Northern District of California Certifies Previously Denied Indirect Purchaser Damages Class in Sutter Health Antitrust Litigation

Lee Brand
Isabella Borges
Pillsbury Winthrop Shaw Pittman LLP

On July 30, 2020, Magistrate Judge Laurel Beeler of the Northern District of California certified a class of indirect purchasers seeking damages for Sutter Health’s alleged violations of the Sherman Antitrust Act, the Cartwright Act, and the California Unfair Competition Law.  Sidibe v. Sutter Health, No. 12-cv-04854-LB, 2020 WL 4368221 (N.D. Cal. Jul. 30, 2020) (“Sidibe II”).  The class consists of entities and individuals throughout California that paid health insurance premiums that were allegedly inflated due to Sutter’s improper tying conduct.  Id. at *1.  Specifically, Plaintiffs allege that Sutter uses its market power in Northern California to force five health insurance plans to accept supra-competitive prices in other geographic markets, and that those prices are passed through to consumers in health insurance premiums.  Id.

In August 2019, Judge Beeler had certified a class of indirect purchasers seeking injunctive relief under Federal Rule of Civil Procedure 23(b)(2) but had declined to certify a Rule 23(b)(3) damages class.  Sidibe v. Sutter Health, 333 F.R.D. 463 (N.D. Cal. 2019) (“Sidibe I”).  At that point, the Court held that a damages class could not be certified because Plaintiffs’ methodology for assessing antitrust injury and damages was not reliable to prove or calculate either the alleged overcharges to the five health plans or how those overcharges were passed through in premiums.  Id. at 493.  In her July 2020 order, however, Judge Beeler found that the indirect purchasers had resolved these issues for the majority of their purported class period by conducting new and more comprehensive data analysis.  Sidibe II, at *12.

Background

Sutter is one of the largest health systems in Northern California, with 24 hospitals and approximately 5,200 patient beds.  Sidibe I, at 472.  Plaintiffs allege that health insurance plans were previously able to negotiate with Sutter on a hospital-by-hospital basis, including some hospitals in the plans’ provider networks while excluding others.  Id.  Beginning in the early 2000s, however, Sutter allegedly began requiring “systemwide” contracts that effectively require a health plan that wants to include one Sutter hospital in its provider network to include all Sutter hospitals and to refrain from incentivizing enrollees to use lower-cost, non-Sutter hospitals.  Id.  Plaintiffs further contend that Sutter can impose these systemwide contracts because health plans are required to offer their enrollees at least one nearby in-network hospital, and Sutter has either the only hospital or the predominant hospital in several Northern California markets.  Id. at 474.  Although the five health plans to which class members belong would likely have to contract with Sutter in the these Northern California tying markets no matter what, Plaintiffs claim that but for Sutter’s anticompetitive systemwide contracts, the plans would pay lower rates overall—and pass through that savings as lower premiums—because they could exclude or threaten to exclude overpriced Sutter hospitals in other tied markets.  Id. at 475.

Classwide Damages Methodology

In Plaintiffs’ initial attempt to certify a damages class, their expert Dr. Tasneem Chipty had used a regression-analysis model to calculate what portion of Sutter’s charges to two of the health plans, accounting for 70% of the class, were attributable to Sutter’s alleged anticompetitive practices.  Id. at 493.  Dr. Chipty suggested that the other three health plans could be addressed through further modeling or extrapolation, but was not sure whether data was available to separately calculate overcharges for those plans and could not explain how to extrapolate from her prior results given that she had calculated significantly different overcharges for the first two plans.  Id. at 493-94.  As such, the Court found that even if Plaintiffs had soundly calculated overcharges for two plans, they had failed to demonstrate calculable classwide overcharges given the absence of a method to address the other 30% of the class.  Id. at 494.  The Court also found that Dr. Chipty had assumed rather than proven that the health plans passed through 100% of Sutter’s overcharges as premiums, and thus found Dr. Chipty’s analyses insufficient to show classwide antitrust injury and damages on that basis as well.  Id. at 496-98.

