HAYS MEDICAL CENTER; MERCY HOSPITAL LEBANON; MERCY HOSPITAL ARDMORE, INC.; NORTH PLATTE NEBRASKA HOSPITAL CORPORATION, d/b/a Great Plains Medical Center; HANOVER HOSPITAL, INC.; KNOX COMMUNITY HOSPITAL; LABETTE COUNTY MEDICAL CENTER; MEMORIAL HOSPITAL OF SWEETWATER COUNTY; NEWMAN MEMORIAL COUNTY HOSPITAL, d/b/a Newman Regional Health; NORTHWESTERN MEDICAL CENTER, INC.; POCATELLO HOSPITAL, LLC, d/b/a Portneuf Medical Center, Plaintiffs - Appellants, and RICHLAND MEMORIAL HOSPITAL, Plaintiff, v. ALEX M. AZAR, II, in his official capacity as Secretary of Health and Human Services, Defendant - Appellee.
No. 17-3232
United States Court of Appeals, Tenth Circuit
April 21, 2020
PUBLISH. Christopher M. Wolpert, Clerk of Court
ALEX M. AZAR, II, in his official capacity as Secretary of Health and Human Services,1
Notes
Appeal from the United States District Court for the District of Kansas (D.C. No. 2:15-CV-09893-JTM-GEB)
Daniel J. Hettich (Elizabeth N. Swayne with him on the briefs), King & Spalding LLP, Washington D.C., for Plaintiffs - Appellants.
Carleen M. Zubrzycki, Attorney, Appellate Staff (Chad A. Readler, Acting Assistant Attorney General, Stephen McAllister, United States Attorney, and Michael S. Raab and Abby Wright, Attorneys, with her on the brief), United States Department of Justice, Washington, D.C., for Defendant - Appellee.
Before TYMKOVICH, Chief Judge, HOLMES and PHILLIPS, Circuit Judges.
HOLMES, Circuit Judge.
Plaintiff-Appellants are eleven rural hospitals (the “Hospitals“). They challenge the methodology that the U.S. Secretary of Health and Human Services (the “Secretary“) uses to calculate their Medicare reimbursements. The question before us is whether that methodology is arbitrary and capricious. We hold that it is not. Accordingly, exercising jurisdiction under
I
This case plunges us into the “labyrinthine world” of Medicare reimbursement. Adirondack Med. Ctr. v. Sebelius, 740 F.3d 692, 694 (D.C. Cir. 2014).2 In order to navigate this world, we first lay out the necessary background, which we do in three steps. First, we outline the basics of the Medicare reimbursement system. Second, we explain the Secretary‘s methodology at issue and trace its application over the years. Third, we recount how this appeal unfolded.
A
The Medicare program provides federally funded healthcare to persons over sixty-five years old, as well as to disabled persons. See Social Security Amendments of 1965, Pub. L. No. 89-97, § 102, 79 Stat. 291 (1965) (codified as amended at
The Secretary reimburses participating hospitals for certain inpatient care they provide to Medicare Part A patients through what is called the “inpatient prospective payment system,” a regime established by Congress in the 1980s aimed at encouraging more cost-efficient management of medical care by hospitals. See Social Security Amendments of 1983, Pub. L. No. 98-21, § 601, 97 Stat. 65, 149 (1983) (codified as amended at
1
At its most basic, a hospital‘s reimbursement for a given Medicare Part A patient is the mathematical product of two figures—the payment rate and the patient‘s diagnosis-related group weight. We turn first to providing an overview of the latter figure, the diagnosis-related group weight (the payment rate is explored in the immediately succeeding subsection infra). Ascertaining this figure begins with the Secretary sorting the Medicare Part A patient into a “diagnosis-related group” (periodically referred to hereinafter as “DRG“). These groups are essentially “categor[ies] of inpatient treatment” that reflect the differing costs of treating various types of diagnoses, with each Part A patient being sorted into a group at the time of discharge, based on that patient‘s diagnosis. Adirondack Med. Ctr., 740 F.3d at 694 n.1; see also
In order to account for changes in resource consumption, Congress recognized that it would be necessary to periodically recalculate the diagnosis-related group weighting factors. Changes to the Inpatient Hospital Prospective Payment System and Fiscal Year 1991 Rates, 55 Fed. Reg. 35,990, 36,008 (“FY 1991 Final Rule“). Accordingly, at Congress‘s direction, the Secretary annually “adjust[s]” the diagnosis-related group “classifications and weighting factors.” Id. (citing
We focus here on the Secretary‘s annual adjustment of the diagnosis-related group weighting factors, or, in the common parlance of regulators, his “recalibration” of the weights. The Secretary starts this process by assembling “a dataset of recent patients.” J.A. at 159 (Mem. in Opp‘n to Pls.’ Mot. for Summ. J. & in Supp. of Def.‘s Cross-Mot. for Summ. J., filed Oct. 20, 2016); see also 55 Fed. Reg. at 36,033 (FY 1991 Final Rule) (“One of the basic issues in recalibration is the choice of a data base that allows us to construct relative DRG weights that most accurately reflect current relative resource use.“). Based on this dataset, the Secretary then calculates the average cost of treating patients across all diagnosis-related groups and “the average cost of treating a patient in each diagnosis-related group.” J.A. at 159. The Secretary then divides the average cost for each diagnosis-related group by the overall average cost. See id. Using this ratio, he then assigns a new weighting factor to each diagnosis-related group. See id.; see also
The recalibration process has the potential of leading to higher payments to hospitals systemwide. However, Congress requires the Secretary to recalibrate “in a manner that assures that the aggregate payments . . . are no greater or less than those that would have been made [pre-recalibration].”
