American Hospital Association v. Thomas Price

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued May 15, 2017 Decided August 11, 2017 No. 17-5018 AMERICAN HOSPITAL ASSOCIATION, ET AL., APPELLEES v. THOMAS E. PRICE, IN HIS OFFICIAL CAPACITY AS SECRETARY OF HEALTH AND HUMAN SERVICES, APPELLANT Appeal from the United States District Court for the District of Columbia (No. 1:14-cv-00851) Joshua M. Salzman, Attorney, U.S. Department of Justice, argued the cause for appellant. With him on the briefs were Mark B. Stern, Attorney, Janice L. Hoffman, Associate General Counsel, U.S. Department of Health & Human Services, and Susan Maxson Lyons, Deputy Associate General Counsel. Catherine E. Stetson argued the cause for appellees. With her on the brief was Morgan L. Goodspeed. Adam K. Levin entered an appearance. Ronald S. Connelly was on the brief for amicus curiae Fund for Access to Inpatient Rehabilitation in support of appellees. 2 Before: GARLAND, Chief Judge, and HENDERSON and WILKINS, Circuit Judges. Opinion for the Court filed by Circuit Judge WILKINS. Dissenting Opinion filed by Circuit Judge HENDERSON. WILKINS, Circuit Judge: Ought implies can. 1 That is, in order for law – man-made or otherwise – to command the performance of an act, that act must be possible to perform. This lofty philosophical maxim, ordinarily relevant only to bright-eyed college freshmen, sums up our reasoning in this case. Congress established an administrative appeals process for denied Medicare reimbursement claims, and directed the U.S. Department of Health and Human Services (“HHS”) to complete that process within a specified timeframe. Buried under an ever-growing backlog of over a half-million appeals, HHS failed – and continues to fail – to comply with the statutorily mandated deadlines. Consequently, the American Hospital Association and three healthcare providers (together, “Healthcare Providers”) sought a mandamus order to force the HHS Secretary to clear the backlog and adhere to the statute’s timeframe. The District Court, in turn, thoughtfully and scrupulously weighed the equities, concluding that the scales tipped in favor of mandamus. 1 This principle is attributed to the 18th century German philosopher, Immanuel Kant. See, e.g., IMMANUEL KANT, CRITIQUE OF PURE REASON 548 (Norman Kemp Smith trans., Macmillan 1953) (1781) (“The action to which the ‘ought’ applies must indeed be possible under natural conditions.”); IMMANUEL KANT, RELIGION WITHIN THE LIMITS OF REASON ALONE 43 (Theodore M. Greene and Hoyt H. Hudson trans., Harper and Row 1960) (1793) (“[D]uty demands nothing of us which we cannot do.”). 3 The District Court was then confronted with the unenviable task of defining the scope and substance of the mandamus order. In an effort to minimize the judiciary’s intrusion on the political branches’ prerogatives, the Court adopted an ends-oriented approach of setting targets for HHS to hit, leaving to the Secretary the choice of means for hitting those targets. But what were the appropriate targets to set? The Healthcare Providers proposed an ambitious four-year timetable. The Secretary criticized that timetable as impossible to achieve lawfully and potentially counterproductive, but offered no alternative. Lacking a competing proposal, the District Court adopted the timetable suggested by the Healthcare Providers. In doing so, however, the Court declined to seriously grapple with the Secretary’s assertion that lawful compliance with such a mandamus order would be impossible. That is, the Court commanded the Secretary to perform an act – clear the backlog by certain deadlines – without evaluating whether performance was possible. We conclude that, notwithstanding the District Court’s earnest efforts to make do with what the parties presented, the failure to seriously test the Secretary’s assertion of impossibility and to make a concomitant finding of possibility was an abuse of discretion. The Court declared that a party ought without regard for whether the party can. I. A. “Medicare provides federally funded health insurance to disabled persons and those aged 65 or older . . . .” Council for Urological Interests v. Burwell, 790 F.3d 212, 215 (D.C. Cir. 2015) (discussing 42 U.S.C. §§ 1395 et seq.). After a healthcare provider (e.g., a hospital) performs a service it believes is covered by Medicare, it submits a claim for 4 reimbursement to the Centers for Medicare and Medicaid Services, an agency within HHS. 42 U.S.C. §§ 1395ff(a)(1)- (2), 1395kk-1(a); 42 C.F.R. §§ 405.904(a)(2), 405.920- 405.928. When a provider is denied reimbursement, or is otherwise “dissatisfied” with the initial determination, it is entitled to a four-level administrative appeals process, followed by judicial review. See generally 42 U.S.C. § 1395ff. We previously described the process in greater detail. See Am. Hosp. Ass’n v. Burwell, 812 F.3d 183, 185-87 (D.C. Cir. 2016) (hereinafter, “AHA I”). From start to finish, the administrative appeals process is designed to take less than one year. To keep things moving, the statute sets specific time frames for each of the four levels of the process: sixty days for the first level, 42 U.S.C. § 1395ff(a)(3)(C)(ii); another sixty days for the second level, id. § 1395ff(c)(3)(C)(i); ninety days for the third level, id. § 1395ff(d)(1)(A); and another ninety days for the fourth level, id. § 1395ff(d)(2)(A). “For years, the administrative appeal process functioned largely as anticipated, with its various stages typically completed within the statutory time frames.” AHA I, 812 F.3d at 186 (citing Am. Hosp. Ass’n v. Burwell, 76 F. Supp. 3d 43, 46 (D.D.C. 2014)). But starting in fiscal year 2011, an unexpected and dramatic uptick in appeals produced a jam in the process. The uptick was attributable to multiple causes, including “a large increase in the number of new beneficiaries as members of the ‘baby boom’ generation began to reach 65 and become eligible for Medicare,” and “a growing sense, among at least some members of the provider community, that it is a good business practice to appeal every denied claim.” Decl. of Ellen Murray, Chief Fin. Officer of the Dep’t of Health and Human Servs., J.A. 91-92. Furthermore, as we stressed in our previous decision, much of the increased workload can be traced back to 5 the congressionally mandated Medicare Recovery Audit Program. AHA I, 812 F.3d at 186-87. Under that program, recovery audit contractors (“RACs”) would review reimbursement claims that have already been paid, “identify[] underpayments and overpayments,” and “recoup[] overpayments.” 42 U.S.C. § 1395ddd(h)(1). When a RAC flags an overpayment, the healthcare provider could either repay the difference or appeal the RAC’s decision through the four-level administrative appeals process, as though the claim were denied at the outset. Id. § 1395ddd(f)(2)(A). Instead of repaying the difference, many providers elected to avail themselves of the administrative process. After the program was implemented, “the number of appeals filed ballooned from 59,600 in fiscal year 2011 to more than 384,000 in fiscal year 2013.” AHA I, 812 F.3d at 187. As those appeals moved through the process, they piled up at the third level, where an administrative law judge (“ALJ”) reviews the matter de novo. Instead of waiting in line, providers stuck at the ALJ level may skip to the next, through a process called “escalation.” 42 U.S.C. § 1395ff(d)(3). But that choice comes at a cost: the provider must forfeit certain procedural rights, such as a hearing before an independent ALJ. Id. § 1395ff(d)(1), (2). Many claimants, therefore, have been reluctant to “escalate” their appeals, and the ALJ backlog continues to grow. As of June 2, 2017, there was a backlog of 607,402 appeals awaiting review at this level. Status Report of Def. Thomas Price at 2, No. 14-cv-851 (June 5, 2017), ECF No. 56. On its current course, the backlog is projected to grow to 950,520 by the end of fiscal year 2021, id., and “some already-filed claims could take a decade or more to resolve,” AHA I, 812 F.3d at 187. This is, of course, far outside the ninety-day timeframe set by statute. 42 U.S.C. § 1395ff(d)(1)(A). 