SAYLOR, District Judge.
This is an action brought under the Administrative Procedure Act ("APA") challenging certain regulations promulgated by the Department of Health and Human Services ("HHS") under the Patient Protection and Affordable Care Act ("ACA"), Pub. L. No. 111-148, 124 Stat. 119 (2010).
Minuteman Health, Inc. is a nonprofit health-insurance provider that offered plans in Massachusetts in 2014, and in both Massachusetts and New Hampshire from 2015 to 2017. In 2014, it was required under HHS and Massachusetts regulations implementing the ACA's risk-adjustment program to pay 71% of its gross premium revenues to the program. In 2015, it was required to pay 40% of its New Hampshire revenues and 39% of its Massachusetts revenues. Perhaps unsurprisingly, it was not able to survive the loss of such a huge percentage of its revenues. It is now in receivership and is not offering plans to subscribers.
In substance, Minuteman challenges the HHS regulations that forced it to make those large transfer payments. It contends that the regulations at issue (1) were arbitrary and capricious, and therefore in violation of the APA, 5 U.S.C. § 706, and (2) were beyond HHS's statutory authority because they contravene the statute providing for risk adjustment, 42 U.S.C. § 18063.
The issues posed in this lawsuit are far from simple. The ACA is a notoriously complex statute, health insurance is notoriously difficult to administer effectively, and the federal health-care bureaucracy is notoriously cumbersome. The implementation of the statute and its regulations can hardly be called an unqualified success, and it appears to have triggered a host of unintended consequences. But the role of this Court is not to sit in judgment on the wisdom of the law, nor is it to judge the actions of HHS with the benefit of hindsight. Rather, it is to consider this specific challenge to certain regulations implemented under the act by HHS, and to analyze that challenge according to a specific legal framework: in essence, to determine whether HHS acted arbitrarily or unreasonably based on the record before it at the relevant time.
The essential facts are not disputed, and both parties have cross-moved for summary judgment. In substance, the Court concludes that HHS acted within the bounds of its authority, even when the consequences of its choices may not always have been optimal. Accordingly, and for the reasons set forth below, defendant's motion will be granted and plaintiff's motion will be denied.
The ACA was passed to regulate health insurance in the United States. Among other things, it "bars insurers from taking a person's health into account when deciding whether to sell health insurance or
Congress recognized, however, that prohibiting insurers from denying coverage to individuals based on their health status, combined with insurers' lack of knowledge of the health status of the anticipated new enrollees, would create a substantial risk of premium volatility. To alleviate the effects of that uncertainty, the ACA established three premium-stabilization programs, colloquially known as the "3Rs": the reinsurance, risk-corridors, and risk-adjustment programs. See generally 42 U.S.C. §§ 18061-18063.
The goal of the risk-adjustment program is to spread the costs of covering higher-risk members across insurers throughout a given state, thereby reducing incentives for insurers to engage in "risk-avoidance" techniques, such as designing or marketing their plans in ways that tend to attract healthier individuals, who cost less to insure. Mark A. Hall, Risk Adjustment Under the Affordable Care Act: Issues and Options 20 KAN. J.L. & PUB. POL'Y 222, 224 (2011). In broad terms, it requires issuers with healthier members to pay into the program, which in turn provides subsidies to issuers with less-healthy members.
The key provisions of the statute are contained within a single, short section. It provides that "each State shall assess a charge on health plans and health insurance issuers ... if the actuarial risk of the enrollees of such plans or coverage for a year is less than the average actuarial risk of all enrollees in all plans or coverage in such State for such year that are not self-insured group health plans," and, correspondingly, "each State shall provide a payment to health plans and health insurance issuers ... if the actuarial risk of the enrollees of such plans or coverage for a year is greater than the average actuarial risk of all enrollees in all plans and coverage in such State for such year that are not self-insured group health plans." 42 U.S.C. § 18063(a) (emphases added).
Congress delegated to HHS the responsibility for administering many of the programs under the ACA, including the risk-adjustment program. See id. § 18063(b) ("The Secretary [of Health and Human Services], in consultation with States, shall establish criteria and methods to be used in carrying out the risk-adjustment activities
Other parts of the ACA relevant to this action include the Consumer Operated and Oriented Plan Program (the "CO-OP" program), 42 U.S.C. § 18042, and the actuarial categorization of plans on the Health Benefit Exchanges, id. § 18022(d).
The CO-OP program, among other things, makes loans available to "qualified nonprofit health-insurance issuers" in order to encourage new entrants and bolster competition in the health-insurance market. Id. § 18042(b)(1); (Pl. Mem. in Supp. Summ. J. Ex. 11); see also 42 U.S.C. § 18042(c)(4) ("[A]ny profits made by the organization are required to be used to lower premiums, to improve benefits, or for other programs intended to improve the quality of health care delivered to its members."). CO-OP loan applicants must submit business plans to HHS, which, if approved, are incorporated into the final loan agreement. (Pl. Mem. in Supp. Summ. J. Exs. 11, 12). Congress appropriated $6 billion in the ACA to assist the launch of the CO-OP program. 42 U.S.C. § 18042(g).
The Health Benefit Exchanges are state-run insurance marketplaces created by the ACA to facilitate consumer choice and competition. 42 U.S.C. §§ 18031-18033. To allow consumers to compare products more easily, health plans sold on the exchanges are regulated in various ways. Id. §§ 18021-18024. Most relevant here, health plans sold on the exchanges are categorized as either catastrophic plans, which are only available to certain groups of enrollees, or one of four "metal levels": platinum, gold, silver, or bronze. Id. § 18022(d), (e). The metal levels correspond to the actuarial value of the plan — that is, the percentage of the full actuarial value of the benefits provided under the plan that the plan will actually cover. A platinum plan has an actuarial value of 90%, gold 80%, silver 70%, and bronze 60%. Id. § 18022(d).
As described by HHS, the risk-adjustment program "is intended to provide payments to health-insurance issuers that attract higher-risk populations, such as those with chronic conditions, and eliminate incentives for issuers to avoid higher-risk
HHS sets the risk-adjustment formula in advance of each benefit year through a notice-and-comment rulemaking process. See 45 C.F.R. §§ 153.100(b)-(c), 153.320. Although HHS goes through separate rulemakings for each benefit year, since 2014, the inaugural year of the program, "the record for each Annual Benefit Rule incorporates the records for each of the preceding Annual Benefit Rules." (See Index to the Rulemaking Record at 3, n.2). HHS sets the parameters ahead of the applicable benefit year, with the intention that insurers will be able to rely on the methodology to price their plans appropriately. Standards Related to Reinsurance, Risk Corridors and Risk Adjustment, 76 Fed. Reg. 41,930, 41,932-33 (July 15, 2011) (proposed rule); see also HHS Notice of Benefit and Payment Parameters 2018, 81 Fed. Reg. 94,058, 94,702 (Dec. 22, 2016) ("2018 Final Rule") (explaining the importance of setting rules ahead of time and describing comments supporting that practice).
To initiate the rulemaking process, HHS publishes a proposed Notice of Benefit and Payment Parameters ("NBPP"), including a proposed risk-adjustment formula, in November or December of the year two years prior to the applicable benefit year — thus, for example, for the 2014 benefit year, HHS issued the proposed NBPP on December 7, 2012. 2014 Proposed Rule, 77 Fed. Reg. 73,118. The public then has an opportunity to comment on the proposed rule. The final rule is published in February or March of the year prior to the applicable benefit year — for example, the final rule for the 2014 benefit year was published on March 11, 2013. 2014 Final Rule, 78 Fed. Reg. 15,410.
The original methodology embodied in the 2014 Benefit Rule is generally as follows.
Since 2011, when the planning for the 2014 benefit year began, HHS has treated the risk-adjustment program as "self-funded" or "budget-neutral," meaning that money collected from low-risk plans is the only source of funding for the payments to high-risk plans. (Def. Mot. for Summ. J. Ex. B at 13-16). Its transfer formula is accordingly designed so that the charges to plans with healthier members will equal the payments to plans with less-healthy members.
