Gresham v. AzarGresham v. Azar
- Reporters:
- Before:
- Jeb
Adrian McGonigal is 40 years old and lives with his brother in Pea Ridge, Arkansas. He used to have a job working in the shipping department of Southwest Poultry, a food-service company located nearby, although he received no medical insurance through his employer. Like many Americans, he has several serious medical conditions. Beginning in 2014, McGonigal was able to receive medical care — including regular doctor visits and numerous prescription drugs — through the state‘s expanded Medicaid program. In mid-2018, however, McGonigal learned that he would be subject to new work requirements, which he would have to report online, as a condition of receiving health benefits. These were imposed by the Arkansas Works Amendments (AWA), approved by the U.S. Secretary of Health and Human Services in March 2018. Despite his lack of access to, and difficulty working with, computers, he was able to report his employment in June 2018, but he did not know he needed to continue to do so each month. As a result, when he went to pick up his prescriptions in October, the pharmacist told him that he was no longer covered, and his medicines would cost him $800. In the absence of Medicaid, he could not afford the cost of the prescriptions and so did not pick them up. His health conditions then flared up, causing him to miss several days of work, and Southwest Poultry fired him for his absences. He thus lost his Medicaid coverage and his job.
Anna Book is 38 years old and lives in Little Rock. She currently rents a room in an apartment but was homeless for most of the last eight years. In July 2018, she got a job as a dishwasher in a restaurant, for which she works about 24 hours each week. Before that, she was unemployed for two years. She nevertheless also had health care provided through Arkansas‘s Medicaid program, which a local pastor helped her sign up for in 2014. Book learned last August that, pursuant to AWA, she would have to report 80 hours each month of employment or other activities to keep that coverage. While she reported her compliance in August and September with the pastor‘s help — she does not have reliable internet access — Book has several health conditions and worries that she will not maintain sufficient hours at her job to keep her coverage.
Russell Cook is 26 and also lives in Little Rock. He is currently homeless. While he has spent time working as a landscaper, he is not presently employed and has minimal job prospects. The state‘s Medicaid program has previously given him access to health care for various health conditions, including a torn Achilles tendon and serious dental problems. Cook, however, does not believe he will be able to comply with the new AWA work requirements, which began applying to him in January 2019. Lacking access to the internet or a phone, he also worries that he will be unable to report compliance with those requirements. He thus expects to lose his Medicaid coverage.
These are three of the ten Arkansans who come to this Court seeking to undo the work requirements the state added in 2018 to its Medicaid program. They sued the Secretary of Health and Human Services in August 2018, arguing that the
Plaintiffs’ suit does not offer an issue of first impression. Indeed, this Court just last summer considered a challenge to the Secretary‘s approval of very similar changes to Kentucky‘s Medicaid program — including work or “community engagement” requirements — in Stewart v. Azar, 313 F. Supp. 3d 237 (D.D.C. 2018) (Stewart I). There, it vacated the agency‘s decision because it had not adequately considered whether the program “would in fact help the state furnish medical assistance to its citizens, a central objective of Medicaid.” Id. at 243. Plaintiffs point to the identical deficiency in the record in this case. Despite the protestations in its (and intervenor Arkansas‘s) briefing, HHS conceded at oral argument that the administrative decision in this case shares the same problem as the one in Stewart I. See Oral Argument Transcript at 6–7. The Court‘s job is thus easy in one respect: the Secretary‘s approval cannot stand.
Yet a separate question remains: what is the proper remedy? In Stewart I, the Court vacated the approval and remanded to the Secretary. Here, however, the Government argues that vacatur is improper both because, unlike Kentucky, AWA is already active and halting it would be quite disruptive, and because any error is easily fixed, just as it has been for Kentucky. The challengers disagree, positing that the deficiency in the approval is substantial and that any resulting disruption is outweighed by the ongoing harms suffered by the more than 16,000 Arkansans who have lost their Medicaid coverage. Given the seriousness of the deficiencies — which, as this Court explains in a separate Opinion issued today, the remand in Kentucky did not cure — and the absence of lasting harms to the Government relative to the significant ones suffered by Arkansans like Plaintiffs, the Court will vacate the Secretary‘s approval and remand for further proceedings.
I. BACKGROUND
As it did in Stewart I, the Court begins with an overview of the relevant history and provisions of the Medicaid Act. See 313 F. Supp 3d. at 243–44. It then turns to Arkansas‘s challenged plan before concluding with the procedural history of this case.
A. Legal Background
1. The Medicaid Act
Since 1965, the federal government and the states have worked together to provide medical assistance to certain vulnerable populations under Title XIX of the Social Security Act, commonly known as Medicaid. See
To be approved, state plans must comply with certain minimum parameters set out in the Medicaid Act. See
Generally, a state must cover all qualified individuals or forfeit its federal Medicaid funding.
