KETHLEDGE

ARGUED:Katherine L. McDaniel, Kirkland & Ellis LLP, New York, New York, for Appellants. Michael W. Kirk, Cooper & Kirk, PLLC, Washington, D.C., for Appellees. ON BRIEF:Katherine L. McDaniel, Kirkland & Ellis LLP, New York, New York, G. Gordon Bonnyman, Jr., Michele M. Johnson, Christopher E. Coleman, Tennessee Justice Center, Nashville, Tennessee, for Appellants. Michael W. Kirk, Nicole J. Moss, Cooper & Kirk, PLLC, Washington, D.C., Linda A. Ross, Office of the Tennessee Attorney General, Nashville, Tennessee, for Appellees.

Before: ROGERS, COOK, and , Circuit Judges.

OPINION

, Circuit Judge.

Consent decrees are not entitlements. Instead, a decree may remain in force only as long as it continues to remedy a violation of federal law. Here, Tennessee's Medicaid program has operated under a federal consent decree for 15 years. In recent years, however, the State moved to vacate the decree on grounds that the State is now compliant with both the decree and the Medicaid statute. The district court conducted a month-long evidentiary hearing to explore those grounds. Afterward, the court issued an exhaustive opinion in which it found that the State has vastly improved its Medicaid program and is indeed compliant with all the relevant provisions of federal law. Thus the court vacated the decree.

The plaintiffs now challenge the court's decision on numerous grounds. Some of those grounds misstate the bases of the court's decision. Other grounds are simply meritless. Our conclusion is therefore the same as the district court's: control of Tennessee's Medicaid program must now return to the State of Tennessee.

I.

Medicaid is a cooperative federal-state program that provides medical care to the poor. States are not required to participate in Medicaid, but those that do must comply with the Medicaid Act, 42 U.S.C. § 1396 et seq. Tennessee participates in Medicaid through a program known as TennCare. SeeTenn.Code § 71–5–102.

In relevant part, the Medicaid Act requires that TennCare administer an Early and Periodic Screening, Diagnosis, and Treatment program for all enrollees under the age of 21. See42 U.S.C. §§ 1396a(a)(43), 1396d(r). As a part of this program, TennCare must provide two basic services: first, provide medical checkups to its enrollees on a regular basis (referred to as “screens” by the Act); and second, diagnose and treat any health problems revealed by those screens. See id. § 1396a(a)(43)(B)-(C). TennCare must also conduct outreach to educate its enrollees about these services. See id. § 1396a(a)(43)(A).

In 1998, the plaintiffs filed a putative class action under 42 U.S.C. § 1983, alleging that TennCare had failed to fulfill these obligations. The parties quickly settled, and the district court entered a consent decree that explained in detail the requirements that TennCare had to meet to “achieve and maintain compliance” with the Medicaid Act. See Consent Decree ¶ 14. The parties expressly based these requirements on the assumption that the Act created rights enforceable under § 1983. See Consent Decree ¶ 15. The decree also included a sunset clause. In relevant part, that clause provides that the decree “shall expire” when TennCare reaches an “adjusted periodic screening percentage” of 80%, and is in “current, substantial compliance” with the rest of the decree. See Consent Decree ¶ 113, at 54.

Eight years later, this court held that one part of the Medicaid Act—42 U.S.C. § 1396a(a)(30)—was unenforceable under § 1983. See Westside Mothers v. Olszewski, 454 F.3d 532, 542 (6th Cir.2006)( Westside Mothers II ). Shortly thereafter, TennCare moved to vacate the consent decree under Federal Rule of Civil Procedure 60(b), arguing that Westside Mothers II had invalidated the parties' assumption that the Medicaid Act created rights enforceable under § 1983. The district court denied the motion. TennCare appealed.

On appeal, we affirmed in part, reversed in part, and remanded. See John B. v. Goetz, 626 F.3d 356 (6th Cir.2010). Like the district court, we rejected TennCare's argument that Westside Mothers II had left the entire Medicaid Act unenforceable under § 1983. But we agreed that certain parts of the Act—such as § 1396a(a)(30)—could not be privately enforced. Thus, we instructed the district court to determine the statutory basis of the decree, and to vacate any paragraphs based on parts of the Act that are not privately enforceable. We also ordered the case reassigned to a new district judge.

Judge Thomas A. Wiseman, Jr. took up the case on remand. The district court familiarized itself with the case's 13–year history and ordered that the parties file supplemental briefs with respect to TennCare's Rule 60(b) motion to vacate the decree. It also held a hearing on that motion. The court later issued a 17–page opinion that examined the statutory basis of every paragraph in the decree. In that opinion, the court held that several of the decree's paragraphs had been based upon statutory provisions or regulations that were themselves unenforceable under § 1983. Thus, the court vacated those paragraphs. But the court held that “the [d]ecree as a whole, and the principal provisions in it, remained enforceable.”

Meanwhile, TennCare filed a second motion to vacate the decree. There, TennCare argued in part that it had fulfilled the terms of the decree's sunset clause by reaching a screening percentage greater than 80% and by achieving current, substantial compliance with the rest of the decree. Thereafter, the district court held an 18–day evidentiary hearing, during which it heard testimony from 31 witnesses and admitted 260 exhibits. The court also received 345 pages of proposed findings of fact and conclusions of law from the parties. The court later issued a 38–page opinion that included a thorough examination of TennCare's compliance with the decree and the Medicaid Act.

The district court began that examination with a discussion of TennCare's outreach efforts. The court found that, after enrolling in TennCare, a family with children learns about the Early and Periodic Screening, Diagnosis, and Treatment program in five different ways. First, Tennessee's Department of Human Services (which handles the enrollment process) tells each family about the program and urges them to take their children to the doctor for a screen. Second, TennCare sends each family a welcome letter that encourages them to schedule an appointment. Third, a managed-care organization ( i.e., a contractor that TennCare hires to manage part of its program) sends each family a member handbook that includes information about screening services. Fourth, a managed-care organization calls each newly enrolled family to urge them to schedule a screen. And fifth, Tennessee's Department of Health also calls each family to tell them about the program, to offer assistance in scheduling a screen, and to offer transportation to the appointment.

