HUMANA INC v. DEPARTMENT OF HEALTH AND HUMAN SERVICES (2025)

MEMORANDUM ORDER AND OPINION

Before the Court are Plaintiffs Humana and Americans for Beneficiary Choice's Motion for Summary Judgment, filed July 29, 2025 (ECF No. 16), Defendant Centers for Medicare and Medicaid Service's Cross-Motion for Summary Judgment, filed August 19, 2025 (ECF No. 33), Plaintiffs’ Combined Response and Reply, filed September 5, 2025 (ECF No. 35), and Defendants’ Reply, filed September 19, 2025 (ECF No. 37). After considering the motions, the briefs, the administrative record, and the applicable law, the Court DENIES Plaintiffs’ motion and GRANTS Defendant's Motion.

I. BACKGROUND

A. Statutory Background

Title XVIII of the Social Security Act, 42 U.S.C. § 1395 (the “Medicare Act”), establishes the Medicare program, a federally funded and administered health insurance program for eligible elderly and disabled persons and certain individuals with end-stage renal disease. See 42 U.S.C. § 1395c. The Secretary administers the Medicare program through Centers for Medicare and Medicaid Service (“CMS”), a component agency of the United States Department of Health and Human Services. Under the Medicare Advantage program, known as Medicare Part C and D, private insurers, known as Medicare Advantage Organizations (“MAOs”), provide coverage that beneficiaries would otherwise receive from the government. See 42 U.S.C. § 1395w-21; § 1395w-101. Humana is an MAO.

CMS rates the quality of Medicare Advantage Plans through its Star Ratings system. 42 U.S.C. §§ 1395w-22(e); 1395w-23(o)(4)(A). The Star Ratings system rates MAO plans on a one-to-five-star scale with five reflecting the highest quality and one reflecting the lowest. CMS analyzes about three dozen quality metrics to calculate a plan's Star Rating. CMS assigns a score of one-star to five-stars for each metric. CMS then averages the star scores across these metrics to give each plan an overall rating. CMS publishes the overall ratings every October to help prospective enrollees compare plans and make an informed decision. The Star Ratings effect the number of enrollees, drive so-called quality bonus payments, revenue, and overall market share. See 42 U.S.C. §§ 1395w-23(o)(1); 1395w-24(b)(1)(C); 1395w-24(b)(1)(C)(v); 42 C.F.R. §§ 422.260; 422.266(a)(2)(ii). If plans score below a 2.5-star rating on multiple contracts, they may be ineligible to provide Medicare Advantage. 42 C.F.R. § 422.502(b)(1)(i)(D).

CMS creates a guidance document known as the Technical Notes to help MAOs understand how they are analyzed.¹ Each year, CMS distributes these Technical Notes to MAOs and publishes them on its website. Among other things, these Technical Notes inform the MAOs about the measures that comprise that year's Star Ratings, how those measures will be weighted, and how CMS assesses each measure.² MAOs may comment on the prospective Star Rating details before that year's Technical Notes are finalized. See 42 C.F.R § 422.164(c), (d); see also 42 U.S.C. § 1395w-23(b)(2).

One way the Star Ratings evaluate MAO call centers is the interpreter availability quality metric. This metric evaluates how well an MAO's call center processes incoming calls from enrollees who have limited English language proficiency. Specifically, the metric “ensure[s] that interpreters are available within 8 minutes of the caller reaching a customer service representative.”³

As explained by the agency's guidance in the Technical Notes, CMS collects data for this metric by placing foreign-language test calls during its Accessibility and Accuracy study. These tests calls are rated as successful or unsuccessful. A test call is successful only if the call center completes four stages—1) dial, 2) connect, 3) introductory question, 4) accuracy—on a single call.⁴ The guidance does not explicitly say that callbacks are not allowed.