The Medicaid Act, in addition to defining who is entitled to coverage, also ensures what coverage those enrolled individuals receive. Under
2. Section 1115 of Social Security Act
Both before and after the passage of the ACA, a state is not entirely locked in; instead, if it wishes to deviate from the Medicaid Act‘s requirements, it can seek a waiver from the Secretary of HHS. See
While the ultimate decision whether to grant § 1115 approval rests with the Secretary, his discretion is not boundless. Before HHS can act on a waiver application, the state “must provide at least a 30-day public notice[-]and[-]comment period” regarding the proposed program and hold at least two hearings at least 20 days before submitting the application. See
B. Factual Background
1. Arkansas Works Amendments
Arkansas‘s Medicaid program dates back to 1970. For most of the program‘s history, the state maintained among the most stringent eligibility thresholds in the nation for adults, covering only the aged, disabled, and parents with very low incomes. See ECF 53-6, Exh. 54 (Ark. Health Care Independence Program Interim Report) at 16. That changed with the passage of the ACA. While states had a choice after NFIB not to expand Medicaid, Arkansas was one of those that opted to do so. Under its expansion program, which began January 1, 2014, Medicaid-eligible persons were given the opportunity to enroll in private insurance plans financed by the state. See AR 71. In its first two years, the program provided health coverage to more than 278,000 newly eligible individuals, helping to lower the uninsured rate from 19% to 11%. See AR 1274. The program became known as Arkansas Works in January 2017.
That month featured another significant change in the political landscape, as the Trump administration took over from President Obama. In March 2017, then-Secretary Thomas Price and CMS Director Seema Verma sent a letter to all 50 governors announcing the administration‘s view that the ACA‘s expansion of Medicaid was “a clear departure from the core, historical mission of the program.” See AR 85. They thus alerted states of the agency‘s “intent to use existing Section 1115 demonstration authority” to help revamp Medicaid. See AR 86. Together they promised to find “a solution that best uses taxpayer dollars to serve” those individuals they deemed “truly vulnerable.” Id. Heeding HHS‘s call, Governor Asa Hutchinson proposed three substantial amendments to Arkansas Works under Section 1115. See AR 2057. First, he proposed to shift income eligibility for the expansion population from 133% to 100% of the Federal Poverty Line. Id. Second, he proposed to “institute work requirements as a condition” of continued Medicaid coverage. Id. Third, he proposed to eliminate retroactive health coverage. Id. The state did not estimate the effects these amendments would have on Medicaid coverage. CMS held a public-comment period from July 11 to August 10, 2017, and numerous organizations offered their views and analysis of the changes.
According to Arkansas‘s Department of Human Services, only a small percentage of the persons required to report compliance with the work requirements actually did so during the first six months of the program. In October, for example, only 12.3% (1687 out of 13653) of persons not exempt from the requirements reported any kind of qualifying activity. See ECF No. 42-1 (Arkansas Works Reports June–November 2018) at 47, 52. Since the program began, more than 16,900 individuals have lost Medicaid coverage for some period of time for not reporting their compliance. Id. at 18, 27, 36, 45. It is not known what percentage of these individuals completed the work requirements but did not report versus those who did not engage in the work itself.
2. Kentucky HEALTH
Arkansas was not the only state interested in the new administration‘s proposal to rethink the Medicaid Expansion. The Commonwealth of Kentucky proposed a demonstration project called Kentucky HEALTH — with similar community-engagement requirements and cutbacks to retroactive coverage. (It also contained other elements not relevant here.) Kentucky, unlike Arkansas, did estimate the coverage effects of its project, explaining that thousands of persons would lose their Medicaid benefits over the course of the project; indeed, their estimate corresponded to about 95,000 persons losing Medicaid for one full year. As it did in Arkansas, the Secretary approved that project on the ground that it was likely to “improv[e] health outcomes” and “increas[e] individual engagement in health care decisions.” Stewart I, 313 F. Supp. 3d at 258 (quoting AR 7).
Before the project took effect, several Medicaid recipients challenged the Secretary‘s approval in this Court. They argued, among other things, that the agency had failed to adequately explain why Kentucky HEALTH promoted the objectives of Medicaid and that approval of the project exceeded HHS‘s statutory authority. The Court concluded that the plaintiffs were right in one central and dispositive respect: “[T]he Secretary never adequately considered whether Kentucky HEALTH would in fact help the state furnish medical assistance to its citizens, a central objective
HHS has since reopened the comment period and subsequently reapproved Kentucky‘s project, offering additional explanation for why the project advances the objectives of the Medicaid Act. The parties have now come back to the Court and filed cross-motions for summary judgment in that case. The Court issues a separate Opinion today resolving those motions, which it will refer to as Stewart II.
C. Procedural History
Several Arkansas residents filed this lawsuit in August 2018. They assert that the Secretary‘s approval of the Arkansas Works Amendments was arbitrary and capricious, in excess of his statutory authority, and in violation of the Take Care Clause of the Constitution. Because it was designated as related to Stewart I, see ECF No. 2, the case was directed to this Court. While Defendants objected to the related-case designation, see ECF No. 17, the Court determined that the cases’ common legal and factual issues militated in favor of its retaining the matter. See Minute Order of Sept. 12, 2018. The State of Arkansas has since intervened as a Defendant, and numerous amici have also joined the fray. Dueling Cross-Motions for Summary Judgment are now ripe.
II. LEGAL STANDARD
The parties have cross-moved for summary judgment on the administrative record. The summary-judgment standard set forth in
The Administrative Procedure Act “sets forth the full extent of judicial authority to review executive agency action for procedural correctness.” FCC v. Fox Television Stations, Inc., 556 U.S. 502, 513 (2009). It requires courts to “hold unlawful and set aside agency action, findings, and conclusions” that are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”
In other words, an agency is required to “examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.” Id. at 43 (quoting Burlington Truck Lines v. United States, 371 U.S. 156, 168 (1962)) (internal quotation marks
III. ANALYSIS
The Court, as it must, first addresses whether there is subject-matter jurisdiction before proceeding to the merits of Plaintiffs’ challenges.
