Opinion for the Court filed by Circuit Judge TATEL.
TATEL, Circuit Judge:
At heart, this case is about an agency caught between two congressionally assigned tasks. Congress has prescribed specific time frames for the Secretary of Health and Human Services to reach decisions on various stages of administrative appeals of Medicare reimbursement claim denials. But Congress has also directed the Secretary to implement the Medicare Recovery Audit Program to detect waste, fraud, and abuse. Although the audit program has recovered billions of dollars in fraudulently or otherwise improperly paid funds, it has also contributed significantly to a volume of appeals that makes compliance with the statutory time frames impossible. Plaintiffs, including several hospitals with a significant amount.of money tied up in the appeals process for far longer than the statute contemplates, seek a writ of mandamus compelling the Secretary to act within those time frames. Although Plaintiffs disclaim any desire or authority to force the Secretary to curtail the audit program or take any other particular action to meet thfe deadlines, the record suggests that absent further congressional action, the Secretary would likely have to drastically curtail that program to comply with such an order. The district court concluded that mandamus relief was unwarranted, noting the political branches’ ongoing efforts to resolve this tension and the audit program’s success in detecting improper payments. For the reasons set forth in this opinion, we reverse and remand with instructions to the district court to consider the problem as it now stands— worse, not better.
I.
After a hospital or other health-care provider performs Medicare-eligible services, it submits a claim for reimbursement to a Medicare Administrative Contractor (MAC). 42 U.S.C. §§ 1395ff(a)(l)-(2), 1395kk-l(a); 42 C.F.R. §§ 405.904(a)(2), 405.920-405.928. The MAC decides whether to pay or deny the claim. If a claim is denied, the Medicare Act provides a four-level administrative appeal process, followed by judicial review. At the first level, the health care provider presents its claim again to the MAC for “redetermination.” 42 U.S.C. § 1395ff(a)(3)(A), (a)(3)(C)(ii). The second level involves “reconsideration” by a Qualified Independent Contractor (QIC).
Id.
§ 1395ff(c). The Centers for Medicare and Medicaid Services (CMS) oversees initial determinations and redeterminations by the MACs, as well as reconsiderations by the QICs.
If the provider remains unsatisfied, and if its claim exceeds $150, it may continue to the third stage: de novo review by an administrative law judge, including a hearing.
Id.
§ 1395ff(b)(l)(E)(i)„ (b)(l)(E)(iii), (d)(1)(A); 42 C.F.R. § 405.1006(b); 80 Fed.Reg. 57,827, 57,827 (2015). This stage of the process is overseen by the Office of Medicare Hearings and Appeals (OMHA), which houses ALJs and their support staff, and which is funded by a separate appro
priation.
See
Medicare Prescription Drug, Improvement, and Modernization Act of 2003, Pub.L. No. 108-173, § 931, 117 Stat. 2066 (requiring the Secretary to create an “administrative office that is organizationally and functionally separate from [CMS]” to “assure the independence of administrative law judges”). The fourth and final administrative stage involves de novo review by the Medicare Appeals Council, a division of the Departmental Appeals Board (DAB). Although the DAB has authority to hold a hearing, it does so only if “there is an extraordinary question of law/policy/fact.” Mot. for Summ. J. Ex. 2, at 118. Finally, after completing the administrative appeal process, providers may seek review in district court of claim denials worth at least $1,500. 42 U.S.C. § 1395ff(b)(l)(E)(i), (b)(l)(E)(iii); 42 C.F.R. § 405.1006(c); 80 Fed.Reg. at 57,-827. We apologize to our readers for all of the acronyms, but this is, after all, a Medicare case, and acronyms seem integral to the parties’ native language.
To prevent appeals from lingering unresolved, the statute includes specific time frames for each step of the process. In particular, redetermination by the MACs “shall be concluded” within sixty days, 42 U.S.C. § 1395ff(a)(3)(C)(ii), and, with exceptions not relevant here, QICs “shall conduct and conclude” reconsiderations within sixty days,
id.