To achieve budget neutrality in aggregate payments, the Secretary takes two steps. The Secretary‘s first step is to “normalize” the recalibrated diagnosis-related group weights to offset any change in the average weight caused by recalibration. 74 Fed. Reg. at 4395 (FY 2010 Final Rule). The Secretary has explained that normalization negates any increase in average diagnosis-related group weights by “equating the average case weight after recalibration to the average case weight before recalibration.” Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 1994 Rates, 58 Fed. Reg. 46,270, 46,291 (Sept. 1, 1993) (“FY 1994 Final Rule“).8 More specifically, the Secretary applies a numerical adjustment “in order to ensure that the average [diagnosis-related group] weight after recalibration is equal to the average [diagnosis-related group] weight prior to recalibration.” Proposed Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 2010 Rates, 74 Fed. Reg. 24,080, 24,184 (May 22, 2009) (“FY 2010 Proposed Rule“).9
Normalization, however, “does not necessarily achieve budget neutrality with respect to aggregate payments to hospitals because payment to hospitals is affected by factors other than average [diagnosis-related group] weight.” 58 Fed. Reg. at 46,291 (FY 1994 Final Rule); see also 74 Fed. Reg. at 43,895 (FY 2010 Final Rule) (noting that “our analysis has indicated that the normalization adjustment does not usually achieve budget neutrality with respect to aggregate payments to hospitals” (emphasis added)). For example, Congress requires the Secretary to apply wage-index adjustments to hospital reimbursement rates to account for varying labor costs of hospitals located in different geographic areas of the country, as well as in urban and rural settings, and to do this “in a manner that assures that aggregate payments to hospitals are not affected by the change in the wage index.” 58 Fed. Reg. at 46,346 (FY 1994 Final Rule); see
Payment to hospitals is also affected by Congress‘s requirement that the Secretary “provide for an additional payment amount for” teaching hospitals “with indirect costs of medical education.”
In light of such factors, normalization of the average diagnosis-related group weights will not necessarily result in budget neutrality of aggregate payments systemwide. That is because those factors affect—and frequently increase—the hospitals’ payment rate and, in computing a hospital‘s reimbursement, “a patient‘s [normalized] diagnosis-related group weight is multiplied by the hospital‘s applicable [payment] rate.” Aplee.‘s Resp. Br. at 7.10
Accordingly, to ensure budget neutrality in aggregate payments, the Secretary takes a second step: he calculates “a budget neutrality adjustment” to account for those other factors, such that the adjustment is in fact “budget neutral.” 58 Fed. Reg. at 46,291 (FY 1994 Final Rule); see 55 Fed Reg. at 36,073–74 (FY 1991 Final Rule) (noting the “interactive effect of the wage index and DRG weights on aggregate payments” and that the Secretary applied a “budget neutrality adjustment factor” to the hospital-specific rates because unless he did so, he would not be able to “meet the statutory requirement that aggregate payments neither increase nor decrease“). “This budget-neutrality adjustment takes the form of a multiplier that the Secretary applies in calculating the hospital‘s base payment rate[.]” Aplee.‘s Resp. Br. at 8.
Since Fiscal Year (“FY“) 1994, the Secretary has applied the budget-neutrality adjustment “in a cumulative manner,” commencing with the FY 1993 adjustment factors for that inaugural cumulative computation. Aplts.’ Opening Br. at 10; see 70 Fed. Reg. at 47,429 (FY 2006 Final Rule). That is to say, he “does not remove the prior year‘s budget-neutrality adjustment before calculating and applying the current year‘s adjustment.” Aplee.‘s Resp. Br. at 8; see id. (“[A]ll diagnosis-related group budget-neutrality adjustments that were applied each fiscal year from FY 1993 onward are permanently incorporated into a later year‘s rates.“); see also 58 Fed. Reg. at 46,346 (FY 1994 Final Rule) (explaining that the “budget neutrality adjustment factor is applied . . . without removing the effects of the FY 1993 budget neutrality adjustment“). The Secretary then applies this cumulative budget-neutrality adjustment factor to the payment rate used in the reimbursement formula. See 74 Fed. Reg. at 43,895 (FY 2010 Final Rule).
2
Having defined and described the term “diagnosis-related group weight,” we turn now to explicating the other major component undergirding a hospital‘s reimbursement: the payment rate. The Secretary reimburses most hospitals using what is called the federal rate. The starting point in calculating the federal rate in a given year is the “average standardized amount,” which is essentially the estimated average per-patient operating costs of inpatient hospital services for all hospitals within a geographic region. See
Although, as discussed, the above-noted federal rate is the mode of reimbursement that is employed for most hospitals, because certain rural hospitals “provide critical services to the underserved and uninsured, Congress has adopted special payment provisions for them.” J.A. at 286 (Dist. Ct. Mem. & Order, dated Aug. 31, 2017) (quoting Adirondack II, 29 F. Supp. 3d at 32). As relevant here, those rural hospitals are “sole community” and “Medicare-dependent” hospitals.11 See, e.g., Cmty. Hosp. v. Sullivan, 986 F.2d 357, 358 (10th Cir. 1993) (noting Congress has evinced special concern for sole community hospitals). It is undisputed that each of the Hospitals here is either a sole community or Medicare-dependent hospital.
Chief among the special payment provisions Congress has adopted for sole community and Medicare-dependent hospitals is the option of selecting reimbursement for covered healthcare based on a “hospital-specific rate,” instead of the federal rate. See
The Secretary‘s starting point for calculating a hospital-specific rate is the hospital‘s “target amount” in a given “base year.” See
As with the federal rate, the Secretary annually adjusts the hospital-specific rate to account for variables such as inflation and wage levels. See
Although both sole community and Medicare-dependent hospitals may select a “payout” based on a hospital-specific rate, the nature of the payout “differs slightly” between the two hospitals. Adirondack Med. Ctr., 740 F.3d at 695 n.2. Reimbursement for sole community hospitals is “fairly straightforward“: they are to be paid the higher of either the hospital-specific rate or the federal rate. Id.; see
B
Having laid out the basics of the Medicare Part A reimbursement system, we can focus on the part of that system at issue here—the Secretary‘s methodology for calculating the hospital-specific rate when Congress adds a new base year. In the following subsections, we explain how that methodology works and how the Secretary has applied it over the years.