A call is unsuccessful when, among other things, (1) the center cannot connect to a live customer service representative during the call; (2) the representative cannot assist with the questions or cannot forward to someone who can assist in answering questions; or (3) the call disconnects before the call center completes all stages of a call.⁵ CMS has discretion to invalidate a test call that would be categorized as unsuccessful if it determines that CMS made the mistake: for example, if CMS dials an incorrect number or called after the call center was closed because of a time zone difference.⁶ In order for an MAO to receive a 5.0 Star Rating for its interpreter availability measure, 100% of the non-invalidated foreign-language tests calls must be successful.⁷

CMS also has minimum standards that MAOs must maintain to participate in the Medicare Advantage program. 42 C.F.R. § 422.502(b)(1)(i)(D). One standard requires interpreters to be “available for 80 percent of incoming calls [that] requir[e] an interpreter within 8 minutes of [the caller] reaching the customer service representative.” 42 C.F.R. §§ 422.111(h)(1) (iii); 423.128(d)(1)(iii). CMS gathers data to ensure compliance with its minimum standards through test calls conducted during its annual Accuracy and Accessibility Study.

B. Factual Background

Plaintiff Humana disputes its 2025 Star Ratings. Specifically, Humana challenges the 3.5 star-rating it received for its call center's performance processing calls from enrollees with limited English language proficiency. Humana received a rating of 3.5 stars because CMS ranked three interpreter availability calls as unsuccessful. For two of these calls, D1100955 and D0900533, the CMS caller connected to a customer service representative, but Humana's call center dropped the call before an interpreter could join.⁸ Humana treats customer calls and CMS test calls differently. For customer calls, Humana representatives are instructed to callback if the customer's call drops. For a CMS test call, Humana does not attempt to callback after a dropped call because CMS has stated that callbacks are not allowed.⁹ Accordingly, Humana did not attempt to call back for either of the two dropped CMS test calls.¹⁰

The third call, C0701002, was completely silent, meaning neither Humana's representative nor the CMS test caller spoke. The CMS test caller believed he was on a silent hold.¹¹ Humana hung up on the caller after about six minutes.¹²

Humana brought challenges to the phone calls last year while it also pursued an administrative appeal with CMS. This Court dismissed the lawsuit for failing to exhaust all administrative remedies.¹³ After exhausting all administrative remedies, Humana returns to challenge the agency's characterization of the three calls. It argues that the agency's decision to rate the three calls as unsuccessful was arbitrary and capricious. The Parties moved for cross motions on summary judgment.

II. LEGAL STANDARD

In a case challenging an agency action under the APA, disputes “are commonly resolved on summary judgment, where district courts sit as an appellate tribunal to decide legal questions on the basis of the administrative record.” Nat'l Ass'n for Gun Rights, Inc. v. Garland, 741 F.Supp.3d 568, 596 (N.D. Tex. 2024) (O'Connor, J.) (citing Amin v. Mayorkas, 24 F.4th 383, 391 (5th Cir. 2022)). The district court applies the APA's “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law” standard to determine whether “the evidence in the administrative record permitted the agency to make the decision it did.” 5 U.S.C. § 706(2); MRC Energy Co. v. U.S. Citizenship & Immigr. Servs., No. 3:19-cv-2003-K, 2021 WL 1209188, at *3 (N.D. Tex. Mar. 31, 2021) (citation omitted).

III. ANALYSIS

Humana challenges the agency's decision to characterize the three calls as unsuccessful as a pure question of pure law. Specifically, Humana argues that the decisions “contradicted [CMS's] governing regulations and guidance” and “were otherwise arbitrary and capricious.”¹⁴

Humana requests that the Court set aside the 3.5 Star Rating it received and remand back to the agency for recalculation.¹⁵ It further asks the Court to set aside the no-callback policy as unlawful.¹⁶

A. The no-callbacks policy is not unlawful.

Humana argues the unsuccessful rating for the two dropped calls depended on a no-callback rule that 1) CMS implemented unlawfully without notice-and-comment required under 5 U.S.C. § 553, 42 U.S.C. § 1396hh(a), and § 1395w-23(b)(2) and 2) is substantively arbitrary.¹⁷ Humana is incorrect. The no-callback policy is lawful and clear from the guidance.