In support of Plaintiffs’ second attempt to certify a class under Rule 23(b)(3), Dr. Chipty applied a common methodology to estimate overcharges to all five class health plans, and Sutter dropped its challenge that Plaintiffs had not reliably shown a method for proving overcharges.  Sidibe II, at *1 n.8.  Sutter continued to dispute, however, the soundness of Plaintiffs’ passthrough methodology.  Id. at *2.  This time, rather than assume a passthrough rate, Dr. Chipity performed a regression analysis of the relationship between premium prices and medical costs, controlling for 14 variables and calculating an overall weighted average passthrough of 97.16%.  Id. at *2.  This new passthrough rate calculation depended on econometric analyses of Medical Loss Ratio (“MLR”) data, statewide data on premium revenues and all incurred medical costs that the class health plans report to ensure compliance with the Affordable Care Act (“ACA”).  Id. at *3.  Dr. Chipty also identified six key features that guided this regression analysis: (1) the single link in the supply chain and single passthrough from plan to customer; (2) the pooling of medical risk by health insurance, meaning a premium reflects the average expected cost of care; (3) the need to include information on all medical costs; (4) that premiums can be a function of both medical costs and product attributes, such as shared costs (e.g., deductibles) and different levels of access to medical care (e.g., HMOs); (5) that the regulatory environment limits flexibility in setting premiums; and (6) the need to control for competitive conditions.  Id. at *3-4.

In opposing Plaintiffs’ second attempt to certify a damages class, Sutter primarily argued that Dr. Chipty inappropriately averaged harm and relied on unverified MLR data.  Id. at *9.  The Court disagreed, finding that this objection did not preclude certification of a Rule 23(b)(3) class.  Id.  While Judge Beeler recognized that courts reject the use of averaging in calculating passthrough rates where experts have not shown sound methodology, she found that Dr. Chipty’s new analysis did not generally suffer from such methodological flaws.  Id. at *10-11.  In particular, Dr. Chipty’s analysis did not ignore variation, sufficiently explained the use of averaging, and corroborated its use of MLR data with other quantitative and qualitative evidence.  Id. at *11.  The one exception was with regard to the initial 2008 to 2010 portion of Plaintiffs’ purported class period, for which MLR data did not exist because the relevant reporting requirement began in 2011 after passage of the ACA.  Id.  For these years, the Court once again declined to certify a class where Plaintiffs contended that they could soundly calculate classwide damages by extrapolating from available data to fill a data gap.  Id.  This time, the Court found that such extrapolation would be unsound because Dr. Chipty had not evaluated whether the ACA had impacted passthrough rates.  Id.

Other Arguments

Sutter also challenged Plaintiffs’ failure to offer a method to allocate premium increases between employers and employees.  Id. at 12.  The Court found that this did not defeat class certification because it went to the allocation rather than calculation of damages, and because it could be evaluated on a percentage basis that reflected actual splits of premium payments.  Id.  In addition, Sutter argued that the Court should strike Dr. Chipty’s reliance on MLR data because it was disclosed after the deadline for expert reports.  Id.  The Court disagreed, explaining that it was permissible for Dr. Chipty to use new data after her previous methodology had been rejected in connection with the first class certification motion.  Id.  Finally, Sutter argued that Dr. Chipty’s model was inappropriately novel and that she had not previously measured passthrough rates of medical costs to insurance premiums.  Id.  Because the Court had already found Dr. Chipty’s methodology sufficient, and Plaintiffs’ did not challenge her overall expertise, the Court rejected this argument as well.  Id.

Conclusion

In sum, the District Court’s opinion offers several key takeaways for antitrust practitioners regarding the econometric analysis sufficient to show classwide impact for purposes of certification under Rule 23(b)(3).  For one, it serves as a reminder that when a class is composed of indirect purchasers, sound analysis must support the alleged passthrough rate, even if that rate is nearly 100%.  In addition, it demonstrates that attempts to address data gaps through extrapolation must also be soundly supported, and that in the absence of a sound basis for extrapolation, data gaps may preclude plaintiffs from showing classwide damages for some or all of an alleged class period.