According to HHS, "[t]he risk-adjustment methodology addresses three considerations: (1) the newly insured population; (2) plan metal levels and permissible rating variation; and (3) the need for inter-plan transfers that net to zero.... Transfers depend not only on a plan's average risk score, but also on its plan-specific cost factors relative to the average of these factors within a risk pool within a state." 2014 Final Rule, 78 Fed. Reg. at 15,417. The rule "[a]djusts payment transfers for plan metal level, geographic rating area, induced demand, and age rating, so that transfers reflect health risk and not other cost differences." Id.
More specifically, the risk-adjustment model calculates the relative actuarial risk of each enrollee as follows.
To calculate the demographic and HCC numerical coefficients, HHS started with a database containing enrollee-specific clinical utilization and expenditures relating to more than 500 million claims from approximately 100 commercial health-insurance payers covering individuals living in all states, aged 0-64. 2014 Proposed Rule, 77 Fed. Reg. at 73,127. It used that data to calculate expenditures for each enrollee and adjusted the figure for metal level to arrive at a predicted plan liability for an enrollee with a given age, sex, HCC, and coverage level. Id. That data was then fed into a statistical regression to calculate the coefficients, which "represent the predicted relative incremental expenditures for each category or HCC." Id. at 73,130.
The sum of the demographic coefficient and the diagnostic coefficient(s) was then multiplied by a cost-sharing reduction ("CSR") adjustment factor. The CSR adjustment factor is designed to account for "increased plan liability due to increased utilization of health care services" by certain low-income and/or Native American enrollees who are eligible for premium subsidies. 2014 Final Rule, 78 Fed. Reg. at 15,421-22 & tbl.1; see 2014 Proposed Rule, 77 Fed. Reg. at 73,138; see also 42 U.S.C. § 18071. The resulting figure is an enrollee's "individual risk score." (Hearing Tr. 60:7-22); see 2014 Proposed Rule, 77 Fed. Reg. at 73,139.
The plan average risk score is essentially a member month-weighted average of the individual risk scores of the enrollees of a given plan, slightly modified to account for the rule that only three children can count toward the build-up of family rates.
"Conceptually, the goal of payment transfers is to provide plans with payments to help cover their actual risk exposure beyond the premiums the plans would charge reflecting allowable rating and their applicable cost factors. In other words, payments would help cover excess actuarial risk due to risk selection." 2014 Final Rule, 78 Fed. Reg. at 15,430. Accordingly, HHS set a given plan's transfer amount to equal the difference between two estimated premiums, which can be thought of as ideal premiums that a plan
In order to get from a plan's risk score — a number representing the cost of providing care to the plan's risk-selected enrollees relative to the cost of providing that same level of care to enrollees with average risk — to a dollar figure representing the estimated premium, HHS chose to use the statewide average premium as a conversion factor. 2014 Proposed Rule, 77 Fed. Reg. at 73,139; 2014 Final Rule, 78 Fed. Reg. at 15,432. The average premium of a given plan is "based on the total premiums assessed to enrollees, including the portion of premiums that are attributable to administrative costs." 2014 Proposed Rule, 77 Fed. Reg. at 73,142. The statewide average premium is "calculated as the enrollment-weighted mean of all plan average premiums of risk-adjustment covered plans in the applicable risk pool in the applicable market in the State." Id.; see 2014 Final Rule, 78 Fed. Reg. at 15,431-32. Thus, both of the estimated premiums in the transfer formula are based on the statewide average premium, which is one of several multipliers making up each term.
HHS, however, also wanted to ensure that transfers would not "reflect liability differences attributed to cost factors other than risk selection." 2014 Final Rule, 78 Fed. Reg. at 15,431; see 2014 Proposed Rule, 77 Fed. Reg. at 73,139 ("Risk-adjustment transfers are intended to reduce the impact of risk selection on premiums while preserving premium differences related to other cost factors, such as the actuarial value, local patterns of utilization and care delivery, local differences in the cost of doing business, and, within limits established by the Affordable Care Act, the age of the enrollee."). Therefore, in calculating the estimated premiums, it included a series of "premium-adjustment terms," detailed below, which (like the numbers assigned in the risk-adjustment model) are "relative measures that compare how plans differ from the market average with respect to the cost factors." 2014 Final Rule, 78 Fed. Reg. at 15,430-31.
To calculate the estimated premium for a given plan with risk selection, the formula multiplies the plan's liability risk score by two premium-adjustment terms: an "induced-demand" factor and a "geographic-cost" factor. Id. at 15,431. The induced-demand factor is meant to "reflect[] differences in enrollee spending patterns attributable to differences in the generosity of plan benefits (as opposed to risk selection)" — in other words, a person with a more generous plan might consume more health care than the same person in a less generous plan. 2014 Proposed Rule, 77 Fed. Reg. at 73,143. The geographic-cost factor accounts for differences in plan costs across geographic areas to prevent transfers from "subsidiz[ing] high-risk plans in high-cost areas at the expense of low-risk plans in low-cost areas." Id. at 73,144. The resulting product is then normalized for the plan's share of overall state enrollment (such that the "`normalized' term would average to 1.0," id. at 73,141) and multiplied by the statewide average premium to arrive at a dollar figure, id.; 2014 Final Rule, 78 Fed. Reg. at 15,431.
To calculate the estimated premium for that plan, assuming it had enrollees of
The output of the payment-transfer formula is a "per member per month ... transfer amount for a plan within a rating area." 2014 Final Rule, 78 Fed. Reg. at 15,431. That amount is then multiplied by the plan's "billable member months," defined as the number of months during the risk-adjustment period that each billable member (excluding children who do not count towards family rates) is enrolled in the plan, to arrive at the plan's total risk adjustment for a given rating area. Id. at 15,431-32.
Since 2014, HHS has maintained a position supporting model stability, and has elected not to rework the program's overarching methodology. See 2014 Final Rule, 78 Fed. Reg. at 15,418 ("Though we anticipate making future adjustments to the model, we seek to balance stakeholders' desire for a stable model in the initial years with introducing model improvements as additional data becomes available."); HHS Notice of Benefit and Payment Parameters for 2015, 79 Fed. Reg. 13,744, 13,753 (Mar. 11, 2014) ("2015 Final Rule") ("We believe it is important to maintain model stability in implementing the risk-adjustment methodology in the initial years of risk adjustment, and therefore do not intend to recalibrate the model in the initial years."); HHS Notice of Benefit and Payment Parameters for 2016, 79 Fed. Reg. 70,674, 70,684 (Nov. 26, 2014) ("2016 Proposed Rule") ("We propose to continue to use the same risk-adjustment methodology finalized in the 2014 Payment Notice, with changes to reflect more current data...."). However, there have been smaller, incremental adjustments over time.
Health plans are required to submit their risk-adjustment data to HHS by April 30 of the year following the benefit year. 45 C.F.R. § 153.730. HHS announces the transfer amounts by June 30. Id.
After the 2014 results were calculated, HHS proposed to update the model to include preventative-care costs in the 2017 rule and sought comment on how the risk-adjustment methodology could more accurately account for partial-year enrollees. HHS Notice of Benefit and Payment Parameters for 2017, 80 Fed. Reg. 75,488, 75,499-500 (Dec. 2, 2015) ("2017 Proposed Rule"). The final rule, published March 6, 2016, incorporated an adjustment for preventative-care costs and indicated that HHS would further explore adjustments for partial-year enrollees. 2017 Final Rule, 81 Fed. Reg. at 12,218-20. Later that month, HHS published a lengthy discussion paper and held a public meeting to discuss possible modifications to the riskadjustment methodology, including adjustments for partial-year enrollment. (Def. Mot. for Summ. J. Ex. C (dated Mar. 31, 2016)); see 2017 Final Rule, 81 Fed. Reg. at 12,216, 12,220. Following the conference, HHS announced in a June 2016 press release that it would incorporate an adjustment for partial-year enrollment for the 2017 benefit year, and it finalized that adjustment in December 2016, as part of the 2018 Final Rule. 2018 Final Rule, 81 Fed. Reg. at 94,071-74. The 2018 Final Rule also stated that HHS
Id. at 94,073.
The 2018 Final Rule made additional changes to take effect beginning in the 2018 benefit year. HHS began to use pharmaceutical data to assign a limited number of HCCs. Id. at 94,074-76. It also "reduce[d] the Statewide average premium in the risk-adjustment transfer formula by 14 percent to account for the proportion of administrative costs that do not vary with claims." Id. at 94,099-100.