1. 42 U.S.C. 1395hh(a)(2) does not apply to the no-callback policy.

Humana argues that the no-callback policy is a substantive legal rule requiring notice-and-comment rulemaking under 42 U.S.C. § 1395hh(a)(2).¹⁸ Section 1395hh(a)(2) requires notice and comment for a “rule, requirement, or other statement of policy ․ that establishes or changes a substantive legal standard governing the scope of benefits, the payment for services, or the eligibility of individuals, entities, or organizations to furnish or receive services or benefits under” the Medicare statute.

First, Humana argues that no-callback policy is a substantive rule because it “defines” the “duties” of MAOs to provide an interpreter within a single call under the minimum standards requirements.¹⁹ Allina Health Services v. Price, 863 F.3d 937, 943 (D.C. Cir. 2017), aff'd, 587 U.S. 566, 139 S.Ct. 1804, 204 L.Ed.2d 139 (2019) (quoting Black's Law Dictionary (10th ed. 2014)). But Humana brings a challenge under the Star Ratings—not a compliance challenge under the minimum standards regulation.

Humana argues that 42 U.S.C. § 1395hh(a)(2) nevertheless applies to the Star Rating quality metrics such as the no-callback policy in the interpreter availability metric.²⁰ This is because the Star Ratings “govern the scope of benefits, the payment for services, [and] eligibility of” MAOs “to furnish benefits” under the Medicare statute. Namely, the Star Ratings influence quality bonus payments, rebates, and when a plan may begin enrolling beneficiaries and, if an MAO receives 2.5 stars or below for two consecutive years, CMS may deny its application to participate in the Medicare Advantage program. 42 U.S.C. § 1395hh(a)(2); 42 C.F.R. § 422.62(b)(15); id. § 422.502(b)(1)(i)(D). Humana is incorrect for several reasons.

First, quality bonus payments and rebates are not “payments for services.” Under § 1395, payments for services refers to the amounts Medicare must pay providers for furnishing covered services. See 42 U.S.C. § 1395w-22(a)(B)(iv) (describing “services” as including chemotherapy, renal dialysis, skilled nursing care, clinical diagnostic laboratory tests, drugs); see also id. § 1395x(u) (“ ‘provider of services’ means a hospital, critical access hospital, rural emergency hospital, skilled nursing facility, comprehensive outpatient rehabilitation facility, home health agency, hospice program.”). Regardless of how much money an MAO receives in quality bonus payments or in rebates because of its Star Rating, its obligations to pay providers of services remains the same. Thus, the Star Ratings do not affect “payment for services.”

Second, the enrollment period does not determine whether an MAO can participate in the Medicare Advantage program at all—only when prospective customers may enroll in their plan.

Third, the fact that the Star Ratings influence eligibility does not mean the Star Ratings metrics govern eligibility under § 1395hh. The Court “interpret[s] regulations in the same manner as statutes, looking first to the regulation's plain language.” Fort Bend Cnty. v. United States Army Corps of Engineers, 59 F.4th 180, 197 (5th Cir. 2023); D.C. Bd. Of Elections & Ethics v. D.C., 866 A.2d 788, 798 n.18 (D.C. 2005) (“In finding the ordinary meaning, ‘the use of dictionary definitions is appropriate in interpreting undefined statutory terms.’ ”). Section 1395hh applies to standards that “govern.” Influencing is not the same as governing. To “govern” means “to prevail or have decisive influence; control.”²¹ To “influence” means to “affect or alter by indirect or intangible means.”²²

The Star Ratings do not govern. The provisions stating that two consecutive years of a 2.5-or lower-Star Rating means CMS may—but is not required to—deny an MAO's application for participating in the Medicare Advantage Program. The discretionary component of the low-performing provision and indirect nature of the Star Ratings metrics mean they are too attenuated to qualify as “governing” eligibility for Medicare Advantage programs. It is only the minimum standards requirements that govern. See 42 C.F.R. § 422.502(b)(1)(i)(D). As explained below, the no-callbacks policy does not alter the minimum standards regulation. Therefore, 42 U.S.C. § 1395hh(a)(2) does not apply to the no-callback policy.

