reimbursement for the "reasonable cost" of their services, 42 U.S.C. Sec. 1395f(b)(1)(1982), defined as "the cost actually incurred," 42 U.S.C. Sec. 1395x(v)(1)(A)(1982). The statute also requires the Secretary to define "reasonable cost" so as to prevent cost-shifting between Medicare and non-Medicare patients. 42 U.S.C. Sec. 1395x(v)(1)(A)(1982). The Malpractice Rule and the procedure used to promulgate it violate these statutory provisions in several respects.

First, as discussed above, the Secretary has not shown that a disproportion existed in the overall G & A pool which would require some remedial measure such as the Malpractice Rule. Second, the Secretary has not shown that a system of reimbursing malpractice premiums based solely on the proportion of malpractice losses will accurately reflect premium costs. The Secretary herself recognizes that a large part of the cost of malpractice insurance is the expense of processing claims. This cost is totally independent of a plaintiff's recovery. Thus, regardless of the loss history for a given group of patients, this cost alone will prevent malpractice premiums for that group from decreasing below a certain level. However, despite numerous criticisms of the Rule on this ground, the Secretary did not address the point.

Third, the Secretary did not discuss any of the alternatives to the Malpractice Rule proposed by the commenters. One promising suggestion was that providers obtain separate policies for Medicare patients. This alternative would clearly result in hospitals' receiving only their actual costs for Medicare malpractice premiums, whereas, as just noted, the Malpractice Rule only roughly approximates those actual costs.

Fourth, the five-year period used by the Malpractice Rule leads to some peculiar results. Some providers will be over-reimbursed, and others under-reimbursed or not at all. For example, Humana of Illinois, Inc., one of the plaintiffs in the St. James case, received no reimbursement for 1980, despite its Medicare patient utilization ratio of 49.38%. Humana had paid only a single $750 malpractice claim during the five-year period, and this claim was to a non-Medicare patient. (If it had been to a Medicare patient, Humana would have received 100% reimbursement for its 1980 premiums.) The Secretary argues that over the long run these discrepancies will even out, but again this point was not discussed in the comments accompanying the final rule.8 The Secretary did not attempt to assess the extent to which such discrepancies would occur or the likelihood that, over time, the Malpractice Rule would balance them out.

The Secretary asks that we direct the District Court to remand the case to her for further rule-making. We decline to do so. The Malpractice Rule is invalid and was so from its inception. Unless special circumstances are present, which we do not find here, prior regulations remain valid until replaced by a valid regulation or invalidated by a court. Abington Memorial Hospital v. Heckler, 750 F.2d 242, 244 (3d Cir.1984). The hospitals are entitled to be paid under the prior regulations, which, not having been validly replaced, remain in effect. We express no view on whether, or the extent to which, a new regulation could be made retroactive.9 The District Court will remand this case to the Provider Reimbursement Review Board for further proceedings consistent with this opinion.