CUMMINGS, Chief Judge.

This appeal arises from the challenge by St. Marys Hospital Medical Center ("St. Marys") to the Secretary of Health and Human Services' (the "Secretary") method of calculating St. Marys' apportionment of laboratory costs between Medicare and non-Medicare patients under Title XVIII of the Social Security Act, 42 U.S.C. § 1395 et seq. ("Medicare"). The district court found in favor of St. Marys. We reverse.

I

St. Marys is a short-term acute-care hospital located in Madison, Wisconsin. In 1972 it introduced a program, later known as Shared Laboratory Services ("SLS"), that was designed to spread the hospital's fixed laboratory costs. Through SLS, St. Marys, in cooperation with Methodist Hospital in Madison, sold laboratory services to non-hospital patients, including private physicians, clinics, and small hospitals. These clients would provide several services that the test-taking procedures required to be done and that St. Marys fulfilled for its own patients, viz., ordering the tests, drawing and processing the sample, recording the test results in the patient's records. SLS provided its own courier service to transport samples to be tested and Methodist Hospital billed the SLS clients. The two hospitals shared the expenses of these two activities. St. Marys states that its SLS fees are lower than those it charges its own patients because of competition with commercial laboratories and allegedly lower costs (Br. 6) St. Marys admits that the costs of the SLS program exceed the revenue derived from it ( id. at 30 n. 21.)

Medicare reimburses St. Marys for the laboratory services provided Medicare patients by determining first the ratio of Medicare charges to total charges, and then multiplying total costs by that ratio. Thus if Medicare patients account for thirty percent of the laboratory's charges, then Medicare will reimburse the hospital thirty percent of its laboratory's total costs. This formula assumes that the ratio of Medicare charges to total charges equals that of Medicare costs to total costs. To protect the integrity of this assumption, the Secretary requires a provider to "gross-up" any discounted charges to the level of the regular charges made to all other paying patients before the hospital calculates the ratio of Medicare charges to other charges. Not doing so would mean that total charges would equal less than would otherwise be the case, so that Medicare charges would comprise a larger percentage of the total, resulting in Medicare's bearing a disproportionately heavy burden of the laboratory's cost.

In 1978, St. Marys attempted to account for SLS' lower charges by revenue offset, subtracting SLS revenues from total laboratory costs and total charges before calculating the ratio of Medicare charges to other charges. The intermediary overseeing St. Marys' Medicare reimbursement refused to accept this method, requiring St. Marys to gross-up SLS charges instead. In 1979, St. Marys calculated the ratio of Medicare charges to total charges by including SLS revenues at their actual and not their grossed-up level. Again the intermediary disallowed this approach and required St. Marys to gross-up the SLS charges. In both years St. Marys justified its action by claiming that the SLS services were not similar to tests provided hospital patients, so that grossing-up was unwarranted. Moreover, St. Marys claimed that SLS services cost less than did the same tests provided to its patients.

St. Marys appealed both the 1978 and 1979 adjustments to the Provider Reimbursement Review Board ("PRRB" or the "Board"). The hospital also retained Ernst Whinney, a national accounting firm, to conduct an industrial engineering work-measurement study to ascertain whether St. Marys could document both of its assertions — that the SLS services were markedly different from, and substantially less expensive than, tests provided its patients. The PRRB rejected the Ernst Whinney study as neither accurate nor verifiable and dismissed St. Marys' objections. The Secretary of course did not appeal this decision, so that it became the final administrative decision. St. Marys appealed to federal district court under 42 U.S.C. § 1395oo(f)(1). The trial court reversed the PRRB in a judgment entered December 30, 1983. Although agreeing that the Ernst Whinney study was inaccurate "in some respects" (Mem. Order of Dec. 24, 1983 at 2), it found that SLS costs in fact were always less than those of tests for hospital patients. The court remanded to the PRRB, ordering it to resolve any inaccuracies in St. Marys' study and to apply a method of cost apportionment that would not require grossing-up SLS charges. The Secretary filed a motion for reconsideration on January 9, 1984, but the trial court reaffirmed its decision on January 17, 1984. Notice of appeal was filed in this Court on March 15, 1984. We reverse.