A. Jurisdiction
Unlike in Stewart I, Defendants do not contest Plaintiffs’ standing to challenge the Secretary‘s approval of the Arkansas Works Amendments as a whole. The Court, nevertheless, has an independent duty to assure that it has subject-matter jurisdiction in this case. See Kaplan v. Cent. Bank of Islamic Repub. of Iran, 896 F.3d 501, 509 (D.C. Cir. 2018). To establish standing under Article III, Plaintiffs must show that they have suffered a concrete injury that is fairly traceable to the challenged conduct and that is likely to be redressed by a favorable judicial decision. See Lujan v. Defs. of Wildlife, 504 U.S. 555, 590 (1992). On review, the Court easily concludes that at least one Plaintiff has established all three elements. Consider, for example, Adrian McGonigal, whom we encountered in this Opinion‘s opening paragraph. He attests that he has lost his Medicaid coverage as a result of the community-engagement requirement and has thus been unable to pay for certain medical bills and prescription drugs. See ECF No. 27-3 (McGonigal Declaration). Or look to Russell Cook, also mentioned in the introduction, who avers that he will be unable to meet the community-engagement requirement once it applies to him and thus believes that loss of his health-care coverage is imminent. See ECF No. 27-7 (Cook Declaration). From these declarations and others submitted with Plaintiffs’ Motion, there is little doubt that at least one Plaintiff has suffered an injury (or will suffer an injury in the future) — the loss of Medicaid coverage — that is attributable to the Secretary‘s approval of AWA, and that a favorable decision from the Court would redress it. See NB ex rel. Peacock v. District of Columbia, 682 F.3d 77, 82–83 (D.C. Cir. 2012).
While standing is thus easily established for their claim challenging the project as a whole, the state of Arkansas attacks Plaintiffs’ standing to make one of their arguments. It specifically says that no Plaintiff may challenge Arkansas Works’ online-only reporting requirements because the state changed its policy before this suit so as to allow reporting by phone or in person. See ECF No. 39 (Arkansas MSJ) at 34. There is no need for the Court to weigh in here. Because it resolves this case based on the challenge to the Arkansas Works Amendments writ large, the Court declines to decide whether certain Plaintiffs have standing to challenge this particular part of the project.
With that threshold issue easily dispatched, the Court turns to the merits. Plaintiffs’ central position is identical to that of the challengers in Stewart I: the Arkansas Works Amendments “fundamentally alter the design and purpose of Medicaid.” ECF No. 27 (MSJ) at 13. They thus assail the Secretary‘s approval of the Amendments on similar fronts. First, with regard to the project as a whole, Plaintiffs assert that HHS did not sufficiently consider whether it would promote the objectives of Medicaid, including how it would affect the provision of medical assistance to the needy. Second, they maintain that the Secretary lacked statutory authority to approve numerous aspects of AWA. Finally, Plaintiffs posit that a letter CMS issued in January 2018 violates the APA because it did not go through notice and comment. As in Stewart I, the Court only needs to consider the first of these contentions: “whether the Secretary acted arbitrarily or capriciously in concluding that [Arkansas Works] was ‘likely to assist in promoting the objectives’ of the Medicaid Act.” Stewart I, 313 F. Supp. 3d at 259 (quoting
Under that deferential standard, the Court “is not empowered to substitute its judgment for that of the agency.” Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 416 (1971). Nor can it “presume even to comment upon the wisdom of [Arkansas‘s] effort at [Medicaid] reform.” C.K. v. N.J. Dep‘t of Health & Human Servs., 92 F.3d 171, 181 (3d Cir. 1996). Still, it is a fundamental principle of administrative law that “agencies are required to engage in reasoned decisionmaking.” Michigan v. EPA, 135 S. Ct. 2699, 2706 (2015) (internal quotation marks omitted). This means that an agency must “examine all relevant factors and record evidence.” Am. Wild Horse Pres. Campaign v. Perdue, 873 F.3d 914, 923 (D.C. Cir. 2017). At minimum, the Secretary cannot “entirely fail[] to consider an important aspect of the problem.” Motor Vehicle Mfrs. Ass‘n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). Rather, he must “adequately analyze . . . the consequences” of his actions. See Am. Wild Horse, 873 F.3d at 932. In doing so, “[s]tating that a factor was considered . . . is not a substitute for considering it.” Getty v. Fed. Savs. & Loan Ins. Corp., 805 F.2d 1050, 1055 (D.C. Cir. 1986). The agency must instead provide more than “conclusory statements” to prove it “consider[ed] [the relevant] priorities.” Id. at 1057.
With that framework in mind, Plaintiffs’ position is simple: “The purpose of [] Medicaid” is to enable states “to furnish health care coverage to people who cannot otherwise afford it.” MSJ at 1, 15. Yet the Secretary, just as in Stewart I, “failed to consider adequately” the impact of the proposed project on Medicaid coverage. See Am. Wild Horse, 873 F.3d at 923. Indeed, he neither offered his own estimates of coverage loss nor grappled with comments in the administrative record projecting that the Amendments would lead a substantial number of Arkansas residents to be disenrolled from Medicaid. Those omissions, they urge, make his decision arbitrary and capricious.
Plaintiffs are correct. As Opening Day arrives, the Court finds its guiding principle in Yogi Berra‘s aphorism, “It‘s déjà vu all over again.” In other words, as the Secretary‘s failures here are nearly identical to those in Stewart I, the Court‘s analysis proceeds in the same fashion. It begins with the basic deficiencies in the Secretary‘s approval in this case and then examines Defendants’ counterarguments.
