Gresham v. Azar

UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF COLUMBIA

CHARLES GRESHAM, et al.,

Plaintiffs,
v. Civil Action No. 18-1900 (JEB)
ALEX M. AZAR II, et al.,

Defendants.

MEMORANDUM OPINION

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

1
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

2
and Human Services in August 2018, arguing that the federal government’s approval of the

state’s new requirements violated the Administrative Procedure Act and the Constitution.

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

3
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.

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 42 U.S.C. § 1396-1. The Centers for Medicare and

Medicaid Services (CMS), a federal agency within the Department of Health and Human

Services, has primary responsibility for overseeing Medicaid programs. Under the cooperative

federal-state arrangement, participating states submit their “plans for medical assistance” to the

Secretary of HHS. Id. To receive federal funding, those plans — along with any material

changes to them — must be “approved by the Secretary.” Id.; see also 42 C.F.R. § 430.12(c).

Currently, all states have chosen to participate in the program.

To be approved, state plans must comply with certain minimum parameters set out in the

Medicaid Act. See 42 U.S.C. § 1396a (listing 83 separate requirements). One such provision

requires state plans to “mak[e] medical assistance available” to certain low-income individuals.

Id. § 1396a(a)(10)(A). Until recently, that group included pregnant women, children, and their

families; some foster children; the elderly; and people with certain disabilities. Id. In 2010,

however, Congress enacted the Patient Protection and Affordable Care Act (ACA), colloquially

4
known as Obamacare, “to increase the number of Americans covered by health insurance.” Nat’l

Fed. of Indep. Business v. Sebelius, 567 U.S. 519, 538 (2012). Under that statute, states can

expand their Medicaid coverage to include additional low-income adults under 65 who would

not otherwise qualify. See 42 U.S.C. § 1396a(a)(10)(A)(i)(VIII).

Generally, a state must cover all qualified individuals or forfeit its federal Medicaid

funding. Id. § 1396a(a)(10)(B). That was originally so for the ACA expansion population as

well. See 42 U.S.C. § 1396c. In NFIB, however, the Supreme Court held that Congress could

not, consistent with the Spending Clause of the Constitution, condition previously appropriated

Medicaid funds on the state’s agreeing to the expansion. See 567 U.S. at 584–85. The result

was that states could choose not to cover the new population and lose no more than the funds that

would have been appropriated for that group. Id. at 587. If, however, the state decided to

provide coverage, those individuals would become part of its mandatory population. Id. at 585–

87 (explaining that Congress may “offer[] funds under the Affordable Care Act to expand the

availability of health care, and requir[e] that States accepting such funds comply with the

conditions on their use”). In that instance, the state must afford the expansion group “full

benefits” — i.e., it must provide “medical assistance for all services covered under the State

plan” that are substantially equivalent “in amount, duration, or scope . . . to the medical

assistance available for [other] individual[s]” covered under the Act. See 42 U.S.C.

§ 1396d(y)(2)(B); 42 C.F.R. § 433.204(a)(2).

The Medicaid Act, in addition to defining who is entitled to coverage, also ensures what

coverage those enrolled individuals receive. Under § 1396a, states must cover certain basic

medical services, see 42 U.S.C. §§ 1396a(a)(10)(A), 1396d(a), and the statute limits the amount

and type of premiums, deductions, or other cost-sharing charges that a state can impose on such

5
care. Id. § 1396a(a)(14); see also id. § 1396o. Other provisions require states to provide three

months of retroactive coverage once a beneficiary enrolls, see id. § 1396a(a)(34), and to ensure

that recipients receive all “necessary transportation . . . to and from providers.” 42 C.F.R.

§ 431.53. Finally, states must “provide such safeguards as may be necessary to assure” that

eligibility and services “will be provided, in a manner consistent with simplicity of

administration and the best interests of the recipients.” 42 U.S.C. § 1396a(a)(19).