II

Plaintiff contends that the Secretary's notice of appeal was untimely because her motion for reconsideration did not seek specific relief, and, even if requisite specificity was present, it was not timely served. Fed.R.App.P. 4(a)(1) permits the government sixty days in which to file an appeal from a judgment. The filing of one of four post-judgment motions tolls this period. Our jurisdiction depends on the adequacy of the motion for reconsideration filed by the Secretary.

Plaintiff misunderstands the substantive requirements for a Rule 59 motion. The motion need only be in writing, specify the grounds on which the motion is based, and request relief. Fed.R.Civ.P. 7(b)(1); Martinez v. Trainor, 556 F.2d 818, 819-820 (7th Cir. 1977). Even if the Federal Rules required that the requested relief as well as the grounds for it be set forth with specificity, the Secretary's instant motion meets these requirements. St. Marys' reliance on Western Transportation Co. v. E.I. Du Pont de Nemours and Co., 682 F.2d 1233 (7th Cir. 1982), is misplaced, for that case deals with the particularity of grounds for relief, not the specificity of the relief requested. In Western Transportation, the disputed motion asked the district court to reexamine its dismissal of some of the moving party's claims, without delineating any reasons whatsoever for the request. In contrast, the motion at issue here identified two specific errors in reasoning that the secretary believed the trial court had made. The motion also requested that if the trial judge did not agree with the Secretary's interpretation, the district court either modify or clarify its remand order. Supp.App. 3-5. As such, the motion detailed both the action it requested and the reasons for its request, thus falling well within the specificity requirements of the Federal Rules.

Since the motion made does request "that the judgment be altered or amended," Western Transportation, 682 F.2d at 1236, it falls within the intent of Fed.R.Civ.P. 59(e). The critical factor in classifying a motion under Rule 59 is whether "the motion questioned the substantive validity of the judgment and not merely its form," St. Paul Fire and Marine Insurance Co. v. Continental Casualty Co., 684 F.2d 691, 694 (10th Cir. 1982). The Secretary fashioned her motion as falling within Fed.R.Civ.P. 52(b) rather than Rule 59(e) (Supp.App. 2 n. 1), but that fact is not dispositive. Browder v. Director, 434 U.S. 257, 261-262 n. 5, 98 S.Ct. 556, 559-560 n. 5, 54 L.Ed.2d 521. "'Any motion that draws into question the correctness of the judgment is functionally a motion under Civil Rule 59(e), whatever its label.'" United States v. City of Chicago, 631 F.2d 469, 474 (7th Cir. 1980) (quoting 9 Moore's Federal Practice ¶ 204.12[1], at 4-67 (2d ed. 1980)).

The more serious objection is that involving service. The Secretary concedes that through inadvertence her motion for reconsideration was not served on opposing counsel, though it was timely filed (Reply Br. 20). We assume arguendo that the motion would have had to have been both filed and served before it would have been timely for purposes of Fed.R.App.P. 4(a)(1). Nonetheless, the exception the Supreme Court enunciated in Thompson v. Immigration and Naturalization Service, 375 U.S. 384, 84 S.Ct. 397, 11 L.Ed.2d 404 ( per curiam), applies here. In Thompson the Rule 59 motion at issue was untimely because it was filed and served more than ten days after entry of judgment. The notice of appeal filed in the case would have been timely had the post-judgment motion tolled the appeals period. Opposing counsel did not object to the motion and the district court explicitly informed the moving party that its motion for a new trial was timely. In these circumstances, the Supreme Court refused to find that the court of appeals lacked jurisdiction and ordered it to hear the case on the merits. Id. at 386-387, 84 S.Ct. at 398-399.

The case before us is very similar. A post-judgment motion was filed that would have been timely had it been served. The Secretary's notice of appeal also would have been timely had the post-judgment motion tolled the appeals period. In addition, the opposing party took no formal action cognizable by the court to oppose the Secretary's motion. The district court gave the Secretary no explicit assurance that her motion was timely, as occurred in Thompson, but explicit assurance is unnecessary. In Textor v. Board of Regents of Northern Illinois University, 711 F.2d 1387, 1390-1391 (7th Cir. 1983), we applied the Thompson exception to a motion that was both filed and served seven days late. There the district court gave the moving party implicit assurance that her motion was timely by considering it without mentioning any timeliness problem, and opposing counsel failed to raise the issue. The district judge in the instant situation gave the Secretary the same implicit assurance by issuing a six-page memorandum and order explaining his ruling denying reconsideration, and by not retracting that order sua sponte once St. Marys' counsel informally advised him of the lack of service.