Before approving a demonstration or pilot project, the Secretary must identify the objectives of Medicaid and explain why the project is likely to promote them. As it did in Stewart I, the Court assumes that the Secretary‘s identification of those objectives is entitled to Chevron deference. That is, in reviewing his interpretation, the Court must first ask whether “Congress has directly spoken to the precise question at issue,” and, if not, whether “the agency‘s answer is based on a permissible construction of the statute.” Chevron U.S.A., Inc. v. Nat‘l Res. Def. Council, Inc., 467 U.S. 837, 842–43 (1984). According such deference is not of much practical significance here, however, because the Secretary agrees with the Court‘s understanding of a “core objective” of the Medicaid Act. See ECF No. 52 (HHS Reply) at 5.
In Stewart I, the Court explained that “one of Medicaid‘s central objectives” is to “furnish medical assistance” to persons who cannot afford it. See 313 F. Supp. 3d at 243, 261, 266, 273. That conclusion followed ineluctably from
Defendants, as mentioned, agree that providing health coverage to the needy is a purpose of the Act. See ECF No. 37 (HHS MSJ) at 12; Ark. MSJ at 13. In Arkansas‘s words, “[T]hat Medicaid coverage is a Medicaid objective is readily apparent from the substantive provisions of the statute.” Ark. MSJ at 13. The Secretary, in fact, refers to the provision of medical care to eligible persons as “Medicaid‘s core objective.” HHS Reply at 5 (emphasis added). HHS nevertheless did not consider whether AWA would advance or impede that objective.
In his approval letter, the Secretary explained that he considered the following objectives of the Medicaid Act: (1) “whether the demonstration as amended was likely to assist in improving health outcomes“; (2) “whether it would address behavioral and social factors that influence health outcomes“; and (3) “whether it would incentivize beneficiaries to engage in their own health care and achieve better health outcomes.” AR 4. Those are substantially the same objectives HHS considered when it first approved the Kentucky program. See Stewart I, 313 F. Supp. 3d at 261–62. What the Court said in that case thus holds true here: “While those may be worthy goals, there [i]s a notable omission from the list” — namely, whether the project would “help or hurt [Arkansas] in ‘funding . . . medical services for the needy.‘” Id. (quoting Alexander v. Choate, 469 U.S. 287, 289 n.1 (1985)). By his own description, the Secretary “entirely failed to consider” this question. See State Farm, 463 U.S. at 43.
The
a. Risk to Coverage
The Secretary‘s approval letter did not consider whether AWA would reduce Medicaid coverage. Despite acknowledging at several points that commenters had predicted coverage loss, the agency did not engage with that possibility. For example, after mentioning that commenters had “expressed concerns that these requirements would . . . create barriers to coverage,” the Secretary responded that “[t]he state has pledged to do beneficiary outreach and education on how to comply” and has created an “easy” online reporting system. See AR 6. He also pointed to exemptions built into the project and to Arkansas‘s assurances that it will allow for “reasonable modifications” for beneficiaries unable to meet the requirements. Id. But those statements did not grapple with the coverage issue. Not only did they fail to address whether coverage loss would occur as predicted, but they also ignored that commenters had projected that such loss would happen regardless of the exemptions and the education and reporting processes; indeed, some comments pinpointed online-only reporting as a source of coverage loss. See, e.g., AR 1272, 1287.
Later, HHS noted again many commenters’ view that community-engagement requirements would “create barriers to coverage for non-exempt people who might have trouble accessing care.” AR 6. Instead of addressing that issue, however, it merely said: “We believe that the community engagement requirements create appropriate incentives for beneficiaries to gain employment.” Id. That position says nothing about the risk of coverage loss those requirements create. The bottom line: the Secretary did no more than acknowledge — in a conclusory manner, no less — that commenters forecast a loss in Medicaid coverage. But “[s]tating that a factor was considered . . . is not a substitute for considering it.” Getty, 805 F.2d at 1055. His decision thus falls short of the kind of “reasoned decisionmaking” the APA requires. See Michigan, 135 S. Ct. at 2706.
Defendants argue that the Secretary did not need to and perhaps was not even able to provide a numeric estimate of coverage loss. See HHS MSJ at 21; Ark. MSJ at 24. While producing an empirical prediction of coverage loss does not seem like too much to ask of the expert agency tasked with supervising Medicaid programs in all 50 states, the Court does not need to decide whether such an estimate is required. Here, numerous commenters predicted that substantial coverage loss would occur; a table cataloguing the relevant comments is included at the end of this Opinion in an Appendix. See, e.g., AR 1269 (Arkansas Advocates noting that requirement “will increase the rate of uninsured Arkansans“); AR 1277 (American Congress Obstetricians and Gynecologists explaining that “[t]he experience of the TANF program . . . demonstrates that imposing work requirements on Medicaid beneficiaries would . . . lead to the loss of health care coverage for substantial numbers of people who are unable to work or face major barriers to finding and retaining employment.“); see also ECF No. 33
The Secretary should explain, for example, whether it agrees with the commenters’ coverage predictions. If so, it might elucidate whether it expects the loss to be minor or substantial, and how that weighs against the advancement of other Medicaid objectives. Nothing close to this appears in the Secretary‘s approval letter. That does not mean that the Government must “recit[e] and refut[e] every objection submitted in opposition to the proposed demonstration.” HHS MSJ at 22. It just means that, at a minimum, the agency cannot “entirely fail[] to consider an important aspect of the problem,” repeatedly raised in the comment period. See State Farm, 463 U.S. at 43.