§ 1395ff(c)(3)(C)(i). Similarly, ALJs “shall conduct and conclude a hearing ... and render a decision” within ninety days,
id.
§ 1395ff(d)(l)(A), although the appealing provider may “waive” this “deadline,”
id.
§ 1395ff(d)(l)(B). And finally, the DAB “shall conduct and conclude a review ... and make a decision or remand the case to the administrative law judge for reconsideration” within ninety days.
Id.
§ 1395ff(d)(2)(A). If all these time periods are met, appeals will work their way through the administrative process within about a year.
The statute also prescribes “consequences of failure to meet” several of the statutory “deadlines.” In a process commonly referred to as “escalation,” a provider that has been waiting for longer than the statutory time limit may advance its appeal to the next stage. Thus, a provider may “escalate” its appeal to the ALJ stage if the QIC fails to act within the required sixty days,
id.
§ 1395ff(c)(3)(C)(ii), to the DAB stage if the ALJ fails to act within the required ninety days,
id.
§ 1395ff(d)(3)(A), and to district court review if the DAB fails to act within the required ninety days,
id.
§ 1395ff(d)(3)(B).
For years, the administrative appeal process functioned largely as anticipated, with its various stages typically completed within the statutory time frames.
American Hospital Ass’n v. Burwell,
76 F.Supp.3d 43, 46 (D.D.C.2014). Then, in 2010, the Secretary fully implemented the Medicare Recovery Audit Program, which Congress had required the Secretary to set up “for the purpose of identifying underpayments and overpayments and recouping overpayments.” 42 U.S.C. § 1395ddd(h)(l). Specifically, Congress directed that the Secretary “shall enter into contracts with recovery audit contractors” (RACs), who must be paid “on a contingent basis for collecting overpay-ments” and “in such amounts as the Secretary may specify for identifying underpayments.”
Id.
§ 1395ddd(h)(l)(B). Although Congress also specified certain other features of the program, such as that it must have “[n]ationwide coverage,”
id.
§ 1395ddd(h)(3), it left the Secretary broad discretion to determine many other program details.
The RAC program has had two primary effects. First, the government has recovered a great deal of improperly paid mon
ey. According to the Secretary, “[i]n 2012, the program identified $2.3 billion in over-payments, and in fiscal year 2013, the recovery auditors identified and corrected $3.65 billion in overpayments.” Appellee’s Br. 8-9 (footnotes omitted). In 2012, the Secretary adds, “only 7% of claims identified by audit contractors as overpayments were challenged and overturned on appeal,” and only 9.3% were in 2013.
Id.
at 9.
But because RAC denials are appealable through the same administrative process as initial denials, the RAC program has contributed to a drastic increase in the number of administrative appeals. Thus, the number of appeals filed ballooned from 59,600 in fiscal year 2011 to more than 384,000 in fiscal year 2013. Mot. for Summ. J. Ex. 7, at 4. Although the Secretary explains that other factors, such as “increased utilization of Medicare-covered services,” have played a role in increasing the number of appeals filed, Appellee’s Br. 9, the government acknowledged at oral argument that 46% of the appeals currently pending before OMHA originated from the RAC program. Oral Arg. Tr. 35.
Between RAC and non-RAC appeals, OMHA currently receives many more cases than it can process in a timely fashion. Indeed, every two months or less, it receives as many appeals as it can process in a full year. Appellants’ Br. 12. As of February 2015, the decisions ALJs were releasing had been pending for an average of 572 days. Appellee’s Br. 10. This number will almost certainly continue to grow as the backlog worsens.
The Secretary has worked to address the backlog and corresponding delays. As a result of various reforms, the number of appeals the average ALJ resolves each year has more than doubled since 2009. Mot. for Summ. J. Ex. 7, at 4. Moreover, the agency secured funding for seven additional ALJs and associated staff in fiscal year 2014 — an increase of about 10% over previous staffing levels.
Id.
at 5.