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 42 U.S.C. § 1315. In enacting the Social Security Act (and, later, the

Medicaid program within the same title), Congress recognized that statutory requirements “often

stand in the way of experimental projects designed to test out new ideas and ways of dealing

with the problems of public welfare recipients.” S. Rep. No. 1589, 87th Cong., 2d Sess. 19,

reprinted in 1962 U.S.C.C.A.N. 1943, 1961–62. To that end, § 1115 of the Social Security Act

allows the Secretary to approve “experimental, pilot, or demonstration project[s]” in state

medical plans that would otherwise fall outside Medicaid’s parameters. The Secretary can

approve only those projects that “in [his] judgment . . . [are] likely to assist in promoting the

[Act’s] objectives.” 42 U.S.C. § 1315(a). As conceived, demonstration projects were “expected

to be selectively approved by the Department and to be those which are designed to improve the

techniques of administering assistance.” Supra S. Rep. No. 1589 at 1962. Once the Secretary

has greenlighted such a project, he can then waive compliance with the requirements of § 1396a

“to the extent and for the period . . . necessary to enable [the] State . . . to carry out such project.”

Id. § 1315(a)(1).

6
 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 42 C.F.R.

§§ 431.408(a)(1), (3). Once a state completes those prerequisites, it then sends an application to

CMS. Id. § 431.412 (listing application requirements). After the agency notifies the state that it

has received the waiver application, a federal 30-day public-notice period commences, and the

agency must wait at least 45 days before rendering a final decision. Id. §§ 431.416(b), (e)(1).

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

7
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.

On March 5, 2018, the Secretary approved the work requirements and limits to

retroactive coverage, concluding that they were “likely to assist in improving health outcomes”

and “incentivize beneficiaries to engage in their own health care.” AR 2–4. Under the new work

requirements, most able-bodied adults in the Medicaid expansion population ages 19 to 49 must

complete each month 80 hours of employment or other qualifying activities — or earn income

equivalent to 80 hours of work. Id. Compliance was required to be reported monthly through an

online portal. See AR 29. Various groups of persons are exempt, including the medically frail,

pregnant women, full-time students, and persons in drug- or alcohol-treatment programs. See

AR 28. Nonexempt individuals who do not report sufficient qualifying hours for any three

months in a plan year are disenrolled from Medicaid for the remainder of that year and not

permitted to re-enroll until the following plan year. See AR 14, 30–31. The work requirements

8
took effect for persons age 30 to 49 on June 1, 2018, and for persons age 20 to 29 on January 1,

2019. See ECF No. 26-3 (Arkansas Works Eligibility and Enrollment Monitoring Plan) at 7–8.

As to retroactive coverage, the Secretary approved a reduction from the three months required by

the Act to one month; the more drastic proposal of eliminating such coverage entirely was

abandoned, as was the Governor’s request to reduce eligibility down to 100% of the FPL. See

AR 12, 22.

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,

9
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 of Medicaid.” Id. at 243. It therefore vacated the Secretary’s

approval and remanded the matter to the agency for further consideration. Id. at 273.

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.

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

10
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 Federal Rule of Civil Procedure 56(c), therefore, “does

not apply because of the limited role of a court in reviewing the administrative record.” Sierra

Club v. Mainella, 459 F. Supp. 2d 76, 89 (D.D.C. 2006); see also Bloch v. Powell, 227 F. Supp.

2d 25, 30 (D.D.C. 2002), aff’d, 348 F.3d 1060 (D.C. Cir. 2003). “[T]he function of the district

court is to determine whether or not as a matter of law the evidence in the administrative record

permitted the agency to make the decision it did.” Sierra Club, 459 F. Supp. 2d. at 90 (quotation

marks and citation omitted). “Summary judgment is the proper mechanism for deciding, as a

matter of law, whether an agency action is supported by the administrative record and consistent

with the [Administrative Procedure Act] standard of review.” Loma Linda Univ. Med. Ctr. v.

Sebelius, 684 F. Supp. 2d 42, 52 (D.D.C. 2010) (citation omitted).