Arkansas maintains that the Secretary did not need to consider any reduction in coverage because it unlike Kentucky did not predict that the project would even cause coverage loss. See Ark. MSJ at 24. But the state‘s failure in that respect does not alter HHS‘s inquiry. Under the Medicaid Act, the Secretary may approve only those demonstration projects that are “likely to assist in promoting the objectives of [Medicaid],” and the parties agree that the provision of health coverage is a “central” objective of the Act. See
In a last attempt to resist this conclusion, the Secretary says that he did not need to consider coverage because he had no obligation to offer any explanation of his decision to approve a demonstration project. See HHS MSJ at 22–23; see also Tr. at 9. For support, HHS points to the regulations governing its approval of demonstration projects, which do not
explicitly require the Secretary to respond to comments or articulate the basis for his decision. See HHS MSJ at 22 (discussing
b. Promote Coverage
At the same time that he failed to consider the risk to coverage, the Secretary identified only one element of the Amendments that might promote health coverage. In a single sentence, he noted that “a more limited period of retroactive eligibility will encourage beneficiaries to obtain and maintain health coverage, even when they are healthy.” AR 8. Little needs to be said on this score. It is well established that “conclusory or unsupported suppositions” do not satisfy the agency‘s obligation to engage in reasoned decisionmaking. See McDonnell Douglas Corp. v. U.S. Dep‘t of Air Force, 375 F.3d 1182, 1187 (D.C. Cir. 2004). That is particularly so in the face of numerous comments taking the opposite position. As the American Congress of Obstetricians and Gynecologists, among others, explained, limiting retroactive coverage may lead “Medicaid-eligible persons [to] wait even longer to have their conditions treated to avoid incurring medical bills they cannot pay.” AR 1279. And when they do eventually arrive for treatment, they will be covered for less time than they would have been before AWA took effect, by definition reducing their Medicaid coverage. See AR 1338 (National Health Law Program describing this risk). HHS‘s brief reference to the potential coverage-promoting effects of the changes to retroactive eligibility thus does not get it across the line.
2. Counterarguments
Defendants offer two separate reasons for the Court to overlook the Secretary‘s failure to consider coverage, neither of which is persuasive. They say first that the Arkansas Works Amendments promote several other important objectives of Medicaid, including the health of Medicaid-eligible persons. Second, Defendants maintain that any deficiency in the administrative record in this case is cured by the agency‘s subsequent approval of Kentucky‘s similar project on remand from the Court‘s decision in Stewart I.
a. Other Objectives
Defendants justify the proposed demonstration project on the ground that, regardless of its effect on Medicaid coverage, it advances other objectives of the Act. HHS specifically insists, as it did in Stewart I, that the Secretary was on solid ground in finding that the project would improve health outcomes, thereby advancing the goals of Medicaid. See HHS MSJ at 17–18. Faced with this argument previously, this Court expressed skepticism that health, generally construed, was properly considered an objective of the Act. See Stewart I, 313 F. Supp. 3d at 266. It ultimately held that the agency‘s “focus on health is no substitute for considering Medicaid‘s central concern: covering health costs” through the provision of free or low-cost health coverage. Id. The Court reached the same conclusion in response to assertions that Kentucky HEALTH promoted independence and self-sufficiency. Id. at 271–72. HHS has offered no argument here that calls those conclusions into question.
Arkansas presses the point in a somewhat different way, asserting that the provision of Medicaid coverage is (1) the purpose only of Medicaid appropriations, not Medicaid, (2) in “irreconcilable tension” with other purposes of the Act, and (3) not applicable to the Medicaid expansion population. See Ark. MSJ at 10–22. At the same time, it concedes, seemingly in conflict with its other contentions, that it is “readily apparent” that providing “Medicaid coverage for Medicaid-eligible people” is “an objective of Medicaid.” Id. at 13. The Court has said this before and will say it again: if, as Arkansas and HHS admit (and
While the Court might stop there, a brief foray into Arkansas‘s arguments is nevertheless worthwhile. As to the first, Medicaid is an appropriations statute enacted pursuant to “Congress‘s power under the Spending Clause.” NFIB, 567 U.S. at 542. What better place could the purpose of a spending program be found than in the provision that sets up the ”purpose” of its appropriations? Arkansas‘s second objection is even more puzzling. The Court does not understand how the objectives of a statute all agree was designed to provide free or low-cost medical care to the needy could nevertheless stand in “irreconcilable tension” with the goal of providing free or low-cost medical care to that population. The third sits on more comprehensible ground, though it yields Arkansas no more success. Addressing the purpose of the Medicaid expansion in Stewart I, the Court explained that “the Medicaid statute — taken as a whole — confirms that Congress intended to provide medical assistance to the expansion population.” 313 F. Supp. 3d at 269. HHS conceded as much in that case. Id. Neither party has offered any reason to retreat from that determination.