1
When Congress adds a new base year, the hospital-specific rate for that new year is calculated according to the following four-step formula:
Step One: Divide the hospital‘s target amount for the new base year “by the number of discharges in the base [year].”
Step Two: Divide the figure from step one by the hospital‘s normalized average diagnosis-related group weight for the new base year. See
Step Three: Apply an update factor to account for specific variables such as inflation and wage levels. See
Step Four: Apply a cumulative budget-neutrality adjustment to the resulting rate figure. See
The amount this formula yields is the hospital-specific rate for a given hospital in the new base year. That rate is then multiplied by each patient‘s diagnosis-related group weight, which yields the hospital‘s reimbursement for each patient‘s inpatient care.
The Hospitals do not challenge steps one and three. And much of steps two and four is also uncontroversial. The Hospitals agree, for instance, “that the average DRG weight used as a divisor [in step two] has been normalized.” Aplts.’ Reply Br. at 17. And they “do not object to the fact that DRG weights are used as a divisor” in step two. Aplts.’ Opening Br. at 9 n.4. Likewise, “[t]he Hospitals do not object to prior-year adjustments being applied once.” Id. at 18. In fact, the Hospitals do not even complain about the cumulative nature of this adjustment. See Aplts.’ Reply Br. at 2 (“The Hospitals explicitly stated on several occasions . . . that their contention is not that the prior-year budget-neutrality adjustments should not be applied once . . . .“). Nor do they “necessarily take issue with the Secretary‘s decision to apply the [cumulative budget-neutrality adjustment] to the rates instead of the weights.” Aplts.’ Opening Br. at 24.
What the parties do vigorously dispute, however, are the consequences of the Secretary applying a cumulative budget-neutrality adjustment in step four. Above all, they contest whether that adjustment is being applied twice—once in step two and then again in step four. Going forward, the focal point of our analysis will be unpacking how those steps interact. Before delving into that disagreement, however, it helps to understand how the Secretary has applied his methodology over the years.
2
As discussed supra, Congress has periodically added new base years. Whenever this happens, the Secretary issues “rebasing” instructions to fiscal intermediaries.14 See, e.g., J.A. at 339–43 (CMS Transmittal A-01-123, dated Sept. 27, 2001) (“2000 Instructions“). These are technical instructions, which tell the intermediaries how to calculate the hospital-specific rates for the new base year.
In 1999, Congress added FY 1996 as a base year for sole community hospitals. See Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act of 1999, Pub. L. No. 106-113, § 405, 113 Stat. 1501 (codified as amended at
the intermediaries were to apply a cumulative budget-neutrality adjustment to the hospital-specific rate for the new base year. The closest the instructions came to addressing the issue was in the statement that the FY 1996 hospital-specific rate was to be multiplied “by a factor of 1.02547 to update it from 1996 to 2000.” Id. at 342.
But that remark is ambiguous. On the one hand, the 1.02547 figure might have included the budget-neutrality adjustments from FY 1993 (the first year for such adjustments) to FY 1996. See Aplts.’ Opening Br. at 11–12 (acknowledging this possibility); Aplee.‘s Resp. Br. at 15 (arguing that the 2000 Instructions directed fiscal intermediaries to “apply the cumulative diagnosis-related group budget-neutrality adjustment from the first year the adjustment was applied“); cf. Adirondack Med. Ctr. v. Sebelius (Adirondack I), 935 F. Supp. 2d 121, 132 (D.D.C. 2013) (explaining that the 2000 Instructions directed intermediaries to apply “cumulative DRG budget neutrality adjustment factors from FY 1993 to FY 2000“). On the other hand, the intermediaries were instructed to update the hospital-specific rate “from 1996 to 2000,” implying the budget-neutrality adjustment was not cumulative. J.A. at 342 (2000 Instructions) (emphasis added).
In 2006, Congress added FY 2002 as a new base year for Medicare-dependent hospitals. See
Two years later, Congress added FY 2006 as a new base year for sole community hospitals. See
But about six weeks later, the Secretary revised the 2008 Instructions in a Joint Signature Memorandum. See J.A. at 350–53 (Joint Signature Mem., dated Nov. 17, 2008). In this memorandum, the Secretary no longer instructed intermediaries to apply the budget-neutrality adjustments prospectively only, as he had in his initially issued 2008 Instructions. Rather, he instructed the intermediaries to apply to the new FY 2006 base year all budget-neutrality adjustments extending back to FY 1993. See id. at 350–51.
Another six months passed before the Secretary issued a proposed rule pertaining to Medicare-dependent hospitals that explained to the public his methodology for FY 2010. 74 Fed. Reg. at 24,183–85 (FY 2010 Proposed Rule). The FY 2010 Proposed Rule reiterated the Secretary‘s policy of applying a cumulative budget-neutrality adjustment “to both the standard Federal rate and hospital specific rates” without “remov[ing] the prior years’ budget neutrality adjustment.” Id. at 24,184. To do otherwise, the Secretary explained, “would not satisfy” Congress‘s mandate to achieve budget neutrality. Id.
In explaining why a cumulative budget-neutrality adjustment was needed, the Secretary said that “[i]f we were to remove this budget neutrality adjustment factor for years prior to the base year, we believe the normalized DRG weights applied to the hospital-specific amounts would be artificially high, thus resulting in higher aggregate payments than permitted under the statute.” Id. This notion of “artificially high” normalized average diagnosis-related group weights was new. For years, the Secretary had explained that normalization negated any increase in diagnosis-related group weights by “equating the average [diagnosis-related group] weight after recalibration to the average [diagnosis-related group] weight before recalibration.” 58 Fed. Reg. at 46,291 (FY 1994 Final Rule). But normalization did not always “achieve budget neutrality with respect to aggregate payments . . . because payment . . . is affected by factors other than average [diagnosis-related group] weight.” Id. (emphasis added). The cumulative budget-neutrality adjustment accounted for these other factors—not “artificially high” average diagnosis-related group weights. 74 Fed. Reg. at 24,184 (FY 2010 Proposed Rule). These longstanding positions of the Secretary were inconsistent with the idea of “artificially high” normalized average diagnosis-related group weights.