The district court also found that TennCare sends at least five reminders to its enrollees about screening appointments each year. Each family receives quarterly newsletters and a postcard around the child's birthday, all of which encourage the family to schedule a screen. Families may receive further reminders from the Department of Health, which runs a community-outreach program designed to target hard-to-reach enrollees, such as pregnant teenagers.

In addition, the district court found that TennCare makes numerous attempts to contact children that have missed a screen. When a child's date for a screen has passed, a managed-care organization sends a reminder to the family. If a child goes an entire year without a screen, the family will receive two more reminders—one from a managed-care organization and one from TennCare. The Department of Health also runs a home-visit program to reach children that have missed a screen. Under this program, the Department sends a community-outreach worker to the home of any child who is overdue for a screen, and urges the family to schedule an appointment. Thus, if a child goes a whole year without a screen, TennCare will contact that child's family at least nine times—four times through quarterly newsletters, one time through a postcard, three times through reminder notices, and one time through a home visit.

Next, the district court examined TennCare's screening services. It found that TennCare provides the four types of screens required by the Medicaid Act: physical, vision, hearing, and dental. See42 U.S.C. § 1396d(r). The court also found that TennCare had complied with the Act by adopting the periodicity schedules ( i.e., schedules that state how often a child should receive each type of screen) recommended by a committee of experts in each field. See, e.g., 42 U.S.C. § 1396d(r)(1)(A)(i). TennCare provides screens to its enrollees free of charge, whether or not the screen is recommended by the relevant periodicity schedule. Thus, parents or guardians can take a child in for as many screens as they like, no matter how many times they have already done so.

The district court also examined TennCare's diagnostic and treatment services. The court looked first at TennCare's policies and found that “TennCare children are entitled to receive, free of charge, all medically necessary covered diagnosis and treatment services.” The court then examined how TennCare's managed-care organizations actually applied those policies. “In practice[,]” the court found, “the vast majority of diagnosis and treatment services are provided to TennCare enrollees automatically, without any medical-necessity review, when the service is ordered by a licensed provider.” And the court found that, even when a managed-care organization does engage in medical-necessity review, the organization usually approves the requested service.

In the comparatively few cases where a managed-care organization denies a requested service, TennCare offers the affected family an exhaustive appeals process. An appeal begins with TennCare asking the managed-care organization to have a second doctor review the request. If that doctor recommends denial, TennCare sends the case to an independent medical consultant. If the consultant also recommends denial, the case goes to an administrative law judge. Thus, on appeal, TennCare will only deny a requested service if two doctors, an independent medical consultant, and an administrative law judge all agree that the service is not medically necessary.

The district court also found that the plaintiffs' own witnesses “largely confirmed that TennCare provides medically necessary diagnostic and treatment services.” Although these witnesses testified that TennCare did not always provide services as quickly as it should, the plaintiffs “did not identify any instance where needed services were not ultimately provided.”

The district court then examined the four primary ways that TennCare monitors its compliance with the Medicaid Act. First, the court found that Tennessee was the first state in the country to require that its managed-care organizations earn full accreditation from the National Committee for Quality Assurance, which is an independent organization dedicated to improving health-care quality. Of TennCare's three managed-care organizations, two have earned the highest overall accreditation rating of “excellent” and the third earned the second-highest rating of “commendable.”

Second, the court found that TennCare requires its managed-care organizations to use a measuring tool known as the Healthcare Effectiveness Data and Information Set (HEDIS). This tool measures the performance of TennCare's managed-care organizations in 75 different areas. See HEDIS & Performance Measurement,National Committee for Quality Assurance, http:// www. ncqa. org/ HEDISQuality Measurement. aspx (last visited Feb. 19, 2013). HEDIS allows TennCare to track its year-to-year performance and to compare itself with other health plans across the country. The court also found that TennCare's HEDIS results “compare[d] favorably to national Medicaid averages[.]” For example, TennCare's results exceeded, or were comparable to, the national average for “access to and availability of care for children, timeliness and frequency of prenatal care, child immunization rates, and effectiveness of behavioral health [.]” According to the court, these results compared “even more favorably to the Southeastern regional averages.”

Third, the court found that TennCare requires its managed-care organizations to use another measuring tool known as Consumer Assessment of Healthcare Providers and Systems (CAHPS). This tool measures the satisfaction of TennCare's enrollees with their medical care. Again, the court found that TennCare's CAHPS results were better than the national averages for Medicaid. For example, between 84% and 86% of TennCare's enrollees stated that they always or usually get the care they need for their children—whereas only 77% of Medicaid recipients nationally said the same.

Fourth, the court found that TennCare had hired Qsource to serve as its External Quality Review Organization. See generally42 C.F.R. § 438.356. In that role, Qsource performs two primary tasks: First, it reviews the policies adopted by TennCare's managed-care organizations to determine whether they comply with federal law. Second, it reviews randomly selected medical files to make sure that TennCare's managed-care organizations are actually implementing those policies in practice.

Finally, the district court examined TennCare's compliance with every paragraph of the decree that the court had not already vacated. It found that TennCare was in substantial compliance with the decree and had therefore fulfilled the terms of the decree's sunset clause. See Consent Decree ¶ 113, at 54.

Based on this extensive examination, the court found that, “compared with its performance in 1998, TennCare ha[d] dramatically improved the provision of medical services to its enrollees in every respect.” In fact, the court found that “no other state's ... program surpasses that of Tennessee in any salient respect.” The court also concluded that TennCare is “fully compliant with the [relevant Medicaid] law and regulations.” It therefore vacated the decree in full and dismissed the case. This appeal followed.

II.

A.

The plaintiffs first challenge the district court's order granting in part TennCare's Rule 60(b) motion to vacate the decree. We review that decision for an abuse of discretion. See Northridge Church v. Charter Twp. of Plymouth, 647 F.3d 606, 613 (6th Cir.2011). The plaintiffs have the burden to prove that the court had no reasonable basis for granting the motion. See Cleveland Firefighters for Fair Hiring Practices v. City of Cleveland, 669 F.3d 737, 740 (6th Cir.2012).