Defendants’ attempts to find refuge in other purposes of the Act and the propriety of Chevron deference as to those purposes are thus all hat, no cattle. Because they agree that the provision of low-cost medical care to Medicaid-eligible persons is a “core” purpose of the Act, see HHS Reply at 5, there is no legally significant dispute over the meaning of the Medicaid Act. What matters, instead, is the question addressed above: whether the Secretary adequately considered this issue. As has been made abundantly clear, he did not. Perhaps understanding as much, HHS largely attempts to justify its approval of the project in this case not on the Arkansas record but on another record entirely.
b. Kentucky Remand
This brings the Court to the argument that leads off the Secretary‘s Reply Brief: that his approval of AWA “is amply justified by the reasoning in his November 20, 2018, approval of Kentucky‘s materially similar project.” HHS Reply at 1. In particular, HHS argues that the project on review here will, like the one approved on remand in Kentucky, help adults “transition from Medicaid to financial independence,” thereby enhancing “the fiscal sustainability of Arkansas‘s Medicaid program” an objective of the Act. Id. at 6. The Government clarified at oral argument that this is not merely a contention against vacatur although it was principally offered as such but also an argument in favor of sustaining the Secretary‘s approval entirely. See Tr. at 8–10. The Court addresses the latter position here, leaving the remedy question for the end. In short, three weighty and independent rationales require rejecting HHS‘s assertion that the Amendments should be approved based on the record in the Kentucky remand proceeding.
First, it runs headlong into the “fundamental rule of administrative law” that a reviewing court “must judge the propriety of such action solely by the grounds invoked by the agency.” SEC v. Chenery Corp., 332 U.S. 194, 196 (1947). Nowhere in the Secretary‘s approval letter does he justify his decision based on concerns about the sustainability of Arkansas‘s Medicaid program, or on a belief that the project will help Medicaid-eligible persons to gain sufficient
HHS‘s argument suffers from a second and equally significant flaw. The demonstration project under consideration in Kentucky involves different considerations from the Arkansas project, and the rationales in favor of approving one may well not apply to approving the other. The Secretary said as much in opposing this case‘s designation as related to the Kentucky one. See ECF No. 17 (“The two cases involve two separate approvals of two distinct projects in two different States.“). Consider the principal arguments the Secretary relies upon on remand in Kentucky. First, he says that the project promotes coverage because in its absence, the expansion population would have no Medicaid coverage. See Stewart v. Azar, No. 18-152, ECF No. 108 (HHS MSJ) at 18–20. A necessary ingredient of this argument appears to be that the Kentucky Governor has conditioned the Commonwealth‘s continued expansion of Medicaid on the Secretary‘s approval of the proposed project. Id. at 19. There is no suggestion that Arkansas‘s Governor has made any similar kind of threat with regard to the Arkansas Works Amendments. Second, the Secretary justifies the Kentucky program on the ground that it advances the fiscal sustainability of the state‘s Medicaid program, which is at risk due to Kentucky‘s dire budgetary situation. Id. at 15–18. Yet there is no assertion that Arkansas is suffering from similar fiscal problems. The Government‘s argument that the Kentucky approval justifies the decision on review in this case is particularly unpersuasive considering these significant differences.
The final reason to reject this argument is the simplest: the justification the Secretary has given for sustaining Kentucky‘s program on remand is insufficient and the Court today rejects it in its latest Opinion in Stewart. See Stewart v. Azar, No. 18-152, Slip Opinion at 3 (Mar. 27, 2019) (Stewart II). If the explanation does not even justify affirmance of Kentucky‘s project, it cannot support upholding a different administrative decision approving a different state‘s project.
* * *
In sum, the Secretary‘s approval of the Arkansas Works Amendments is arbitrary and capricious because it did not address despite receiving substantial comments on the matter whether and how the project would implicate the “core” objective of Medicaid: the provision of medical coverage to the needy. Neither his consideration of other Medicaid Act objectives nor his subsequent approval of Kentucky‘s separate demonstration project cure that deficiency. This failure infected the Secretary‘s approval of AWA as a whole, such that those Amendments are invalid. The Court will thus grant Plaintiffs full relief on their arbitrary-and-capricious claim, removing any need to address their separate statutory-authority, APA notice-and-comment,
C. Remedy
That leaves only the question of the proper remedy, which in these circumstances is not small beer. When a court concludes that agency action is unlawful, “the practice of the court is ordinarily to vacate the rule.” Ill. Pub. Telecomms. Ass‘n v. FCC, 123 F.3d 693, 693 (D.C. Cir. 1997); Reed v. Salazar, 744 F. Supp. 2d 98, 119 (D.D.C. 2010) (“[T]he default remedy is to set aside Defendants’ action.“); Sierra Club v. Van Antwerp, 719 F. Supp. 2d 77, 78 (D.D.C. 2010) (“[B]oth the Supreme Court and the D.C. Circuit Court have held that remand, along with vacatur, is the presumptively appropriate remedy for a violation of the APA.“). “[A]lthough vacatur is the normal remedy, [courts] sometimes decline to vacate an agency‘s action.” Allina Health Servs. v. Sebelius, 746 F.3d 1102, 1110 (D.C. Cir. 2014). That decision depends on the “seriousness of the order‘s deficiencies (and thus the extent of doubt whether the agency chose correctly) and the disruptive consequences of an interim change.” Allied-Signal, Inc. v. U.S. Nuclear Reg. Comm‘n, 988 F.2d 146, 150-51 (D.C. Cir. 1993) (citation omitted); see also Standing Rock Sioux Tribe v. U.S. Army Corps of Engineers, 282 F. Supp. 3d 91, 103 (D.D.C. 2017) (declining to vacate when agency “largely complied” with statute and could likely substantiate prior conclusions on remand).