6 B. In 2014, the Healthcare Providers filed suit seeking a mandamus order to compel the HHS Secretary to clear the backlog and comply with the ninety-day statutory timeframe for ALJ hearings. The Healthcare Providers moved for summary judgment, and the Secretary simultaneously moved to dismiss for lack of subject-matter jurisdiction. Am. Hosp. Assoc. v. Burwell, 76 F. Supp. 3d 43, 45 (D.D.C. 2014). The District Court first grappled with whether it faced a jurisdictional question – i.e., whether, pursuant to 28 U.S.C. § 1361, the threshold mandamus requirements were met, United States v. Monzel, 641 F.3d 528, 534 (D.C. Cir. 2011) – or a merits question – i.e., whether mandamus would be equitable, Telecomms. Research & Action Ctr. v. FCC, 750 F.2d 70, 80 (D.C. Cir. 1984). The Court concluded that the jurisdictional and equitable merits inquiries were one and the same (“merged”), and so resolved the summary judgment and dismissal motions together. Am. Hosp. Assoc., 76 F. Supp. 3d at 49-50. Based on this merged analysis, the District Court granted the Secretary’s motion to dismiss, reasoning that “HHS’s budgetary constraints, its competing priorities, and its incipient efforts to resolve the issue together dictate that mandamus is not warranted.” Id. at 56. “Congress,” furthermore, was “aware of the situation and [was] in a position to address the problem.” Id. On appeal, we reversed the District Court’s dismissal. AHA I, 812 F.3d at 194. We first clarified that “the distinction between the jurisdictional inquiry and the equitable merits inquiry matters, especially because it affects our standard of review.” Id. at 190. As for the jurisdictional inquiry, we held that the Healthcare Providers “ha[d] demonstrated that the threshold requirements for mandamus jurisdiction [were] met.” 7 Id. at 192. We then left the equitable merits inquiry to the District Court to consider but, in an effort to help guide the Court’s “difficult decision,” we “set out the factors that weigh most strongly for and against mandamus in this case.” Id. Counseling for mandamus, we highlighted the backlog’s real impact on human health and welfare, and, “critically to our thinking,” the Secretary’s substantial discretion over the RAC program, which contributed significantly to the backlog. Id. at 193. Counseling against mandamus, we highlighted the risk of “infringing on the authority and discretion of the executive branch;” the legislative branch’s awareness of the problem and its capacity to furnish a comprehensive solution; the Secretary’s incipient but good-faith efforts to reduce the backlog; and the availability of some, albeit incomplete, alternative relief in the form of “escalation.” Id. at 192-93. Ultimately, “the clarity of the statutory duty,” we remarked, “likely will require issuance of the writ if the political branches have failed to make meaningful progress within a reasonable period of time – say, the close of the next full appropriations cycle.” 2 Id. at 193. 2 Fun fact: Even though we refer to the “writ” of mandamus, both in past decisions and here, the writ was technically abolished. FED. R. CIV. P. 81(b). As a matter of convenience and habit, we continue to refer to the “writ” because the remedy continues to exist in character, if not in name. 28 U.S.C. § 1361 (“The district courts shall have original jurisdiction of any action in the nature of mandamus to compel an officer . . . .” (emphasis added)); see also 33 CHARLES ALAN WRIGHT & CHARLES H. KOCH, JR., FEDERAL PRACTICE & PROCEDURE § 8299, at 41 (2006) (“Although [Rule 81] abolishe[d] the remedy formally known as mandamus, mandamus in character was not abolished by the rule change.”); id. at 42 (“[C]ourts in interpreting [§ 1361] brought over all the old mandamus restrictions and applied them in § 1361 actions.”). Consequently, when we refer to the “writ of mandamus” in this opinion, we mean the remedy provided for in 28 U.S.C. § 1361. 8 On remand, the District Court balanced the equities to determine whether mandamus was appropriate. After considering our guidance regarding the factors that counseled for and against the writ’s issuance, the District Court evaluated the political branches’ progress – and potential for progress – toward a solution. But by the Court’s estimation, the current measures were unlikely to yield meaningful progress, and so it concluded that the equities weighed in favor of mandamus. Having concluded that some relief was warranted, the District Court ordered further briefing and a status conference to determine the scope and substance of that relief. The Healthcare Providers proposed two sets of options: either a means-oriented plan requiring the Secretary to take specific actions, or an ends-oriented plan setting a timetable for clearing the backlog. The District Court opted for a timetable, reasoning that such an approach would “intrude as little as possible on the Secretary’s specific decisionmaking processes and operations.” Mem. Op. at 5, No. 14-851 (D.D.C. Dec. 5, 2016), ECF No. 48 (hereinafter, “Mandamus Op.”). Because it adopted the ends-oriented approach, the Court believed that it “need[ed] not dive into the parties’ debate over” the means. Id. Arguing against the Healthcare Providers’ proposed timetable, the Secretary advanced three contentions relevant here. First, although this Court indicated that curtailment or complete suspension of the RAC program would go a long way to clearing the backlog, AHA I, 812 F.3d at 193, the facts had since changed: few of the newly generated appeals were RAC- related. Second, since even dramatic changes to the RAC program would not enable compliance with the timetable, hitting the court-ordered targets would be impossible without settling unsubstantiated claims en masse, which the Secretary alleged would violate the Medicare statute. Third, the 9 timetable would only exacerbate the backlog: hard deadlines would counterproductively incentivize claimants to file meritless appeals and hold out for settlement. The District Court brushed aside the Secretary’s contentions. According to the Court, it “need[ed] not dive into the parties’ debate” over the “legality and propriety” of the reforms necessary to comply with the timetable, since it was not ordering any particular reforms. Mandamus Op. at 5. Furthermore, compliance with the timetable would not require violations of the Medicare statute, but rather “simply demand[ed] that the Secretary figure out how to undertake proper claim substantiation within a reasonable timeframe.” Id. (internal quotation marks omitted). Since the Secretary refused to engage with the premise of setting a timetable at all, proposing no alternative targets, the District Court adopted the Healthcare Providers’ four-year plan: the Secretary was ordered to reduce the current backlog of cases pending at the ALJ level by 30% by December 31, 2017; 60% by December 31, 2018; 90% by December 31, 2019; and 100% by December 31, 2020. After filing an unsuccessful motion for reconsideration, the Secretary appealed the District Court’s order. II. “Our consideration of any mandamus petition ‘starts from the premise that issuance of the writ is an extraordinary remedy, reserved only for the most transparent violations of a clear duty to act.’” In re Core Commc’ns, Inc., 531 F.3d 849, 855 (D.C. Cir. 2008) (quoting In re Bluewater Network, 234 F.3d 1305, 1315 (D.C. Cir. 2000)); accord Power v. Barnhart, 292 F.3d 781, 784 (D.C. Cir. 2002) (“The remedy of 10 mandamus is a drastic one, to be invoked only in extraordinary circumstances.” (internal quotation marks omitted)). We previously explained that the decision to issue mandamus relief involved two distinct inquiries: one jurisdictional, and one regarding the equitable merits. AHA I, 812 F.3d at 190. In that previous appeal, we settled the former question, holding that the threshold requirements for mandamus jurisdiction were met, id. at 192, although one of our sister circuits has since thought otherwise, Cumberland Cnty. Hosp. Sys., Inc. v. Burwell, 816 F.3d 48, 52-57 (4th Cir. 2016). We, of course, do not revisit our previous conclusion regarding mandamus jurisdiction. See LaShawn A. v. Barry, 87 F.3d 1389, 1395 (D.C. Cir. 1996) (en banc) (“One three- judge panel . . . does not have the authority to overrule another three-judge panel.”). Instead, we focus now on the equitable merits inquiry, along with the relief that the inquiry produced. We review this part of the District Court’s analysis for abuse of discretion. In re Medicare Reimbursement Litig., 414 F.3d 7, 10 (D.C. Cir. 2005). And “[a] district court by definition abuses its discretion when it makes an error of law.” Koon v. United States, 518 U.S. 81, 100 (1996). We conclude that since the Secretary represented that lawful compliance with the mandamus order was impossible, it was an error of law, and therefore an abuse of discretion, to nonetheless order the Secretary to render that performance without first finding that lawful compliance was indeed possible. Once the District Court determined that an ends-oriented approach of setting targets was the best course of action, it adopted the timetable proposed by the Healthcare Providers. Because it was mandating the ends, not the means, the Court 11 believed