Despite these additional resources and significant improvements, the Secretary and OMHA find themselves in an untenable position. OMHA still has the capacity to process only about 72,000 appeals per year, a far cry from the almost 400,000 appeals it received in fiscal year 2013, or from the over 800,000 appeals that composed its backlog in July 2014.
Id.
These figures suggest that at current rates, some already-filed claims could take a decade or more to resolve. Bowing to this reality, in December 2013, OMHA’s Chief ALJ sent a memorandum informing various hospitals that OMHA had temporarily suspended assigning appeals to ALJ dockets, that the suspension would last “at least 24 months,” and that the agency “expect[ed] post-assignment hearing wait times [would] continue to exceed 6 months.” Mot. for Summ. J. Ex. 3, at 1. The DAB stage is also plagued by delays, although not quite to the same degree.
E.g.,
Mot. for Summ. J. Ex. 7, at 107, 111.
Congress is fully aware of both the backlog and its connection to the RAC program. The Senate Finance Committee has held multiple hearings on the issue, dating back to at least July 2013. Indeed, although at a 2015 hearing, Senator Orrin Hatch, the committee chairman, expressed concern over the lengthy delays, he recognized that OMHA “has also taken steps to address its backlog, but there is only so much the agency can do with their current authorities and staffing.”
See
Hatch Statement at Finance Hearing on Medicare Audit and Appeals (Apr. 28, 2015), http://www.finanee.senate.gov/chairmans-news/hatch-statement-at-finance-hearing-on-medieare-audit-and-appeals.
Moreover, the Senate is considering a bill known as the “AFIRM Act,” which
would provide $125,000,000 in additional annual funding for OMHA, as well as make other reforms to the appeal process designed to address the backlog. AFIRM Act, S. 2368, 114th Cong. (2015). If enacted, this legislation might go some way toward resolving the problems. As of yet, however, the bill remains only a bill, and the delays continue.
If the vast majority of these delayed appeals were ultimately denied, they might amount to little more than an unfortunate nuisance. The record suggests, however, that many have merit. Hospitals responding to a survey conducted in 2014 by one of the plaintiffs in this case, the American Hospital Association, reported that they had appealed 52% of RAC denials, and that 66% of these appeals that had been completed were successful. Mot. for Summ. J. Ex. 5, at 55. The Secretary quibbles with the details of this statistic— and we acknowledge the obvious self-selection and bias problems — but even government counsel conceded at oral argument that 43% of ALJ appeals (including from RAC and non-RAC denials) succeed. Oral Arg. Tr. 37. This reversal rate is hardly negligible.
The delays at the ALJ stage are especially harmful to hospitals because HHS recoups funds after the QIC stage. 42 U.S.C.- § 1395ddd(f)(2)(A). Given hospitals’ frequent success at the ALJ level, this means that they are often deprived of access to significant funds to which they are entitled. This problem takes a particular toll on hospitals with a large share of patients who rely on Medicare.
Plaintiffs in this case, three such hospitals or hospital systems and the American Hospital Association (collectively, the “Association”), filed suit in United States district court seeking relief in the nature of mandamus under 28 U.S.C. § 1361 (which, for ease of reference, we refer to simply as “mandamus”) to “compel the Secretary ... to meet the statutory deadlines for administrative review of denials of claims for Medicare reimbursement.” Compl. at 1. The three hospitals or hospital systems are (1) Baxter Regional Medical Center, a 268-bed regional hospital in Arkansas that derives 65% of its gross revenue from Medicare, Holleman Decl. ¶¶ 5-7; (2) Covenant Health, a community-owned health system of hospitals in Tennessee that derives 55% of its gross revenue from Medicare, Geppi Deck ¶¶ 5, 8; and (3) Rutland Regional Medical Center, a community-owned 188-bed hospital in Vermont that derives 47% of its revenues from Medicare, Wallace Deck ¶¶ 6, 10. All three allege that they have significant funds tied up in the Medicare appeals process — including in appeals that have already exceeded the statutory time frames — and that their inability to access these funds makes a number .of essential activities, such as replacing ICU beds, difficult or impossible. All three report that the inability to access the money makes it more difficult to provide adequate care, and at least Baxter and Covenant say they may stop offering certain services if the system is not fixed!