In Stewart I, the Court concluded that both factors supported vacatur. The Government‘s failure to consider an objective of Medicaid was a “major shortcoming” going “to the heart” of his decision. See 313 F. Supp. 3d at 273. And vacatur was not overly disruptive because the project had “yet to take effect” and the plaintiffs could suffer “serious harm[s]” were Kentucky HEALTH allowed to be implemented pending further proceedings. Id. While the journey is somewhat different in this case, the Court arrives at the same destination.
1. Seriousness of Deficiencies
The first factor does not favor the Government. For starters, in Stewart I, the Court concluded that the same legal error was a “major shortcoming” going “to the heart of the Secretary‘s decision.” 313 F. Supp. 3d at 273. It explained that the D.C. Circuit has “repeatedly vacated agency actions with that flaw.” Id. Defendants respond that the Secretary has cured the error identified in Stewart I on remand, so it will assuredly be able to cure this one upon remand, too. See HHS MSJ at 28-29; see also Ark. MSJ at 37-38. Not so. As explained at length in Stewart II, the Court finds that the remand has not cured this “major shortcoming.” See Slip Op. at 3, 14–45. Because the agency failed to provide a legally sufficient rationale upon remand from Stewart I, the Court is even less sanguine that it will be able to do so in this case than when it vacated the Secretary‘s Kentucky approval the first time.
This does not mean it will be impossible for the agency to justify its approval of a demonstration project like this one. The Court‘s decision does not go that far. But after at least two attempts for Kentucky, it has yet to do that analysis. Indeed, HHS may find it more difficult to offer a sufficient rationale in its second attempt in this case than in Kentucky. Arkansas does not appear to face the kind of fiscal issues asserted in Kentucky; instead, the state‘s data suggest that the Medicaid expansion has reduced the amount Arkansas will spend on health care for this population between 2017 and 2021. See ECF No. 53-6, Exh. 55 (Final Report of Arkansas Health Reform Legislative Task Force) (explaining that if Arkansas rejects Medicaid expansion,
2. Seriousness of Disruption
The second factor is a closer call. Arkansas began implementing its demonstration project in June 2018, imposing work requirements on adults ages 30–49 and implementing the changes to retroactive coverage; it began enforcing work requirements as to adults ages 19–29 in January 2019. HHS and Arkansas assert that any interruption in the project would be enormously disruptive because it would interfere with the “State‘s data collection efforts,” HHS Reply at 22, and “undermine” its “extensive efforts to educate Arkansas Works beneficiaries” on the work requirements. See Ark. MSJ at 38–39. They emphasize that, because the Kentucky program had not yet taken effect at the time of its vacatur, these concerns were not present in Stewart I. Id. The Court is not insensitive to the practical concerns Defendants raise about pausing enforcement of the Amendments, nor does it take lightly the effect of its ruling upon the state today. For the reasons that follow, however, it finds that the probable disruptions are not so significant as to require deviation from the ordinary rule of vacatur.
Consider first the nature and extent of the disruptions. If the Court vacates the Secretary‘s approval of AWA, the state would no longer condition certain Medicaid recipients’ coverage on reporting 80 hours of qualifying activities each month and would restore the number of months of retroactive coverage to three. In other words, vacatur would return matters to the way they were before the project was approved. Both changes, HHS asserts, will disrupt the state‘s data-collection efforts. See HHS MSJ at 29. If Arkansas — as the party responsible for collecting and analyzing data from the project — has concerns about data collection in the event of vacatur, it does not say as much. See Ark. MSJ at 38–40 (mentioning only disruptive effects on education and outreach); ECF No. 45 (Ark. Reply) (same). Indeed, one amicus points out that the Secretary approved this project without “a proposed evaluation design.” See Amicus Brief of Deans, Chairs, and Scholars at 19–20.
The Court assumes, however, that vacatur would interrupt the state‘s efforts to collect data on the effects of the work requirements and changes to retroactive coverage. While such concerns are not insignificant, they are tempered in the context of this case. Experimental projects are intended to help states like Arkansas “test out new ideas” for providing medical coverage to the needy, thereby influencing the trajectory of the federal-state Medicaid partnership down the line. See supra S. Rep. No. 1589 at 1961. If, after further consideration or after prevailing on appeal, the Secretary and Arkansas wish to move ahead with work requirements, they will remain able to do so in the future. And if they are dissatisfied with the data gathered from the initial months of the project because of the interruption caused by vacatur, Defendants could extend the project for an additional period of time to collect more information. This is not to minimize the importance of data collection in the context of an experimental project; it is just to say that vacatur will have little
Defendants also maintain that vacatur will harm “Arkansas‘s education and outreach efforts.” Ark. MSJ at 39. In that regard, they explain that a decision invalidating the work requirements will be confusing to Medicaid recipients who have just recently been informed that they have to meet those requirements. Id. at 38–39. The Court grants that vacatur of work requirements that have already been implemented may send mixed messages. But any disruption in this respect is not sufficiently significant to avoid vacatur. For one thing, Defendants have expressed confidence throughout this case that they can communicate with Medicaid recipients regarding the terms of the work requirements. See HHS MSJ at 8; Ark MSJ at 27, 34–35. If that is so, they should be able to inform them that the requirements are paused for now and, if later reapproved, that they are put back into effect. It bears mentioning here, however, that the State‘s outreach efforts may well be falling severely short. Notably, only 12.3% of persons not exempt from the requirements reported any kind of qualifying activity. See Arkansas Works Reports June–November 2018 at 47, 52. The numbers are even lower for several other months. Id. Arkansas might use the time while the program is paused to consider whether and how to better educate persons about the requirements and how to satisfy them. Admittedly, vacatur could make such outreach complicated. Ultimately, however, the Court finds that the harms to prior and ongoing education do not tip the scales against vacatur.