Plaintiff Minuteman Health, Inc. was created under the CO-OP program. (Pl.
Minuteman offered health plans in Massachusetts starting in 2014, the first year the ACA came into effect, and entered the New Hampshire market in 2015. (Id. Ex. 10 ¶ 31). Its goal was to provide affordable health insurance in the individual and small-group markets by contracting with high-quality, low-cost health-care providers and excluding high-priced hospital systems from its network. (Id. Ex. 10 ¶¶ 12, 24-26). Minuteman priced its premiums significantly lower than many other issuers in the Massachusetts and New Hampshire markets. (Id. Ex. 10 ¶¶ 32-34).
In all the benefit years at issue here for which results are available, Minuteman's enrollees have collectively been healthier than average, and Minuteman has accordingly been required to pay into the risk-adjustment program. For the 2014 benefit year, Minuteman was required to pay 71% of its gross premium revenue into the risk-adjustment program. (Pl. Reply in Supp. Summ. J. Ex. M-25 pt. 2 at 11). For the 2015 benefit year, Minuteman was required to pay 40% of its gross New Hampshire premium revenue to the risk-adjustment program, and 39% ($6,110,676) of its gross Massachusetts premium revenue. (Pl. Mem. in Supp. Summ. J. Ex. 13 at 5).
The payment of such huge amounts — up to 71% of its gross revenue — has had a deleterious effect on Minuteman's business. Minuteman is now in receivership and will not be offering plans for the 2018 benefit year. (Hearing Tr. 3:21-4:4; Pl. Opp'n to Mot. to Strike Ex. A).
Minuteman filed the complaint in this action on July 29, 2016. The amended complaint asserts one count for "Violations of Section 1343 of the ACA and the APA, 5 U.S.C. § 706." (Am. Compl. ¶¶ 208-15). It alleges that "[t]he Risk-Adjustment methodology developed and implemented by CMS, at the direction of HHS, is arbitrary, capricious, and unlawful" and that "HHS and CMS have gone beyond the bounds of their statutory directive, injecting unauthorized factors into the Risk-Adjustment methodology, and failing to create a methodology that effects the directive of Congress." (Id. ¶ 215).
The parties have cross-moved for summary judgment. At the summary-judgment hearing, Minuteman dropped its challenge to the 2018 benefit rule. (Hearing Tr. 3:23-4:4). The government then filed a motion to strike all materials related to the 2018 benefit year (and certain other materials never presented to the agency in any benefit year) as materials that are outside the record.
Summary judgment is ordinarily appropriate when the pleadings and evidence show that "there is no genuine dispute as to any material fact and [that] the movant is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(a). However, in cases involving review of agency action under the APA, the traditional Rule 56 standard does not apply due to the limited role of a court in reviewing the administrative record. See Int'l Junior Coll. of Bus. & Tech., Inc. v. Duncan, 802 F.3d 99, 106 (1st Cir.2015) ("The summary judgment `rubric' also `has a special twist in the administrative law context.'" (quoting Associated Fisheries of Me., Inc. v. Daley, 127 F.3d 104, 109 (1st Cir. 1997))). Rather, when administrative action is challenged
The Administrative Procedure Act provides that a reviewing court should "hold unlawful and set aside agency action, findings, and conclusions found to be ... arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law," or "in excess of statutory jurisdiction, authority, or limitations, or short of statutory right." 5 U.S.C. § 706.
In Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), the Supreme Court established a two-step analysis for reviewing an agency's construction of a statute that it administers. Id. at 842-43, 104 S.Ct. 2778. The analysis begins with "whether Congress has directly spoken to the precise question at issue." If Congress's intent is clear, "the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress." Id. at 842-43, 104 S.Ct. 2778. "[I]f the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency's answer is based on a permissible construction of the statute." Id. at 843, 104 S.Ct. 2778. If Congress has explicitly left a gap for the agency to fill, the agency's interpretation is "given controlling weight unless [it is] arbitrary, capricious, or manifestly contrary to the statute." Id. at 843-44, 104 S.Ct. 2778; see also Household Credit Servs., Inc. v. Pfennig, 541 U.S. 232, 239, 124 S.Ct. 1741, 158 L.Ed.2d 450 (2004).
Under the second step, the agency's construction is accorded substantial deference. Chevron, 467 U.S. at 844, 104 S.Ct. 2778; see also United States v. Mead Corp., 533 U.S. 218, 227-28, 121 S.Ct. 2164, 150 L.Ed.2d 292 (2001). "This broad deference is all the more warranted when ... the regulation concerns `a complex and highly technical regulatory program,' in which the identification and classification of relevant `criteria necessarily require significant expertise and entail the exercise of judgment grounded in policy concerns.'" Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512, 114 S.Ct. 2381, 129 L.Ed.2d 405 (1994) (quoting Pauley v. BethEnergy Mines, Inc., 501 U.S. 680, 697, 111 S.Ct. 2524, 115 L.Ed.2d 604 (1991)). The court should not simply substitute its judgment for that of the agency. See Mead, 533 U.S. at 229, 121 S.Ct. 2164 ("[A] reviewing court has no business rejecting an agency's exercise of its generally conferred authority to resolve a particular statutory ambiguity simply because the agency's chosen resolution seems unwise.").
In determining whether agency action is arbitrary and capricious under the APA, the court must examine the evidence relied on by the agency and the reasons given for its decision. The 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.'" Motor Vehicle Mfrs. Ass'n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983) (quoting Burlington Truck Lines v. United States, 371 U.S. 156,
Under the APA, an agency is required to respond to "relevant" and "significant" comments raised in the rulemaking process. Public Citizen, Inc. v. Fed. Aviation Admin., 988 F.2d 186, 197 (D.C. Cir. 1993). But this requirement is not "particularly demanding." Id. "[I]t is settled that `the agency [is not required] to discuss every item of fact or opinion included in the submissions made to it in informal rulemaking.'" Id. (second alteration in original) (quoting Auto. Parts & Accessories Ass'n v. Boyd, 407 F.2d 330, 338 (D.C. Cir. 1968)). "Instead, the agency's response to public comments need only `enable [the reviewing court] to see what major issues of policy were ventilated... and why the agency reacted to them as it did.'" Id. (citing Auto. Parts, 407 F.2d at 335); see also Del. Dep't of Nat. Res. & Envtl. Control v. Envtl. Prot. Agency, 785 F.3d 1, 15 (D.C. Cir. 2015). "[C]omments which themselves are purely speculative and do not disclose the factual or policy basis on which they rest require no response." Home Box Office, Inc. v. Fed. Commc'ns Comm'n, 567 F.2d 9, 35 n.58 (D.C. Cir. 1977).
A court reviewing agency action must judge that action by the reasons given by the agency; it is not permitted to supply its own reasoned basis not present in the administrative record. Bowman Transp. Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281, 285-86, 95 S.Ct. 438, 42 L.Ed.2d 447 (1974) ("[W]e may not supply a reasoned basis for the agency's action that the agency itself has not given...." (citing Sec. & Exch. Comm'n v. Chenery Corp., 332 U.S. 194, 196, 67 S.Ct. 1760, 91 S.Ct. 1995 (1947))); Sec. & Exch. Comm'n v. Chenery Corp., 318 U.S. 80, 87-88, 63 S.Ct. 454, 87 S.Ct. 626 (1943) ("The grounds upon which an administrative order must be judged are those upon which the record discloses that its action was based."). Additionally, the agency's action may only be judged against the information available to the agency at the time — namely, the materials in the administrative record. Camp v. Pitts, 411 U.S. 138, 142, 93 S.Ct. 1241, 36 L.Ed.2d 106 (1973). The reviewing court may, however "uphold a decision of less than ideal clarity if the agency's path may be reasonably discerned." Bowman Transp., 419 U.S. at 286, 95 S.Ct. 438; see Encino Motorcars, LLC v. Navarro, ___ U.S. ___, 136 S.Ct. 2117, 2125, 195 L.Ed.2d 382 (2016).