Our applying the Thompson exception does not prejudice the plaintiff or greatly impose on the policies underlying the federal rules. The Secretary's motion itself met all substantive requirements of specificity and timeliness of filing. Cf. Martinez v. Trainor, supra (court rejected party's attempt in effect to extend filing deadline by filing a timely, skeletal motion that specified no grounds for the motion and supplementing the motion with specific grounds after the ten-day deadline had passed). Here the judge below ruled on the motion promptly, and he decided in the plaintiff's favor. The notice of appeal was filed only sixteen days after the deadline that would have applied had the post-judgment motion not been made. If St. Marys wished to object to timeliness, it should have raised that objection within sixty days of entry of the judgment, while the Secretary still had an opportunity either to file a notice of appeal or to request an extension of time in which to file an appeal. Plaintiff's counsel's attempt to capitalize on the Secretary's reliance he helped induce by seeking no relief in his letter to the district judge (see n. 8 supra) is hardly commendable. Considering all the circumstances, we hold that the decisions in Thompson and Textor govern. Because the district court's judgment was final under 28 U.S.C. § 1291 and the notice of appeal was timely under Thompson and Textor, the motion to dismiss the appeal is denied.

III

The issue for review is the appropriateness of the intermediary's requiring St. Marys to gross-up its SLS charges to the level of its other laboratory charges before calculating the ratio of Medicare charges to total charges. As explained by the district judge,

"grossing-up" means that patient charges which are lower than those charged to regular hospital patients, including Medicare patients, are raised to the level of the regular patient charges before apportioning the laboratory charges between Medicare beneficiaries and others * * * so that Medicare reimbursement truly reflects the costs incurred by the provider for Medicare patients.

Mem. Order of Dec. 24, 1983 at 3-4.

We were informed at oral argument that the issue is one of first impression for the circuit courts of appeals.

The Medicare statute permits reimbursement of a participating hospital's reasonable direct and indirect costs necessary to the efficient delivery of health services to Medicare beneficiaries. 42 U.S.C. § 1395f(b)(1). The reimbursed costs should be actual costs, but the statute gives the Secretary wide latitude in developing methods of determining costs. Id. § 1395x(v)(1)(A).

The Secretary's regulations must receive great deference as actions of an administrative agency. A court may not invalidate them unless they are unsupported by substantial evidence in the record or are "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. § 706(2)(A), applicable to Medicare by virtue of 42 U.S.C. § 1395oo(f)(1). Methods of reimbursement that shift costs properly allocable to the care of Medicare patients to non-Medicare patients, and vice versa, are impermissible. 42 U.S.C. § 1395x(v)(1)(A). If the Secretary's decision meets these requirements, we cannot overturn it merely because we might have decided differently. Northwest Hospital, Inc. v. Hospital Service Corp., 687 F.2d 985, 990 (7th Cir. 1982).

Medicare regulations do permit providers, under certain circumstances, to use alternatives to orthodox methodology. The alternative advocated must be demonstrably "more sophisticated." 42 C.F.R. § 405.453(d)(2)(ii). The burden is on the provider to produce verifiable and auditable cost data to persuade the intermediary to accept the alternative methodology. Id. § 405.453(a), (c).

The grossing-up requirement has its genesis in 42 C.F.R. § 405.452(d)(6) (1982) (recodified without change in 1983 as § 405.452(b)): "Charges refer to the regular rates for various services which are charged to both beneficiaries and other paying patients who receive the services. Implicit in the use of charges as the basis for apportionment is the objective that charges for services be related to the cost of the services." The Administrator of the Health Care Financing Administration reminded intermediaries in March 1980 in Intermediary Letter 80-8 that hospitals should gross-up discounted laboratory charges such as those for the SLS services St. Marys provides in order to avoid cross-subsidization. Supp.App. 96. This directive assumes that the cost of SLS-type services approximates that of the corresponding testing services the hospital provides its regular patients. The hospital's decision to charge its clients less, for reasons of competition or for any other reason, should not affect the cost apportionment process.

