Osinek v. Kaiser Permanente

• Court: District Court, N.D. California
• Decided: May 5, 2022
• Docket: 3:13-cv-03891
• Status: Unknown

Opinion

1 2 3 4 UNITED STATES DISTRICT COURT 5 NORTHERN DISTRICT OF CALIFORNIA 6 7 UNITED STATES OF AMERICA ex rel. Case No. 13-cv-03891-EMC RONDA OSINEK, 8 CONSOLIDATED MEMBER CASES Plaintiff,

9 v. Case No. 16-cv-01558-EMC 10 Case No. 16-cv-05337-EMC PERMANENTE MEDICAL GROUP, INC, Case No. 18-cv-01347-EMC 11 et al., Case No. 21-cv-03124-EMC Case No. 21-cv-03894-EMC 12 Defendants. ORDER GRANTING IN PART AND 13 DENYING IN PART DEFENDANTS’ MOTION TO DISMISS 14 Docket No. 141 15 16 17 18 The above cases are all predicated on allegations that various Kaiser entities1 submitted 19 false claims for payment to the federal government as part of the Medicare Part C program, which 20 is also called Medicare Advantage. Osinek was the first-filed case and was followed by five other 21 cases: Taylor, Arefi, Stein, Bryant, and Bicocca.2 The cases were consolidated in June 2021. See 22 Osinek, Docket No. 61 (order). In July 2021, the United States filed a notice that it was 23 24 1 The Kaiser entities are, generally speaking, various Health Plans, Hospitals, and Medical 25 Groups. These entities “publicly hold themselves out and do business collectively as an integrated healthcare provider called “Kaiser Permanente.” U.S. Compl. ¶ 28. 26

2 Taylor was initially filed in the District of Colorado in 2014; Arefi in the Central District of 27 California in 2015; Stein in the Central District of California in 2016; Bryant in the Northern 1 intervening in part and declining to intervene in part.3 See Osinek, Docket No. 64 (notice of 2 election). 3 Currently pending before the Court is Defendants’ motion to dismiss based on the first-to- 4 file bar in the False Claims Act (“FCA”). The relevant FCA provision states as follows: “When a 5 person brings an action under this subsection, no person other than the Government may intervene 6 or bring a related action based on the facts underlying the pending action.” 31 U.S.C. § 7 3730(b)(5) (emphasis added). The purpose of the first-to-file bar is twofold: (1) “to promote 8 incentives for whistle-blowing insiders” and (2) “[to] prevent opportunistic successive plaintiffs.” 9 United States ex rel. Lujan v. Hughes Aircraft Co., 243 F.3d 1181, 1187 (9th Cir. 2001). 10 Defendants argue that, with limited exceptions, the claims presented by the cases that follow 11 Osinek are barred. The Arefi plaintiffs have filed a statement of nonopposition with respect to the 12 motion to dismiss their case. See Docket No. 143 (nonopposition). The plaintiffs in all other 13 cases have opposed dismissal. 14 Having considered the parties’ briefs, as well as the oral argument of counsel, the Court 15 hereby GRANTS in part and DENIES in part Defendants’ motion. 16 I. FACTUAL & PROCEDURAL BACKGROUND 17 A. United States’ Complaint-in-Intervention 18 Although the United States’ Complaint-in-Intervention is not at issue in the pending 19 motion, the Court begins with this pleading as it provides a good overview of the Medicare 20 background. 21

22 3 The notice stated as follows:

23 Specifically, the United States intervenes on the allegations that defendants Kaiser Permanente; Kaiser Foundation Health Plan, Inc.; 24 Kaiser Foundation Health Plan of Colorado; The Permanente Medical Group, Inc.; Southern California Permanente Medical 25 Group, Inc.; and Colorado Permanente Medical Group, P.C.; submitted, or caused to be submitted, false claims for risk- 26 adjustment payments based on diagnoses improperly added via addenda under Medicare Part C from the years 2009 until present. 27 The United States declines to intervene on all other allegations. 1 “Medicare is a federally operated health insurance program.” U.S. Compl. ¶ 52. It has 2 four parts: 3 • Part A covers inpatient and institutional care. 4 • Part B covers outpatient care. 5 • Part C is the Medicare Advantage program at issue in this case. 6 • Part D covers prescription drugs. 7 See U.S. Compl. ¶ 52. 8 Parts A and B are “traditional” Medicare.