Although the FY 2010 Proposed Rule introduced some uncertainty to the regulatory landscape, it also shed substantial light on the lingering question surrounding application of the cumulative budget-neutrality adjustment. As explained above, the 2000 Instructions had left in doubt whether the Secretary‘s instructions required application of a cumulative budget-neutrality adjustment to calculate the hospital-specific rates for the FY 1996 base year. The FY 2010 Proposed Rule removed that doubt. The proposed rule clarified that “the instructions for implementing both the FY 1996 and FY 2006” base years applied “cumulative budget neutrality adjustment factors” including each year‘s adjustment “since FY 1993.” Id. And those instructions were in keeping with the Secretary‘s “established policy.” Id.
The FY 2010 Proposed Rule brought the 2006 Instructions in line with that declared policy. Recall, those instructions had directed intermediaries to apply “budget adjustment factors” prospectively for the new FY 2002 base year—that is, “for FYs 2003 through 2007.” Id. The failure to include the budget-neutrality adjustments from “FYs 1993 through 2002,” the Secretary opined, was “inconsistent with [his] stated policy of applying a cumulative budget neutrality adjustment” and led to about a “1.74 percent” overpayment to Medicare-dependent hospitals using the FY 2002 base year. Id. So going forward, the Secretary announced, intermediaries were to once again apply a cumulative budget-neutrality adjustment for all base years. Id. Of course, he admitted, this change would potentially “lower the hospital-specific rate” for some Medicare-dependent hospitals using the FY 2002 base year, to the point that the federal rate would result in higher payments than the hospital-specific rate. Id. at 24,185. But the cumulative adjustment was necessary to achieve budget neutrality and maintain “a meaningful comparison between payments under the Federal rate, which is adjusted by the cumulative budget neutrality factor, and payments based on the hospital-specific rate.” Id. at 24,184.
The FY 2010 Proposed Rule drew many public comments. Several commenters argued that a cumulative budget-neutrality adjustment for new base years was unnecessary. Indeed, they believed this adjustment was already baked into step two of the Secretary‘s methodology—i.e., by dividing the hospital‘s average per-patient expenses in the new base year by its normalized average diagnosis-related group weight. See J.A. at 312-13 (Rural Referral Center/Sole Community Hospital Coalition, Comment Letter on FY 2010 Proposed Rule (June 25, 2009)). After all, these commenters pointed out, the Secretary had admitted that those average diagnosis-related group weights were “artificially high“—even after normalization. Id. at 312 (quoting 74 Fed. Reg. at 24,184 (FY 2010 Proposed Rule)). And using these inflated weights as a divisor, the commenters explained, resulted in a hospital-specific rate that was lower than it would have otherwise been. See id. at 313. The same was not true for the federal rate, however, because the Secretary did not use the inflated weights as a divisor in calculating that rate. Hence, they argued, a cumulative budget-neutrality adjustment was appropriate for the federal-rate hospitals, but it was unnecessary—indeed, duplicative—for sole community and Medicare-dependent hospitals. Id. at 314-15.
The Secretary responded to these comments in the preamble to the FY 2010 Final Rule. See 74 Fed. Reg. at 43,895–96 (FY 2010 Final Rule). Retreating from the “artificially high” language in the proposed rule, the Secretary rejected the claim that the normalized average diagnosis-related group weights were too high. See id. at 43,896. Rather, he explained that “[b]ecause the weights are normalized,” they were no higher than they were pre-recalibration. Id. For that reason, the Secretary denied that “the cumulative budget neutrality adjustment” was meant to “offset[] an average [DRG] weight increase due to recalibration.” Id. Thus, he reasoned that “[t]he cumulative budget neutrality adjustment [was] not already being accounted for” in step two. Id. Satisfied that he put down the confusion, the Secretary published the FY 2010 Final Rule.
C
This case unfolded against that backdrop. After the publication of the FY 2010 Final Rule, the Hospitals took issue with the Secretary‘s methodology for calculating the hospital-specific rate for new base years. And dissatisfied with their reimbursements under that methodology, the Hospitals filed administrative appeals with the Provider Reimbursement Review Board, an independent panel authorized to hear appeals from the Secretary‘s final determinations. See, e.g., J.A. at 354–81 (Hannover Hosp. Appeal Request, dated May 7, 2010); see
The Hospitals then sued the Secretary in the district court. See J.A. at 50–108 (First Am. Compl., filed Feb. 16, 2016). Their complaint alleged that the Secretary‘s methodology violated the Administrative Procedure Act (the “APA“). See id. at 59–62. Specifically, the Hospitals argued that the Secretary‘s application of the cumulative budget-neutrality adjustment to the hospital-specific rate for new base years was “arbitrary, capricious, . . . or otherwise not in accordance with law.”15 Id. at 61 (quoting
Both parties moved for summary judgment. The Hospitals argued that the Secretary‘s methodology was arbitrary and capricious for three reasons. See id. at 133–47 (Pls.’ Mem. of P. & A. in Supp. of Mot. for Summ. J., filed Sept. 6, 2016). First, they claimed that the Secretary applied the same cumulative budget neutrality adjustment twice—once by using inflated normalized diagnosis-related group weights as a divisor in step two and then again in step four. See id. at 133-43. Second, the Hospitals contended that the Secretary‘s methodology yielded different payments than “would have been made had [he] . . . applied the budget-neutrality adjustments to the DRG weights themselves.” Id. at 132. Third, they claimed that the Secretary acted arbitrarily and capriciously by not calculating the hospital-specific rate for new base years “based on 100 percent” of a hospital‘s base-year “target amount.”16
The Secretary cross-moved for summary judgment. See J.A. at 150-84 (Def.‘s Mem. in Opp‘n to Pls.’ Mot. & in Supp. of Cross-Mot. for Summ. J, filed Oct. 20, 2016). First, the Secretary denied that his methodology applied the budget-neutrality adjustment twice. See id. at 181. That argument, the Secretary noted, turned on the false premise that the normalized average diagnosis-related group weights were “too high.” Id. Instead, normalization guaranteed that the weights were exactly the same as they were pre-recalibration. Id. at 181 & n.10; see id. at 181. That the FY 2010 Proposed Rule had used the phrase “artificially high” was an unfortunate and imprecise “way of stating that without a cumulative budget-neutrality adjustment the payments under the hospital-specific rate would be higher.” Id. at 181 n.10. Second, the Secretary pointed out that the Hospitals had conceded that the Medicare statute did not compel him to apply the budget-neutrality adjustment to the weights rather than to the rates. As a result, whether his methodology yielded a different payment than if he had applied the adjustment to the weights was irrelevant. See id. at 176–77. Lastly, the Hospitals’ “based on 100 percent” argument failed, too, the Secretary observed: that argument wrongly assumed that a new base year was “a kind of reset” entitling the Hospitals to 100% of their operating costs rather than a starting point from which the Secretary must make adjustments. Id. at 178. Thus, the Secretary argued that he was entitled to summary judgment.