In order to avoid grossing-up, St. Marys must establish that its SLS services in fact differ from laboratory services provided its own patients, and that the SLS services are less costly. Even if the SLS tests differ from the tests St. Marys provides its own patients, as long as their cost to the laboratory equals that of the hospital's tests of its patients, grossing-up maintains a constant cost-to-charge ratio. With this ratio the same, the ratio of Medicare charges to total charges will remain constant, and Medicare reimbursement will not be disturbed. Requiring grossing-up in that situation would be rational, and the Secretary's decision could be justified. Only when the cost of a dissimilar service is less than the hospital's service would grossing-up unfairly minimize the amount of reimbursement to which St. Marys would be entitled. Consequently the factor determinative of the acceptability of the Secretary's requiring St. Marys to gross-up its SLS charges is whether a cost differential exists. St. Marys' failure to prove that such a differential existed with an accurate and verifiable cost study justified the adverse PRRB decision. The district court erred in asserting that the PRRB had in fact accepted the existence of a cost differential but was quibbling over minor inaccuracies in St. Marys' study. This logic misinterprets the thrust of the administrative decision.

The crux of the PRRB's decision was its conclusion "that the Provider's study did not establish the cost of providing services to non-Provider patients for the years 1978 and 1979" (PRRB Dec. 83-D42 at 6). The Board supported this conclusion by noting several discrepancies in the study: both St. Marys and its accounting firm indicated "a lack of confidence that the cost study" accurately reflected the SLS costs; the pathologists' fees were not properly allocated in the cost computation initially submitted to the Board; the Provider never used the results of the study to adjust its charges to either inpatients or SLS patients; the SLS charges had been established on an arbitrary basis, without consistent price mark-ups in later years; and the accounting firm's testimony indicated that SLS revenues were insufficient to meet the costs of the service.

Other evidence in the record also supported the PRRB's decision. Thus the Provider's study was conducted specifically to prove that SLS costs were less than laboratory testing St. Marys provided its own patients, suggesting an initial bias in the study. It focused on only one blood test, the SMAC, that accounted for forty-three to forty-six percent of total SLS test volume, thus ignoring the remaining tests that together accounted for the majority of the SLS test volume. It evidently did not properly allocated the costs of the SLS courier service. Ernst Whinney conducted its study for one week in January four years after the accounting years in issue, from which limited figure it purported to project annual costs accurately. Moreover, the study was never validated by subsequent reviews. The PRRB concluded that the cost computations "would have to be accurate and verifiable before their use could be considered as a more sophisticated method of cost apportionment" ( id. at 7). This conclusion, coupled to the earlier declaration that the study failed to establish SLS costs, provides a clear negative inference that the Ernst Whinney study was neither accurate nor verifiable.

The district court agreed with the PRRB that inaccuracies did inhere in the Ernst Whinney study. The court noted: the costs of the pathologist who supervised the laboratory operation and the costs of supervising the SLS courier services may have been improperly allocated; the costs of collecting specimens may have been overstated; the period of the study may not have been representative of the accounting years at issue; and the costs of tests done on St. Marys' patients at an outside hospital may have been improperly allocated. the court then observed that although "it is arguable that St. Marys' witnesses resolved some or all of these questions, the Board was not obliged to accept the testimony" (Mem. Order of Dec. 24, 1983 at 10). The judge concluded "that the questions were raised and that the evidence of possible inaccuracy is in the record. Thus, the Board's decision concerning inaccuracy cannot be overturned." ( Id.)

Despite this conclusion, the district court believed that the PRRB had made an implicit finding that SLS costs were cheaper, "in fact were significantly cheaper, to perform than the same tests done for hospital patients" ( id. at 11). Ernst Whinney had broken down the one test they examined into twelve steps, of which three steps were technical ones performed by laboratory personnel. St. Marys argues that this single fact establishes that the SLS tests are different and cheaper to perform. Yet the hospital has not provided any information substantiating this assertion. The bulk of the cost of the tests may very well lie in their technical components; the breakdown of the test into component parts alone is not dispositive of cost. St. Marys assiduously avoids mentioning that it shares the costs of four additional steps through the SLS courier service and the hospital's billing arrangements, thus being responsible for the costs of seven of the necessary steps. Until the hospital can prove that such costs as ordering and collecting specimens comprise a significant percentage of the test's total cost and that St. Marys does not in any way bear these costs or their equivalent when SLS testing is involved, the hospital cannot assert that a cost differential exists. Without a cost differential, grossing-up is of course entirely proper.

