Opinion
Case No. 17-cv-00183-VC
12-23-2019
RACHEL CONDRY, et al., Plaintiffs, v. UNITEDHEALTH GROUP, INC., et al., Defendants.
ORDER GRANTING IN PART AND DENYING IN PART THE MOTION FOR CLASS CERTIFICATION
Re: Dkt. No. 222, 244, 245, 246, 247
The motion for class certification is granted in part and denied in part. This ruling assumes that the reader is familiar with the facts of the case, the parties' arguments, and the Rule 23 requirements, as well as this Court's earlier rulings on the cross-motions for summary judgment as to the named plaintiffs, the prior class certification motion, and the motion to intervene. A case management conference is scheduled for January 22, 2020 at 10:00 a.m. to discuss next steps. A joint case management statement is due January 15, 2020.
Denial-Letter Class
The Court ruled at summary judgment that United Healthcare, when it denied five named plaintiffs' claims for reimbursement of out-of-network lactation services, violated ERISA's requirement that the plan administrator "write a denial in a manner calculated to be understood by the claimant." 29 C.F.R. § 2560.503-1(g); 29 U.S.C. § 1133; Booton v. Lockheed Medical Benefit Plan, 110 F.3d 1461, 1463 (9th Cir. 1997). The plaintiffs now seek certification of a class of ERISA plan participants who received the same denial letters as the five named plaintiffs, with an eye towards a court order requiring United Healthcare to send class members new letters that explain the basis for denial in a comprehensible fashion (which would, in turn, allow participants to meaningfully assess whether to contest the denial). United Healthcare's primary argument against certification of this class is that some class members may have had subsequent communications with United Healthcare, and those subsequent communications may have resolved the dispute between plan and participant (or at least rendered understandable a denial that was initially incomprehensible). Although it's safe to assume that some members of the class will fit this description, that's not a reason to deny the motion for class certification. United Healthcare engaged in the same conduct with respect to each of the proposed class members—sending an incomprehensible denial letter. Although subsequent communications may have resolved disputes about benefits, it does not change the fact that United Healthcare's denial letters to these class members violated ERISA in the same way as to each participant. An appropriate remedy, therefore, is to order United Healthcare to send a follow-up letter to each class member. The new letter can be worded so as to emphasize that if a participant believes her dispute with the company was mooted by activity or communications subsequent to the initial denial letter, she need not take further action in response to the new letter. Accordingly, this class is certified pursuant to Federal Rule of Civil Procedure 23(b)(2). See, e.g., Wit v. United Behavioral Health, 317 F.R.D. 106, 138 (N.D. Cal. 2016); cf. Kartman v. State Farm Mutual Automobile Insurance Co., 634 F.3d 883, 893 (7th Cir. 2011). The parties should include a stipulated draft letter to class members with the next case management statement. If they cannot agree on language after a good-faith effort, they should submit competing draft letters.
Claims Reprocessing Class
In contrast, the plaintiffs' request to certify a nationwide class of people who were denied coverage for out-of-network lactation services, for the purpose of ordering United Healthcare to reprocess all those claims under the correct standard, is an overreach.
The plaintiffs actually seek certification of two classes—a class of people denied lactation coverage who were members of ERISA plans, and a similar class of people denied coverage who were members of non-ERISA plans. Because the issues are the same for both proposed classes, this ruling discusses them together.
To be sure, there is overwhelming evidence that United Healthcare's efforts to ensure that participants would receive coverage for lactation services as required by the Affordable Care Act were woefully inadequate. The company seemingly made no effort to compile comprehensive lists of in-network lactation providers, thus making it difficult for plan participants to determine whether such services were available in-network. Moreover, United Healthcare sometimes miscommunicated with participants who called to inquire about coverage, telling them as a blanket matter that out-of-network services were not included. Several documents, such as the Coverage Determination Guide that United Healthcare prepared for providers to use when submitting claims, stated that there was no obligation to cover out-of-network lactation services, which was true in situations where in-network services were available to the participant but false in situations where in-network services were unavailable. United Healthcare appeared to be operating on the assumption that in-network lactation services would be available to participants in its plans without inquiring whether that was actually so, and without communicating adequately to plan participants about their right to coverage for out-of-network lactation services if in-network services were unavailable. Moreover, as the internal emails reflect, the company was aware of these problems yet chose not to address them in a meaningful way. As a result, United Healthcare undoubtedly caused a significant number of mothers and their newborn babies to lose out on coverage for lactation services that they should have received under the ACA. Indeed, this happened to some of the named plaintiffs in this case. United Healthcare's misconduct, which appears to be ongoing, would presumably support a classwide claim for prospective relief—specifically, an injunction requiring the company to adopt reforms to better ensure coverage for lactation services in the future.