The district court agreed with the Secretary. It viewed the Secretary‘s choice “to apply pre-base year budget neutrality adjustments” as a “consistent and rational” one. Id. at 295, 297 n.8. And the district court thought the Secretary had adequately explained in the FY 2010 Final Rule why his methodology did not in fact apply the budget-neutrality adjustment twice. See id. at 297–98. Likewise, the court agreed with the Secretary‘s view that “an amount based on 100 percent of the hospital‘s target amount,”
II
The Hospitals now appeal from the final judgment. Their appeal raises one issue: Is the Secretary‘s methodology for calculating the hospital-specific rate for a new base year arbitrary and capricious under
In Medicare-reimbursement cases, the APA supplies the standard we use to review the Secretary‘s actions. See
But APA review is narrow. And the arbitrary-and-capricious standard “is very deferential to the agency.” W. Watersheds, 721 F.3d at 1273 (quoting Hillsdale Envtl. Loss Prevention, Inc., v. U.S. Army Corps of Eng‘rs, 702 F.3d 1156, 1165 (10th Cir. 2012)). Indeed, we presume that an agency action is valid unless the party challenging the action proves otherwise. See Dine Citizens Against Ruining Our Env‘t v. Bernhardt, 923 F.3d 831, 839 (10th Cir. 2019). “Our deference to the agency is ‘especially strong where the challenged decisions involve technical or scientific matters within the agency‘s area of expertise.‘” Id. (quoting Morris v. U.S. Nuclear Regulatory Comm‘n, 598 F.3d 677, 691 (10th Cir. 2010)). Above all, we may not “substitute our judgment for that of the agency” on matters within its expertise. Judulang v. Holder, 565 U.S. 42, 53 (2011) (quoting Motor Vehicle Mfrs. Ass‘n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)).
The parties agree that
III
With the standard of review laid out, we can turn to the question before us: Is the Secretary‘s methodology for calculating the hospital-specific rate for a new base year arbitrary and capricious?
The Hospitals answer that question in the affirmative, for three reasons. First, they argue that the “Secretary acted arbitrarily and capriciously in applying the same budget-neutrality adjustments twice.” Aplts.’ Opening Br. at 6. Second, the Hospitals claim that the Secretary‘s methodology is arbitrary and capricious and contrary to Congress‘s statutory command because it “yields a different payment than had he budget-neutralized the weights themselves.” Id. Third, they contend that the Secretary‘s methodology is arbitrary and capricious because it “would not reimburse a hospital its full base-year allowable operating costs.” Id.
The Secretary disagrees. He denies that applying a cumulative budget-neutrality adjustment to a new base year mistakenly applies that adjustment twice. The Secretary likewise refutes the claim that he must apply the cumulative budget-neutrality adjustment to the diagnosis-related group weights themselves. Finally, the Secretary disagrees that his methodology must be arbitrary and capricious because it would not reimburse the Hospitals for 100% of their base-year operating costs.
We agree with the Secretary. That is to say, for the reasons below, we hold that the Secretary‘s methodology is not arbitrary and capricious. Thus, we affirm the district court‘s judgment.
A
The centerpiece of the Hospitals’ position is their argument that the Secretary mistakenly applies the cumulative budget-neutrality adjustment twice when calculating the hospital-specific rate for a new base year. This double application is, the Hospitals say, “arbitrary and capricious on its face.” Id. at 27. “[B]ut at a minimum,” they argue, this error proves that the Secretary misunderstood his own methodology and its consequences. Id. For those reasons, the Hospitals claim that the Secretary‘s methodology is arbitrary and capricious.
We disagree. We explain our disagreement in three sections. First, we recount the Hospitals’ double-application argument in more detail. Second, we conclude that this argument rests on flawed premises. Third, we reject the Hospitals’ suggestion that the Secretary misunderstood his own methodology.
1
To understand the Hospitals’ double-application argument, recall three features of the Medicare-reimbursement system:
Recalibration: Congress requires the Secretary to “recalibrate” (i.e., adjust) the diagnosis-related group weights every year. See
Normalization: Achieving budget neutrality is a two-step process. To that end, the Secretary‘s first step is to “normalize” the recalibrated weights. Normalization guarantees “that the average [diagnosis-related group] weight after recalibration is equal to the average . . . weight prior to recalibration.” 58 Fed. Reg. at 46,291 (FY 1994 Final Rule). For example, in 1990, recalibration increased the average diagnosis-related group weight by 1.35%. See Changes to the Inpatient Hospital Prospective Payment System and Fiscal Year 1990 Rates, 54 Fed. Reg. 36,452, 36,470 (Sept. 1, 1989) (“FY 1990 Final Rule“). So to offset the 1.35% increase in average weight caused by recalibration, the Secretary normalized the post-recalibration weights. See id.