9 [T]he Government reimburses healthcare providers using a fee-for- service system, in which providers submit claims to CMS [Centers 10 for Medicare and Medicaid Services] for healthcare services actually rendered, such as a provider officer visit or hospital stay. CMS then 11 pays the providers directly for each service based on payment rates predetermined by the Government. 12 13 U.S. Compl. ¶ 53. 14 A Medicare beneficiary can opt out of traditional Medicare and enroll instead in a 15 Medicare Advantage plan managed by a Medicare Advantage Organization (“MAO”). See U.S. 16 Compl. ¶ 54. “CMS reimburses [Medicare Advantage] plans differently than traditional 17 Medicare.” U.S. Compl. ¶ 58. Specifically, Medicare Advantage uses a “‘capitation’ payment 18 system.” United States ex rel. Silingo v. Wellpoint, Inc., 904 F.3d 667, 672 (9th Cir. 2018). Under 19 that system, “private health insurance organizations provide Medicare benefits in exchange for a 20 fixed monthly fee per person enrolled in the program – regardless of actual healthcare usage.” Id. 21 The fixed monthly fee for an enrollee is set as follows. First, there is a predetermined base 22 payment for each enrollee in a Medicare Advantage plan. See U.S. Compl. ¶ 57. Second, the base 23 payment is then adjusted “to account for (1) demographic factors such as age and gender (among 24 others) and (2) health status. This is known as risk adjustment.” U.S. Compl. ¶ 58. 25 Risk adjustment is accomplished by assigning each beneficiary a risk score, which “acts as 26 a multiplier that is applied to the [Medicare Advantage] plan’s base rate to determine the overall 27 monthly payment for the beneficiary.” U.S. Compl. ¶ 58. A beneficiary’s risk score is determined 1 as indicated above, is based on the patient’s demographic factors and health status. See U.S. 2 Compl. ¶ 59. With respect to health status, the model relies on diagnosis codes from the 3 International Classification of Diseases (“ICD”). See U.S. Compl. ¶ 60. “ICD diagnosis codes are 4 alphanumeric codes used by healthcare providers, insurance companies, and public health 5 agencies to represent medical conditions; every disease, injury, infection, and symptom has its 6 own code.” U.S. Compl. ¶ 62.

7 The ICD diagnosis codes included in the CMS-HCC model are grouped into categories of clinically related medical diagnoses that 8 comprise the HCCs (i.e., the categories). For example, various cancer diagnosis codes are grouped together (e.g., colorectal and 9 bladder cancers). The CMS-HCC model organizes related conditions into hierarchies based on disease severity and expected 10 cost. For example, various cancer HCCs are in the same hierarchy, with the HCC associated with metastatic cancer diagnosis codes as 11 the most severe. If a patient is diagnosed with conditions (diagnosis codes) that correspond to more than one HCC in a hierarchy, only 12 the most severe HCC is kept and any lower-ranking HCCs are dropped. 13 14 U.S. Compl. ¶ 63.

15 Each HCC is assigned a coefficient. CMS calculates a beneficiary’s risk score by adding the coefficients associated with each of the 16 beneficiary’s applicable demographic characteristics (such as age and gender) and the applicable HCCs, if any, that apply to the 17 beneficiary. A risk score of 1.0 reflects the average expected Medicare-incurred expenses. A risk score of 0.75 reflects expected 18 costs for a particular beneficiary that are 25% less than the estimated average costs for enrollees in the MA plan, and a risk score of 1.25 19 reflects expected costs that are 25% greater than the estimated average costs for enrollees in the MA plan. 20 21 U.S. Compl. ¶ 65. 22 The CMS-HCC model is prospective in the sense that it uses diagnoses made in a base 23 year (the ‘service year’), along with demographic information (such as age and gender, among 24 others), to predict costs for Medicare benefits and adjust payments for the following year (the 25 ‘payment year’).” U.S. Compl. ¶ 60. 26 “To combat the ‘incentive for [Medicare Advantage] organizations to potentially over- 27 report diagnoses,’ Medicare regulations require risk adjustment data to be produced according to 1 the diagnosis codes that MA Organizations submit to CMS for risk- adjustment purposes must be: 2 a. established by a qualified physician; 3 b. based on a face-to-face medical visit between the patient and 4 physician[4];