HHS has moved to strike all materials relating to the 2018 benefit year on the ground that it is "black-letter administrative law" that an agency action can only be judged on the materials before it at the time it made its decision. See Hill Dermaceuticals, Inc. v. Food & Drug Admin., 709 F.3d 44, 47 (D.C. Cir. 2013); Massachusetts v. Hayes, 691 F.2d 57, 60 (1st Cir. 1982) ("`Simple fairness to those who are engaged in the tasks of administration, and to litigants, requires as a general rule that courts should not topple over administrative decisions unless the administrative body not only has erred but has erred
Plaintiff's principal response is simply to contend that "except for discrimination against bronze plans, all of Minuteman's challenges to the 2014-2017 rules were raised either by Minuteman, by other commenters, or by the agency itself during the relevant rulemaking proceedings for those years, and are thus preserved for judicial review." (Pl. Opp'n to Mot. to Strike at 3). But even if true, that does not mean that this Court may judge the actions of the agency in 2014-2017 based on the content of the comments made in 2018; it can only do so based on what was before the agency at the relevant time. Therefore, while the existence of prior comments raising similar issues may provide the Court with a basis to determine whether the particular agency decisions contested by plaintiff in this action are arbitrary and capricious, it is not a reason to consider the 2018 materials.
Plaintiff further argues that in its 2018 comment, it petitioned for "reconsideration" of the 2014-2017 rules pursuant to 5 U.S.C. § 553(e); that HHS denied its petition; that when it denied the petition, the record before the agency included the 2018 materials; and that therefore the administrative record on review should include those materials. See Am. Horse Prot. Ass'n, Inc. v. Lyng, 812 F.2d 1, 5 (D.C. Cir. 1987).
The complaint, however, makes no allegation of an improper denial of any petition under § 553(e). Indeed, the first time plaintiff made any mention of that theory of recovery was in its summary judgment reply memorandum. (Pl. Reply in Supp. Summ. J. at 5-6). Therefore, that argument is untimely, and the Court need not consider it. See Torres-Rios v. LPS Labs., Inc., 152 F.3d 11, 15-16 (1st Cir. 1998).
In any event, even considered on its merits, that theory is untenable. Although plaintiff calls its comment a petition for "reconsideration," there is no such thing. In substance, plaintiff was seeking a modification of a rule with retroactive effect. (See Pl. Opp'n to Mot. to Strike at 3). The APA provides that "[e]ach agency shall give an interested person the right to petition for the issuance, amendment, or repeal of a rule." 5 U.S.C. § 553(e). The APA defines "rule" as "the whole or a part of an agency statement of general or particular applicability and future effect designed to implement, interpret, or prescribe law or policy ...." Id. § 551(4) (emphasis added). That statute, by its terms, does not provide an interested party with the right to seek amendment to past rules with retroactive effect. Indeed, agencies are generally prohibited from promulgating retroactive rules, absent express permission from Congress. Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 208, 109 S.Ct. 468, 102 L.Ed.2d 493 (1988).
Plaintiff cites to no authority suggesting that interested parties have a right to petition agencies for retroactive rule changes, and the cases it cites are not on point. In particular, Beverly Hospital v. Bowen, 872 F.2d 483 (D.C. Cir. 1989), stands for the proposition that when a plaintiff has successfully challenged a regulation and it is ruled to be "void ab initio," the general prohibition on retroactive rulemaking does not prevent the agency from making "corrective adjustments" to afford plaintiff relief reflecting what the court has found to be illegal. Id. at 485-86. Beverly Hospital does not suggest that the agency has the
Plaintiff further argues that its 2018 comment, which was submitted on October 6, 2016, was not requesting retroactive changes as to the 2016 and 2017 rules, because the 2016 and 2017 risk-adjustment transfers had yet to be calculated. (See Pl. Mem. in Supp. Summ. J. Ex. 13). But the HHS regulations require setting the payment-transfer formula in advance, so that issuers can set their premiums with the transfer methodology in mind. The fact that the 2016 transfer amounts would not be calculated until June 2017 is of no moment — as of October 2016, issuers could hardly change their 2016 premiums, and therefore changing the transfer formula at that time would implicate the same policy concerns as the prohibition on retroactive rulemaking. See 2018 Final Rule, 81 Fed. Reg. at 94,073 (explaining the importance of finalizing rules in time to allow issuers to incorporate the methodology in their rate setting).
Plaintiff has a somewhat stronger position with respect to the 2017 Final Rule, which was published on March 8, 2016 (prior to plaintiff's comment, which was submitted on October 6), but modified on December 22, 2016, to include adjustments for partial-year enrollment. See 2017 Final Rule, 81 Fed. Reg. 12,204; 2018 Final Rule, 81 Fed. Reg. 94,058. That timing superficially suggests that at the time plaintiff submitted its comment for 2018, it was not too late for HHS to consider its comments in making changes to the 2017 rule. But the particular modification lately made to the 2017 rule had been subject to notice and comment by way of a March 2016 discussion paper and public meeting and had been previously announced in a June 2016 press release. (Def. Mot. for Summ. J. Ex. C); see 2017 Final Rule, 81 Fed. Reg. at 12,216, 12,220. No other changes to the 2017 rule were noticed in those proceedings. The 2018 notice of proposed rulemaking did not generally request comment relating to the 2017 rule, and plaintiff's comments seeking changes going back to 2014 were certainly outside the scope of that proposal. Without notice of its intent to change the 2017 rule, HHS could not have otherwise altered it in response to comments on the 2018 rule. See Am. Fed'n of Labor & Congress of Indus. Orgs. v. Donovan, 757 F.2d 330, 338 (D.C. Cir. 1985) (explaining that the final rule must be a "logical outgrowth" of the proposed rule and that "if the final rule deviates too sharply from the proposal, affected parties will be deprived of notice and an opportunity to respond to the proposal" (quoting Small Refiner Lead Phase-Down Task Force v. Envtl. Prot. Agency, 705 F.2d 506, 547 (D.C. Cir. 1983))); 2018 Final Rule, 81 Fed. Reg. at 94,072-73 (summarizing concerns from commenters about making changes after the time for rate setting). Therefore, the notice of the proposed modification for partial-year enrollment did not effectively open the 2017 rule to the incorporation of any comment submitted for the 2018 rule.
To the extent that plaintiff contends that its October 2016 comment (which, again, was submitted as part of the 2018 benefit year notice-and-comment process and requested changes to the 2014-2017 benefit years) could have been considered a petition to amend the 2017 Final Rule, the Court cannot fault the agency for failing to treat it that way. It is true that
Accordingly, the Court will generally decline to consider the 2018 materials in deciding whether HHS's rules for the 2014-2017 benefit years were arbitrary and capricious. However, because the 2018 Final Rule modified a portion of the 2017 Final Rule to include accounting for partial-year enrollees, the comments related to that adjustment appear to be properly part of the record. Furthermore, to the extent that plaintiff's 2018 comment and the documents submitted with it contain relevant background information about plaintiff, the Court will consider those materials.
Defendants also contend that that certain additional materials — mostly newspaper articles, but also testimony from various state insurance commissioners before the U.S. Senate Committee on Homeland Security and Governmental Affairs; a court transcript from a District of Columbia District Court decision; and the Commonwealth Health Insurance Connector Authority ("Health Connector") decision denying plaintiff's request for review of its 2014 risk-adjustment payment — should be struck because they were not before HHS in any benefit year. (Def. Mem. in Supp. Mot. to Strike Ex. A). Plaintiff admits that "these materials are not necessary for [its] arguments that the risk-adjustment rules violated the APA." (Pl. Opp'n to Mot. to Strike at 15). Its only justification for including these materials is that the Court may take judicial notice of them.
Judicial notice is an evidentiary doctrine that may be used to admit "a fact that is not subject to reasonable dispute."
Defendant's motion to strike will therefore be granted in part and denied in part.
Standing is a threshold question in every case. "If a party lacks standing to bring a matter before the court, the court lacks jurisdiction to decide the merits of the underlying case." United States v. AVX Corp., 962 F.2d 108, 113 (1st Cir. 1992). To satisfy the case-or-controversy requirement of Article III of the U.S. Constitution, a plaintiff bears the burden of establishing that it (1) has suffered an "injury-in-fact," (2) that the injury is "`fairly traceable' to the actions of the defendant[s]," and (3) "that the injury will likely be redressed by a favorable decision." Bennett v. Spear, 520 U.S. 154, 162, 117 S.Ct. 1154, 137 L.Ed.2d 281 (1997) (quoting Lujan v. Defenders of Wildlife, 504 U.S. 555, 560-61, 112 S.Ct. 2130, 119 L.Ed.2d 351 (1992)). These elements must be proved "with the manner and degree of evidence required at the successive stages of the litigation." Lujan, 504 U.S. at 561, 112 S.Ct. 2130. The redressability element of standing requires that the requested relief directly redress the injury alleged. Steel Co. v. Citizens for a Better Env't, 523 U.S. 83, 105-09, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998). Plaintiff must establish that it is "likely," as opposed to merely "speculative," that its claimed injuries will be redressed by a favorable decision. Lujan, 504 U.S. at 561, 112 S.Ct. 2130.