In fact, the structure of the Amendments, considered with the timing of this Opinion, renders vacatur less disruptive that might be expected. As mentioned before, Arkansas Works recipients only lose coverage after three months of non-compliance with the work requirements. See AR 31. And the three-month clock starts over at the beginning of the calendar year. Id. Because fewer than three months have elapsed in 2019, the work requirements have not yet resulted in anyone‘s being disenrolled, as such actions cannot take place until April 1. As a consequence, vacatur of the Amendments will not require Arkansas to re-enroll persons who have lost their coverage, with the administrative and communication-related headaches that might entail. Instead, it just requires them to communicate to providers that they should not disenroll persons moving forward on account of the requirements. The bottom line: “This is not a case in which the ‘egg has been scrambled,’ and it is too late to reverse course.” Allina Health, 746 F.3d at 1110–11 (quoting Sugar Cane Growers Co-op of Fla. v. Veneman, 289 F.3d 89, 97 (D.C. Cir. 2002)).
Finally, the Court emphasizes that the disruptions to Arkansas‘s administration of its Medicaid program must be balanced against the harms that Plaintiffs and persons like them will experience if the program remains in effect. Cf. A.L. Pharma, Inc. v. Shalala, 62 F.3d 1484, 1492 (D.C. Cir. 1995) (explaining that vacatur inappropriate because “nothing in the record suggests that significant harm would result from allowing the approval to remain in effect pending the agency‘s further explanation“); see also Tr. at 13 (conceding that court should consider harms to Plaintiffs as part of equitable inquiry into vacatur). Arkansas‘s own numbers confirm that in 2018, more than 16,000 persons have lost their Medicaid. Defendants offer no reason to think the numbers will be different in 2019; indeed, once the requirements apply
IV. CONCLUSION
For the foregoing reasons, the Court will grant Plaintiffs’ Motion for Summary Judgment and deny Defendants’ Cross-Motions. A separate Order consistent with this Opinion will issue this day, remanding the matter to HHS.
/s/ James E. Boasberg
JAMES E. BOASBERG
United States District Judge
Date: March 27, 2019
APPENDIX A
| Arkansas Health Plan Component | Comments |
|---|---|
| Community-Engagement Requirement | AR 1269 (Arkansas Advocates for Children & Families) (noting that the requirement “will increase the rate of uninsured Arkansans” based on comparable effect in TANF program) AR 1277 (American Congress of Obstetricians and Gynecologists, et al.) (“The experience of the TANF program . . . demonstrates that imposing work requirements on Medicaid beneficiaries would . . . lead to the loss of health care coverage for substantial numbers of people who are unable to work or face major barriers to finding and retaining employment.“); AR 1285 (Families USA) (“The presence of the requirement itself will be a barrier to enrollment, causing some eligible working individuals to forego applying for coverage, and will make it more difficult for some statutorily eligible individuals to maintain coverage.“); AR 1291 (AARP) (expressing concern that requirements would “present an unnecessary barrier to health coverage for a sector of Arkansas‘s population for whom coverage is critical“); AR 1294 (Cystic Fibrosis Foundation) (“We are concerned that this definition [of medically unfit] does not specify what will qualify an individual for exemption, and that people with cystic fibrosis may lose coverage because they are unable to satisfy the requirement due to health status.“); AR 1308 (Arkansas Hospital Association) (“These proposed changes . . . will likely lead to increases in churn, gaps in coverage, uninsurance and uncompensated care for hospitals and other providers.“); AR 1326 (Legal Aid of Arkansas) (noting that the requirement “would exclude individuals . . . who are partially employable but suffer due to chronic health conditions“); AR 1337 (National Health Law Program) (“The end result of this policy will likely be fewer people with Medicaid coverage and more uninsured people delaying treatment.“); AR 1341 (Nat‘l Alliance on Mental Illness) (“NAMI Arkansas is concerned that the implementation of mandatory work requirements could cause substantial numbers of people with mental illness to lose health coverage, making it difficult to access mental health care.“); AR 1364–65 (Urban Institute Study) (detailing “coverage losses” as consideration for pending Medicaid work-related requirements nationwide and noting “potential adverse impacts on enrollees who have high health care needs but who do not qualify for disability benefits“); AR 1402 (Medicaid and CHIP Payment and Access Commission) (listing an impact on coverage as implication of Medicaid work requirement and noting almost every state proposing requirement had estimated a coverage loss). AR 1421 (Kaiser Family Foundation Issue Brief) (arguing that based on the TANF experience, “a work requirement might result in eligible people losing |
| Retroactive Eligibility | AR 1292 (AARP) (warning lack of retroactive coverage would increase debt obligations on previous beneficiaries and would “increase the burden of uncompensated care on providers“); AR 1297 (Human ARC) (“Gaps of time without medical coverage for the low-income population that are eligible and applying for Medicaid will be significant.“); AR 1307 (Arkansas Hospital Association) (“AHA is concerned that the waiver of retroactive eligibility will result in unanticipated and avoidable gaps in coverage and healthcare debt.“); AR 1320 (Cancer Action Network) (stating waiver of retroactive eligibility “could place a substantial financial burden on enrollees and cause significant disruptions in care“); AR 1338 (National Health Law Program) (“The entirely predictable result will be . . . more individuals experiencing gaps in coverage when some providers refuse to treat them.“). |