that it “need[ed] not dive into the parties’ debate over” the “legality and propriety” of the reforms necessary to clear the backlog. Mandamus Op. at 5. But this was a misstep. Although true that the Court was mandating no particular reforms, the Secretary would, of course, need to adopt some reforms to meet the mandated timetable. After all, that was the point of mandamus relief. But if, as the Secretary insisted, no lawful reforms could be implemented to meet the timetable, then it was an error of law to order the timetable met. The Secretary first contends that, given changing patterns in appeals, the tools within his discretion – most notably, curtailment or suspension of the RAC program – are not enough to clear the backlog. A major reason, according to the Secretary, is that the RAC program is no longer the principal cause of the backlog: only 9.5% of new appeals in 2016 were RAC-related, compared to more than 50% in 2013 and 2014. Appellant’s Br. at 18. This contention is, at best, suspect. Those statistics coincide with a two-year suspension of most of the RAC program, which was instituted while new contracts were being negotiated. See Suppl. Decl. of Ellen Murray, Chief Fin. Officer of the Dep’t of Health and Human Servs., J.A. 140-41 (“RAC activity decreased temporarily while [the Centers for Medicare and Medicaid Services] was negotiating a new Statement of Work (SOW) with the RACs, but several other changes took place that are expected to make lasting and continuing reductions to RAC-related appeal receipts.”); see also U.S. GOV’T ACCOUNTABILITY OFFICE, GAO-16-366, MEDICARE FEE-FOR-SERVICE: OPPORTUNITIES REMAIN TO IMPROVE APPEALS PROCESS 38 n.64 (2016) (explaining that the RAC program was temporarily suspended); id. at 38 (“HHS reported that it expects the number of incoming appeals to increase again when the new [RAC] contracts are awarded and 12 the [RAC] program resumes full operation.”). We are not sold on the Secretary’s suggestion that concerns regarding the RAC program are behind us, and the District Court should scrutinize that claim on remand. We also share the District Court’s skepticism of the Secretary’s assertion that he has done all he can to reduce RAC- related appeals. As the Court explained, there are “around 300,000 RAC-related appeals pending ALJ review, which constituted a sizable portion – 31% – of all pending . . . appeals.” Mem. Op. at 13, No. 14-851 (D.D.C. Sept. 19, 2016), ECF No. 38. “Yet the only RAC-related action the Secretary reports to be undertaking or planning to undertake consist of three modifications to RAC contracts that will reduce the number of appeals that reach [the ALJ level] by [fiscal year] 2020 by just 22,000.” Id. The Secretary’s RAC-related interventions appear to be curiously weak medicine for an agency facing mandamus. Nevertheless, the record supports the Secretary’s principal contention that reform of the RAC program and other programmatic tweaks may not be enough. At oral argument, the Healthcare Providers conceded that ALJs currently have the capacity to review only about 90,000 appeals per year. Oral Arg. at 27:05, Am. Hosp. Assoc. v. Price (May 15, 2017) (No. 17-5018). Even in the years when the RAC program was temporarily suspended, HHS received between 200,000 and 250,000 appeals. Therefore, although more reforms of the RAC program may help, even a complete suspension is likely to leave an annual disposition gap of over 100,000 appeals – appeals that will be piled onto the existing backlog, frustrating HHS’s efforts to comply with the statute’s timeframe and the Court’s mandamus order. 13 So what could the Secretary do to close the disposition gap and clear the backlog of over a half-million pending appeals? There appears to be no dispute that mass settlements would play a central role. But the Secretary repeatedly insisted that the type of mass settlement necessary to comply with the Court’s timetable would be illegal. Specifically, the Secretary argued that the Healthcare Providers’ proposal required him “to make payment on Medicare claims regardless of the merit of those claims,” which would “squarely conflict with the Medicare statute.” Def.’s Mot. for Summ. J. at 23, No. 14-cv- 851 (Nov. 7, 2016), ECF No. 41 (discussing 42 U.S.C. §§ 1395f, 1395g(a), 1395y(a)(1)(A)). The Court declined to seriously grapple with the Secretary’s contention, explaining matter-of-factly that the timetable “simply demands that the Secretary figure out how to undertake proper claim substantiation within a reasonable timeframe.” Mandamus Op. at 5 (internal quotation marks omitted). But that response gave short shrift to the Secretary’s proffer that “proper claim substantiation within a reasonable timeframe” was impossible. The Secretary essentially asserted that the timetable placed him between a rock and a hard place: either violate the Medicare statute by settling reimbursement claims en masse without regard for their merit, or violate the Court’s mandamus order by missing the court-ordered deadlines. By declining to evaluate the Secretary’s claims, the Court was, in effect, saying: “hit the targets by any means necessary.” But if the necessary means were unlawful, the Court could not have mandated them; equity courts, like any other, may not order parties to break the law. See INS v. Pangilinan, 486 U.S. 875, 883 (1988) (“[I]t is well established that courts of equity can no more disregard statutory and constitutional requirements and provisions than can courts of law.” (alterations and internal quotation marks omitted)). 14 But if only lawful reforms were implemented, the Secretary claimed, compliance with the timetable would be impossible. And just as a court may not require an agency to break the law, a court may not require an agency to render performance that is impossible. See Ala. Power Co. v. Costle, 636 F.2d 323, 359 (D.C. Cir. 1979); NRDC v. Train, 510 F.2d 692, 713 (D.C. Cir. 1974). A century ago, we explained that “[t]he writ of mandamus will not issue to compel the performance of that which cannot be legally accomplished.” United States ex rel. Newman v. City & Suburban Ry. of Wash., 42 App. D.C. 417, 420-21 (D.C. Cir. 1914). The reasoning is simple and intuitive: it is not appropriate for a court – contemplating the equities – to order a party to jump higher, run faster, or lift more than she is physically capable. This principle extends to cases where the impossibility is the result of insufficient congressional appropriations. See, e.g., Morton v. Ruiz, 415 U.S. 199, 230-31 (1974) (recognizing that an agency may balance competing statutory commands to cope with insufficient appropriations); Ala. Power, 636 F.2d at 359 (same); Train, 510 F.2d at 710-14; see also In re Aiken Cnty., 725 F.3d 255, 259 (D.C. Cir. 2013) (“Under Article II of the Constitution and relevant Supreme Court precedents, the President must follow statutory mandates so long as there is appropriated money available and the President has no constitutional objection to the statute.” (emphasis added)). For example, in Train, we considered the prospect that practical challenges, such as resource constraints, might prevent the EPA Administrator from meeting a statutory deadline for publishing certain guidelines. 510 F.2d at 710-14. In light of those challenges, like the possibility that “budgetary commitments and manpower” would render performance “beyond the agency’s capacity or would unduly jeopardize the implementation of other essential programs,” we remarked that “courts cannot responsibly mandate flat guideline deadlines 15 when the Administrator demonstrates that additional time is necessary.” Id. at 712. “The sound discretion of an equity court,” we concluded, “does not embrace enforcement through contempt of a party’s duty to comply with an order that calls him ‘to do an impossibility.’” Id. at 713 (quoting Maggio v. Zeitz, 333 U.S. 56 (1948)). By extension, where a party insists that resource constraints render lawful compliance with a court’s order impossible, an equity court must examine that claim and, prior to issuing the order, find that lawful compliance is indeed possible. The District Court made no such finding. The Court also did not evaluate the Secretary’s assertion that the timetable would increase, not decrease, the number of backlogged appeals. The Secretary posited that because strict deadlines would require settlements en masse, the timetable would generate an incentive for claimants to file additional appeals and hold out for big payouts. By the Secretary’s account, the mandamus relief would prove counterproductive; the relief would exacerbate the risk of the Court’s order amounting to a command to do the impossible. The Court did not