Counsel for plaintiff clarified at the hearing that its only claim as to the 2014 benefit year — when it was operating only in Massachusetts and Massachusetts was running its own risk-adjustment program — is that the use of the statewide premium was arbitrary and capricious. (Hearing Tr. 45:1-5). Plaintiff alleges, and defendants do not dispute, that HHS regulations required Massachusetts to use the statewide average premium, and thus HHS regulations caused its alleged injury. See 2014 Proposed Rule, 77 Fed. Reg. at 73,147; 2014 Final Rule, 78 Fed. Reg. at 15,436. Therefore, the only standing challenge the Court need address is defendants' argument that because the relevant Massachusetts agency is not a party and would not be otherwise compelled to make retroactive changes to its rules, any action taken by this Court as to HHS's rulemaking process would not redress plaintiff's harm with respect to that year.
Plaintiff, however, cites no authority suggesting that Massachusetts agencies may make retroactive rule changes upon petition — or, for that matter, that this Court could order the agency to do so.
(Id.).
It is thus true that a ruling in plaintiff's favor from this Court would not directly redress its injury with respect to the 2014 benefit year — even with a favorable ruling, it would still have to file a petition with the Health Connector. It seems likely to this Court that such a petition would be futile. Nonetheless, while the Health Connector, as described above, has already denied the relief plaintiff seeks in the context of plaintiff's request for reconsideration, that is not controlling as to how it might respond to a Chapter 30A, § 4 petition — indeed, the Health Connector's response could not have been otherwise, as its regulations specifically prohibit it from considering changes in methodology in connection with requests for reconsideration. 956 C.M.R. § 13.06 ("The risk adjustment methodology cannot be the subject of a request for reconsideration."). And if, as noted, Massachusetts was required by HHS to use the statewide average premiums for the 2014 benefit year, it follows that plaintiff cannot
Although it is less than clear that plaintiff's alleged 2014 injury — namely, its payment to the Massachusetts risk-adjustment program for the 2014 benefit year — can be redressed by a favorable decision in this lawsuit, that is not the only possible basis for this Court to review the 2014 rule. Because of the way that the rulemaking proceeded here — specifically, the fact that HHS set the rule in 2014 and then declined to substantially change it for at least the 2015 and 2016 benefit years — the Court finds that plaintiff's 2015 injury is sufficient to provide it with standing to challenge the 2014 rule.
Plaintiff contends that HHS's risk-adjustment methodology is flawed for multiple reasons. First, it contends that using the statewide average premium as the baseline figure for pricing premiums in the transfer formula unreasonably inflates the charges to issuers with lower prices. Second, it contends that HHS's formula systematically underestimates the costs of insuring members who are not diagnosed with a condition associated with an HCC. Third, it contends that the model misses enrollees who are eligible for HCC classification because (a) until 2018 it did not use prescription-drug data to assign HCCs and (b) until 2017 it did not account for members who were enrolled for a short period of time during which they were not diagnosed. Fourth, it contends that the new risk-adjustment formula has made lower-cost bronze plans economically unviable, in spite of the requirement of the ACA that they be offered.
The response of HHS is that it has taken all of these concerns into consideration and that there are no fatal flaws regarding its methodology. It contends that lower-cost plans such as the bronze plans predominantly sold by plaintiff generally attract healthier enrollees. According to HHS, plaintiff, with its high percentage of bronze plans, has simply attracted many healthier-than-average
Plaintiff's challenges will be taken in turn. Although plaintiff has presented its arguments against all benefit years together, the Court may consider the actions of HHS only with regard to the materials it had before it at the time. The Court will proceed by evaluating whether HHS's initial decisions were arbitrary or capricious in light of the comments and data available in 2014, and then evaluate whether HHS's decisions remained permissible in each successive year as new information was presented to the agency.
Plaintiff makes two distinct arguments against the use of the statewide average premium: first, that it is contrary to the unambiguous wording of the statute, and second, that it is arbitrary and capricious. The first argument requires a two-step Chevron analysis, under which the Court must decide whether the statute is unambiguous and then, if not, whether the agency's interpretation of the statute is reasonable. The second argument requires the Court to determine whether the particular action taken by the agency in these rules violated the APA because it was not the product of reasoned decision-making.
Notwithstanding that analytical distinction, in the context of this case, there is little practical difference between the second Chevron step and the APA analysis. See 1 RICHARD J. PIERCE JR., ADMINISTRATIVE LAW TREATISE § 7.4, at 604 (5th ed. 2010) ("[T]he question whether an agency engaged in reasoned decision-making within the meaning of State Farm often is identical to the question a court must answer under step two of the test announced in Chevron...."); see, e.g., Judulang v. Holder, 565 U.S. 42, 52 n.7, 132 S.Ct. 476, 181 L.Ed.2d 449 (2011); Animal Legal Def. Fund, Inc. v. Glickman, 204 F.3d 229, 234 (D.C. Cir. 2000). Therefore, in the analysis that follows, the Court will treat those two issues together.
Plaintiff first contends that the use of the statewide average premium in the transfer formula is contrary to law because the risk-adjustment statute "unambiguously provide[s] that risk-adjustment assessments cannot be based on factors other than actuarial risk." (Pl. Mem. in Supp. Summ. J. at 16-17). The relevant portion of the statute is sparse: it simply provides that a state shall "assess a charge" if "the actuarial risk of the enrollees of such plans or coverage for a year is less than the average actuarial risk of all enrollees in all plans or coverage in such State for such year." 42 U.S.C. § 18063(a)(1). It also directs HHS to "establish criteria and methods to be used in carrying out the risk-adjustment activities under this section." 42 U.S.C. § 18063(b).
The basic problem is that under the statute, HHS must both determine "actuarial risk" and assess a "charge." HHS must therefore devise some way to transform a plan's relative "actuarial risk" into a "charge," which necessarily requires establishing "criteria and methods" to convert risk to dollars. Because the statute does not require a particular methodology, it necessarily requires HHS to exercise its discretion in doing so. HHS elected, as a part of its methodology, to use the statewide average premium in the transfer formula. Indeed, even plaintiff does not suggest that HHS should have relied on the risk score alone; rather, it complains that HHS should have used a plan's own premium to scale the risk.
HHS's decision to use the statewide average premium was the result of extensive debate. In a white paper published in 2011, the Center for Consumer Information and Insurance Oversight ("CCIIO") proposed four distinct options for establishing a "baseline premium" to use in the risk-adjustment model, and noted the features of each. The two relevant here are the following:
(Def. Mot. For Summ. J. Ex. B at 14-15).
Later in the same paper, HHS presented the results of some modeling experiments estimating the effects of using certain combinations of baseline-premium options and balancing options. Use of the statewide average premium did not require balancing in order to ensure that the charges the program collected could cover the payments due. However, the agency modeled three balancing options in combination with use of a plan's own premium: decreasing payments, increasing charges, and splitting the shortfall by both decreasing payments and increasing charges. (Id. Ex. B at 36-37). It concluded that using a plan's own premium with a balancing strategy that decreased payments would distribute the shortfall "across all the above-average-risk plans in proportion to the risk payment each would have received before balancing" and would cause those plans' premiums to rise. (Id. Ex. B at 36). Conversely, using a plan's own premium with a balancing strategy that increased charges
The agency compared the results of using the statewide average premium to using the plan's own premium with "split-the-shortfall" balancing as follows:
(Id. Ex. B at 37-38).