Under Rule 60(b)(5), a court may vacate a consent decree if, among other things, “a significant change ... in law renders [its] continued enforcement detrimental.” Northridge Church, 647 F.3d at 613 (quotation marks omitted). A change in law satisfies that test when the parties “based their agreement on a misunderstanding” of the law. Rufo v. Inmates of Suffolk Cnty. Jail, 502 U.S. 367, 390, 112 S.Ct. 748, 116 L.Ed.2d 867 (1992); accord Doe v. Briley, 562 F.3d 777, 782–83 (6th Cir.2009). Here, the district court held that the parties based three clusters of paragraphs on a misunderstanding of the law. We consider each in turn.

1.

The first cluster includes paragraphs 43, 58, 60(v)-(vi), 61(ii), and 71(ii) of the consent decree. Broadly stated, these paragraphs address the adequacy of TennCare's provider network. Paragraph 43 requires that TennCare “ensure that [its managed-care organizations'] networks are adequate ... to properly screen children in conformity with the requirements of ... the Medicaid statute[.]” Paragraph 58 requires that all utilization-review and prior-authorization decisions “be made only by qualified personnel with education, training, or experience in child and adolescent health.” Paragraph 60(v) requires that TennCare's network “include providers with cultural and linguistic competency ... as may be needed for the effective treatment of children from ethnic minorities[.]” Paragraph 60(vi) requires that TennCare's managed-care organizations “have a sufficient array of services and specialists to meet the medical and behavioral health needs” of TennCare's enrollees. Paragraph 61(ii) requires that TennCare's “provider networks currently comply with the ‘Terms and Conditions for Access' ” document issued by the Health Care Financing Administration (now the Centers for Medicare & Medicaid Services). And paragraph 71(ii) requires that TennCare “[p]rovide a comprehensive and appropriate scope of geographically accessible child and adolescent behavioral health services[.]”

The district court held that these network-adequacy paragraphs were all based on § 1396a(a)(30)(A) of the Medicaid Act, which likewise addresses network adequacy. Specifically, that subsection provides that TennCare must use procedures, including “utilization review,” to assure that its provider network is comparable in size to the private network in the geographic area. See42 U.S.C. § 1396a(a)(30)(A). The court also held that the parties based these paragraphs on a misunderstanding of subsection 30(A)—namely, that it created a right enforceable under § 1983. See Consent Decree ¶ 15. Westside Mothers II proved that assumption wrong, so the district court vacated these paragraphs.

The plaintiffs now object to this holding on five grounds. First, they argue that the court's decision violated the law-of-the-case doctrine. In support, they point to a 2001 opinion in which Judge John T. Nixon (who later recused himself from the case) held that TennCare had violated the consent decree. See John B. v. Menke, 176 F.Supp.2d 786, 802–06 (M.D.Tenn.2001). In that opinion, the plaintiffs say, the court concluded that the entire decree was based on § 1396a(a)(43)of the Medicaid Act, which of course is enforceable pursuant to § 1983. They therefore contend that, under the law-of-the-case doctrine, the district court was required to find that subsection 43 was the statutory basis of the paragraphs at issue.

The plaintiffs overread Judge Nixon's opinion. The law-of-the-case doctrine only applies to issues the court actually decided. See United States v. Cunningham, 679 F.3d 355, 376–77 (6th Cir.2012). And Judge Nixon did not actually decide the statutory basis of the consent decree. True, Judge Nixon held that TennCare had to comply with subsection 43 in order to fulfill its obligations under the decree. But that is not the same as a holding that the entire decree—including the network-adequacy paragraphs at issue here—was based on subsection 43. Indeed, the plaintiffs have not pointed to any part of Judge Nixon's opinion that even mentions the network-adequacy paragraphs, much less determines their statutory basis. So the law-of-the-case doctrine did not bind the district court here.

Second, the plaintiffs argue that the decree's primary purpose was to remedy violations of subsection 43. They note that their complaint was “primarily based on” subsection 43 and that the decree identifies subsection 43 as its “overall basis.” Thus, the plaintiffs argue, the parties must have based the network-adequacy paragraphs on subsection 43. But the decree's purpose does not determine the statutory basis of each individual paragraph. Indeed, if the plaintiffs were correct, we could not have ordered the district court to vacate a substantial portion of the decree the last time this case was before us—which of course we did. See John B., 626 F.3d at 363. So this argument too is meritless.

Third, the plaintiffs argue that none of the network-adequacy paragraphs explicitly mention subsection 30(A). That is true enough, but none of those paragraphs explicitly mention subsection 43 either. Moreover, a paragraph can be based on subsection 30(A) without mentioning it. For example, paragraph 61(ii) incorporates by reference a document titled “Terms and Conditions for Access[.]” And that document repeatedly says that TennCare must provide “access [to its networks] that is equal to or greater than the currently existing practice in the fee-for-service system”—an implicit reference to subsection 30(A)'s requirement that TennCare have a provider network that is comparable in size to the private network in the geographic area. That the network-adequacy paragraphs do not mention subsection 30(A), therefore, ultimately does not matter here.

Fourth, the plaintiffs argue that TennCare needs adequate networks to comply with subsection 43's screening and treatment requirements. Thus, they say, the network-adequacy paragraphs must be based on that subsection. But the plaintiffs overlook that TennCare also needs adequate networks to comply with subsection 30(A)'s geographic-comparability requirement. So this argument likewise goes nowhere.

Fifth, the plaintiffs argue that the language of the network-adequacy paragraphs shows that they are based on subsection 43, rather than subsection 30(A). For most of these paragraphs—namely, paragraphs 43, 60(v)-(vi), 61(ii), and 71(ii)—the plaintiffs' argument is utterly conclusory: they merely quote the language of each paragraph, and then announce that it does not refer to subsection 30(A). Suffice it to say that the plaintiffs' bare assertions as to these paragraphs are unconvincing.