The Association sought summary judgment in the district court, and the Secretary moved to dismiss for lack of jurisdiction. The district court concluded that the jurisdictional and merits questions merged, and thus resolved both motions at once. The district court denied the Association’s motion for summary judgment and granted the Secretary’s motion to dismiss for lack of jurisdiction, concluding that the agency’s delay was not so unreasonable as to justify mandamus. In doing so, the district court concluded as follows: “The Court hopes that the Secretary and Congress will continue working together toward a solution and that OMHA will
receive the. resources necessary to fulfill its obligations. Hospitals that are owed reimbursement should not be indefinitely deprived of funds. The Court cannot predict whether, over time, if HHS and Congress cannot adequately address the overflow of appeals, the [analysis] might shift toward Plaintiffs.”
American Hospital Ass’n,
76 F.Supp.3d at 56.
This appeal followed.
II.
“The remedy of mandamus, is a drastic one, to be invoked only in extraordinary circumstances.”
Power v. Barnhart,
292 F.3d 781, 784 (D.C.Cir.2002) (internal quotation marks omitted). To show entitlement to mandamus, plaintiffs must demonstrate (1) a clear and indisputable right to relief, (2) that the government agency or official is violating a clear duty to act, and (3) that no adequate alternative remedy exists.
United States v. Monzel,
641 F.3d 528, 534 (D.C.Cir.2011). These three threshold requirements are jurisdictional; unless all are met, a court must dismiss the case for lack of jurisdiction.
See In re Medicare Reimbursement Litigation,
414 F.3d 7, 10 (D.C.Cir.2005) (internal quotation marks and alteration omitted). “Even when the legal requirements for mandamus jurisdiction have been satisfied, however, a court may grant relief only when it finds compelling equitable grounds.”
Id.
“The party seeking mandamus has the burden of showing that its right to issuance of the writ is clear and indisputable.”-
Power,
292 F.3d at 784 (internal quotation marks omitted).
Mandamus claims that, like this one, target agency delay, turn on “whether the agency’s delay is so egregious as to warrant mandamus.”
In re Core Communications, Inc.,
531 F.3d 849, 855 (D.C.Cir.2008) (internal quotation marks omitted). In such cases, courts are guided by the
“TRAC
factors,” so named because they come from our decision in
Telecommunications Research & Action Center v. FCC.
These factors are as follows:
(1) the time agencies take to make decisions must be governed by a rule of reason; (2) where Congress has provided a timetable or other indication of the speed with which it expects the agency to proceed in the enabling statute, that statutory scheme may supply content for this rule of reason; (3) delays that might be reasonable in the sphere of economic regulation are less tolerable when human health and welfare are at stake; (4) the court should consider the effect of expediting delayed action on agency activities of a higher or competing priority; (5) the court should also take into account the nature and extent of the interests prejudiced by delay; and (6) the court need not find any impropriety lurking behind agency lassitude in order to hold that agency action is unreasonably delayed.
Telecommunications Research & Action Center v. FCC (TRAC),
750 F.2d 70, 80 (D.C.Cir.1984) (internal quotation marks and citations omitted). Although these factors provide guidance by setting out “the hexagonal contours of a standard,” we have been careful to emphasize that they are “hardly ironclad,”
id.,
and that “[e]aeh case must be analyzed according to its own unique circumstances,”
Air Line Pilots Ass’n v. Civil Aeronautics Board,
750 F.2d 81, 86 (D.C.Cir.1984).
We have never squarely addressed the interplay of the three threshold mandamus requirements — clear duty, clear. right to relief, and absence of an adequate alternative remedy — and the six
TRAC
factors. Because these factors function not as a hard and fast set of required elements, but rather as useful guidance as to whether a delay is “so egregious as to warrant man
damus,”
TRAC,
750 F.2d at 79, their roles may differ depending on the circumstances. For example, in situations where plaintiffs allege that agency delay is unreasonable despite the absence of a specific statutory deadline, the entire
TRAC
factor analysis may go to the threshold jurisdictional question: does the agency’s delay violate a clear duty? By contrast, in situations where the statute imposes a deadline or other clear duty to act, the bulk of the
TRAC
factor analysis may go to the equitable question of whether mandamus
should
issue, rather than the jurisdictional question of whether it
could.