Then, in the 2014 Proposed Rule, HHS explained:
2014 Proposed Rule, 77 Fed. Reg. at 73,139.
2014 Final Rule, 78 Fed. Reg. at 15,432.
HHS thus considered several options in detail — including the very option plaintiff promotes — before arriving at its conclusion. Ultimately, HHS chose to use the statewide average premium, subject to certain cost-factor adjustments. It did so because it concluded that such an approach would result in balanced transfers, was a "straightforward and predictable benchmark," and would best compensate plans for liability differences due to risk selection, as opposed to other cost factors. Those articulated reasons have a clearly rational connection to HHS's choice. Thus, the record demonstrates that the proposal to use a plan's own premium was actively considered and rejected on rational grounds.
Plaintiff further contends that the statute does not require the risk-adjustment program to be budget-neutral; that HHS has never explained why the program should be budget-neutral; that even if it did have to be budget-neutral, that would not require the use of the statewide average premium; and that therefore budget-neutrality is "not a legitimate reason to depart from using the issuer's own premium." (Pl. Mem. in Supp. Summ. J. at 23, 36; Hearing Tr. 15:19-20).
Although the statute does not require the program to be budget-neutral, it does not prohibit the program from being budget-neutral, either. Even plaintiff does not argue that the statute forbids budget-neutrality — at most it argues that the statute's silence, when read in conjunction with 42 U.S.C. § 18061(b)(1)(B), which requires the reinsurance program to be budget-neutral, "suggests" that the risk-adjustment program should be operated differently. (Pl. Mem. in Supp. Summ. J. at 24; see also Hearing Tr. 12:5-6 ("[T]he risk-adjustment statute doesn't necessarily have to be budget neutral.")).
The question then becomes whether HHS's decision to attempt to operate the risk-adjustment program in a budget-neutral way was unreasonable or arbitrary. It was not. According to HHS, the goal of the program is to spread the risk of insuring unhealthy enrollees among all insurers and eliminate incentives for plans to engage in risk selection. See, e.g., Premium Stabilization Rule, 77 Fed. Reg. at 17,230; 2014 Final Rule, 78 Fed. Reg. at 15,411. The modeling outlined in the 2011 white paper showed that use of a plan's own premium without balancing ran the risk of a shortfall, where there was no money to shore up plans that took on less-healthy patients.
Plaintiff suggests that HHS's general appropriations could be used to fund any shortfall, that insurers could sue for the shortfall in the Court of Claims, or that HHS could reduce payments out of the program. But a showing that there are other ways a budget-neutral program might have been achieved is not a showing that what HHS actually did was unreasonable or arbitrary. The risk-adjustment statute may be reasonably read as intending to level the playing field by spreading the risk among insurers, not by having the government subsidize the costs of insuring less-healthy people through HHS's general appropriations budget or the Judgment Fund.
Furthermore, as HHS points out, the risk-adjustment program was meant to be run by the states. Congress could not, under the Constitution, require the states to use their own money to fund a federal program. See Printz v. United States, 521 U.S. 898, 929-30, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997); New York v. United States, 505 U.S. 144, 175-76, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992). It therefore stands to reason that absent any appropriation, Congress expected the states to run budget-neutral risk-adjustment programs, and for HHS to set its federal regulations to allow it to certify such programs. See 45 C.F.R. § 153.310. It was not unreasonable or arbitrary for HHS to attempt to design the program to pay for itself, as opposed to exposing its own general appropriations to float the program for any state that chose not to run its own program. Nor was it unreasonable or arbitrary to implement the program in such a way as to avoid having the participants sue annually in federal court for the program to function.
To the extent that plaintiff complains that HHS did not adequately explain its decision to run the program in a budget-neutral way, the claim must likewise fail. There is no evidence that there was any significant comment on the topic that HHS was required to address in 2014. In any event, as set forth in the 2011 white paper, HHS considered the effects of non-budget-neutral methodologies and rationally chose to operate a budget-neutral program. Therefore, HHS's decision to use the statewide average premium in the risk-adjustment formula was not contrary to law, unreasonable, or arbitrary or capricious.
Plaintiff also attacks the stated rationale that using the statewide average premium is a "straightforward and predictable benchmark" on the ground that, for smaller issuers, the average premium is a "black box." (Pl. Mem. in Supp. Summ. J. at 27). But, assuming that it is reasonable to operate the program in a budget-neutral way, plaintiff's proposal appears to be even less predictable. Under plaintiff's proposal, a plan — knowing its own premium and its own risk pool — presumably would be able to predict how much it would owe or receive under the transfer formula. But even if HHS adopted a particular methodology for addressing any shortfall in advance,
Plaintiff further complains that HHS's statement that using a plan's own premium would cause "unintended distortions in transfers" is vague and unsupported, and therefore does not meet its obligation to engage in reasoned decision-making. 2014 Final Rule, 78 Fed. Reg. at 15,432. But the Court finds that HHS's meaning here may be "reasonably discerned" from the record. Bowman Transp., 419 U.S. at 286, 95 S.Ct. 438. Specifically, HHS explained in the white paper that "[f]or plans with a sicker than average risk mix, a lower premium plan would receive less in payments than a higher premium plan, even if the two plans have the same risk level. This could create disincentives for high-risk plans to operate efficiently or set lower prices." (Def. Mot. for Summ. J. Ex. B at 14-15). HHS also explained that using the average premium "disconnects each plan's risk-adjustment compensation from its choice of what premium to charge." (Id. Ex. B at 37-38). It seems clear that HHS was concerned that issuers would price their plans to game the risk-adjustment calculation or that individual plans' pricing decisions could have an outsize effect on the total funds available or required for the risk-adjustment program in a given year. In any event, the other reasons HHS gave for choosing the statewide average premium are sufficient to show that it considered the relevant comments and rationally chose to use the average premium based on the evidence before it.
Finally, plaintiff complains that use of the statewide average premium affords issuers with dominant market positions too much control, allowing those issuers to punish smaller issuers who have found ways to keep costs down and charge less, and thus perversely rewards price-increasing behavior. However, plaintiff has not cited to any materials before the agency in the 2014 rulemaking process that show the catastrophic consequences plaintiff asserts occurred or would occur — all those materials are from later rulemaking processes. Plaintiff points to the 2011 white paper as evidence that HHS knew that using the
For the 2017 benefit year, HHS received multiple comments protesting the use of the statewide average premium. (Pl. Reply in Supp. Summ. J. Ex. M-25 pt. 1 at 6-7 (Minuteman 2017 Comment); Id. Ex. M-18 at 3 (New Mexico Health Connections 2017 Comment); Id. Ex. M-21 at 2 (Evergreen Health 2017 Comment); Id. Ex. M-24 at 5 (Land of Lincoln Health 2017 Comment)). HHS responded, "We did not propose changes to the transfer formula, and therefore, are not addressing comments that are outside the scope of this rulemaking." 2017 Final Rule, 81 Fed. Reg. at 12,230.
That response is at least somewhat troublesome in this context. By 2017, the use of statewide average premium had apparently resulted in serious problems, and it might have been prudent for HHS to address those comments directly. At some point, if an agency's rules have become obviously unreasonable or unworkable, the agency surely has some obligation to reconsider its actions, or at least respond to comments pointing out that fact, even if technically outside the scope of its proposed rulemaking. See Nat'l Mining Ass'n v. Mine Safety & Health Admin., 116 F.3d 520, 549 (D.C. Cir. 1997) (suggesting that agencies are not required to address comments outside the scope of the rulemaking as long as they do not cast doubt on the reasonableness of the agency's position); Sherley v. Sebelius, 776 F.Supp.2d 1, 22 (D.D.C. 2011) (same).
In this case, however, HHS had already considered alternatives to using the statewide average premium, and the comments at issue did not raise any concerns that HHS had not considered before; they merely pointed out that a plan whose premiums are well below a state's average will find that its risk-adjustment charge is inflated. HHS was aware of that aspect of its methodology at least as early as 2011, and
For the reasons discussed above, it was not unreasonable or irrational for HHS to use the statewide average premium in the first instance. Because the 2017 comments raised nothing new, HHS was not required to re-justify its choice. Therefore, the continued use of the statewide average premium was not arbitrary or unreasonable under the circumstances.