The plaintiffs develop this argument only for paragraph 58, which requires that “utilization review and prior authorization decisions be made only by qualified personnel[.]” Although paragraph 58 and subsection 30(A) both discuss “utilization review[,]” the plaintiffs argue that “there is virtually no relationship between the two [provisions].” In support, the plaintiffs contend that “[subsection] 30(A) mandates that [TennCare] employ utilization review ... [while] [p]aragraph 58 prohibits [TennCare] from using [it] ... inappropriately.” (Quotation marks omitted.) But the plaintiffs have tried to create a contradiction where none exists. There is nothing inconsistent about requiring that TennCare use utilization review, and then explaining that it must employ qualified personnel when it does so. Instead that makes perfect sense. This argument is meritless.

2.

The district court also vacated paragraph 84 of the consent decree. In relevant part, that paragraph provides:

The Department of Children's Services shall ensure that the case planning and case review required under the relevant portions of the Adoption Assistance and Child Welfare Act[, 42 U.S.C. § 670 et seq.,] for TennCare children in DCS custody ... shall identify and provide for the treatment of the behavioral health and medical needs of these children in accordance with [the Act.]
The district court held that the parties based this paragraph on the Adoption Act. It also held that the parties based this paragraph on the belief that the Act creates rights enforceable under § 1983. See Consent Decree ¶ 15. But the Act does not do so, see John B., 626 F.3d at 363; and thus the district court vacated this paragraph.

The plaintiffs again disagree with the court's statutory-basis determination. They argue that the parties based this paragraph on 42 U.S.C. § 1396a(a)(43), rather than the Adoption Act, because the paragraph enforces “only the parts of the [Adoption Act] that overlap with” subsection 43. But this argument proves only that the Adoption Act and subsection 43 both address children's medical care; it does not prove which of those provisions the parties based paragraph 84 on. Moreover, the text of paragraph 84 shows that the parties based it on the Adoption Act. That paragraph says that the Department of Children's Services must care for TennCare children “in accordance with” the Adoption Act; and it notably does not say that the Department must act in accordance with subsection 43 as well. The district court did not abuse its discretion on this point.

3.

The third cluster includes paragraphs 78–83, all of which address TennCare's coordination with other government agencies. The district court held that the parties based these paragraphs on a federal Medicaid regulation, 42 C.F.R. § 441.61(c). That regulation has two main requirements. First, it provides that TennCare “must make appropriate use of State health agencies, State vocational rehabilitation agencies, and Title V grantees.” Second, it says that TennCare “should make use of other public health, mental health, and educational programs and related programs ... to ensure an effective child health program.” The court further held that § 441.61(c) did not create rights enforceable under § 1983. It therefore vacated these paragraphs.

The plaintiffs argue that the court erred when it held that § 441.61(c) is not enforceable under § 1983. To that end, the plaintiffs contend that, “if a statute is privately enforceable, so too are its implementing regulations.” And because the district court held that § 441.61(c) implements a statutory provision that is privately enforceable—namely, 42 U.S.C. § 1396a(a)(43)(C)—the plaintiffs say that this regulation is enforceable as well.

That a statutory provision is privately enforceable, however, does not necessarily mean that a regulation that implements the provision is privately enforceable as well. To the contrary, an implementing regulation is not privately enforceable—even if its controlling statute is—when it “imposes an obligation or prohibition that is not imposed generally by the controlling statute.” Ability Ctr. of Greater Toledo v. City of Sandusky, 385 F.3d 901, 906 (6th Cir.2004). Section 441.61(c) does precisely that: it requires that TennCare coordinate its services with other agencies, even though subsection 43(C) says nothing of the sort. Consequently, the district court did not abuse its discretion in vacating these paragraphs.

4.

Finally, the plaintiffs argue that the district court should not have vacated any of the decree's paragraphs because they all “further[ ] the objectives” of subsection 43. In support, the plaintiffs rely on Local No. 93, International Ass'n of Firefighters, AFL–CIO C.L.C. v. City of Cleveland, which says that a consent decree must “further the objectives of the law upon which the complaint was based.” 478 U.S. 501, 525, 106 S.Ct. 3063, 92 L.Ed.2d 405 (1986). This argument confuses a necessary condition with a sufficient one. It is true that a decree must further the objectives of the federal law that formed the basis of the complaint. But that does not mean that a decree is enforceable simply because it does so; the decree must still be based on a provision that is privately enforceable. See John B., 626 F.3d at 362. And because the parties failed to base the paragraphs at issue here on a privately enforceable provision, the court was within its discretion to vacate them. See Briley, 562 F.3d at 782–83. Indeed, we expressly instructed the court to do so. See John B., 626 F.3d at 362–63. The plaintiffs' reliance on Firefighters is unpersuasive.

In sum, the plaintiffs had the burden of proving that the district court abused its discretion when it vacated paragraphs 43, 58, 60(v)-(vi), 61(ii), 71(ii), and 78–84. See Cleveland Firefighters for Fair Hiring Practices, 669 F.3d at 740. The plaintiffs have not met that burden. We therefore reject their challenge to the court's order granting in part TennCare's Rule 60(b) motion to vacate the decree.

B.

Next, the plaintiffs challenge the district court's order granting TennCare's motion to vacate the decree under the sunset clause. See Consent Decree ¶ 113, at 54. Notably absent from the plaintiffs' challenge, however, is any effort to contest the district court's factual findings regarding TennCare's delivery of services to its enrollees. Those findings came after the court heard 18 days of testimony and considered hundreds of exhibits; and the findings include that TennCare reminds parents and guardians in virtually every way imaginable to obtain screens for their children free of charge; that TennCare is a national leader in implementing programs to ensure its compliance with federal law; that TennCare has dramatically improved its provision of services since 1998; and that no other state's Medicaid program surpasses TennCare in any salient respect.

What the plaintiffs argue, rather, is that the district court should have considered more evidence than it did during the evidentiary hearing, and that the court misinterpreted certain paragraphs of the decree. We consider these arguments in turn.

1.