Budget-Neutrality Adjustment: Normalization, however, does not always “achieve budget neutrality” because “factors other than average [diagnosis-related group] weight” affect payments. 58 Fed. Reg. at 46,291 (FY 1994 Final Rule). So the Secretary‘s second step to achieve budget neutrality is to calculate and apply “a budget neutrality adjustment” to account for those other factors. Id.
Now recall the Secretary‘s four-step methodology for calculating the hospital-specific rate for a new base year:
Step One: Divide the hospital‘s target amount for the new base year by the number of discharges that year.
Step Two: Divide the figure from step one by the hospital‘s normalized average diagnosis-related group weight for the new base year.
Step Three: Apply an update factor to account for specific variables such as inflation and wage levels.
Step Four: Apply the cumulative budget-neutrality adjustment.
The nub of the dispute is whether the Secretary effectively applies the budget-neutrality adjustment twice—once in step two and then again in step four. The Hospitals argue that they can prove that step two includes the cumulative budget-neutrality adjustment. This argument rests on two premises. First, that the normalized diagnosis-related group weights are “too high to be budget neutral.” Aplts.’ Opening Br. at 2–3. Second, that these weights are inflated by the precise amount necessary to achieve budget neutrality. And, because these inflated weights are used as a divisor in step two, the hospital-specific rate for a new base year is reduced by the exact percentage that the Secretary later reduces the rates through application of the cumulative budget-neutrality adjustment in step four.
Consider the following hypothetical that the Hospitals say proves their double-application argument.19 Imagine that Congress adds FY 2002 as a new base year. And assume that the Secretary calculates a cumulative budget-neutrality adjustment factor of 0.986. (This figure would be the product of budget-neutrality adjustments from 1996 to 2002.) When the Secretary multiplies this adjustment factor by the federal rate and does the same for the hospital-specific rate in step four, he reduces those rates by 1.4%. And, from this “1.4 percent budget-neutrality adjustment,” the Hospitals reason that “the DRG weights have increased by 1.4 percent” from 1996 to 2002. Aplts.’ Reply Br. at 19. Put
differently, the Hospitals think that—even after normalization—the average diagnosis-related group weights for the new FY 2002 base year are 1.4% too high. See id. at 20–21. And by using these inflated weights as a divisor in step two, the Secretary applies the same 1.4% budget-neutrality reduction to the hospital-specific rate that he later applies in step four. See id.
Here is how this hypothetical would play out under the Hospitals’ assumptions. Assume that in 1996 Hospital X‘s per-patient target amount was $1,000, and its average diagnosis-related group weight that year was 1.0. So Hospital X would have “receive[d] $1,000 for treating its average patient” in 1996. Id. at 19. And assume that nothing changed from 1996 to 2002; that is, the per-patient target amount in 2002 remained $1,000, and the patients that Hospital X treated in 2002 had identical ailments to those treated in 1996. Despite this lack of change, Hospital X‘s average diagnosis-related group weight for FY 2002—even after normalization—would have increased by 1.4% to 1.014. See id. So when the Secretary in step two divides Hospital X‘s per-patient target amount for 2002 ($1,000) by its inflated average diagnosis-related group weight (1.014), he gets $986.19. Id. Then, in step four, “the Secretary . . . would mistakenly believe that he had to apply another 1.4 percent reduction [in the form of a cumulative budget-neutrality adjustment of 0.986] to the hospital‘s 2002 hospital-specific rate of $986.19, leading to a hospital-specific rate of $972.38.”
Id. at 20. This rate is about 2.8% less than the 1996 rate. Thus, the Hospitals posit that the Secretary mistakenly applied the 1.4% cumulative budget-neutrality adjustment twice—once in step two and then again in step four.
2
The Hospitals’ double-application argument, however, rests on flawed premises. The Hospitals assume that the normalized average diagnosis-related group weights used as a divisor in step two are “artificially high” or “too high to be budget neutral.” Aplts.’ Opening Br. at 24 & n.12. This assumption turns on the validity of the claim “that a hospital that had an average DRG weight of 1.0 in 1996 would have a [normalized average] DRG weight of 1.014 in 2002 if it treated the same type of patients.” Aplts.’ Reply Br. at 23; see id. at 21 (noting that “the one point of disagreement between the parties regarding this hypothetical” pertains to “whether the hospital‘s DRG weight in 2002 would necessarily be 1.014 ... as the Hospitals contend“). According to the Hospitals, it is “necessarily” true that the DRG weight would be 1.014 “because the 1.4 percent budget neutrality adjustment means that, based solely on DRG recalibrations, the [normalized] DRG weights have increased by 1.4 percent.” Id. at 19.
These assumptions evince a misunderstanding of the role of normalization and the budget-neutrality adjustment. Recall that after the Secretary recalibrates the diagnosis-related group weights, he normalizes them. The parties seem to agree that normalization “ensure[s] that the average DRG weight is the same after recalibration as it was before.” Aplts.’ Opening Br. at 10. It should ineluctably follow, then, that under the Hospitals’ hypothetical, the average diagnosis-related group weight of Hospital X in FY 1996 should be the same as in FY 2000—that is, 1.0. See
Correcting these misunderstandings exposes the logical hole in the Hospitals’ double-application argument. They assume that the existence of a “1.4 percent budget neutrality adjustment means that, based solely on DRG recalibrations, the [normalized average] DRG weights have increased by 1.4 percent.” Aplts.’ Reply Br. at 19. But, by its very nature, normalization makes that impossible. And in any event, the Hospitals’ reasoning begs the question of what the cumulative budget-neutrality adjustment accounts for. Underlying their reasoning appears to be their answer: the Hospitals repeatedly assert that “the entire purpose of the budget-neutrality exercise is simply to budget neutralize the DRG weights.” Id. at 26 (emphasis added); see id. at 1 (“The budget-neutrality adjustments at issue here are meant to keep the recalibrated DRG weights budget neutral and ensure that Medicare payments neither increase nor decrease based simply on the adjustments to the DRG weights.“); see also Aplts.’ Opening Br. at 19 (noting that “since the budget-neutrality adjustments were instituted to fulfill the Congressional mandate to budget-neutralize the DRG weights“); id. at 31 (noting that “the entire point of the [budget-neutrality] adjustment is to render DRGs weights budget neutral“). But the Hospitals are wrong.