Plaintiff further contends that the HHS transfer formula underestimates the cost of healthier enrollees who do not qualify for an HCC. Under the HHS formula, those enrollees will receive an individual risk score based only on their age and sex. But, plaintiff argues, they may still consume health care (for example, if they use preventative care services, experience a catastrophic injury, suffer from chronic low back pain, need joint replacement surgery, or become at risk for developing a condition covered by an HCC code and undergo aggressive clinical therapy to prevent that illness). (Pl. Mem. in Supp. Summ. J. at 34-35). Plaintiff contends that "[f]or these reasons, the HHS formula underestimates health care costs for enrollees without an HCC by 10%-35%," and that accordingly, to the extent that premiums from healthy members are being paid out for such care, those funds are not available to offset payments made either within the plan to predictably sicker enrollees or in risk-adjustment transfers to other plans with predictably sicker enrollees. (Id. at 35).
The Court finds it difficult to discern from the briefing precisely what feature of the transfer formula plaintiff is complaining about. Clearly, plaintiff takes issue with the result that premiums from healthier enrollees are not balancing out losses from sicker enrollees. But, in explaining why that is, plaintiff mostly points to the possibility of random, expensive events that could happen to any enrollee, regardless of his or her overall health. Although the Court understands the point that plans may spend money on otherwise healthy enrollees with bad luck, the Court does not see that plaintiff is suggesting that risk adjustment should incorporate HCCs for random events.
Plaintiff points to only one comment for the 2014 benefit year related to the issue of undercompensating plans for enrollees without HCCs, which was from the Blue Cross Blue Shield Association ("BCBSA"). (Pl. Reply in Supp. Summ. J. Ex. M-1 (submitted Apr. 30, 2013)). That comment in fact postdated the 2014 Final Rule — it was submitted in response to a proposed interim rule addressing certain amendments to the 2014 rule relating to risk corridors and cost-sharing reductions. (Id.); see Amendments to the HHS Notice of Benefit and Payment Parameters for 2014, 78 Fed. Reg. 15,541 (Mar. 11, 2013) (Interim Final Rule with Comment).
It is not at all clear that the potential problem of undercompensating for members without HCCs was raised to HHS in in the notice-and-comment period for the rule finalizing the risk-adjustment transfer formula for 2014.
For the 2017 benefit year, HHS proposed to recalibrate the risk-adjustment model by recalculating the weights assigned to various HCCs and demographic factors using more recent data and incorporating preventive services in its simulation of plan liability. 2017 Proposed Rule, 80 Fed. Reg. at 75,499. The proposal explained that HHS would identify claims related to preventive care in the model data sets and then "adjust[] plan liability by adding 100 percent of preventive services covered charges to simulate plan liability for all metal levels" so that the "[t]otal adjusted simulated plan liability is the sum of preventive-services covered charges, and non-preventive services simulated plan liability." Id. Comparing the new model to the old one, HHS found that "[a]djusting for preventive services increases age-sex coefficients relative to HCC coefficients, especially in the higher cost-sharing metal tiers (bronze and silver), and in age/sex ranges with high preventive-services expenditures (for example, young adult females)." Id. HHS predicted that "the risk scores of healthy enrollees (whose risk scores are based solely on model age-sex coefficients) will likely rise relative to the risk scores of the less healthy (whose risk scores include one or more HCC coefficients in addition to an age-sex coefficient), especially in bronze and silver plans." Id.
Plaintiff points to three comments HHS received related to this topic in 2017. (Pl. Reply in Supp. Summ. J. Ex. M-25 pt. 1 at 4-5 (Minuteman 2017 Comment); Id. Ex. M-18 at 2 (New Mexico Health Connections 2017 Comment); Id. Ex. M-30 at 8487, 8489 (Health New England 2017 Comment)). All three comments pointed out that the risk-adjustment transfer formula undercompensates plans for enrollees without HCCs while overcompensating for enrollees with HCCs. And they all attached the same November 4, 2015 white paper written by Consumers for Health Options, Insurance Coverage in Exchanges in States ("CHOICES"). (See Pl. Reply in Supp. Summ. J. Ex. M-25 pt. 2).
In response, HHS finalized its proposal to include preventive services in the plan-liability simulation. 2017 Final Rule, 81 Fed. Reg. at 12,206. It explained: "We have attempted to address the range between enrollees without HCCs and those with HCCs by finalizing the incorporation of preventive services into our simulation of plan liability. While overall this is not a very large effect, it does have a noticeable effect on certain demographic subgroups, resulting in more accurate payments for enrollees without HCCs." Id. at 12,218.
Plaintiff basically complains that HHS did not do enough to address the problem.
Furthermore, even plaintiff's own comment proposed only "short-term solutions" to stabilize premiums while "HHS convenes the appropriate experts to reexamine the methodology and make appropriate changes to mitigate unintended consequences." (Pl. Reply in Supp. Summ. J. Ex. M-25 pt. 1 at 7). None of those short-term proposals were specific to the problem of bias against non-HCC enrollees, and it seems that the source of this bias was not well understood (and apparently still is not, as plaintiff's brief does not satisfactorily explain how the formula should have been different). (See id. Ex. M-25 pt. 1 at 6-7).
(Pl. Reply in Supp. Summ. J. Ex. M-25 pt. 2 at 4). Because it was not clear why the formula was biased the way it was, it was not unreasonable for HHS to proceed cautiously, and not simply adopt a short-term fix.
Finally, plaintiff argues that HHS's modification of the transfer formula to include preventive care (and, beginning in 2018, administrative costs) is an admission that the prior-year formulas were erroneous, and therefore HHS should have made retroactive changes to the formula for those years. But, as explained above, HHS is generally prohibited from retroactive rulemaking, and could not spontaneously change its prior-year formulas. Furthermore, an adjustment to a rule, made with the benefit of results from 2014 is not an "admission," or conclusive evidence, that the prior rule was arbitrary or capricious at the time it was made. The 2014 rule was not arbitrary or capricious, in light of the information available at the time.
Next, plaintiff argues that the transfer formula fails to adequately assign HCCs to all eligible enrollees, with the result that certain plans appear to cover healthier populations than they actually do. Plaintiff identifies two specific deficiencies: first, that HHS's plan fails to account for partial-year enrollment, and second, that it ignores prescription-drug-utilization data.
In order to qualify for an HCC, an enrollee must be diagnosed with an HCC-qualifying condition during the time he or she is enrolled in the plan. See 2014 Proposed Rule, 77 Fed. Reg. at 73,127-28 (explaining the concurrent model). Plaintiff complains that when members switch plans part way through the year, the second plan may pay out for things related to that enrollee's diagnosis, but may not be credited with that enrollee's HCC if the enrollee does not visit a doctor during the part of the year he or she is enrolled in the second plan. Plaintiff also argues that the formula's calculation of risk on a per-month basis inappropriately spreads the costs of insuring a patient throughout the year, when in fact plans may incur the full amount of liability for a patient with an HCC in only a few months of enrollment. (To use a simple example, a woman could be only enrolled for two months, during which she gives birth and incurs high medical costs).
The HHS commentary to the 2014 Proposed Rule acknowledged that partial-year enrollment might be a problem. HHS explained that it was using a concurrent model because "we anticipate that enrollees may move between plans, or between programs. A concurrent model would be better able to handle changes in enrollment than a prospective model because individuals newly enrolling in health plans may not have prior data available that can be used in risk adjustment." 2014 Proposed Rule, 77 Fed. Reg. at 73,128.
The 2014 Final Rule shows that HHS continued to consider methods of addressing the partial-year-enrollment concern. It reiterated its reasoning in favor of the concurrent model, and further stated:
2014 Final Rule, 78 Fed. Reg. at 15,420, 15,421.
Plaintiff does not identify a specific recommendation that HHS should have adopted to fix the partial-year-enrollment problem, apart from the use of prescription-drug data, discussed below. The single comment identified by plaintiff suggested only (1) shortening the minimum length of enrollment required for an enrollee before diagnostic information can be considered by the risk-adjustment formula and (2)
In the 2017 Proposed Rule, HHS sought comments on how to make its formula more predictive for partial-year enrollees.
2017 Proposed Rule, 80 Fed. Reg. at 75,500. Several comments were submitted in response. (E.g., Pl. Reply in Supp. Summ. J. Ex. M-29 at MH1564-65). HHS gave a detailed summary of those comments and their varied proposed solutions in its final rule. 2017 Final Rule, 81 Fed. Reg. at 12,220. HHS noted: "We appreciate commenters' substantive feedback on accounting for partial-year enrollment in future recalibrations and will continue to analyze this issue and include our findings in the White Paper for discussion at the March 31, 2016 risk-adjustment conference." Id.