But as discussed in prior rulings, none of the named plaintiffs in this case has standing to seek prospective injunctive relief. See Order Denying Motion for Class Certification, Dkt. 213 at 4-5. Thus, these plaintiffs are left to seek certification of a "reprocessing class"—that is, a class consisting of all people denied coverage for out-of-network lactation services for the purpose of ordering that United Healthcare reprocess their claims. They contend that certification of such a class is warranted because United Healthcare had a blanket nationwide policy of adjudicating claims for out-of-network coverage without reference to whether in-network services were available to plan participants.
Presumably, if United Healthcare indeed applied a uniform standard or took a uniform approach to claims for out-of-network lactation services, class certification would be appropriate. See Wit v. United Behavioral Health, 317 F.R.D. 106, 138 (N.D. Cal. 2016); Des Roches v. California Physicians' Serv., 320 F.R.D. 486, 510 (N.D. Cal. 2017). But the evidence submitted in this case (in connection with both rounds of class certification briefing as well as the cross-motions for summary judgment as to the named plaintiffs) does not demonstrate a uniform standard or practice. If anything, the evidence undermines the plaintiffs' assertion that a uniform standard or approach existed with respect to coverage for out-of-network lactation services.
As a preliminary matter, the plaintiffs have proposed a class that includes two distinct groups of people: (i) those who received out-of-network lactation services, submitted claims, and were denied; and (ii) those who received out-of-network services but never submitted claims at all. It makes no sense to include the second group when the asserted basis for class certification is that the wrong standard was applied to claims, and when the asserted purpose of class certification is to obtain reprocessing of those claims. Moreover, each of the named plaintiffs falls in the first group; there is no named plaintiff who contends her right to coverage was violated even though she didn't submit a claim for coverage. So there is no named plaintiff whose claim is typical of the second group, or who could adequately represent that group. After all, the primary question relating to liability would be different for each group—for the first group, it would involve whether United Healthcare applied a uniform (and uniformly incorrect) standard or practice to their claims; for the second group, it would presumably involve inquiry into whether participants were adequately informed of their potential right to coverage for out-of-network services.
But even taking the first group—people whose claims for out-of-network lactation services were denied—the plaintiffs have not met their burden of demonstrating that United Healthcare applied a uniform standard or practice to those claims. Indeed, the plaintiffs' own presentation raises more questions than answers in this regard. The plaintiffs have identified what they believe to be the universe of out-of-network lactation claims submitted to United Healthcare during the class period. They note that roughly 88 percent of the claims were partially or fully denied (which would be contrary to the ACA if no in-network services were available), meaning that roughly 12 percent of them were fully granted (as the ACA would require if no in-network services were available). See Dkt. 222 at 19. If United Healthcare applied a uniform policy to all these claims—and if the uniform policy was to deny out-of-network claims without regard to the availability of in-network services—why were 12 percent of the claims fully granted? The plaintiffs mostly leave the Court to speculate on this question. That itself is a problem, because the burden is on the plaintiffs to demonstrate that class certification would be appropriate. But in any event, there is reason to believe, based on the limited evidence the parties have presented, that the varying results are the product of varying practices rather than a uniform one.
The plaintiffs' dataset includes claims from the beginning of the class period in 2012 through December 31, 2018, although the class period is ongoing. Dkt. 222-1 at 2.