address this claim, perhaps because, as a counterfactual, such an assertion is difficult to test. But the claim was plausible enough that, as a matter of crafting an equitable remedy, the Court should address it. On remand, the Court should determine in the first instance whether, in fact, lawful compliance with the timetable is impossible. We note, however, that the Secretary bears the “heavy burden to demonstrate the existence of an impossibility.” Ala. Power, 636 F.2d at 359 (discussing Train, 510 F.2d at 713). The burden serves to prevent an agency from shirking its duties by reason of mere difficulty or inconvenience. As we explained before, although “[a]n equity court can never exclude claims of inability to render absolute 16 performance, . . . it must scrutinize such claims carefully since officials may seize on a remedy made available for extreme illness and promote it into the daily bread of convenience.” Train, 510 F.2d at 713. Therefore, on remand, if the Court finds that the Secretary failed to carry his burden of demonstrating impossibility, it could potentially reissue the mandamus order without modification. But given the Secretary’s claim of impossibility, the Court must make the predicate finding of possibility. Our dissenting colleague believes such a finding is unnecessary, Dissenting Op. at 10-17, and amounts to a hyper- technical procedural requirement for the District Court to “incant magic words,” id. at 2. But possibility is a necessary and antecedent condition for the writ’s issuance, according to both our precedent and the collected wisdom of our sister courts. See, e.g., Newman, 42 App. D.C. at 420-21 (“The writ of mandamus will not issue to compel the performance of that which cannot be legally accomplished.” (emphasis added)); 52 AM. JUR. 2D § 24 (2017 Update) (“To warrant the issuance of a writ of mandamus, the act sought to be performed must be capable of being performed. Mandamus will not issue if the performance of the requested action is impossible, or beyond the physical, mental, or financial power of the respondent.” (emphasis added)); 55 C.J.S. Mandamus § 19 (2017 Update) (“The writ of mandamus will not lie where performance of the duty is impossible. Thus, the court will not grant the writ unless the law afford the means by which the officer may discharge the prescribed duty.” (emphasis added)); id. § 20 (“[A] public officer or public body will generally not be required to do an act when it is impossible through a want of funds and inability to raise them.”). We are also not asking for a magic incantation. There is nothing mystical or punctilious about the judiciary giving due 17 consideration to an executive agency’s central argument – made repeatedly and emphatically across three sets of motions, not solely with allegations but with proffers of evidence 3 – before issuing extraordinary relief with multi-billion-dollar stakes, Suppl. Decl. of Ellen Murray ¶ 6, J.A. 170 (amount-in- controversy for pending appeals is approximately $6.6 billion). Such a requirement is neither onerous nor trivial, and the interests of alacrity and expedition do not excuse its satisfaction. See Albemarle Paper Co. v. Moody, 422 U.S. 405, 416 (1975) (“That the court’s discretion is equitable in nature hardly means that it is unfettered by meaningful standards or shielded from thorough appellate review.” (citation omitted)). In sum, it was an abuse of discretion to tailor the mandamus relief without tackling the Secretary’s claims that 3 See, e.g., Def.’s Mot. for Summ. J. at 23, No. 14-cv-851 (Nov. 7, 2016), ECF No. 41 (“It is telling that Plaintiffs are unable to identify a remedy that would not cause the Secretary to violate her other obligations under the Medicare statute.”); id. at 1, 7-8, 11-24; Def.’s Reply in Support of Summ. J. at 9, No. 14-cv-851 (Nov. 23, 2016), ECF No. 45-1 (“Plaintiffs’ deadlines indeed would be impossible for the Secretary to meet absent augmentation of her resources and authorities, which only Congress can provide.”); id. at 1-2, 7-11; Def.’s Opp’n to Pls’ Mot. for Summ. J. at 1, 7-8, 11-24, No. 14-cv- 851 (Nov. 7, 2016), ECF No. 42; Def.’s Mot. for Recons. at 1, No. 14-cv-851 (Dec. 15, 2016), ECF No. 49 (“Specifically, the ruling errs in ordering scheduled percentage reductions in the Medicare appeals backlog that the Secretary cannot achieve unless she were to pay pending claims without regard to their merit, which would violate her statutory obligation to protect the Medicare Trust Funds.”); id. at 2-3; Def.’s Reply in Support of Mot. for Reconsideration at 2, No. 14-cv-851 (Dec. 23, 2016), ECF No. 51 (“And notably, Plaintiffs do not and cannot deny that it is impossible for the Secretary to comply with the benchmarks set forth in this Court’s [timetable] unless she offers settlements without regard to the merits of the claims . . . .”). 18 lawful compliance would be impossible. We emphasize, however, that the District Court was assigned an exceptionally difficult project. The Secretary presented a flurry of arguments as to what cannot be mandated, but a paucity of proposals regarding what can be. With little assistance from the party best positioned to furnish crucial information, the Court needed to craft workable relief while negotiating both the on-the- ground realities and the guidance offered in our past decision. An unenviable task. 4 Difficult as it was, however, courts must ensure that it is indeed possible to perform the act being commanded. Ought, after all, implies can. *** For the foregoing reasons, we vacate the mandamus order and the order denying reconsideration, and remand to the District Court to evaluate the merits of the Secretary’s claim that lawful compliance would be impossible. So ordered. 4 We note that if, despite his burden, the Secretary fails to offer information that would aid the crafting of mandamus relief, the Court has options. See, e.g., FED. R. CIV. P. 56(e) (providing that courts may order parties to address facts or “issue any other appropriate order”); FED. R. CIV. P. 53(a) (authorizing courts to appoint special masters). KAREN LECRAFT HENDERSON, Circuit Judge, dissenting: Just 18 months ago, we reversed the district court for holding that it lacked jurisdiction to compel the Department of Health Human Services (HHS), via mandamus, to comply with statutory deadlines for resolving Medicare reimbursement appeals. Am. Hosp. Ass’n v. Burwell, 812 F.3d 183, 192 (D.C. Cir. 2016) (AHA I). Further, we indicated that mandamus would “likely” be “require[d]” by the end of September 2017 unless HHS, with congressional assistance if necessary, made “meaningful progress” toward reducing a backlog of hundreds of thousands of appeals filed with the agency’s administrative law judges (ALJs). Id. at 193. Seeing no such progress, the district court on remand issued a mandamus order directing HHS to eliminate the backlog by December 31, 2020, and to meet reduction targets in the interim. Today my colleagues overturn the district court again, this time concluding that it abused its discretion in too readily imposing a schedule for statutory compliance. Why the change of direction? It is not because the district court miscalculated the equities. Maj. Op. 2 (court “thoughtfully and scrupulously weighed the equities”). It is not because the court lacked a basis for issuing the writ. Maj. Op. 16 (court “could potentially reissue the mandamus order”). It is not even because HHS cannot lawfully comply with the court’s order; impossibility is HHS’s primary argument but my colleagues do not consider it. Maj. Op. 15 (reserving issue for district court). Instead they send the case back because of a perceived procedural error: “since [HHS] represented that lawful compliance with the mandamus order was impossible, it was an error of law, and therefore an abuse of discretion, to nonetheless order the [agency] to render that performance without first finding that lawful compliance was indeed possible.” Maj. Op. 10 (emphasis altered). On both law and fact, I disagree. 2 A district court need not make a finding of possibility as a precondition to mandamus relief unless the agency makes a strong threshold showing of impossibility. HHS has not met its burden: lawful compliance with the mandamus order here, even if difficult, is not demonstrably impossible. The district court expressly found as much in rejecting HHS’s impossibility claim. And by necessary implication, it found as much in its careful equities analysis. Remanding so that the court can incant magic words—“lawful compliance [is] indeed possible,” Maj. Op. 10 (emphasis in original)—will tell us only what we already know and almost certainly produce a third appeal. The process will waste time, punishing blameless Medicare providers who await billions of dollars of delayed payments essential to their operations. I. BACKGROUND The majority recounts much of the legal, factual and procedural background, Maj. Op. 2-9, but I offer some additional context. A. THE BACKLOG Under Title XVIII of the Social Security Act—formally named the Health Insurance for the Aged Act, Pub. L. No. 89- 97, 79 Stat. 286 (July 30, 1965), and better known as the Medicare Act (Act), 42 U.S.C. §§ 1395 et seq.