Here, the district court recognized the unsettled relationship between jurisdictional and merits questions in mandamus suits.
American Hospital Ass’n,
76 F.Supp.3d at 49-50. Guided in part by our precedent outside of the agency-delay context, the district court concluded that the jurisdictional and merits inquiries “merge[d],” and that “the dual nature of the inquiry” allowed it to “resolve Plaintiffs’ Motion for Summary Judgment together with Defendant’s Motion to Dismiss for Lack of Jurisdiction.”
Id.
at 50 (internal quotation marks omitted). Accordingly, after analyzing the
TRAC
factors, the district court denied Plaintiffs’ summary judgment motion and granted the Secretary’s motion to dismiss for lack of jurisdiction.
In our view, however, the distinction between the jurisdictional inquiry and the equitable merits inquiry matters, especially because it affects our standard of review. We review the threshold requirements for mandamus jurisdiction de novo.
In re Medicare Reimbursement Litigation,
414 F.3d at 10. But we review “the equities” for abuse of discretion.
Id.
Accordingly, we first consider the threshold jurisdictional question, and then turn to the equities.
A.
At the outset, we must determine whether, as the Association argues, the statutory time frames are mandatory deadlines. According to the Association, the statute imposes mandatory duties by providing that certain actions “shall” occur within specified time frames. As the Secretary sees it, however, the opportunity for providers to escalate appeals deprives the district court of jurisdiction to issue mandamus, either because escalation demonstrates the lack of a statutory duty or because it provides an adequate alternative remedy. In the unique circumstances of this case, we agree with the Association.
To begin with, as to clear duty, the statute uses the typically mandatory “shall.”
E.g.,
42 U.S.C. § 1395ff(d)(l)(A) (“[A]n administrative law judge
shall
conduct and conclude a hearing ... and render a decision on such hearing” within ninety days, (emphasis added)). To be sure, as the Secretary points out, context can dictate that “shall” take a directory rather than a mandatory meaning. But here, context only reinforces a mandatory reading. The statute itself repeatedly refers to the time frames as “deadlines.”
E.g., id.
§ 1395ff(d)(l)(B). And the provision permitting “[w]aiver of deadline by party seeking hearing,”
id.,
would lack meaning if the agency had no obligation to comply with the deadline in the first place.
The Secretary argues that by permitting escalation, Congress acknowledged that the time frames would sometimes remain unmet, thus suggesting that Congress did not view them as mandatory. Appellee’s Br. 19-20. The Secretary’s premise fails to support her conclusion. Merely providing a consequence for noncompliance does not necessarily undermine the force of a command.
The argument that escalation provides an adequate alternative remedy is somewhat stronger. If delays occurred only in isolated or occasional cases, escalation might suffice. Indeed, we agree with the Secretary that Congress’s inclusion of the remedy in the statutory scheme indicates that Congress anticipated that violations might occur with some measure of regularity. That said, nothing suggests that Congress intended escalation to serve as an adequate or exclusive remedy where, as here, a systemic failure causes virtually all appeals to be decided well after the statutory deadlines.
In some circumstances, of course, distinguishing between violations that escalation can adequately address and those it cannot might be difficult. The systemic failure at issue in this case, however, presents no such line-drawing dilemma. Escalation from the ALJ stage to the DAB stage is unlikely to provide a timely hearing. Not only does the DAB itself have a backlog, but it holds hearings only where an “extraordinary question” is involved.
Nor does further escalation to district court suffice. As the government acknowledged at oral argument, district court review would be deferential, Oral Arg. Tr. 28, hardly an adequate substitute for a de novo hearing before an administrative law judge.