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.” 5 U.S.C. § 706(2). Agency action is arbitrary and capricious if, for

example, the agency “entirely failed to consider an important aspect of the problem, offered an

explanation for its decision that runs counter to the evidence before the agency, or is so

implausible that it could not be ascribed to a difference in view or the product of agency

11
expertise.” Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S.

29, 43 (1983).

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 omitted). Courts, accordingly, “do not defer to the agency’s

conclusory or unsupported suppositions,” United Techs. Corp. v. Dep’t of Def., 601 F.3d 557,

562 (D.C. Cir. 2010) (quoting McDonnell Douglas Corp. v. Dep’t of the Air Force, 375 F.3d

1182, 1187 (D.C. Cir. 2004)), and “agency ‘litigating positions’ are not entitled to deference

when they are merely [agency] counsel’s ‘post hoc rationalizations’ for agency action, advanced

for the first time in the reviewing court.” Martin v. Occupational Safety & Health Review

Comm’n, 499 U.S. 144, 156 (1991) (citation omitted). Although a reviewing court “may not

supply a reasoned basis for the agency’s action that the agency itself has not given,” a decision

that is not fully explained may, nevertheless, be upheld “if the agency’s path may reasonably be

discerned.” Bowman Transp., Inc. v. Arkansas-Best Freight System, Inc., 419 U.S. 281, 285-86

(1974) (citation omitted).

III. ANALYSIS

The Court, as it must, first addresses whether there is subject-matter jurisdiction before

proceeding to the merits of Plaintiffs’ challenges.

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

12
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

13
Amendments writ large, the Court declines to decide whether certain Plaintiffs have standing to

challenge this particular part of the project.

Merits

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 42

U.S.C. § 1315(a)).

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

14
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.

15
 1. The Secretary’s Consideration of Medicaid’s Objectives

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 § 1396-1 of the Act, which provides that

Congress appropriated Medicaid funds “[f]or the purpose of enabling each State, as far as

practicable under the conditions in such State, to furnish (1) medical assistance . . . [to]

individuals[] whose income and resources are insufficient to meet the costs of necessary medical

services, and (2) rehabilitation and other services to help such families and individuals attain or

retain capability for independence or self-care.” Case law discussing the program’s objectives

confirms as much. See, e.g., Schweiker v. Hogan, 453 U.S. 569, 571 (1982) (explaining that

Congress established Medicaid “for the purpose of providing federal financial assistance to

States that choose to reimburse certain costs of medical treatment for needy persons”); W. Va.

Univ. Hosps. Inc. v. Casey, 885 F.2d 11, 20 (3d Cir. 1989) (“[T]he primary purpose of

16
[M]edicaid is to achieve the praiseworthy social objective of granting health care coverage to

those who cannot afford it.”).

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 Government conceded as much at oral argument, stating that HHS’s Arkansas

approval letter no more addresses the program’s effects on Medicaid coverage than the Kentucky

approval letter before the Court in Stewart I. See Tr. at 6–7. Because this is a separate

administrative decision on review in a separate case, however, a brief assessment of the

17
deficiency is instructive. To “adequately analyze” the issue of coverage, Am. Wild Horse, 873

F.3d at 932, the Secretary needed to consider whether the demonstration project would be likely

to cause recipients to lose coverage and whether it would cause others to gain coverage. He did

neither.

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

18
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 (Amicus Brief of

Deans, Chairs, and Scholars) at 14. Under these circumstances, the agency must grapple with

the risk of coverage loss. See Nat’l Lifeline Assoc. v. FCC, 915 F.3d 19, 30-31 (D.C. Cir. 2019).

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

19
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 42 U.S.C. § 1315(a); HHS MSJ at 12–

13; Ark. MSJ at 13. Whether a state gives the Secretary excellent data or no data at all about

coverage, his duty remains the same: to determine whether the proposed project will promote the

objectives of the Act, including whether it advances or hinders the provision of health coverage

to the needy. If it were otherwise, HHS could approve a project that would decimate Medicaid

coverage without so much as addressing the issue where the state did not submit its own estimate

of coverage loss. Even putting to one side the agency’s affirmative obligation to address

coverage loss, however, the Secretary unquestionably has a duty to consider that issue where

multiple commenters provide credible forecasts that it will occur. See, e.g., AR 1269, 1277,

1285, 1294–95. Here, as has been said, the agency had and neglected that duty.