The plaintiffs argue that the district court improperly refused to consider TennCare's past violations of the consent decree when it granted the motion to vacate. It is true, of course, that the district court told the parties that the scope of its 18–day evidentiary hearing was limited to whether TennCare was in “current, substantial compliance” with the decree. We review that limitation for an abuse of discretion. See Gonzales v. Galvin, 151 F.3d 526, 534–535 (6th Cir.1998). And that limitation was plainly within the court's discretion. Present compliance, not past, was the ultimate issue before the court; and in deciding that issue the court was not required to conduct the equivalent of an archaeological dig. Moreover, the plaintiffs' argument is incorrect even when considered on its own terms. The district court did consider the case's long history—which is why the court took the trouble to become familiar with it. And the court's opinion expressly stated that TennCare had violated the decree in 2001 and that, as a result, TennCare now bore “the burden of proving ... that [it was] in substantial compliance with the [d]ecree.” See John B. v. Emkes, 852 F.Supp.2d 957, 961 (M.D.Tenn.2012).

The plaintiffs further contend that the court's decision to limit the scope of the hearing prevented them from impeaching TennCare's witnesses with evidence of past noncompliance with the decree. But the district court allowed the plaintiffs to do precisely that. For example, the court allowed the plaintiffs to cross-examine Dr. Wendy Long about TennCare's problems tracking referrals—a problem that reached back to 2005. Thus, this argument simply mischaracterizes what happened in the district court.

The plaintiffs also contest several of the court's evidentiary decisions, which we likewise review for an abuse of discretion. See United States v. Stepp, 680 F.3d 651, 660 (6th Cir.2012). First, the plaintiffs argue that the court wrongly excluded a group of 100 documents on the ground that they were “too old.” But again the plaintiffs mischaracterize the court's decision. The court excluded those documents not because they were old, but because the plaintiffs tried to admit them en masse, at the end of the hearing, when TennCare's witnesses could no longer respond to them. Second, the plaintiffs contend that the court excluded as “dated” the minutes from a 2008 meeting of the Tennessee Children's Care Coordination Steering Panel. Yet the plaintiffs fail to mention that the court also based that decision on hearsay grounds: the plaintiffs offered the document for the truth of the matter asserted, and failed to identify an applicable hearsay exception. They have not done any better here, so the document was inadmissible in any event. Third, the plaintiffs object to the court's exclusion of certain reports that court-appointed monitors wrote about TennCare in 2007. True, the court did exclude these reports as beyond the scope of the hearing; but the reports were already part of the record in the case, so they did not need to be admitted during the hearing for the court to consider them. The plaintiffs' evidentiary arguments are meritless.

2.

The plaintiffs next challenge the merits of the district court's order to vacate the decree under the sunset clause. That clause provides, in relevant part, that the decree “shall expire” when TennCare reaches an “adjusted periodic screening percentage” of 80% and is in “current, substantial compliance” with the decree's other requirements. See Consent Decree ¶ 113, at 54. The district court found that TennCare had achieved both goals, so it vacated the decree. In doing so, the plaintiffs contend, the court misinterpreted the decree.

We review the district court's interpretation of the consent decree de novo. See Sault Ste. Marie Tribe of Chippewa Indians v. Granholm, 475 F.3d 805, 810 (6th Cir.2007). For these purposes, we interpret the consent decree as a contract. See id. And under Tennessee law, which guides our interpretation of the decree here, our primary goal is to give effect to the parties' intent as expressed in the decree itself. See DePasquale v. Chamberlain, 282 S.W.3d 47, 53 (Tenn.Ct.App.2008).

a.

The plaintiffs first argue that the district court misinterpreted paragraph 39 of the decree, which describes TennCare's outreach obligations. That paragraph provides that Tennessee “shall adopt any policies and procedures necessary to ensure that TennCare rules and guidelines ... require compliance with ... each specific outreach and informing requirement under federal law [.]” (Emphasis added.) The district court held that TennCare had in fact adopted such “policies and procedures,” and thus had complied with paragraph 39. See John B., 852 F.Supp.2d at 976.

But the plaintiffs contend that paragraph 39 requires more than the adoption of certain policies. Rather, they say, TennCare must show that its outreach efforts are actually “effective.” But that is not what the decree says. Paragraph 39 says that TennCare must adopt policies and procedures that themselves require compliance with federal outreach requirements. TennCare has done so; and that means it has complied with paragraph 39.

b.

The plaintiffs also argue that the district court misinterpreted paragraphs 53 and 54, which describe TennCare's diagnostic and treatment obligations. Paragraph 53 provides that

[TennCare] shall establish and maintain a process for reviewing the practices and procedures of [its managed-care organizations and the Department of Children's Services], and require such modifications of those practices and procedures as are necessary to ensure that children can be appropriately referred from one level of screening or diagnosis to another, more sophisticated level[.]
Paragraph 54 provides that “[TennCare] shall ensure that, within their respective spheres of responsibility, TennCare, [its managed-care organizations] and [the Department of Children's Services] provide children all medically necessary ... services[.]” In sum, these two paragraphs require that TennCare provide to its enrollees all medically necessary diagnostic and treatment services.

According to the plaintiffs, the district court held that TennCare complied with these paragraphs because it “adopt[ed] policies that require [its managed-care organizations] to provide [the necessary] services”—what they refer to as a “have-a-policy” standard. But again the plaintiffs misrepresent the court's holding. The court did not say that TennCare had complied with paragraph 53 simply because TennCare “had a policy.” Instead, the court held that TennCare had actually reviewed its contractors' practices and ordered the contractors to change them when necessary. See John B., 852 F.Supp.2d at 979. For example, the court found that Qsource—TennCare's External Quality Review Organization—regularly reviews the policies and practices of the managed-care organizations. Qsource then works with TennCare's Quality Oversight Unit “to identify opportunities for improvement and to develop a quality strategy[.]” Id. at 973.