HHS did exactly that, devoting several pages of analysis to that topic in its white paper. (Def. Mot. for Summ. J. Ex. C at 35-39). That paper presented results from several modeling simulations exploring the effects of adopting various solutions, and explained the steps HHS was taking to evaluate the feasibility of those options (and possible combinations of options). (Id.). After reviewing the feedback from the white paper, HHS announced on June 8, 2016, that "[it] intended to propose that, beginning for the 2017 benefit year, the risk adjustment model include adjustment factors for partial year enrollees." 2018 Final Rule, 81 Fed. Reg. at 94,072. It then finalized that adjustment, beginning for the 2017 benefit year, in the 2018 Final Rule. Id.
Plaintiff complains that HHS's "repeated failure to consider the problems and solutions raised by commenters constitutes arbitrary and capricious conduct." (Pl. Mem. in Supp. Summ. J. at 41). But the record shows that HHS did consider those comments, determined that further investigation of possible solutions was necessary, and engaged in that further investigation. Again, that response was not arbitrary or capricious.
Plaintiff also suggests that HHS did not move fast enough to rectify the problem. HHS finally calculated the results of the 2014-benefit-year risk transfers at the end of June 2015. It proposed the 2017 benefit rule in December 2015 and finalized it in March 2016. Weeks later it produced a white paper detailing proposed model simulations to fix the problem going forward. It seems clear from that timeline — and the fact that HHS requested comments in the 2017 proposed rule on possible future solutions to the partial-enrollment problem — that HHS was not ready with a specific
Plaintiff also complains that because enrollees do not always receive a documented HCC diagnosis during their enrollment periods, it was arbitrary and capricious for HHS not to use prescription-drug data to determine whether an enrollee has an HCC-qualifying condition. Plaintiff argues that using prescription-drug data is more reliable and efficient than using claims data, and points to several comments in the 2014 rulemaking supporting use of such data.
HHS did consider using prescription-drug data to assign HCCs from the very beginning of the process, but consistently expressed concern that doing so would incentivize doctors to over-prescribe medications. Its 2011 white paper noted that "[i]ncluding prescription data in a risk-adjustment model ... could offer powerful incentives to steer treatment toward pharmaceutical therapy in order to identify risk of the enrolled population." (Def. Mot. for Summ. J. Ex. B at 9). HHS's proposed rule for 2014 explained:
2014 Proposed Rule, 77 Fed. Reg. at 73,128. The final rule provided: "HHS is finalizing its proposal to exclude prescription drugs for the initial HHS risk-adjustment models, but will consider how prescription drugs could be included in future HHS risk-adjustment models." 78 Fed. Reg. 15,419.
Plaintiff argues that strategic prescribing behavior is not a genuine concern and therefore it was not a rational basis to reject using prescription data. (Pl. Reply in Supp. Summ. J. at 20-21; Hearing Tr. 34:3-36:9). But the record before the agency included a report on the use of pharmacy data from RTI International, a consulting firm hired to help HHS research issues related to implementing the ACA. That report determined that "[t]he most salient concern with tying risk-adjustment payments to drug usage is the likely distortion of provider decisions toward pharmaceutical therapies. This distortion would create real costs: not only the costs for of [sic] the drugs themselves, but also the health outcomes that would be diminished by any deviation from clinical best practices." (Def. Mot. for Summ. J. Ex. D at 4). It cited a study in support of that conclusion, and explained how the incentives are worst for cheap drugs with higher risk-adjustment payments, and can penalize doctors for cost-effective prescribing habits. (Id.). It explained how other health systems that use pharmacy data have attempted to avoid this problem, but warns that "a model that relies on drug data will need to be carefully specified," because while "[b]ad incentives and gaming can be avoided, for example, by including only those drugs and therapeutic classes for which there is nearly universal
After 2014, HHS continued to seek comments on and investigate the feasibility of including prescription-drug data in its risk-adjustment model.
2017 Proposed Rule, 80 Fed. Reg. at 75,499-500. As with the partial-year-enrollment issue, the 2017 Final Rule extensively summarized the many comments submitted on the topic (at least one of which supported implementation only after a separate notice-and-comment process specifying the methodology in detail) and promised to "explore the incorporation of prescription drugs in the risk-adjustment model in the White Paper and at the conference in March 2016." 2017 Final Rule, 81 Fed. Reg. at 12,219. HHS stated:
Id. at 12,219-20.
As promised, the 2016 white paper addressed this topic in considerable detail. In thirty single-spaced pages, HHS outlined the benefits and concerns associated with including prescription-drug data, provided a framework for analyzing that data, proposed a classification system, presented an
HHS's decision to leave prescription-drug data out of the 2017 rule and consider the matter further was not unreasonable. Plaintiff points to a variety of comments extolling the virtues of such data. (See Pl. Reply in Supp. Summ. J. Ex. M-25 pt. 1 at 4; id. Ex. M-27 at 5). But HHS's concerns about using it remained valid (and, as evidenced by the detailed 2016 white paper, included more than just the risk of gaming). HHS showed that it was open to using such data and was actively considering the change. It was not required to re-justify its position in response to every comment.
As with the preventive-care changes, the fact that HHS ultimately included prescription-drug data in its model is not evidence that its prior model was arbitrary and capricious. Rather, it shows that the agency was appropriately responsive to concerns and willing to refine its model as the results of its past performance became available.
Plaintiff's final challenge is that HHS's transfer formula unfairly discriminates against bronze plans. The crux of the challenge is that the collection of flaws it identified in this lawsuit resulted in a formula that disproportionately harmed bronze plans to the point of driving them out of business. That result, according to plaintiff, must be illegal, because the ACA clearly intended consumers to have the option of choosing bronze plans. See 42 U.S.C. § 18022(d)(1)(A).
Plaintiff acknowledges that this issue was not raised before the agency in the 2014-2017 benefit years. (Pl. Reply in Supp. Summ. J. at 4 n.8; Pl. Opp'n to Mot. to Strike at 4). It does not contend that that outcome — to the extent it is not merely a function of bronze plans attracting healthier enrollees and therefore confounded with the very problem that the risk-adjustment program seeks to remedy — was so apparent as to render the 2014-2017 rulemakings arbitrary and capricious at the time they occurred. Therefore, because the Court has rejected plaintiff's contention that its 2018-benefit-year comment functioned as a petition for rulemaking under 5 U.S.C. § 553(e) which the agency denied, it need not address this objection. (See also Hearing Tr. 36:10-14 (admitting that the effect of risk adjustment on bronze plans is "data driven from hindsight")).
In summary, and for the foregoing reasons, the challenged regulations were not arbitrary and capricious, and therefore did not violate the APA, 5 U.S.C. § 706; nor did they contravene the statute providing for risk adjustment, 42 U.S.C. § 18063. Defendants' motion to strike portions of the administrative record is GRANTED in part and DENIED in part, as set forth above. Defendants' motion for summary judgment is GRANTED, and plaintiff's motion for summary judgment is DENIED.
The notice-and-comment period for the 2018 benefit year took place somewhat earlier — the rule was proposed in September 2016 and finalized in December 2016. HHS Notice of Benefit and Payment Parameters for 2018, 81 Fed. Reg. 61,456 (Sept. 6, 2016) ("2018 Proposed Rule"); 2018 Final Rule, 81 Fed. Reg. 94,058 (Dec. 22, 2016).
It appears that HHS had ambitions to provide a proposed rule in October of the year two years prior to the applicable benefit year and the final rule in the following January. See Standards Related to Reinsurance, Risk Corridors and Risk Adjustment, 76 Fed. Reg. 41,930, 41,933 (July 15, 2011) (proposed rule).
(Id. (footnote omitted)).
Note that the BCBSA's comment to the 2014 proposed rule, submitted on December 28, 2012, says the opposite (albeit in the reinsurance context): "[D]ue to the imperfect correlation of risk scores to actual health-care expenses, BCBSA does not believe that over-compensation for high-risk individuals will occur. Generally, risk-adjustment models tend to underestimate costs for high-risk claimants and therefore, in our opinion, it is unlikely that reinsurance payments will result in overcompensation." (Def. Reply in Supp. Summ. J. Ex. K at 52).