Congress‘s statutory directive to the Secretary is not to budget neutralize the DRG weights per se but rather to ensure that the process of recalibration of those weights is performed “in a manner that assures that the aggregate payments” are budget neutral. See
In fact, the Hospitals’ own hypothetical (discussed above) undercuts their argument. In that scenario, we assumed that the average diagnosis-related group weights for the new FY 2002 base year were 1.4% higher than the weights from 1996. See Aplts.’ Reply Br. at 19 (“[B]ased solely on DRG recalibrations, the DRG weights have increased by 1.4 percent.“). But we now know that normalization guarantees that the 2002 weights are exactly the same as the 1996 weights. See Aplts.’ Opening Br. at 10 (explaining that normalization “ensure[s] that the average DRG weight is the same after recalibration as it was before“). Even so, the cumulative budget-neutrality adjustment is needed to offset a 1.4% increase in aggregate payments caused by “other Medicare payment adjustments.” Aplee.‘s Resp. Br. at 24. So, in our hypothetical, we may derive Hospital X‘s 2002 hospital-specific rate based on the following formula: its per-patient target amount for 2002 ($1,000) divided by its normalized average diagnosis-related group weight (1.0) and then multiplied by the cumulative budget-neutrality adjustment (0.986). The result is a hospital-specific rate for the new FY 2002 base year of $986.00 (i.e., $1,000 ÷ 1.0 × 0.986 = $986.00). And that rate is 1.4% less than the 1996 hospital-specific rate. Simply put, the Secretary‘s methodology does not apply the cumulative budget-neutrality adjustment twice.20
3
That said, the confusion that seemingly underlies the Hospitals’ double-application argument is understandable. In fact, the Secretary arguably invited that confusion. Recall that in the FY 2010 Proposed Rule the Secretary reiterated that he “normalize[s] DRG weights by an adjustment factor in order to ensure that the average [DRG] weight after recalibration is equal to the average . . . weight prior to recalibration.”
In the FY 2010 Final Rule, the Secretary confronted the confusion apparently wrought by his imprecise wording. He wrote, “[s]ome commenters [had] asserted that the application of a cumulative budget neutrality adjustment factor ... doubles the impact of this adjustment on the hospital-specific rates.”
The Secretary then disabused the commenters of this notion. The Secretary repeated his longstanding position that “the recalibrated DRG weights are normalized each year . . . so that the national average [DRG] weight after recalibration is equal to the . . . average [DRG] weight before recalibration.” Id.; see
The FY 2010 Final Rule should have stifled the Hospitals’ double-application argument. After all, the Secretary refuted the central premise of that argument—that the normalized average diagnosis-related group weights were inflated. Those weights, the Secretary had explained, were not inflated. What‘s more, he had clarified that the cumulative budget-neutrality adjustment was not meant “to offset any increase or decrease in the . . . average [DRG] weight due to recalibration.” Id. In other words, the Secretary had effectively refuted the Hospitals’ later assumption in its briefing here that a budget adjustment of 1.4% means that, “if the hospital‘s average DRG weight was 1.0 in 1996, then its average DRG weight for treating the same patient in 2002 would necessarily be 1.4 percent higher.” Aplts.’ Opening Br. at 43.
Nevertheless, the Hospitals are immovable in their belief that the Secretary has everything “exactly backwards” and fails to understand his own methodology. Id. Indeed, the Hospitals argue that the Secretary “acted arbitrarily and capriciously” by straying from his “prior practice” and failing “to consider an important aspect of the issue“—i.e., the ostensibly inflated DRG weight. Id. at 27.
Once again, however, an examination of the record belies this argument.21 Though the FY 2010 Final Rule itself is proof enough, looking to earlier rules confirms that the Secretary has long understood his methodology. Take the FY 1994 Final Rule. In that rule, the Secretary explained that the recalibrated “weights are normalized by an adjustment factor, so that the average [DRG] weight after recalibration is equal to the average . . . weight prior to recalibration.”
neutrality adjustment to assure the requirement of [budget neutrality in aggregate payments] is met.” Id. Simply put, since 1994, the Secretary has understood that—contrary to the Hospitals’ claims now—normalization offsets any increase in the recalibrated weights, but other factors necessitate a cumulative budget-neutrality adjustment to ensure that the statutory mandate of budget neutrality in aggregate payments is achieved. That explanation guts the assumptions on which the Hospitals’ double-application argument rests. And the Secretary has reiterated this same explanation over the years.22
To be sure, imprecise language has sometimes plagued the Secretary. The FY 2010 Proposed Rule, for example, included the unfortunate phrase “artificially high.”
With this understanding, one can see the 2006 Instructions and the initial 2008 Instructions for what they were—mistakes. Recall that in those instructions, the Secretary did not apply prior years’ budget-neutrality adjustments to the new base years. See J.A. at 328 (2006 Instructions); id. at 349 (2008 Instructions). But just six weeks after issuing the 2008 Instructions, the Secretary corrected his error in those instructions by including the budget-neutrality adjustments from the years before the new base year. See id. at 351 (Joint Signature Mem.). The FY 2010 Proposed Rule did the same for the 2006 Instructions.
In summary, the Secretary does not apply the cumulative budget-neutrality adjustment twice. The Hospitals’ argument to the contrary rests on flawed assumptions. And the Secretary has long understood his methodology and explained it to the public. Thus, the Hospitals’ first argument for why the Secretary‘s methodology is supposedly arbitrary and capricious fails.