Similarly, the court did not hold that TennCare had complied with paragraph 54 merely because it adopted certain policies and procedures. True, the court reviewed TennCare's policies and found that its enrollees “[were] entitled to receive, free of charge, all medically necessary covered diagnosis and treatment services.” Id. at 970. But more to the point, the court examined how TennCare provided those services “[i]n practice[.]” See id. And the court found, for example, that “the vast majority of diagnosis and treatment services are provided to TennCare enrollees automatically, without any medical-necessity review[.]” Id. at 970.

The court's conclusion that TennCare complied with paragraphs 53 and 54, therefore, was based on its finding that TennCare actually provides to its enrollees the services that those paragraphs require TennCare to provide. And in making that finding, the court did not—as the plaintiffs assert here—“brush aside undisputed evidence” that TennCare had failed to comply with the decree. Instead, the court considered the plaintiffs' evidence and found it unpersuasive. Indeed, the court found that the plaintiffs' own witnesses “largely confirmed that TennCare provides medically necessary diagnostic and treatment services.” Id. at 985. And the court found that the plaintiffs had not “identif[ied] any instance”—not a single one—“where needed services were not ultimately provided.” Id. at 985. The plaintiffs have not even argued, much less proved, that these findings were clearly erroneous. The district court did not err in any respect in concluding that TennCare had complied with paragraphs 53 and 54 of the decree.

c.

The plaintiffs next argue that the district court misinterpreted paragraph 96, which provides that “[TennCare] shall establish ... an ongoing process for monitoring and reporting [its] compliance with the [decree's] requirements[.]” The court held that TennCare had complied with this paragraph in two ways. First, TennCare used several third-party measurements to track its performance. See id. at 971–74. For example, TennCare required that its managed-care organizations obtain accreditation from the National Committee for Quality Assurance, which is a nonprofit organization dedicated to improving the quality of health care. Second, TennCare adopted several internal reporting methods, including the filing of semiannual compliance reports with the court and the plaintiffs. See id. at 974–75.

The plaintiffs contend that none of these monitoring processes “fully or adequately assess [TennCare's] performance.” They therefore accuse the district court of “sidestepp[ing] the evidentiary record” and “effectively read [ing] out of the [d]ecree the requirement that [TennCare] specifically monitor its actual provision of [medical] services.” (The latter point is the supposed misinterpretation of the paragraph.) But these assertions again mischaracterize the court's decision. The court did not sidestep the evidentiary record; instead it sifted through the record to make five pages' worth of detailed findings about TennCare's monitoring processes. And the court did not read the monitoring requirement out of the decree. Instead it concluded—based upon factual findings that the plaintiffs do not venture to challenge here—that TennCare met that requirement. The plaintiffs' argument that the court misinterpreted paragraph 96 is meritless.

d.

Finally, the plaintiffs argue that the district court misinterpreted paragraph 46. That paragraph describes a three-step process that TennCare uses to calculate the “adjusted periodic screening percentage” mentioned in the decree's sunset clause. First, TennCare calculates a “screening ratio [,]” which is the number of “periodic screens” that TennCare provided in the past year, divided by “the number of ... screens that should have occurred” during that time. Second, TennCare multiplies the screening ratio by 100 to get a “periodic screening percentage.” Third, TennCare adjusts that screening percentage by conducting a “medical record review.” Only the first step of this process is at issue here.

In 2010, TennCare reported an adjusted periodic screening percentage of 91.3%, well above the 80% target in the sunset clause. TennCare therefore argued during the evidentiary hearing that it had satisfied the sunset clause's requirement. The plaintiffs responded that TennCare had improperly calculated both the numerator and denominator of the screening ratio. The district court admitted expert testimony from both sides and held that TennCare had calculated the ratio correctly.

On appeal, the plaintiffs first object to TennCare's method of counting the number of screens that it provided its enrollees in the past year ( i.e., the screening ratio's numerator). To determine that number, TennCare relies on a set of codes that doctors use to describe and bill for the services they provide. Doctors use several codes to bill TennCare for the screens they perform. Thus, every time that TennCare receives a bill using one of these screening codes, TennCare counts it toward the numerator.

The plaintiffs argue that this counting method inflates the numerator. They say a checkup should only count as a “periodic screen” if it was required by TennCare's periodicity schedule, which states how often each child should receive a screen. Thus, the plaintiffs contend, the periodicity schedule should limit the number of screens that TennCare can count per child. For example, if a child received 10 screens in a year where the periodicity schedule required only seven, then TennCare should only count the seven screens towards the screening ratio.

The problem with this argument is that paragraph 46 expressly refutes it. That paragraph lists several billing codes that “will be the primary determinants of which [checkups] are counted as periodic screens.” And TennCare uses precisely those billing codes when counting the number of screens it provided in a particular year—which means that it calculates the screening ratio's numerator in precisely the manner that the decree says it should. Moreover, paragraph 46 says nothing about using the periodicity schedule to limit the number of screens that TennCare may count per child. Finally—and germane to the question of the parties' intent—the plaintiffs do not explain how, as a practical matter, TennCare could even apply such a limit, given that the billing codes themselves say nothing about whether a screen was required by the periodicity schedule.

That said, the plaintiffs try to bolster their argument in two ways. First, they contend that TennCare's calculation method conflicts with the instructions for the CMS–416 form, which TennCare uses to report its screening ratio to the federal government. But those instructions do not apply to calculation of the numerator. Although paragraph 46 incorporates them for certain purposes, it does not do so when explaining how to count the number of screens that TennCare performed. Second, the plaintiffs point to a decision from the Northern District of Illinois, which held that a counting method similar to TennCare's was “misleading” and “overstate[d] the actual level of ... services provided.” Memisovski ex rel. Memisovski v. Maram, No. 92 C 1982, 2004 WL 1878332, *28, *53 (N.D.Ill. Aug. 23, 2004). But that case is inapposite for the simple reason that it did not involve the decree that we construe here. The plaintiffs' objections to TennCare's calculation of the screening ratio's numerator, therefore, are meritless.