—a healthcare provider (e.g., a hospital) that treats a Medicare patient may seek government reimbursement by filing a claim with an HHS contractor overseen by the agency’s Center for Medicare and Medicaid Services (CMS). 42 U.S.C. § 1395ff(a)(1)-(2); 42 C.F.R. § 405.904(a)(2). If the initial contractor denies reimbursement, the hospital can seek further review from other CMS contractors. 42 U.S.C. § 1395ff(a)(3), (c); 42 C.F.R. § 405.904(a)(2). At the end of the CMS process, a dissatisfied hospital may seek a de novo hearing before an ALJ in HHS’s 3 Office of Medicare Hearings and Appeals (OMHA). 42 U.S.C. § 1395ff(d)(1); 42 C.F.R. § 405.904(a)(2). Under the Act, the OMHA ALJ “shall”—not merely “ought” to, Maj. Op. 2-3, 18 (emphasis omitted)—“render a decision” within 90 days of the hospital’s request for a hearing. 42 U.S.C. § 1395ff(d)(1)(A). “The word ‘shall’ is ordinarily the language of command.” Alabama v. Bozeman, 533 U.S. 146, 153 (2001) (some internal quotations omitted). And so it is here. Because the Act uses mandatory language and refers to the 90-day time limit as a “[d]eadline[],” 42 U.S.C. § 1395ff(d), the time limit is a “congressionally imposed mandate[]” that HHS “must obey,” AHA I, 812 F.3d at 193. The point is obvious but critical. The deadline is not a guideline. HHS has been violating it every day for years on end in failing to timely decide administrative appeals. It now takes an OMHA ALJ an average of nearly three years—more than eleven times longer than permitted—to process a Medicare appeal. HHS, Office of Medicare Hearings and Appeals: Workload Information and Statistics—Average Processing Time by Fiscal Year (May 24, 2017), www.hhs.gov/about/agencies/omha/about/current-workload/ average-processing-time-by-fiscal-year/index.html. As a result, appeals are badly backlogged. At last count, more than 600,000 of them are pending ALJ review. Status Report, Dkt. No. 56, Ex. at 2 (June 5, 2017). HHS projects that the backlog will grow worse with time, snowballing to nearly one million appeals by the end of September 2021. Id. at 8. Absent drastic action, then, it will soon take an ALJ significantly longer than three years to resolve a Medicare appeal. 4 Granted, the ALJ delays and backlog are not a simple matter of agency lassitude. As the majority explains, Maj. Op. 4-5, the number of appeals has risen sharply since the 2011 fiscal year, largely because (1) much of the baby boom generation has reached age 65 and enrolled in Medicare, Joint Appendix (JA) 84, 91, and (2) the Congress authorized implementation of an allegedly cost-saving but time- consuming program under which Recovery Audit Contractors (RACs) identify and recoup alleged overpayments of Medicare reimbursements, 42 U.S.C. § 1395ddd(h)(1). The government pays RACs “on a contingent basis for collecting overpayments.” Id. § 1395ddd(h)(1)(B)(i). A hospital can appeal a finding of overpayment through the same process by which it appeals a denial of reimbursement. Not surprisingly, statistics show that the contingent pay structure gives RACs strong incentive to find overpayments that do not exist. See, e.g., JA 47-48 (in first quarter of 2014, surveyed hospitals collectively reported 66 per cent success rate in appealing adverse RAC decisions). And the incentive to find overpayments has produced a counter-incentive to appeal. B. THE PLAINTIFFS Even the most conservative statistics show that a considerable number of all appeals, not only appeals from adverse RAC decisions, are “[f]ully” meritorious. 1 HHS, Office of Medicare Hearings and Appeals: Workload Information and Statistics—Decision Statistics (Jan. 27, 2017) (53 per cent of appeals in fiscal year 2012 resulted in fully favorable disposition; in fiscal 2013, number was 44 per cent; in fiscal 2014, 37 per cent; in fiscal 2015, 34 per cent; in fiscal 1 If there is a “growing sense” among providers “that it is a good business practice to appeal every denied claim,” Maj. Op. 4 (quoting JA 92), it is probably because they often prevail. 5 2016, 26 per cent; and so far in fiscal 2017, 25 per cent), www.hhs.gov/about/agencies/omha/about/current-workload/ decision-statistics/index.html. The upshot is that hospitals are forced to wait years to receive reimbursements to which the law entitles them now. And they are waiting for quite a lot of money. HHS acknowledges that “[t]he combined billed amounts of the outstanding claims total approximately $6.6 billion.” Br. of Appellant 2. Breaking that number down a bit, plaintiff American Hospital Association (AHA) in 2014 surveyed more than 1,000 of its approximately 5,000 member hospitals and learned that “[t]he value of appealed . . . RAC-denied claims” for those hospitals alone exceeded $1.8 billion. JA 48. Considering the appellate success rate, especially in RAC cases, I believe it is fair to say that on any given day the AHA hospitals represented in this lawsuit are collectively out of pocket nearly a billion dollars of their own money. The delays are causing real-world problems. For plaintiff Baxter Regional Medical Center in Arkansas, Medicare reimbursements represent about two-thirds of gross revenue. In 2014, with millions of dollars tied up in Medicare appeals delayed at the ALJ level, Baxter lacked the cash for essentials such as “[p]urchasing . . . beds for its intensive care unit,” “[r]eplacing a failing roof over its surgery department” and “[r]eplacing its twenty-year-old catheterization laboratory.” JA 38. Plaintiffs Rutland Regional Medical Center in Vermont and Covenant Health in Tennessee face similar problems. For them, Medicare reimbursements represent about half of gross revenue. Collectively, they too have millions of dollars tied up in appeals delayed at the ALJ level. Partly because of the delays, Rutland has had to eliminate 32 jobs and Covenant is contemplating whether to cut back patient services. 6 C. AHA I Based on HHS’s serial statutory violations and with no end in sight, AHA, Baxter, Rutland and Covenant sought relief in district court under the Mandamus Act, 28 U.S.C. § 1361. The court dismissed the complaint for lack of mandamus jurisdiction. 76 F. Supp. 3d 43, 56 (D.D.C. 2014). As noted, we reversed. AHA I, 812 F.3d at 189-94. My colleagues summarize the decision in AHA I, Maj. Op. 6-7, and they take pains not to disturb it, Maj. Op. 10 (“One three-judge panel . . . does not have the authority to overrule another three- judge panel[.]” (quoting LaShawn A. v. Barry, 87 F.3d 1389, 1395 (D.C. Cir. 1996) (en banc))). Without purporting to provide a full summary of my own, I recap three important points. First, we held that the plaintiffs have a “right to demand . . . compliance” with the Act’s “mandatory” “deadlines.” AHA I, 812 F.3d at 190, 192. We acknowledged HHS’s argument that it “lacks the resources to render decisions within the statutory time frames.” Id. at 191. But we observed that “however modest [an agency’s] personnel and budget resources may be, there is a limit to how long it may use these justifications to excuse inaction in the face of a statutory deadline.” Id. (internal quotation omitted). Second, we pointed out that the Act gives HHS “substantial discretion” to limit the scope of the RAC program. AHA I, 812 F.3d at 193. Because “congressionally imposed mandates . . . trump discretionary decisions,” we held that HHS “will have to curtail the RAC program or find some other way to meet” the 90-day ALJ-level deadline. Id. Third, although recognizing the district court’s “broad discretion in weighing the equities,” we also suggested that “the unique circumstances of this case” and “the clarity of the 7 statutory duty likely will require issuance of the writ” absent “meaningful progress” by “the close of the next full appropriations cycle”—i.e., by the end of September 2017. AHA I, 812 F.3d at 193. D. THE MANDAMUS ORDER We issued our mandate in April 2016. Mistaking defeat as victory, HHS sought an 18-month stay of the proceedings on remand so that it could “continue to make meaningful progress in resolving the OMHA backlog.” Def.’s Mot. for Stay, Dkt. No. 30 at 2 (May 25, 2016). Discerning no such progress, 209 F. Supp. 3d 221, 227-30 (D.D.C. 2016), and carefully weighing the equities that would also guide its mandamus decision, id. at 225-26; see id. at 225 (noting that “stay and mandamus inquiries . . . overlap[]”), the district court denied the stay motion, id. at 230. It was unmoved by the agency’s modest restructuring of RAC contracts, its tinkering with appeals procedures, its proposal to “facilitate settlement conferences” and its recall of retired ALJs. Id. at 227-28. Even with those initiatives, the agency projects that its ALJ-level backlog will exceed 800,000 appeals by the end of fiscal year 2020. Status Report, Dkt. No. 56, Ex. at 3. Without the initiatives, the backlog would likely balloon to nearly two million appeals by the end of fiscal 2020. 