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

20
explicitly require the Secretary to respond to comments or articulate the basis for his decision.

See HHS MSJ at 22 (discussing 42 C.F.R. § 431.416). The APA, however, requires more.

Where an agency decision is judicially reviewable, as the Court has already held this one is, see

Stewart I, 313 F. Supp. 3d at 254–56, the Government “must give a reason that a court can

measure . . . against the ‘arbitrary or capricious’ standard of the APA.” Kreis v. Sec’y of Air

Force, 866 F.2d 1508, 1514–15 (D.C. Cir. 1989); see also Coburn v. McHugh, 679 F.3d 924, 934

(D.C. Cir. 2012) (“At the very least, the Board must ‘provide an explanation that will enable the

court to evaluate the agency’s rationale at the time of decision.’”) (quoting Pension Benefit Guar.

Corp. v. LTV Corp., 496 U.S. 633, 654 (1990)). HHS’s regulations — which require CMS to

maintain and publish an administrative record of public comments, any CMS responses, and a

written approval or disapproval letter — are fully consonant with this axiomatic administrative-

law requirement. See 42 C.F.R. § 431.416(f). The argument that no explanation for the

Secretary’s decision is required thus does not save it.

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

21
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

22
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 this Court has found), ensuring

Medicaid coverage for the needy is a key objective of the Act, the Secretary’s failure to consider

the effects of the project on coverage alone renders his decision arbitrary and capricious; it does

not matter that HHS deemed the project to advance other objectives of the Act.

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

23
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.

24
 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

financial resources to be able to purchase private insurance. And the Court “may not accept []

counsel’s post hoc rationalizations for agency action.” State Farm, 463 U.S. at 50; see also

Burlington Truck Lines, 371 U.S. at 168–69 (“Chenery requires that an agency’s discretionary

order be upheld, if at all, on the same basis articulated in the order by the agency itself.”). The

Government responded at oral argument that the Secretary did not need to provide any basis for

his decision approving Arkansas’s proposed project, so it does not matter on what justification

his decision is judicially upheld. See Tr. at 9–10. The Court has already explained why that

assertion is inconsistent with the APA, see supra at 20–21, and it will not spill more ink on the

matter here.

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

25
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,

26
removing any need to address their separate statutory-authority, APA notice-and-comment, and

constitutional arguments.

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

27
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,

“the negative impact to the state budget is approximately $438 [million]” during this time

frame). It stands to reason that the state will have an uphill climb making the case that the

28
expansion has pressed its annual budget, such that eligible persons should be pushed off the rolls.

Such fiscal considerations would, in any event, need to be balanced against the more than 16,000

persons who have already lost their coverage because of the new requirements. See Arkansas

Works Reports at 18, 27, 36, 45. The upshot is that the road to cure the deficiency in this case is,

at best, a rocky one, strongly weighing in favor of vacatur.

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

29
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 lasting impact on HHS’s or Arkansas’s

interests. That distinguishes this case from others in which the D.C. Circuit has declined to

vacate on account of irreversible harms that such a remedy would inflict on the status quo. See

Allied-Signal, 988 F.2d at 151.

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

30
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

31
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 to

persons aged 19–29, they seem likely to rise. See Arkansas Works Reports at 18, 27, 36, 45.

Weighing the harms these persons will suffer from leaving in place a legally deficient order

against the disruptions to the State’s data-collection and education efforts due to vacatur renders

a clear answer: the Arkansas Works Amendments cannot stand.

32
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

33
 APPENDIX A

Arkansas Health Comments
Plan Component
Community- AR 1269 (Arkansas Advocates for Children & Families) (noting that
Engagement the requirement “will increase the rate of uninsured Arkansans” based
Requirement 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 coverage”).

34
Retroactive AR 1292 (AARP) (warning lack of retroactive coverage would
Eligibility 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.”).

35