The ACA mandates free preventive services. So if an insurer does not have providers in its network who are available to provide lactation services, and a plan participant receives those services out of network, the ACA regulations not only require that the insurer cover those out-of-network services, but furthermore forbid the insurer from imposing "cost-sharing requirements" in which the participant has to pay some portion of the charges through a deductible, co-pay, coinsurance, or similar mechanism. See 29 C.F.R. 2590.715-2713(a)(3)(ii). In other words, if lactation services are not available free of charge in-network, then the insurer can't make its customers pay anything when they receive lactation services out-of-network.
Incidentally, the admissibility of the plaintiffs' analysis is hotly contested, and there are reasons to suspect that it's not admissible. The Court will assume its admissibility for purposes of this ruling only.
To understand why this unexplained 12/88 percent split undermines the plaintiffs' efforts to certify a class, it's necessary to consider how United Healthcare's plans are supposed to work when a plan participant seeks lactation services. At least some (and perhaps all) of the plans in which the proposed class members participated offer mechanisms for people to seek authorization to receive coverage for out-of-network care—mechanisms that would presumably apply in situations where ACA-required lactation services are not available in-network. For example, the plan for one of the named plaintiffs, Felicity Barber of San Francisco, states: "If medically appropriate care from a qualified provider cannot be provided within the network, we will arrange for the required care with an available and accessible non-network provider." See Dkt. 102-13 at 27. The plan goes on to say that in this situation, the participant's "network physician" will notify United Healthcare of the need for out-of-network services, in which case the company "will work with you and your network physician to coordinate care through a non-network provider." Id. at 29. And the plan sets forth a process for submitting claims for services from a non-network provider. Id. at 78.
Most of the mechanisms appear to involve processes to obtain approval for receiving out-of-network services—what the parties call a "gap exception"—prior to receiving the services. Although it is not clear from the record, it's at least possible that some mechanisms may involve after-the-fact approval.
With that in mind, let's return to the plaintiffs' analysis of claims for out-of-network lactation services. There are at least three possible explanations for the 12/88 percent split. First, perhaps the 12 percent had their claims granted because they successfully used their plans' mechanisms for obtaining coverage for out-of-network services, while the 88 percent had their claims denied either because the mechanisms revealed in-network services or because these plan participants failed to utilize the mechanisms in the first place. In this scenario, perhaps class certification would be appropriate on the plaintiffs' "uniform policy" theory, but it would not result in a reprocessing of claims. Instead it would result in a grant of summary judgment for United Healthcare and against the class, because this would presumably reflect compliance with the ACA's requirement that lactation services be covered out-of-network when in-network services are unavailable.
This scenario assumes that the mechanisms for obtaining approval of coverage for out-of-network services are reasonable. An ineffective or overly burdensome mechanism would presumably not comply with the ACA, but as discussed later in this ruling, the plaintiffs have not shown that United Healthcare's mechanisms are uniformly ineffective.
A second possibility is that the 12 percent had their claims granted for some reason unrelated to the availability of in-network lactation services, while the 88 percent had their claims denied without regard to the availability of in-network lactation services. Under this scenario, all claims were considered without regard to the availability of in-network services, suggesting a uniform (and uniformly incorrect) policy that would support certification of a class and an order requiring United Healthcare to reprocess denied claims under the correct standard. But the plaintiffs merely speculate that the 12 percent had their claims granted for reasons unrelated to the availability of in-network services, without presenting any evidence of what those reasons might have been. For example, at oral argument plaintiffs' counsel speculated, without evidence to back it up, that perhaps the 12 percent had particularly generous plans that automatically provide coverage for out-of-network services, without regard to the availability in-network services. This sort of speculation is insufficient to support a conclusion that all claims were considered without reference to the availability of in-network services.
The third possibility, which seems at least as realistic as the other two, is that the systems worked adequately for some proposed class members and not for others. Imagine the following scenarios, each of which is plausible and none of which can be ruled out on the evidence presented:
• A plan participant seeks pre-approval for out-of-network services, promptly receives that approval, and ultimately receives reimbursement for her claim. This would put her in the 12 percent, and it would be an example of United Healthcare applying an ACA-compliant standard to her claim.
• A participant seeks pre-approval for out-of-network services but is denied, even though in-network services are unavailable. This would put her in the 88 percent, and it would be an example of United Healthcare applying a noncompliant standard to her claim.