209 F. Supp. 3d at 228. Noting the statistics, the court reasoned that “significant progress toward a solution . . . has to mean real movement toward statutory compliance,” not merely “that things get worse more slowly than they would otherwise.” Id. (internal quotation marks omitted). The court was especially “concern[ed]” about HHS’s RAC proposals, which are projected to “reduce the number of appeals . . . by just 22,000” before fiscal year 2020. Id. at 228-29. The court also saw no legislative fix on the horizon. It emphasized that the Congress, although armed with “ample knowledge of the backlog,” has repeatedly failed 8 to provide significant assistance through appropriations or other means. Id. at 230. The plaintiffs moved for summary judgment and proposed specific methods for eliminating the backlog. They suggested the district court compel HHS to offer broad-based settlement of pending claims, defer repayment of overpayments and penalize RACs for high reversal rates. Alternatively, they proposed that the court order a four-year timetable for eliminating the backlog. HHS opposed all of the plaintiffs’ suggestions. As relevant here, it argued that their proposed timetable “conflict[s] with the Medicare statute” by requiring the agency “to make payment on Medicare claims regardless of . . . merit.” Def.’s Opp., Dkt. No. 41 at 23 (Nov. 7, 2016) (citing 42 U.S.C. §§ 1395g(a), 1395y(a)(1)(A)). HHS proposed no methods of its own that can eliminate the backlog absent legislative action. Nor did it offer an alternative timetable. Instead it touted the modest measures it is already taking and claimed that they will improve the situation “if Congress increases HHS’ authorities and funding.” Id. at 6. In December 2016, the district court granted summary judgment to the plaintiffs, concluding that HHS did “not provide enough evidence of progress” to alter the court’s earlier calculation of the equities. 2016 WL 7076983, at *2 (D.D.C. 2016); see id. (“[HHS] does not point to any categorically new administrative actions and, critically, continues to promise the elimination of the backlog only ‘with legislative action’—a significant caveat.” (quoting Def.’s Opp. 6)). Recognizing that the plaintiffs “could have chosen to demand immediate relief,” the court commended them for instead “offer[ing] a thoughtful and reasonable four-year plan for this complex problem.” Id. at *3. Also, it expressly rejected HHS’s argument that lawful compliance with the plaintiffs’ timetable is not possible. Id. It reasoned that the 9 timetable “demands . . . ‘proper claim substantiation’ within a reasonable timeframe” but does not require the agency to “‘make payment on Medicare claims regardless of . . . merit.’” Id. (quoting Def.’s Opp. 22-23). Accordingly, and in an attempt to “intrude as little as possible on [HHS’s] specific decisionmaking processes and operations,” the district court adopted the plaintiffs’ proposed timetable but did not dictate any particular method for eliminating the backlog. 2016 WL 7076983, at *3. The court mandated a 30 per cent reduction of the backlog by December 31, 2017; a 60 per cent reduction by December 31, 2018; a 90 per cent reduction by December 31, 2019; and 100 per cent elimination by December 31, 2020. Id. The court added that, “if [HHS] fails to meet the above deadlines, [p]laintiffs may move for default judgment or to otherwise enforce the writ of mandamus.” Id. (citing FED. R. CIV. P. 55(d)). II. ANALYSIS We review issuance of the writ of mandamus for abuse of discretion, AHA I, 812 F.3d at 190, which means “[w]e must give the benefit of every doubt to the judgment of the trial judge,” Gasperini v. Ctr. for Humanities, Inc., 518 U.S. 415, 438-39 (1996) (internal quotation omitted). I see no abuse here, nor any legal error tantamount to one, contra Maj. Op. 10 (citing Koon v. United States, 518 U.S. 81, 100 (1996)). A. THE DISTRICT COURT DID NOT HAVE TO MAKE A FINDING OF POSSIBILITY. My colleagues hold that the district court had to, and failed to, make a “finding” that HHS can lawfully comply with the mandamus order. Maj. Op. 10. Their primary authority appears to be NRDC v. Train, 510 F.2d 692 (D.C. Cir. 1974). But to the extent that Train applies, it supports the district 10 court’s judgment. At most it requires a finding of possibility if an agency makes a strong threshold showing of impossibility. HHS has not done that. 1. HHS’s burden At issue in Train were statutory deadlines by which the Environmental Protection Agency (EPA) had to issue guidelines on “the quantity of pollutants that may be discharged into the nation’s waters.” 510 F.2d at 695. It had to issue guidelines for what we called “Group I” sources of pollutants by October 18, 1973. Id. at 704-05. It had to issue guidelines for certain “Group II” sources by December 31, 1974. Id. at 706-11. The first deadline came and went “without the publication of a single . . . guideline.” Id. at 704. The National Resources Defense Council sued in an effort to compel the EPA to act. Id. at 695. In November 1973, the district court ordered the EPA “to comply with a detailed timetable for publication of guidelines” for both Group I and Group II sources “beginning on January 15, 1974, and ending on November 29, 1974.” Id. at 697-98. We vacated and remanded in part. Train, 510 F.2d at 705, 713-14. We first held that the district court “acted reasonably” as to Group I sources: In light of the failure of the agency to meet its acknowledged duty [as to Group I sources], the District Court’s decision to incorporate a timetable into the order constituted a reasonable step to facilitate supervision of the decree and to assure early efforts by the delinquent defendant toward eventual discharge of its statutory responsibility. . . . The authority to set enforceable deadlines both of an ultimate and an intermediate nature is an appropriate procedure 11 for exercise of the court’s equity powers to vindicate the public interest. Id. at 704-05 (footnotes omitted). Turning to the December 1974 statutory deadline for certain Group II sources, we found—before the deadline had passed—“no present failure on the part of the [EPA] Administrator to meet his responsibility.” Train, 510 F.2d at 711. Seeing “no violation of a statutory duty,” we thought it appropriate “to give the Administrator latitude to exercise his discretion in shaping the implementation of the [statute],” id. at 711-12, and we vacated the district court’s timetable insofar as it required early issuance of guidelines for Group II sources, id. at 705, 714. We acknowledged the EPA’s “apprehension” that “manpower or methodological constraints” might prevent compliance as to Group II sources. Id. at 712-13. We instructed the district court that it could not “responsibly mandate flat . . . deadlines” for those sources on remand if “the Administrator demonstrates that additional time is necessary.” Id. at 712 (emphasis added). Our reasoning was that a court’s equitable discretion does not include the authority to punish, through contempt, a party who has “demonstrated that he [is] powerless to comply” with a court order. Id. at 713 (emphasis added). In the majority’s telling, because HHS “represented” that compliance is impossible, the district court had to make a finding of possibility before it could impose a timetable on the agency. Maj. Op. 10; see id. at 15 (district court must make finding of possibility when agency “insists” compliance is not lawfully possible). But nowhere in Train did we suggest that, before ordering phased fulfillment of a statutory obligation an agency is then violating, a court must make a finding of possibility if the agency merely asserts impossibility and 12 makes no threshold “demonstrat[ion]” of it. 510 F.2d at 712- 13. Not only is the majority’s requirement a new one; it is contrary to sound practice. True, a district court should not hold agency officials in contempt “for omitting an act [they are] powerless to perform.” Maggio v. Zeitz, 333 U.S. 56, 72 (1948); see Maj. Op. 15. But private parties have a “right to demand . . . compliance” with a statute’s “mandatory” “deadlines.” AHA I, 812 F.3d at 190, 192. It follows that a court is ordinarily entitled to presume, absent a contrary showing, that the agency can comply. After all, the doctrine of administrative impossibility is a narrow one reserved for “extreme” circumstances. Train, 510 F.2d at 713. If the agency cannot at the threshold meet its “heavy burden to demonstrate . . . impossibility,” Ala. Power Co. v. Costle, 636 F.2d 323, 359 (D.C. Cir. 1979), the court should be permitted to act immediately. Saddling it with a finding requirement— which in some cases might also impose a hearing requirement, see Oral Arg. Recording 21:22-21:35—simply slows the process for issuing time-sensitive relief. In the meantime, faultless private parties bear the costs of the agency’s ongoing statutory violations. 