Contrary to the assertion of the court below, the existence of a cost differential was very much in issue. St. Marys' claim depends on the cost question, because the hospital justified its argument that the services were dissimilar with the assertion that a cost differential existed. Compare, e.g., St. Marys Br. 25-26 (SLS tests less expensive because fewer steps involved) with id. at 28 (tests dissimilar because few steps in common). See also id. at 37 ("uniform charges result in an accurate allocation of costs only when there are uniform costs"); A.R. 248 (Mr. LaFrombois, St. Marys' Director of Fiscal Services, testified that St. Marys' position was that cost differential justified charge differential). The necessity of establishing that a cost differential exists rests with the hospital, not with the Board. The district court reasoned that the PRRB reached the question of the accuracy of St. Marys' alternative methodology because the procedures were in fact dissimilar and SLS costs were substantially less. But the PRRB stated explicitly that the cost study "did not establish" SLS costs. This conclusion carries the opposite implication, that St. Marys failed to prove the cost differential on which its argument depends. Therefore, the district court's reasoning must be regarded with suspicion. Nor do we find that the PRRB made the implicit finding that a cost differential existed, as the district court suggests. Once the PRRB found the cost study to be inadequate, no further inquiry was necessary. St. Marys had failed to carry its burden of proving that a cost differential between SLS costs and regular laboratory costs existed. As a result, the agency had no need to address whether the services were similar or dissimilar. Its failure to do so does not indicate that the Board believed a cost differential existed.

St. Marys' difficulty lies in its concession that the costs of the SLS program exceed the revenue SLS produces (Br. 30 n. 21). This fact means that revenue offset will not accurately account for SLS costs, as some SLS costs will remain even after the total SLS revenues have been applied. Multiplying the ratio of Medicare charges to total charges by the remaining costs, including some unspecified amount of SLS costs, will result in Medicare's reimbursing St. Marys for some costs for which the program is not responsible. The method of reimbursement St. Marys proposed for 1979 would similarly distort the reimbursement process. The inequality in the cost-to-charge ratio for SLS charges would impermissibly minimize total charges, thereby increasing the ratio of Medicare charges to total charges and causing Medicare again to overreimburse St. Marys. The two remaining alternatives — cost offset and constant cost-to-charge ratio — reduce to the same thing (Secretary's Br. 15-16). Cost offset would require St. Marys to subtract SLS charges and SLS costs from total charges and total costs, respectively, before calculating the Medicare charge to total charge ratio. The second method grosses-up SLS charges until the ratio of SLS charges to SLS costs equals that of other charges to other costs (Secretary's Br. 28). Because both of these approaches rely on the exact cost of SLS services, an inability to determine that figure with sufficient accuracy renders them useless for cost apportionment purposes. St. Marys would be entitled to avoid grossing-up SLS charges and rely on either of these alternatives only when it could prove the extent to which SLS costs were less than those of St. Marys' other laboratory tests.

Substantial evidence in the record supports the PRRB's decision that St. Marys failed to carry its assigned burden of proof. A good illustration is one of the examples the PRRB cited as evidence that St. Marys' management lacked requisite confidence in the accuracy of the Ernst Whinney study. Board member Houdek was questioning Edward LaFrombois, St. Marys' Director of Fiscal Services, concerning the implications for reimbursement when SLS costs exceed SLS revenues, as even the Ernst Whinney study indicated they did. Mr. LaFrombois initially refused to accept that the revenue offset method would cause Medicare to cross-subsidize the SLS program (A.R. 216). When asked to justify his position, Mr. LaFrombois finally admitted that cross-subsidization would occur, "[a]ssuming that those costs [that the study reflected for 1978 and 1979] are adequately representing the cost of SLS" ( id. at 217). This qualification caused Mr. Houdek to question Mr. LaFrombois about his confidence in the accuracy of the accountants' cost study, a question Mr. LaFrombois would never answer directly and to which he finally responded "[w]e feel it is an accurate reflection within the state of the art" ( id. at 217-218). The Board correctly realized that whether the study represented "the state of the art" was not the question. The issue was the study's accuracy, and St. Marys gave the PRRB sufficient reason to refuse to rely on its cost study. St. Marys would vindicate its study by contending "that a number of different tests * * * established that the study was accurate within 1%" (St. Marys' Br. 40). But the testimony St. Marys cites to support this claim does not support its assertion. George Whetsall, the management specialist at Ernst Whinney responsible for the cost study, testified that manipulating their collected data in two different ways yielded results within one percent of each other (A.R. 330-331). He did not testify that later studies verified the cost study, and he did not testify that Ernst Whinney's underlying assumptions and categorization of the data were correct. Because a study's results are only as reliable as the way in which the study is set up and conducted, Mr. Whetsall's statements are little more than self-serving manipulations of data that may have been fundamentally flawed.