• A participant seeks pre-approval for out-of-network services, she is denied because in-network services are truly available, she nonetheless obtains services from an out-of-network provider whom she prefers, and she seeks reimbursement but United Healthcare denies the claim. This also would put her in the 88 percent, but it would be an example of United Healthcare applying an ACA-compliant standard to her claim.
• A participant does not seek pre-approval and goes straight to an out-of-network provider, despite the fact that in-network providers are truly available. (Two of the named plaintiffs lost at summary judgment because they fit this description based on the evidence presented.) This participant would also fall in the 88 percent in the plaintiffs' analysis, but there is no indication that United Healthcare applied a noncompliant standard to her claim.The plaintiffs have not presented evidence that would allow the Court to reach a conclusion, or even to make an estimate, about how many participants fit into each category identified above.
• A participant does not seek pre-approval and goes straight to an out-of-network provider, later demonstrates that lactation services were not available in-network, and is denied reimbursement because she did not seek pre-approval. This person too would fall in the 88 percent. But perhaps it would be appropriate to deny her claim because she failed to utilize the pre-approval mechanism.
United Healthcare has offered its own analysis, which provides limited support for the possibility that at least some proposed class members fit this description. United Healthcare's analysis is for both in-network and out-of-network lactation services. It suggests that the large majority of claims for lactation services were for services provided in-network. Although this analysis may be incomplete and therefore may be of limited value, it provides at least a potential explanation for why, in the plaintiffs' analysis, only 12 percent of claims for out-of-network services were granted. If in-network lactation services were available to most participants, that could help explain why most claims for out-of-network services were denied. Indeed, the results of United Healthcare's analysis of out-of-network claims is comparable to that of the plaintiffs: approximately 17 percent of claims for out-of-network services were fully granted. See Dkt. 231 at 3.
Compounding the problem is the fact that the plaintiffs have proposed a nationwide class that includes participants in many different health plans. Although all these plans are administered by United Healthcare or one of its affiliates, they are for different employers and cover different geographic regions. While the parties have submitted the plan documents for a few of the plans at issue, they have not submitted the documents for others. They have not even presented evidence as to the total number of plans implicated by the proposed class. The incomplete and murky nature of this evidentiary submission is likely enough, on its own, to defeat the motion for class certification. That's because the experiences of proposed class members may have varied by plan, or by region, or both.
Consider, for instance, the evidence presented in connection with the cross-motions for summary judgment as to the named plaintiffs. Neither of the two named plaintiffs from the San Francisco Bay Area—Felicity Barber or Rachel Condry—tried to find an in-network provider of lactation services before seeking care outside their plans' network. See Dkt. 146 at 4. And there was unrebutted evidence that their plans had in-network lactation providers near them. Id. Thus, there is no reason to believe that the pre-approval mechanisms would have failed these plaintiffs had they attempted to use them. In contrast, one named plaintiff from Philadelphia—Jance Hoy—tried to find an in-network provider of lactation services but could not do so. Id. And United Healthcare presented no admissible evidence that in-network providers were in fact available to Hoy at the relevant time. Id. A United Healthcare representative spoke with Hoy and confirmed the unavailability of in-network providers, but asserted nonetheless that Hoy would not be covered for out-of-network services. Id. And when Hoy sought reimbursement for out-of-network services, United Healthcare denied her claims. Id. A similar fate was suffered by Teresa Harris, also from Philadelphia, whose motion to intervene as a named plaintiff was denied in a prior ruling. Harris was denied coverage for out-of-network services and appealed, making a facially compelling case for entitlement to reimbursement pursuant to the ACA's mandate. Dkt. 226-2 at 5. But United Healthcare rejected her appeal without providing any meaningful explanation, without addressing Harris's contention that in-network services were unavailable, and without responding to Harris's assertion that the ACA required reimbursement under these circumstances. Id. at 18-22. Thus, although it's risky to reach any firm conclusion based on the experiences of a few people, particularly where the evidence presented about those individual experiences was so sparse, there is a real possibility that different approaches to ACA compliance were taken for different plans or in different regions.