The plaintiffs also object to TennCare's method of counting the number of periodic screens that should have occurred during the past year ( i.e., the screening ratio's denominator). To determine that number, TennCare uses a formula again found in paragraph 46 of the decree. That formula incorporates TennCare's periodicity schedule—the more screens the schedule requires, the larger the screening ratio's denominator.

In 2010, TennCare used a periodicity schedule that required 30 total screens for outreach purposes. When calculating its screening ratio, however, TennCare used a schedule that required only 24 screens. The plaintiffs argue that TennCare's use of the 24–screen schedule artificially decreased the number of screens that should have occurred in the past year, thereby increasing the screening ratio.

To determine which periodicity schedule TennCare should have used when calculating the screening ratio, we start with the text of the decree. Paragraph 46 begins by saying that TennCare must use “HCFA 416 mathematical methodology” to determine “the number of periodic screens that should have occurred” in the baseline year. (HCFA 416 mathematical methodology was a calculation method promulgated by the Health Care Financing Administration. That method is now promulgated by the Centers for Medicare & Medicaid Services and is known as CMS–416 methodology.) Paragraph 46 then says that periodic screening percentages in later years “will be calculated using methodology identical to that used in calculation of the baseline periodic screening percentage.”

The parties disagree about the meaning of the words “methodology identical” as used in this paragraph. The plaintiffs contend that the methodology used to calculate the screening ratio in a particular year must be “identical” to the “HCFA 416 mathematical methodology” for that year. In contrast, TennCare argues that the words “methodology identical” require it to use a periodicity schedule “identical” to the one it used in the baseline year ( i.e., October 1, 1995 through September 30, 1996).

The plaintiffs have the better reading of this paragraph. The periodicity schedule is merely an input for the screening-ratio methodology, not a part of the methodology itself. Thus, when TennCare changes its periodicity schedule, the denominator of its screening ratio should reflect that change. Moreover, TennCare has not consistently followed its own proposed interpretation of “methodology identical”: in the baseline year, TennCare used a 20–screen schedule to calculate its screening ratio, but in 1999 it began using a 24–screen schedule. If paragraph 46 actually required TennCare to use the same periodicity schedule that it used in the baseline year, TennCare presumably would not have made that change.

Paragraph 46 therefore requires that TennCare follow the CMS–416 methodology each year. That methodology is explained in the instructions to the CMS–416 form. In relevant part, those instructions require that TennCare use its “most recent periodicity schedule” when it completes the form. The issue therefore becomes what the instructions mean by “most recent periodicity schedule[.]”

States use the CMS–416 form to comply with 42 U.S.C. § 1396a(a)(43)(D), which directs each state to “report[ ] to the Secretary [of Health and Human Services] ... information relating to early and periodic screening, diagnostic, and treatment services provided under the [state's] plan [.]” Section 1396d(r), in turn, defines the phrase “early and periodic screening, diagnostic, and treatment services” to include screening services that are provided “at intervals which meet reasonable standards of medical ... practice, as determined by the State after consultation with recognized medical ... organizations involved in child health care[.]” 42 U.S.C. § 1396d(r)(1)(A)(i). Thus, when the CMS–416 instructions refer to the “most recent periodicity schedule,” they refer to the schedule that the State adopted “after consultation with recognized medical ... organizations involved in child health care[,]” in compliance with § 1396d(r).

Here, the district court found that TennCare had adopted the 30–screen schedule to comply with this part of § 1396d(r). See John B., 852 F.Supp.2d at 967;see also TennCare Rule 1200–13–13–04(b)(8) (adopting the “latest” periodicity schedule recommended by the American Academy of Pediatrics, which currently requires 30 total screens). Thus, when TennCare calculated its screening ratio, it should have used a 30–screen periodicity schedule rather than a 24–screen one. The court therefore misinterpreted the portion of paragraph 46 that prescribes the denominator of the screening ratio.

III.

And so the district court, in the course of making literally dozens of interpretive decisions with respect to a notoriously complex statute and decree, made a single technical mistake. We now consider whether that error was harmless. SeeFed.R.Civ.P. 61. An error is harmless if it “do[es] not affect any party's substantial rights.” Id.

TennCare's primary argument in the district court was that it had satisfied the terms of the consent decree's sunset clause. In the alternative, however, TennCare argued that its program had “complie[d] fully with the governing provisions of the Medicaid statute[,]” and that the district court should vacate the decree on that basis. In support, TennCare cited the Supreme Court's decision in Horne v. Flores, 557 U.S. 433, 129 S.Ct. 2579, 174 L.Ed.2d 406 (2009). There, the Court held, in determining whether to grant a Rule 60(b)(5) motion in institutional litigation, that the district court and court of appeals alike must determine whether “ongoing enforcement of the original order [is] supported by an ongoing violation of federal law [.]” Id. at 454, 129 S.Ct. 2579. Thus, we think it fair to construe TennCare's alternative argument as one for relief under Rule 60(b)(5) on the ground that ongoing enforcement of the decree would not remedy an ongoing violation of federal law. Moreover, we can affirm on any basis supported by the record, see Taylor v. KeyCorp, 680 F.3d 609, 616 (6th Cir.2012); so we consider whether the Supreme Court's decision in Horne requires affirmance here.

Under Rule 60(b)(5), a party can ask a court to vacate a consent decree “if a significant change ... in factual conditions ... renders [its] continued enforcement detrimental.” Northridge Church, 647 F.3d at 613 (quotation marks omitted). This rule “serves a particularly important function in ... institutional reform litigation.” Horne, 557 U.S. at 447, 129 S.Ct. 2579 (quotation marks omitted). In such cases, we must take a “flexible approach” to these motions so that “responsibility for discharging the State's obligations is returned promptly to the State and its officials when the circumstances warrant.” Id. at 450, 129 S.Ct. 2579 (quotation marks omitted).

In applying this flexible approach, we must answer two questions: first, whether the state has achieved compliance with the federal-law provisions whose violation the decree sought to remedy; and second, whether the State would continue that compliance in the absence of continued judicial supervision. See id.; Bd. of Educ. of Oklahoma City Pub. Sch., Indep. Sch. Dist. No. 89, Oklahoma County, Okl. v. Dowell, 498 U.S. 237, 247, 111 S.Ct. 630, 112 L.Ed.2d 715 (1991). If the State has indeed implemented a “durable remedy[,]” then “continued enforcement of the [decree] is not only unnecessary, but improper.” Horne, 557 U.S. at 450, 129 S.Ct. 2579.