A very different situation appeared in St. Mary's Medical Center (Knoxville, Tenn.) v. Blue Cross Association/Blue Cross Blue Shield of Tennessee, PRRB Dec. No. 79-D79 [1980 Transfer Binder] Medicare and Medicaid Guide (CCH) ¶ 30,337. The controversy there centered on laboratory services quite similar to the SLS program at issue before us. In that case the Board disapproved of grossing-up. As in the instant situation, the cost study showed that hospital laboratory personnel conducted only two of nine steps involved. The court below used this fact to assert that the cost study in the instant case necessarily proved that SLS services were less expensive than other laboratory tests. The Tennessee case, however, will not withstand such an extension. Two other facts of much greater importance controlled the result reached there. The PRRB explicitly found the cost study to be accurate, a finding completely contrary to the one it reached with regard to the Ernst Whinney study. The study there also demonstrated that revenue from the program exceeded costs, so that revenue offset was appropriate and a precise cost determination was unnecessary in order to allocate the laboratory's total costs accurately. The district court conceded that the PRRB "correctly" determined "that costs exceeded revenues, thus precluding a revenue offset" (Mem. Order of Dec. 24, 1983 at 13). It relied on St. Mary's Medical Center (Knoxville) because it believed that in both cases the hospitals had established a cost differential. The Tennessee hospital was able to do so; St. Marys has failed to do so in the instant case.

Much closer in point than St. Mary's Medical Center (Knoxville) is Lake Region Hospital Corp. v. Heckler, 602 F. Supp. 109 (D.Minn. 1983). Again the facts are substantially the same. The main difference is that the PRRB identified an error factor in the Lake Region cost study of as much as twenty-five percent. Although the Board did not quantify an error factor in the instant case, its failure to do so is not fatal. For reasons that are unclear, the district judge determined that the "unverifiable portions" of St. Marys' study were not significant and that "St. Marys clearly established the existence of a cost difference, while Lake Region did not" (Mem. Order of Jan. 16, 1984 at 3, 2). He apparently assumed from his own assessment of the record that only portions of St. Marys' study were unverifiable, that those portions were insignificant, and that a cost differential had been established. Some evidence in the record might support such an assessment. But the real question is not the assessment the trial judge might be inclined to make but whether substantial evidence in the record supports the determinations made by the PRRB. The failure by the court below to address meaningfully this latter question constitutes reversible error. The PRRB unequivocally stated that the cost study did not establish costs, so that the presence or absence of a cost differential could not have been established either. Substantial evidence in the record indicates that the cost study contained serious flaws. Therefore, the PRRB's refusal to accept it is amply justified. What seems to have influenced the district judge the most in deciding to the contrary was testimony by the intermediary's witness that the intermediary would require St. Marys to gross-up charges for tests with the same name, regardless of their respective costs and regardless of whether the services provided were the same or not. Even if this position of the intermediary were error, considering it fails to address the question whether the cost study was sufficiently reliable. St. Marys has not met the burden of proof required by 42 C.F.R. § 405.453(a), (c). Accordingly the district court's judgment is reversed.

* * * * * *

(c) Adequacy of cost information. Adequate cost information must be obtained from the provider's records to support payments made for services rendered to beneficiaries. The requirement of adequacy of data implies that the data be accurate and in sufficient detail to accomplish the purposes for which it is intended. Adequate data capable of being audited is consistent with good business concepts and effective and efficient management of any organization, whether it is operated for profit or on a nonprofit basis. It is a reasonable expectation on the part of any agency paying for services on a cost-reimbursement basis.