2. HHS’s arguments HHS has not made a sufficient threshold “demonstrat[ion]” of impossibility to trigger any finding on the matter. Train, 510 F.2d at 712-13. For starters, its burden is “especially heavy” because it seeks “prospective exemption” from a four-year timetable “based upon [its] prediction” of impossibility. Ala. Power, 636 F.2d at 359 (emphasis added). Its prediction rests on two premises: (1) it “has a statutory obligation to ensure that non-meritorious claims are not paid,” Br. of Appellant 20 (citing 42 U.S.C. § 1395y(a)); and (2) 13 “curtailment of the RAC program cannot resolve the backlog,” id. at 18. The first point is true as far it goes and the second may prove true eventually. But HHS oversells the importance of both points. First, although HHS is not authorized to reimburse providers for items and services that are not medically “reasonable and necessary,” 42 U.S.C. § 1395y(a)(1)(A)-(E), its own regulations permit it to settle claims that are less than certain to prove meritorious on a case-by-case basis. For example, CMS may “compromise” some kinds of claims, including ones relating to overpayment, based on “[l]itigative probabilities” and related considerations. 42 C.F.R. §§ 401.613(c)(2), 405.376(d), (h). HHS offers no good reason to reject the statutory interpretation embodied in the regulations. Nor does it cite any other relevant authority prohibiting it from adopting the plaintiffs’ primary proposal: offering systematic settlements based on the provider, the type of claim or both. The proposal is a sound one. It does not ask HHS to authorize payments without regard to merit. It asks HHS to evaluate merit “at a higher level of generality” based on statistical sampling. Br. of Appellees 23. In district court, HHS’s then-chief financial officer acknowledged that the agency can “resolve pending appeals at OMHA by applying an individualized payment percentage” based on the specific provider’s “historic success rate.” JA 152. Similarly, HHS acknowledges in this Court that it has “globally settled” claims before, dispatching some 380,000 of them based on type without case-by-case adjudication. Br. of Appellant 9. HHS says the remaining claims in the backlog are not appropriate for “bulk settlements” because they do not appear “homogeneous” enough. Br. of Appellant 26. The agency needs to look more closely. HHS’s chief financial officer admitted that a single durable medical equipment supplier is 14 responsible for “more than 24% of all pending appeals” at the ALJ level. JA 138. Suppose, hypothetically, that the supplier’s historic success rate with the ALJs is 50 per cent. 2 As far as I can tell, nothing but obstinance is stopping HHS from offering the supplier 50 cents on the dollar—or less, to account for the time, litigation expenses and uncertainty it spared the supplier—to resolve the supplier’s claims. Just that one approach to one repeat claimant has the potential to dispatch nearly a quarter of the appeals from HHS’s backlog. If the agency were to take the same approach with other high- volume claimants based on each claimant’s historic success rate, the record manifests that it would put a tremendous dent in the backlog. See id. (noting that “top ten appellants” within backlog “comprise more than 40% of all pending appeals”). Granted, bulk settlement poses hazards. If HHS, under the thumb of a mandamus order, looks too willing to settle outside the parameters of case-by-case adjudication, it weakens its bargaining position; enables “hospitals with below-average success rates [to] accept at disproportionately high rates”; and may incidentally encourage the filing of baseless claims. Reply Br. of Appellant 9; see Status Report, Dkt. No. 55, Ex. at 4 (Mar. 5, 2017) (according to acting chief financial officer, some claimants are delaying settlement “in the anticipation that relief mandated by the [district court] will yield a higher payout”). Indeed, the percentage of fully successful ALJ appeals has declined in recent years, see HHS, Decision Statistics, supra pp. 4-5, which might indicate that some bad 2 The chief financial officer’s declaration suggests the rate is higher, JA 138, but at oral argument HHS said the declaration is not “artfully phrased” in that regard, Oral Arg. Recording 12:45-13:41. The particular number is immaterial for my purpose; the agency can offer the supplier a settlement commensurate with the supplier’s historic success rate, whatever it is. 15 actors have injected dubious claims into the backlog in hopes of artificially favorable settlements imposed by the courts. But bargaining power is a two-way street. Subjecting the average claimant to a waiting period more than eleven times longer than the statute permits—and thereby choking off cash flow for basic operational needs—unfairly weakens the claimant’s position, giving it every incentive to settle for only a fraction of what it might win after years of litigation. In other words, the backlog undermines both parties. Thus, if done right, a bulk settlement could well be negotiated in near equipoise. HHS has offered nothing better than rank speculation for concluding otherwise. And the burden, remember, is on HHS. In any event, the district court did not mandate settlement, let alone dictate particulars. HHS is therefore free to mitigate hazards based on its expertise and experience. One way to do so is to focus on high-volume claimants: as mentioned, just a handful account for hundreds of thousands of appeals in the backlog, JA 138, and each is presumably a sizable company with a long-term reputational stake and a predictive success rate rooted in a meaningful sample size. To further ensure a level playing field and to dissuade bad actors, the plaintiffs reasonably suggest that HHS “requir[e] a provider to settle all eligible appeals and . . . extend[] an offer only to those claims pending at a particular date.” Br. of Appellees 26. The agency has undertaken such measures before. CMS, Frequently Asked Questions—Hospital Appeals Settlement for Fee-for-Service Denials Based on Patient Status Reviews for Admissions Prior to October 1, 2013, at 3, https://goo.gl/YH5cC6. Second, although major RAC reform might not alone resolve the backlog, see Br. of Appellant 18, it will go further 16 than HHS lets on. The program was mostly dormant for the better part of two years from 2014 to 2016 while the agency negotiated new RAC contracts. See Maj. Op. 11 (citing, inter alia, U.S. GOV’T ACCOUNTABILITY OFFICE, GAO-16-366, MEDICARE FEE-FOR-SERVICE: OPPORTUNITIES REMAIN TO IMPROVE APPEALS PROCESS 38 n.64 (2016) (GAO Report)). Citing statistics from the lull, HHS misleadingly suggests that RAC appeals are no longer a major contributor to the backlog. See, e.g., Br. of Appellant 18 (“In 2015, . . . RAC appeals comprised just 14.1% of new appeals to OMHA, and in 2016, this figure fell to 9.5% (fewer than 16,000 appeals).”). In 2016, however, HHS itself “reported that it expects the number of incoming appeals to increase again when new [RAC] contracts are awarded and the [RAC] program resumes full operation.” GAO Report 38. As of April 2016, as many as 300,000 appeals from adverse RAC decisions were pending ALJ review. True, that number dipped to about 155,000 by the end of September 2016, before new RAC contracts took effect in October. But even that artificially low number represents a good chunk of the backlog, which supports the supposition—voiced in our earlier decision—that cutting back the discretionary RAC program will go some distance toward statutory compliance. AHA I, 812 F.3d at 185, 193. And if simple contractual reasons are enough justification to put the RAC program on hold, even more so is compliance with a statutory deadline. Id. at 193 (deadline “trump[s]” RAC program and “dictate[s] that [HHS] will have to curtail the RAC program or find some other” method of compliance). Like the majority, Maj. Op. 12, I view HHS’s RAC-related efforts to date as “weak medicine for an agency facing mandamus.” Perhaps HHS should suspend the RAC program altogether until it eliminates the backlog. The district court 17 properly left that decision to the agency. 