Here, the district court found that TennCare has achieved compliance with all of the provisions of federal law whose violation the decree sought to remedy. To review: TennCare “provid[es] or arrang[es] for the provision of ... screening services in all cases where they are requested” in compliance with 42 U.S.C. § 1396a(a)(43)(B). TennCare provides all four types of screens required by the Medicaid Act: physical, vision, hearing, and dental. See42 U.S.C. § 1396d(r). TennCare has also adopted for outreaches purposes the periodicity schedules recommended by a committee of experts in each field. See, e.g., id. § 1396d(r)(1)(A)(i). And TennCare provides these screens free of charge to its enrollees whenever they request them.

TennCare also “arrang[es] for ... corrective treatment the need for which is disclosed by such child health screening services” in compliance with section 1396a(a)(43)(C). TennCare enrollees are entitled to receive all medically necessary diagnostic and treatment services. When these services are ordered by a licensed provider, TennCare provides most of them without engaging in a medical-necessity review. When TennCare does perform that review and denies a request, it has an exhaustive appeals process available if a family wishes to take advantage of it. This process works well: during the evidentiary hearing, the plaintiffs failed to “identify any instance where needed services were not ultimately provided.” See John B., 852 F.Supp.2d at 985.

In addition, TennCare “inform[s] all persons in [Tennessee] who are under the age of 21 and who have been determined to be eligible for [Medicaid], of the availability of early and periodic screening, diagnostic, and treatment services” in compliance with section 1396a(a)(43)(A). TennCare contacts all newly enrolled families at least five times to tell them about the program and to encourage them to schedule a screening appointment. TennCare reminds each family about these services four times a year in quarterly newsletters and once per year in a postcard. If a child goes a year without a screen, TennCare sends three reminder notices and sends a community-outreach worker to the child's home to urge the family to schedule an appointment. On this point the law does not require anything further: nine reminders to bring a child in for a screen are enough.

In summary, the court found that, “compared with its performance in 1998, TennCare ha[d] dramatically improved the provision of medical services to its enrollees in every respect”; that “no other state's ... program surpasses that of Tennessee in any salient respect”; and that TennCare is “fully compliant with the [relevant Medicaid] law and regulations.”

The court also found that TennCare will continue to comply with federal law in the absence of judicial supervision. During the 18–day evidentiary hearing, TennCare's director testified that TennCare will continue to use “independent, nationally recognized third-party monitoring and oversight tools ... to ensure that children continue to receive the services to which they are entitled under federal law.” John B., 852 F.Supp.2d at 975. The district court found this testimony credible because TennCare used those same monitoring tools for adults, even though federal law does not require it to do so. See id. The court also found credible the director's testimony that TennCare has no plans to cut services to its enrollees. See id. at 970.

The court also rejected the plaintiffs' arguments that TennCare would stop complying with federal law if the decree were vacated. The plaintiffs had argued, for example, that TennCare might eliminate the Quality Services Review process, which it used to comply with paragraph 88 of the decree. The court acknowledged that theoretical possibility, but pointed out that “[TennCare's] obligation, going forward, is not to remain in compliance with each precise term of the [d]ecree[.]” Id. at 983. Rather, the court said, TennCare had to “remain in compliance with federal law[.]” Id. And other than “speculation as to what the future holds[,]” the court saw no reason to believe that TennCare would fail to remain in compliance. See id. Neither do we. The record shows, instead, that TennCare has implemented a durable remedy for its past violations of the Medicaid Act.

In Horne, the Supreme Court held that, in determining whether to terminate a consent decree, the courts must go beyond “an inquiry into whether the original order [ i.e., the decree] ha[s] been satisfied.” 557 U.S. at 454, 129 S.Ct. 2579. Instead, the Supreme Court said, a “Court of Appeals need[s] to ascertain whether ongoing enforcement of the original order [is] supported by an ongoing violation of federal law[.]” Id. Here, the district court's findings make clear that TennCare has implemented durable remedies to comply with the provisions of federal law that the decree was intended to enforce. Upon this record, therefore, “continued enforcement of the [decree] is not only unnecessary, but improper.” Horne, 557 U.S. at 450, 129 S.Ct. 2579.

TennCare's mistake with respect to the calculation of the denominator for its screening ratio does not change this result. That mistake at most amounts to a technical violation of 42 U.S.C. § 1396a(a)(43)(D); and the plaintiffs themselves argued in the district court that none of the consent decree's provisions were based on subsection 43(D), and indeed that subsection 43(D) is not enforceable under § 1983 in any event. See John B., 852 F.Supp.2d at 947. Thus, TennCare's violation of subsection 43(D), to the extent there is one, does not provide us with any lawful basis to continue enforcement of the decree. Instead, given the district court's undisputed findings with respect to the sincerity of TennCare's directors, we are confident that TennCare will promptly remedy any technical violation of subsection 43(D) on its own.

Finally, our decision in Gonzales v. Galvin, 151 F.3d 526 (6th Cir.1998) is not contrary to our decision here. For two reasons: first, to the extent of any conflict between the two cases, Horne trumps Gonzales; and second, Gonzales concerned termination of a consent decree sua sponte, rather than by motion.

The district court's error thus did not affect the substantial rights of the parties, and was harmless.

* * *

The district court's handling of this case after our remand last year was exemplary. The court conducted an exhaustive evidentiary hearing, reviewed 345 pages of proposed findings of fact and conclusions of law from the parties, and familiarized itself with thousands of pages of evidence already in the record. And on the basis of all of that evidence, the court found, in a thorough and carefully reasoned opinion, that TennCare had vastly improved its delivery of services to enrollees, and indeed become a national leader in its compliance with the Medicaid statute.

The court's conclusions were sound. Its judgment is affirmed.