Alternatively, the Secretary argues that she has no clear duty because the action the Association seeks mandamus to compel is discretionary rather than ministerial. As the Association points out, however, although both the content of the administrative appeal decisions and the means the Secretary uses to ensure -they are reached in a timely fashion are discretionary, the Association formally seeks to compel neither. Rather, it simply seeks to compel the Secretary to make decisions within the statutory time frames. Appellants’ Reply Br. 6. The Secretary insists that the Association’s request constitutes a “programmatic attack” on the way her department manages its resources, that the department lacks the resources to render decisions within the statutory time frames, and that even if it had the necessary resources, we should hesitate to reorder agency priorities in such a manner. The Supreme Court, however, has distinguished impermissible “programmatic attack[s]” from “the failure to ... take some decision by a statutory deadline,”
Norton v. Southern Utah Wilderness Alliance,
542 U.S. 55, 63-64, 124 S.Ct. 2373, 159 L.Ed.2d 137 (2004), the very failure the Association challenges here.
In making her “programmatic attack” argument, the Secretary emphasizes that many agency delay cases involve one or a small number of decisions, rather than the countless administrative appeals at issue in this case. Appellee’s Br. 27. In the presence of a clear statutory deadline, however, the scope of the program involved goes to “the equities” of granting mandamus rather than to the threshold jurisdictional question of whether a clear duty exists. Indeed, this court has made clear — albeit in a case denying mandamus relief — that “[hjowever many priorities the agency may have, and however modest its personnel and budgetary 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.
In re United Mine Workers of America International Union,
190 F.3d 545, 554 (D.C.Cir.1999). And in another case, we strongly suggested that we would have granted mandamus to require an agency to make over a dozen delayed decisions had the agency not convinced us that it had resolved the underlying issues and was quickly working through its backlog.
In re American Fed
eration of Government Employees, AFL-CIO,
790 F.2d 116 (D.C.Cir.1986).
Finally, our decisions in
In re Barr Laboratories, Inc.,
930 F.2d 72 (D.C.Cir.1991), and
Mashpee Wampanoag Tribal Council, Inc. v. Norton,
336 F.3d 1094 (D.C.Cir.2003) — both relied on by the Secretary— are not to the contrary. In these cases, we rejected mandamus claims that would have had the effect of allowing the plaintiffs to jump the line, functionally solving their delay problem at the expense of other similarly situated applicants. To be sure, the complaint in this case does seek this type of relief,
see
Compl. Prayer for Relief (b)(i), (b)(ii), which we agree our precedent forecloses. But the complaint also requests the broader relief of “requiring HHS to otherwise comply with its statutory obligations in administering the appeals process for all hospitals.”
Id.
(b)(iii). The line-jumping cases of
Barr Labs
and
Mashpee Wampanoag Tribal Council
neither speak to nor preclude such relief.
We thus conclude that the statute imposes a clear duty on the Secretary to comply with the statutory deadlines, that the statute gives the Association a corresponding right to demand that compliance, and that escalation — the only proposed alternative remedy — is inadequate in the circumstances of this case. Because the Association has demonstrated that the threshold requirements for mandamus jurisdiction are met, and because the Secretary’s other jurisdictional arguments fail, we reverse the district court’s dismissal for lack of jurisdiction.
B.
On remand, the district court should determine whether “compelling equitable grounds” now exist to issue a writ of mandamus. The court appears to have considered this question as it stood in late 2014 as part of its merged jurisdictional and merits inquiry, but the record on appeal makes clear that the situation has worsened — something the district court will need to account for as it applies the
TRAC
factors. Although the difficult decision of when to issue the extraordinary writ rests in the first instance with the district court, given the large number of federal agencies within our jurisdiction and the importance of ensuring the application of uniform mandamus standards, we think it helpful to set out the factors that weigh most strongly for and against mandamus in this case.
