The Missouri Department of Social Services appeals from the final decision of the Administrator of the Health Care Financing Administration rejecting retroactive application of proposed changes in Missouri's Medicaid program. We affirm.

The state of Missouri participates in the Medicaid program, under which the federal government reimburses a portion of the cost of caring for eligible patients. Each state administers its own program but must submit to the Health Care Financing Administration (HCFA) a state plan, and any subsequent amendments, describing how the state will administer the program. If an amendment proposes a "significant ... change in [the state's] methods and standards for setting payment rates for services," it may not take effect until after the state has published notice of the change. 42 C.F.R. § 447.205(a), (d) (1990); see also 42 C.F.R. § 447.253(f) (1990).

On September 1, 1989, the Missouri Department of Social Services (Missouri) submitted an amendment to its plan. The amendment proposed adjusting the amount paid for inpatient health care services by applying a new "trend factor adjustment," essentially an inflation index. The new trend factor adjustment was different from the trend factor adjustment Missouri had used in the past. Missouri proposed that the amendment take effect as of July 1, 1989, even though it did not publish notice of the proposed change until September 1, 1989.

The HCFA held that the amendment could not become effective before September 2, 1989, because it proposed a significant change in the methods or standards of setting payment rates which could not become effective until after publication of notice. Concerning whether the change altered a "method or standard," the HCFA noted that neither the amendment nor the plan described how Missouri sets the trend factor. Thus the new trend factor represented an alteration of the equation used to calculate payment rates; as such, it altered the methods or standards for setting rates.

Concerning whether the change was "significant," the HCFA reasoned first that the amendment would result in an approximate 1.7% increase in Medicaid payments, totalling about $3 million. Second, the change would affect all Medicaid patients receiving inpatient care at participating Missouri hospitals. Finally, since it had already found that the amendment changed the methods or standards for calculating payment rates, the HCFA concluded that the change was ipso facto significant.

Missouri challenges the HCFA's conclusion that the amendment changes a method or standard. It argues that the amendment does not create a new equation for calculating payment rates; instead, it simply inserts a new number into an already existing equation. Missouri also contends that the HCFA erred in finding that the change is significant.

The HCFA's interpretation of its own regulations deserves considerable deference, and we will not set it aside unless it is arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with the law. Creighton Omaha Regional Health Care Corp. v. Bowen, 822 F.2d 785, 789 (8th Cir.1987).

Missouri argues that the HCFA acted arbitrarily in deciding that the amendment changed Missouri's "methods or standards" for setting payment rates. Missouri maintains that the amendment did not alter the methodology by which it computes payment rates. Instead, it simply substituted the new trend factor for the old one. Since the equation itself was not changed, the amendment did not alter the methods or standards for setting payment rates.

We disagree. Missouri does not reveal its methodology for determining the size of the trend factor adjustment. The trend factor adjustment operates as a sort of "black box": Missouri offered a new number in the amendment but did not explain the methodology it used to arrive at that number. Hence the HCFA may reasonably view the new trend factor adjustment as an alteration of Missouri's methodology, not as new data plugged into an unchanged equation. So viewed, the amendment may reasonably be construed as a change in Missouri's methods or standards.

To support its position, Missouri relies on Massachusetts Federation of Nursing Homes, Inc. v. Massachusetts, 772 F. Supp. 31 (D.Mass.1991). That case involved changes to the "management minute ranges," a series of categories reflecting the intensity of care patients required. Facilities received government funding, in part, based on the categories into which their patients fit. Certain nursing facilities argued that the state had acted contrary to the federal regulations by not including changes to the management minute ranges in the plan the state had filed with the HCFA. The court determined that the changes did not alter the methods or standards for setting rates and therefore did not need to be filed as part of the state's plan.

The instant case differs from Massachusetts Federation. First, the HCFA in that case had approved the state plan even though it had not included the management minute ranges. Since "[t]he HCFA certainly has more expertise in this complicated area of the law than the courts," the HCFA's approval of the plan absent the management minute ranges indicated that the ranges were not methods or standards for setting payment rates. Id. at 39. In contrast, the HCFA in the instant case decided that the new trend factor adjustment did alter Missouri's methods or standards. This decision is entitled to considerable deference.

Second, the amendment to the management minute ranges in Massachusetts Federation altered the data used in the state's payment equation, not the equation itself. As noted above, the HCFA in the present case could reasonably view the new trend factor adjustment as a new sub-equation within Missouri's overall methodology. Since the nature of Missouri's proposed change differs from the change in Massachusetts Federation, the force of that decision's analysis does not apply here.

Missouri also contests the HCFA's conclusion that the amendment is a "significant" change to Missouri's methods or standards. Missouri first contends that the percentage increase in payment rates resulting from the amendment is really 0.9%, not 1.7%. Both parties agree that the amendment would increase Medicaid payments by approximately $3 million in fiscal year 1990. The parties differ, however, concerning what number should serve as the denominator when calculating the percentage increase. The HCFA reached its 1.7% figure by using the actual expenditures for fiscal year 1989. Missouri contends that the projected fiscal year 1990 budget should serve as the denominator.

Although Missouri casts this argument in terms of the sufficiency of the evidence, we construe it as an attack on the reasonableness of the HCFA's choosing one means of performing a percentage change calculation over another. Missouri does not argue, for example, that the fiscal year 1989 expenditures that the HCFA used were in fact inaccurate, or that the HCFA performed its long division improperly. Instead, Missouri contends that the HCFA used an arbitrary calculation method to determine whether the change was significant. Thus, we must consider whether the HCFA's decision is reasonable.

Missouri argues that its method of calculating the percentage change is the only correct one because the former version of the regulation required prior notice only for changes that were "expected" to alter Medicaid inpatient services payments by 1% or more. 42 C.F.R. § 447.205(a) (1981). Since the current regulation was intended to allow states greater flexibility, see 46 Fed.Reg. 47,967 (1981), the HCFA should continue to focus on expected changes under the new regulation and should therefore use the 1990 budget as the denominator. Hence, goes the argument, the HCFA's decision to use the actual expenditures from 1989 was necessarily unreasonable.

We do not find this argument persuasive. The reference in the former regulation to expected changes plus the current regulation's commitment to flexibility does not mean that the HCFA should have calculated percentage changes by relying on Missouri's budget estimate. To the contrary, the fiscal year 1990 budget Missouri wished to use as the denominator included items not relating to inpatient hospital services, so the calculation's result may not have accurately represented the change in inpatient services payments had the 1990 budget figures been used. In addition, one may reasonably infer that a calculation based on actual expenditures would be more accurate than one based on projections. It was, therefore, reasonable for the HCFA to calculate the percentage increase by dividing the expected increase in expenditures by fiscal year 1989's actual expenditures.

Having found the HCFA's method of calculating the percentage change to be reasonable, we next address whether its conclusion that the amendment was a significant change was likewise reasonable. The amendment would result in an increase in Medicaid expenditures of approximately $3 million, about 1.7% more than the previous year. The increase would have affected all recipients of inpatient Medicaid services in Missouri. In addition, Missouri had in previous years published timely notice of any new trend factor adjustment. This indicates that Missouri itself perceives the change as significant, since non-significant changes need not be published. Given the magnitude of the change, the scope of its impact, and the inference arising from Missouri's past practice, we cannot say that the HCFA acted arbitrarily in finding that the change was significant for purposes of the relevant regulations.

The decision of the HCFA is affirmed.