B
The Hospitals next argue that the Secretary‘s methodology is arbitrary and capricious and contrary to Congress‘s statutory command because it “yields a different payment than had he budget-neutralized the weights themselves.” Aplts.’ Opening Br. at 6. The logic behind this argument is as follows. “The budget-neutrality adjustments have one purpose: to budget-neutralize the DRG weights.” Id. at 27. The Secretary‘s methodology is valid only if it “yields the same payment as had [he] merely budget-neutralized the DRG weights themselves (rather than the payment rates).” Id. at 19. The Secretary‘s “methodology does not yield the same payment” as the one he would have had he “budget-neutralized the DRG weights themselves.” Id. at 20. Therefore, the Secretary‘s methodology is arbitrary and capricious.
Largely for the reasons we already have discussed in Part III.A.2, supra, this argument fails. Congress‘s statutory directive to the Secretary is not to budget neutralize the DRG weights per se but instead to ensure that the process of recalibration of those weights is performed “in a manner that assures that the aggregate payments” are budget neutral. See
C
For their final argument, the Hospitals claim that the Secretary‘s methodology is arbitrary and capricious because, if applied to a new base year itself, it would reimburse them for less than 100% of their actual base-year costs. This argument relies on language from
The Hospitals, however, come up short again. To start, their argument evinces a fundamental misunderstanding about the nature of the current Medicare-reimbursement system. Under the original Medicare Act, the Secretary paid hospitals for “the reasonable cost” of treating each Medicare patient. Social Security Amendments of 1965 §§ 1814(b), 1861(v)(1). But this system quickly became unwieldy; “[t]he more [hospitals] spent, the more they would receive” in reimbursement. Billings Clinic v. Azar, 901 F.3d 301, 303 (D.C. Cir. 2018). Consequently, Congress replaced the old system with the inpatient prospective payment system. See Social Security Act Amendments of 1983 § 601. A central premise of this current system is that hospitals “are reimbursed at a fixed amount per patient, regardless of the actual operating costs they incur in rendering these services.” Auburn Reg‘l Med. Ctr., 568 U.S. at 149 (emphasis added). Against this backdrop, the Hospitals’ claim to 100% of their actual base-year costs looks—at the very least—suspect.
Closer inspection of the Hospitals’ argument makes clear that it is flawed. The whole exercise of inquiring as to whether a given methodology would reimburse the Hospitals for 100% of their actual base-year costs is odd under the inpatient prospective payment system. By statutory mandate, the Secretary never uses the hospital-specific rate he calculates for a base year to reimburse a hospital for discharges from that base year itself. Thus, for example, he never uses the rate calculated for the FY 2002 base year to reimburse a hospital for its 2002 discharges. Rather, that rate applies only for “discharges occurring on or after October 1, 2006.”
Moreover, the text on which the Hospitals rely further discredits their argument. The transitive verb “base” ordinarily means “use as a point from which (something) can develop.” Base, NEW OXFORD AMERICAN DICTIONARY 134 (2d ed. 2005). And as a noun, “base” often means “a main or important element or ingredient to which other things are added.” Id. So when Congress said that the payment using the hospital-specific rate “shall be” an “amount based on 100 percent of the hospital‘s target amount,”
And contrary to the Hospitals’ protestations, reading “based on” according to its ordinary meaning does not “render[] the phrase ‘100 percent’ surplusage.” Aplts.’ Opening Br. at 39. Under this commonsense reading, the phrase “100 percent” clarifies that the starting point for the Secretary‘s calculations must be 100%—not 90%, 50%, or some other percentage—of the hospital‘s target amount. For example, if Hospital X‘s base-year target amount was $1,000, the Secretary could not begin with $900, divide that amount by the normalized average diagnosis-related group weight, and then apply the budget-neutrality adjustment. Without the phrase “100 percent,” the provision would instruct—without any specificity—the Secretary to arrive at the hospital-specific rate using “an amount based on . . . the hospital‘s target amount.” On that wording, it would be unclear whether the Secretary must use an amount based on all or may use only a portion of the hospital‘s target amount. Cf. Anna Jaques Hosp. v. Sebelius, 583 F.3d 1, 5 (D.C. Cir. 2008) (agreeing that a provision instructing the Secretary to calculate “the wage index . . . ‘on the basis of‘” survey data was ambiguous because it was “silent about whether she must use all of the survey data“). The phrase “100 percent” has the effect of clearing up the matter. And so the Hospitals are mistaken in asserting that an interpretation of “based on” as meaning the starting point renders “100 percent” surplusage.27
Aside from its conceptual and linguistic shortcomings, the Hospitals’ argument is logically inconsistent with the Hospitals’ overall approach. Throughout their briefing, the Hospitals stress that they agree “that the Secretary should apply all budget-neutrality adjustments once.” Aplts.’ Reply Br. at 12 n.10. But if the Secretary had to pay hospitals 100% of their actual base-year costs, he could not apply any budget-neutrality adjustment at all. For example, say Hospital X‘s actual base-year costs were $1,000 and the budget-neutrality adjustment was 0.9. If the Secretary applied that adjustment to the base year, he would pay Hospital X $900 for its base-year discharges. Under the logic of the Hospitals’ argument, the Secretary could not apply that adjustment at all because doing so would be “an immediate reduction in their base-year costs.” Aplts.’ Opening Br. at 39. Yet, the Hospitals have repeatedly asserted that the Secretary is not prohibited from applying a budget-neutrality adjustment to the hospital-specific rate for a new base year. Thus, if—as the Hospitals agree—the Secretary may apply a budget-neutrality adjustment to the hospital-specific rate, then the Secretary cannot also be required to pay a hospital 100% of its base-year operating costs. Simply put, the Hospitals’ third argument—like its first two—fails to establish that the Secretary‘s methodology is arbitrary and capricious.
IV
For the foregoing reasons, we hold that the Secretary‘s methodology is not arbitrary and capricious and AFFIRM the district court‘s judgment.