2016 WL 7076983, at *3. It suffices to say that HHS’s ability to limit the program, see AHA I, 812 F.3d at 186, 193, combined with its ability to negotiate bulk settlements, see 42 C.F.R. §§ 401.613(c)(2), 405.376(d), (h), means the agency has not shown that it cannot lawfully comply with the court’s timetable. 3 B. EVEN IF THE DISTRICT COURT HAD TO MAKE A FINDING, IT DID. Assume arguendo the district court was required to make a finding of possibility. The point of such a requirement, I 3 I recognize that, if not for today’s remand, the timetable’s first deadline would be only a few months away. 2016 WL 7076983, at *3 (mandating 30 per cent reduction of backlog by December 31, 2017). Complying would no doubt be a heavy lift. But much of the difficulty in meeting the first deadline is HHS’s own doing. It offered the district court no alternative to the plaintiffs’ phased reduction targets. Nor does it offer us one: primarily it proposes to give the district court status updates in which it will “attest to its ongoing attention to the backlog.” Br. of Appellant 30; see Oral Arg. Recording 12:15-12:19 (“[W]e certainly couldn’t tie ourselves or commit to a specific date . . . .”). And HHS has fought the mandamus order tooth and nail from the moment it issued in December 2016. Had the agency instead committed immediately to bulk settlements and major RAC reform, it would now be much closer to a 30 per cent reduction. The district court did not have to assume in December 2016 that the agency would use fewer than all viable methods. Because the agency did not make a threshold showing of impossibility when the first deadline was still 13 months away, the court was not required to make an express finding of possibility. See Old Chief v. United States, 519 U.S. 172, 182 n.6 (1997) (“It is important that a reviewing court evaluate the trial court’s decision from its perspective when it had to rule and not indulge in review by hindsight.”). 18 take it, is to ensure that the court does not punish agency officials with contempt until it satisfies itself that they can rightly be blamed for failing to do something they are both obligated and able to do. Maj. Op. 14-15; see Maggio, 333 U.S. at 72; Train, 510 F.2d at 713. For two reasons, any worry about rashly punishing blameless officials is absent here. First, and most importantly, the district court in fact made a finding of possibility. What else could it have meant in expressly rejecting HHS’s claim of impossibility? Specifically, the court rebuffed HHS’s contention that the timetable requires the agency to “‘make payment on Medicare claims regardless of . . . merit’” and therefore “‘conflict[s] with the Medicare statute.’” 2016 WL 7076983, at *3 (quoting Def.’s Opp. 22-23). Granted, other language in the order suggests that the court—apart from concluding that the timetable does not demand reimbursement regardless of merit—did “‘not dive into the parties’ debate’ over the ‘legality and propriety’ of the reforms necessary to comply with the timetable.” Maj. Op. 9 (quoting 2016 WL 7076983, at *3). But because we are to “give the benefit of every doubt to the judgment of the trial judge” in this realm of broad district court discretion, Gasperini, 518 U.S. at 438-39 (internal quotation omitted); see AHA I, 812 F.3d at 193, I can only conclude that the court considered and rejected the agency’s impossibility claim, see Sprint/United Mgmt. Co. v. Mendelsohn, 552 U.S. 379, 386 (2008) (on abuse-of-discretion review, “[a]n appellate court should not presume that a district court intended an incorrect legal result when the order is equally susceptible of a correct reading”). Other aspects of the district court’s decision reinforce the point. My colleagues reaffirm that a district court “contemplating the equities” “may not require an agency,” on pain of contempt, “to render performance that is impossible.” 19 Maj. Op. 14; see Train, 510 F.2d at 713 (it is “unjust” to do so). At the same time, they recognize that the district court here “thoughtfully and scrupulously weighed the equities” and only then “conclud[ed] that the scales tipped in favor of mandamus.” Maj. Op. 2. If impossibility goes to the equities and if a court carefully contemplates the equities before issuing the writ, we have no reason to assume it has erroneously required an impossibility. See Sprint/United Mgmt. Co., 552 U.S. at 386. That goes double where, as here, the district court repeatedly characterizes its own demands as “reasonable.” 2016 WL 7076983, at *3 (expressing “appreciat[ion]” for plaintiffs’ “thoughtful and reasonable four-year plan”); id. (noting timetable requires HHS “to undertake proper claim substantiation within a reasonable timeframe” (internal quotation omitted)). Second, even if HHS violates the timetable, the order does not make contempt an automatic consequence. At that point, rather, the “[p]laintiffs may move for default judgment or to otherwise enforce the writ of mandamus.” 2016 WL 7076983, at *3 (citing FED. R. CIV. P. 55(d)). The order thus includes a failsafe for revisiting impossibility as developments dictate. That is precisely how the writ is supposed to work: “The court’s injunction should serve like adrenalin, to heighten the response and to stimulate the fullest use of resources. This may run the risk of overstimulating the organism, but palliative measures may be taken . . . if indicated at a later date.” Train, 510 F.2d at 712. Under the order here, if HHS were truly unable to comply despite deploying its “fullest . . . resources”—which it has not yet done—the court in its discretion could undertake “palliative measures” well short of contempt. Id. Again, we have no reason to assume the court would abuse its discretion. To the contrary, its prudence to date should give us the strongest confidence. 20 C. EVEN IF THE DISTRICT COURT FAILED TO MAKE A REQUIRED FINDING, VACATUR AND REMAND ARE UNWARRANTED. In my estimation, even if the majority were right to find error, its remedy would nonetheless be wrong. Because we review a district court’s judgment, not its rationale, we can sometimes “sustain a ‘right-result, wrong-reason’ decision.” People’s Mojahedin Org. of Iran v. Dep’t of State, 182 F.3d 17, 23 n.7 (D.C. Cir. 1999). In some cases, for example, “[t]here is . . . no need for remand if an intelligent review of the record can be made.” Am. Emp’rs Ins. Co. v. Am. Sec. Bank, 747 F.2d 1493, 1498 (D.C. Cir. 1984). This case strikes me as just such a case. My colleagues acknowledge that HHS will at some point have the “heavy burden” of proving impossibility. Maj. Op. 15 (quoting Ala. Power, 636 F.2d at 359). I see no reason to postpone the inevitable. For all the record-based reasons already discussed, we can say here and now that the agency has not met its burden. We should do so. Much like the district court, we are to consider the equities in arriving at our disposition. 28 U.S.C. § 2106 (appellate court may affirm or, inter alia, vacate and remand with an eye toward what is “just under the circumstances”). Today’s remand gets the equities backwards: it punishes providers with further delay and rewards an obdurate agency. It cannot be the result we had in mind when we suggested that “the clarity of the statutory duty likely will require issuance of the writ” absent “meaningful progress” the political branches still have not made. AHA I, 812 F.3d at 193. ***** This case is not about the difference between ought and cannot. It is about the difference between shall and will not. 21 The district court correctly rejected the agency’s assertion of impossibility. Accordingly, I respectfully dissent. 4 4 In response, the majority repeats with string parentheticals that (1) the district court cannot, by mandamus, demand the impossible, Maj. Op. 16; and (2) HHS asserted that compliance with the court’s timetable is impossible, Maj. Op. 17 n.3. I do not dispute either component of that analysis. But an assertion is not a “demonstrat[ion].” Train, 510 F.2d at 712-13. And by no means do I suggest that, if an agency in fact makes the requisite threshold showing, it is “hypertechnical” to require a ruling on impossibility. Contra Maj. Op. 16. My point is that HHS did not make a threshold showing and, in any event, the court did rule on impossibility. The needless technicality here is requiring the court to be any more specific than it was.