2. The APA notice-and-comment requirement does not apply to the no-callback policy.

Second, Humana argues that the no-callback policy needed to be promulgated through the APA's notice-and-comment rulemaking because “[a]s a practical matter, the no-callbacks rule amends 42 C.F.R. §§ 422.111(h)(1)(iii) and 423.128(d)(1)(iii) [the minimum standards regulations] with an additional obligation that requires MAOs to furnish a connection to an interpreter within eight minutes on a single call.”²³ See 5 U.S.C. § 553; Flight Training International v. FAA., 58 F.4th 234, 241 (5th Cir. 2023). In Humana's view, the Star Ratings are used to enforce the minimum standards regulations, and so any requirement for the Star Ratings is a requirement for minimum standards regulation.²⁴ Humana is incorrect.

The Star Ratings do not enforce the minimum standards. The Star Ratings assess quality beyond the minimum requirements and “compare a contract's performance on a variety of metrics against its peers.”²⁵ This is evident from the different standards required for an MAO to receive a 5.0 Star Rating and for an MAO to be compliant. To receive a 5.0 Star Rating, an MAO must receive a 100% successful ranking on all its non-invalidated foreign-language interpreter calls. By contrast, an MAO would not be subject to any consequences for non-compliance until fewer than 80% of incoming foreign-language calls did not connect to an interpreter. This means a call center could be compliant with the minimum standards requirement of providing an interpreter for 80% of incoming calls while receiving a Star Rating lower than 5.0. Thus, the Star Ratings do not enforce compliance with the minimum standards, and so the APA notice-and-comment requirement does not apply to the no-callback policy.²⁶

3. The no-callback policy did not require notice-and-comment under 42 U.S.C. 1395w-23(b)(2) because the policy is clear from the guidance.

Third, Humana contends that the no-callback policy unlawfully changed the Star Ratings methodology because 1) the language in the guidance does not specify that CMS will not accept callbacks and 2) CMS never announced or sought comments on a requirement that a caller be connected to an interpreter within a single call in the guidance used for updating Star Ratings methodologies.²⁷ 42 U.S.C. § 1395w-23(b)(2). Humana is incorrect because the no-callback policy is clear from the guidance.

The Court reviews the guidance according to the plain meaning of the text as it would for reviewing an agency regulation or statute. See United States v. Yeatts, 639 F.2d 1186, 1189 (5th Cir. 1981). The guidance for rating the interpreter availability measure consistently refers to a call in the singular: as examples, “a call is classified as unsuccessful for any of the following reasons,”²⁸ and a call center must “ensure CSRs stay on the telephone when a foreign-language interpreter joins the call.”²⁹ If the policy allowed callbacks, it would not use the singular form. The language of “joins the call” clearly indicates that all elements must be satisfied on a single call.

Further, the language also addresses the CMS test caller as “caller”—if Humana calls back, that person becomes the callee and is no longer the caller.³⁰

Additionally, Humana has a separate internal policy not to call back CMS test callers because CMS does not accept callbacks. Humana would not have this policy unless it knew CMS required an interpreter to be placed on the line in a single call. Humana's knowledge further demonstrates that the no-callback policy is evident from the guidance.

Accordingly, the no-callback policy is clear from the guidance and is not a substantive new rule requiring notice-and-comment under § 1395w-23(b)(2).

4. The no-callback policy is not substantively arbitrary.

Finally, Humana argues that the no-callback policy is substantively arbitrary.³¹ It is not.