Perhaps counseling most heavily against mandamus is the writ’s extraordinary and intrusive nature, which risks infringing on the authority and discretion of the executive branch. These risks are especially salient here because mandamus would, in effect, probably require the agency to make major changes to its operations and priorities, including drastically limiting the scope of a statutorily mandated program that has recovered billions of dollars in incorrectly paid funds. Moreover, as the district court properly noted, Congress’s awareness of and attention to the situation counsel against issuance of the writ.
American Hospital Ass’n,
76 F.Supp.3d at 56. So too, we think, does the fact that Congress has provided escalation as a remedy, even though that remedy may offer less than full relief. Escalation might inform the district court’s analysis of whether the delay is egregious enough to warrant the grant of mandamus relief, even if it does not preclude mandamus jurisdiction altogether. Finally, the district court also correctly concluded that the Secretary’s good faith efforts to reduce the delays within the constraints she faces — such as .by implementing reforms that have doubled ALJ efficiency — push in the same direction.
Id.
at 55-56. The backlog and delays have their origin in the
political branches, and ideally the political branches should resolve them.
On the other hand, several significant factors counsel in favor of mandamus. To begin with, the record demonstrates that the delays are having a real impact on “human health and welfare.”
TRAC,
750 F.2d at 80. For example, one plaintiff, Baxter Regional Medical Center, submitted a declaration explaining that having money tied up in the appeals process beyond the statutory deadlines makes it much more difficult to purchase replace-' ment ICU beds, replace (rather than patch) a roof over its surgery department, and replace a twenty-year-old catheterization lab which will “soon need to be shut down.” Holleman Decl. ¶¶ 14-15. Likewise, amicus the Fund for Access to Inpatient Rehabilitation reports that the delays have led at least one rehabilitation hospital to “avoid admitting certain types of patients, regardless of whether its staff believes the patients meet the coverage criteria for rehabilitation hospital care, if those patients have indicia within their medical records that are likely to trigger an audit.” Amicus Br. 28. These consequences— none of which the government challenges — are unsurprising; common sense suggests that lengthy payment delays will affect hospitals’ willingness and ability to provide care.
Moreover, and critically to our thinking about this case, although Congress directed the Secretary to establish the RAC program, it has left her with substantial discretion to implement it and determine its scope. 42 U.S.C. § 1395ddd(h). True, Congress seems to approve of the way the Secretary has implemented the program, and the agency is entitled to some leeway to resolve the tension between competing priorities. If it fails to do so, however, and if Congress fails to act, either by providing the Secretary sufficient resources to comply with the clear statutory deadlines it has already enacted or by relieving her of the obligation to do so, these deadlines dictate that the Secretary will have to curtail the RAC program or find some other way to meet them. Federal agencies must obey the law, and congressionally imposed mandates and prohibitions trump discretionary decisions.
All that said, we reiterate that the district court has broad discretion in weighing the equities and deciding “whether the agency’s delay is so egregious as to warrant mandamus.” Taking the above factors into account, the district court — more than a year after its first denial and with the problem only worsening — might find it appropriate to issue a writ of mandamus ordering the Secretary to cure the systemic failure to comply with the deadlines. On the other hand, if the. district court determines on remand that Congress and the Secretary are making significant progress toward a solution, it might conclude that issuing the writ is premature. If so, it could consider such action as ordering the agency to submit status reports updating the court on the level of appropriations, the progress of the AFIRM Act, and ány other relevant information.
In the end, although courts must respect the political branches and hesitate to intrude on their resolution of conflicting priorities, our ultimate obligation is to enforce the law as Congress has written it. Given this, and given the unique circumstances of this case, the clarity of the statutory duty 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.
Cf. In re Aiken County,
725 F.3d 255, 258-59 (D.C.Cir.2013) (granting mandamus after Congress failed to take advantage of a previous order holding the case in abeyance to give
Congress the chance to “clarify” potentially conflicting signals on whether it wanted the Nuclear Regulatory Commission to expend appropriated funds on activities related to storing nuclear waste at Yucca Mountain).
III.
We reverse the district court’s dismissal for lack of jurisdiction and remand for further proceedings consistent with this opinion.
So ordered.