The plaintiffs contend that the various possibilities about what actually happened to a participant don't matter, because the pre-approval mechanisms could not have been adequate to comply with the ACA's lactation coverage mandate, particularly when United Healthcare did an otherwise terrible job of communicating with its members about how to obtain lactation services either in-network or out-of-network. It's certainly true, as already noted, that United Healthcare did an exceedingly poor job complying with the ACA. But the plaintiffs have presented no evidence to support their contention that the pre-approval mechanisms were uniformly inadequate under the ACA. To be sure, the existence of a mechanism to obtain pre-authorization for out-of-network coverage would only allow a plan to achieve compliance with the ACA if it were an effective way of ensuring that participants receive lactation coverage. If the mechanism were inadequate, then United Healthcare would presumably violate the ACA by denying a subsequent claim for coverage merely because the participant failed to utilize it. And if pre-approval mechanisms were uniformly inadequate across plans, then perhaps this would support the contention that United Healthcare participants were uniformly subject to a noncompliant policy or standard. But from an evidentiary standpoint, the plaintiffs have not offered anything about how these mechanisms work in practice, or about whether the mechanisms work the same across plans or regions. And recall that, at least for the few plans whose documents the parties have placed in the record, participants are informed as a general matter (that is, not just for lactation services) that they can obtain approval for out-of-network services when those services are not available in-network. Indeed, jumping through some extra administrative hoops to obtain out-of-network coverage is a familiar concept, at least for people accustomed to using health insurance. And the plan documents themselves do not appear to foreclose the possibility of using pre-approval processes for lactation services (although admittedly United Healthcare did not describe the ACA's requirements accurately in other documents, such as the Coverage Determination Guide). Thus, there is reason to believe that a well-functioning pre-approval mechanism could be adequate to fulfill an insurer's obligations under the ACA's lactation services mandate, and such a mechanism may have been available here for at least some proposed class members.
Similarly, while an insurer could presumably rely on other types of mechanisms (besides pre-approval) to comply with the ACA, those mechanisms could only permit compliance with the ACA if they were effective. For example, United Healthcare makes repeated reference in its brief to the appeal processes available in its plans. But clearly an insurer could not comply with the ACA by denying all out-of-network claims and relying solely on its appeals process to identify claims that should have been granted.
The plaintiffs speculate that it will be difficult for mothers to utilize pre-approval mechanisms when they have newborn babies and are struggling with lactation problems. This is certainly a plausible theory, but it would presumably depend on how easy or difficult it is to invoke any given pre-approval process. Because the plaintiffs have not offered any evidence on this issue, the plaintiffs' assertions about difficulties in utilizing a pre-approval mechanism are not a basis to conclude that the proposed class members have been subject to a uniform practice nationwide.
Similarly, the plaintiffs point out that according to United Healthcare's data, only a very small number of "gap requests" were approved during the class period, suggesting that the preapproval mechanisms were inadequate across the board. But this evidence is inconclusive on the current evidentiary record: the small number of gap exceptions might be explained by United Healthcare's contention that most new mothers were able to obtain lactation services innetwork.
In sum, the data and evidence the plaintiffs have provided doesn't come close to proving that United Healthcare failed to comply with the ACA in a uniform way. See Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011) (plaintiffs' claims "must depend upon a common contention . . . capable of classwide resolution"); cf. Wit v. United Behavioral Health, 317 F.R.D. 106, 127 (N.D. Cal. 2016) ("The harm alleged by Plaintiffs—the promulgation and application of defective guidelines to the putative class members—is common to all of the putative class members."); Des Roches v. California Physicians' Service, 320 F.R.D. 486, 499 (N.D. Cal. 2017) ("[I]t is clear that the Guidelines were used in a similar way for all class members . . . ."). This precludes a finding, on this record, that the members of the proposed nationwide class had their claims denied due to a uniform standard or practice. This aspect of the motion for class certification is therefore denied.
United Healthcare's administrative motion to file portions of its Exhibit F under seal is granted. The motions to exclude experts Hanley, McGlone, and Morton are denied as moot. --------
IT IS SO ORDERED. Dated: December 23, 2019
/s/_________
VINCE CHHABRIA
United States District Judge