A court must set aside an agency rule as arbitrary and capricious if it bears no “rational relationship” to the goals of the rule. 5 U.S.C. § 706(2)(A); Mexican Gulf Fishing Co. v. United States Department of Commerce, 60 F.4th 956, 973 (5th Cir. 2023). CMS argues that its no-callback policy is rationally related to the Star Ratings’ goal of evaluating the service quality provided by an MAO's call center. Specifically, it contends that answering questions in a single call demonstrates better service because it avoids the risks of inconveniently timed follow-up calls or an inability to reach those who dialed from private numbers.³² CMS has articulated a rational connection to satisfy arbitrary and capricious review. Humana's arguments to the contrary rest on its erroneous premise that the Star Ratings enforce the minimum standards requirements.

B. The government's determination of the calls was not arbitrary and capricious.

Humana argues that CMS's characterization of the three calls as unsuccessful contradicts its guidance and was “otherwise arbitrary and capricious.”³³ An agency's decision will be arbitrary and capricious if the agency relied on factors that Congress did not intend it to consider, failed to consider an important aspect of the problem, or offered an explanation counter to the evidence. Amin v. Mayorkas, 24 F.4th at 393 (citing Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983)). An agency's decision is also arbitrary if it contradicts the agency's regulations. See Encino Motorcars, LLC v. Navarro, 579 U.S. 211, 221-222, 136 S.Ct. 2117, 195 L.Ed.2d 382 (2016).

1. The outcome for the two dropped calls not arbitrary and capricious.

Humana's argument that the two dropped calls should be overturned rests on its argument that the no-callback policy is unlawful. Humana agreed that if it failed as a matter of law, the record supports finding that the decision is supported.³⁴ Because the no-callback policy is lawful, the agency's decision to rate the dropped calls as unsuccessful was not arbitrary and capricious.

2. The outcome for the silent call was not arbitrary and capricious.

Humana argues that the silent phone call's characterization as unsuccessful was contrary to the agency's guidance and therefore arbitrary and capricious. In support, it points to UnitedHealthcare Benefits of Texas v. Centers for Medicare and Medicaid Services et al., where a court overturned an unsuccessful test call rating because the CMS caller failed to ask the introductory question at stage three.³⁵ No. 6:24-cv-357-JDK, 2024 WL 4870771 (E.D. Tex. Nov. 22, 2024). Humana misreads the guidance and the precedent.

For an interpreter availability test call to be scored as successful, a call center must complete four stages—1) dial, 2) connect, 3) introductory question, 4) accuracy. Stage two requires the call center to connect the caller to a live customer service representative. Stage three requires the call center to provide an interpreter within eight minutes after the CMS test caller asks an introductory question.³⁶ The record does not show that Humana got to stage three. Rather, there is no proof that Humana connected the caller to a live customer service representative.³⁷ The CMS caller thought he was on a silent hold, and the call was silent for just over six minutes before Humana hung up on the caller.³⁸ Thus, the call was rated as unsuccessful at stage two, not at stage three as in UnitedHealthcare.

Humana further argues that because the call did not connect with an interpreter within the meaning of the guidelines, the call should be invalidated rather than ranked as unsuccessful.³⁹ But CMS's guidance states that “a call is classified as unsuccessful for any of the following reasons ․ [s]urvey could not continue, Call Center disconnected call (including hanging up).”⁴⁰ And, CMS may rate a call as “unsuccessful if we are not able to connect to a live CSR at the plan during that schedule call.”⁴¹

Because the CMS caller did not hear anything, CMS could not determine if it connected to a live CSR and could not confirm that the call connected for purposes of the guidelines. CMS thus ranked the call as unsuccessful in accordance with the agency's guidance at stage two. The agency's determination is in accordance with its guidelines and is not arbitrary or capricious.

IV. CONCLUSION

For the reasons set out above, CMS's Motion is GRANTED and Humana's Motion is DENIED.

SO ORDERED this 14th day of October, 2025.

FOOTNOTES

1.   Administrative Records, 113, ECF No. 26-1.

2.   See generally id. at 113–326.

3.   Id. at 351.

4.   Id., 36.

5.   Id. at 78.

6.   Administrative R., 102, ECF No. 26-1.

7.   Id. at 431.

8.   Earlier, Humana claimed that these calls were dropped because of third-party connectivity issues. See AR 372, ECF No. 26-1; Pl.’s Consol. Reply 25—26, Case No. 4:24-cv-1004-O. However, they now agree that the call center was responsible for dropping the calls. See Pl.’s Combined Resp. and Reply, 4, ECF No. 36.

9.   Administrative R., 26, ECF No. 26-1.

10.   CMS raises legitimate factual issues about whether Humana knew these calls were from CMS. Given the posture of cross motions for summary judgment, the Court infers that Humana did know the calls were from CMS. See Pl.’s App. 13—15, ECF No.19-2.

11.   Administrative R., 1, 454, ECF No. 26-1.

12.   Id. at 459.

13.   Order 5, ECF No. 56, Humana Inc. et al. v. U.S. Dep't. Health and Hum. Serv., No. 4:24-cv-1004-O (N.D. Tex. 2024) (O'Connor, J.).

14.   Pl.’s Br. Supp. Mot. Summ. J., 16, ECF No. 17.

15.   Pl.’s Br. Supp. Mot. Summ. J., 4. ECF No. 17.

16.   Id.

17.   Id. at 26.

18.   Id. at 19—22.

19.   Id. at 20.

20.   The Court does not address whether the no-callback policy would count as substantive for purposes of 42 U.S.C. § 1395hh(a)(2) because § 1395hh(a)(2) does not apply. That question is immaterial because, as explained below, the no-callback policy is clear from the text.

21.   Govern, merriamwebster.com, https://www.merriam-webster.com/dictionary/govern (emphasis added) (last visited Oct. 5, 2025).

22.   Influence, merriamwebster.com, https://www.merriam-webster.com/dictionary/influence (emphasis added) (last visited Oct. 5, 2025).

23.   Pl.’s Br. Supp. Mot. Summ. J., 23, ECF No. 17.

24.   Id. at 22.

25.   Def.’s Br. Supp. Summ. J., 32, ECF No. 34.

26.   Furthermore, even if the minimum standards requirement did apply, the language of the regulation clearly prohibits callbacks. Call centers must make an interpreter available “within 8 minutes” of an “incoming call.” 42 C.F.R. §§ 422.111(h)(1)(iii) (emphasis added.) “Incoming call” requires that interpreters connect the caller to an interpreter on the same call that they received. If a call center called back, it would be providing an interpreter on an outgoing call, not for the incoming call. See incoming, merriamwebster.com, https://www.merriam-webster.com/dictionary/incoming (emphasis added) (last visited Oct. 5, 2025) (defining “incoming” as “coming in; arriving”).

27.   Pl.’s Br. Supp. Mot. Summ. J., 24—26, ECF No. 17.

28.   Administrative R., 81, ECF No. 26-1.

29.   Id. at 48.

30.   See Caller, merriamwebster.com, https://www.merriam-webster.com/dictionary/caller, (defined as “one that calls”); Administrative R., 71, ECF No. 26-1 (“interpreters are available within 8 minutes of a caller reaching a CSR”) (emphasis added).

31.   Pl.’s Br. Supp. Mot. Summ. J., 26, ECF No. 17.

32.   Def.’s Reply, 41, ECF No. 37.

33.   Pl.’s Br. Supp. Mot. Summ. J., 16, ECF No. 17.

34.   Id. at 28 (“If a call center drops a call and fails to call back ․ the call of course should be scored unsuccessful.”)

35.   Pl.’s Br. Supp. Mot. Summ. J., 31, ECF No. 17.

36.   Administrative R., 36, ECF No. 26-1.

37.   Administrative R., 79, ECF No. 26-1.

38.   Id. at 459.

39.   Pl.’s Br. Supp. Mot. Summ. J., 31, ECF No. 17.

40.   Administrative R., 81, ECF No. 26-1.

41.   Id. at 78.

Reed O'Connor, CHIEF UNITED STATES DISTRICT JUDGE