5 c. documented in the medical record; and

6 d. coded in compliance with the ICD [Official Guidelines for Coding and Reporting], including the limitation that the 7 condition must have required or affected patient care, treatment, or management for the visit.[5] 8 9 U.S. Compl. ¶ 87 (emphasis added); see also Silingo, 904 F.3d at 673 (also noting best practices). 10 “[I]t is an express condition of payment that a Medicare Advantage organization ‘certify 11 (based on best knowledge, information, and belief) that the [risk adjustment] data it submits . . . 12 are accurate, complete, and truthful.’” Id. (quoting 42 C.F.R. § 422.504(l)(2)). 13 B. Osinek Complaint 14 The Court turns next to the Osinek Complaint as it provides the baseline for the Court – 15 i.e., the Court will have to compare the Osinek Complaint with the complaints in the other cases to 16 determine whether the cases are related. See 31 U.S.C. § 3730(b)(5) (providing that, “[w]hen a 17 person brings an action under this subsection, no person other than the Government may intervene 18 or bring a related action based on the facts underlying the pending action”).6 19 The allegations below all come from Ms. Osinek’s original complaint filed in 2013. (Ms. 20 4 For example, 21

even if an MA organization knows that a patient was diagnosed in a 22 prior year with a chronic condition that tends not to go away, the MA organization may not submit the diagnosis for payment for the 23 current year unless the physician has a face-to-face visit with the patient in the current year and the chronic condition required or 24 affected care, management, or treatment during that patient visit.

25 U.S. Compl. ¶ 85.

26 5 “In other words, only those conditions that specifically mattered to the patient care, treatment, or management that the physician actually provided at the visit could be submitted to CMS for 27 payment.” U.S. Compl. ¶ 5. 1 Osinek filed an amended complaint in 2021. For the reasons discussed below, it is Ms. Osinek’s 2 original complaint that matters for purposes of the pending motion.) 3 Ms. Osinek has sued “Kaiser Permanente,” “a private provider of Medicare Advantage 4 insurance under Medicare Part C.” Osinek Compl. ¶ 2. Ms. Osinek describes “Kaiser 5 Permanente” as follows:

6 Kaiser Permanente is a California corporation with its principal place of business [in] Oakland, California 94612. Kaiser is one of 7 the largest Medicare Advantage organizations in the country and has more enrollees in its Medicare Advantage Plans than any other 8 organization in California. At all times relevant, Kaiser conducted business in California, including but not limited to providing 9 healthcare services through Medicare Advantage plans and to the general public in California. 10 11 Osinek Compl. ¶ 6. 12 According to the complaint, starting around 2007, Kaiser Permanente began a “scheme to 13 upcode diagnoses to ensure Medicare payments for reimbursable, high-value conditions.” Osinek 14 Compl. ¶ 2. Not surprisingly, Kaiser Permanente “focuses . . . on high value conditions” so that it 15 “can maximize its reimbursement from Medicare.” Osinek Compl. ¶ 25. High-value disease 16 conditions included, e.g., chronic kidney disease, congestive heart failure, depression, chronic 17 respiratory failure, cachexia/protein calories malnutrition, severe obesity, and seizure. See Osinek 18 Compl. ¶ 25. 19 Kaiser Permanente effectuated its upcoding scheme in various ways. For example: 20 • Data mining. Kaiser Permanente used “algorithms to identify [high-value] disease 21 conditions for data mining.” Osinek Compl. ¶ 25. “Kaiser identified the higher 22 value HCCs and then determined the diagnoses its doctors would need to make to 23 support the HCCs Kaiser wanted to submit for Medicare reimbursement.” Osinek 24 Compl. ¶ 25. 25 • Refreshing. Although not clearly described in the complaint, refreshing appears to 26 be a process related to chronic conditions. See Osinek Compl. ¶ 37 (alleging that 27 “Kaiser tracks and rewards physicians based on the percentage of chronic 1 Advantage plans are compensated based on medical conditions diagnosed in the 2 previous payment year. Therefore, if a patient has a chronic condition, then that 3 condition must be rediagnosed each year – i.e., refreshed. Presumably, Kaiser 4 Permanente used refreshing “to increase its billings for high value . . . HCCs,” 5 Osinek Compl. ¶ 24, because a doctor would be told to include the chronic 6 condition as a diagnosis for a visit even if that condition was not at issue in the 7 patient visit.7 8 • Guidance and policies. Kaiser Permanente provided guidance or policies that 9 supported upcoding. For example, “Kaiser told its physicians to diagnose chronic 10 kidney disease instead of the lower value nephritis or nephropathy.” Osinek 11 Compl. ¶ 26. As another example, “when CMS announces that HCCs are 12 eliminated (and no longer reimbursable by Medicare), Kaiser tells its physicians to 13 change coding practices to reflect new reimbursable codes. . . . In response to 14 CMS’s notification that HC 131 will be eliminated, Kaiser promptly sent materials 15 to its staff to begin prompting physicians to code diagnoses for acute kidney injury 16 instead of chronic kidney disease stage 1, 2, or 3, which will be included in the 17 2014 HCC list and reimbursable by Medicare.” Osinek Compl. ¶ 27. 18 • Addenda. In theory, “[a]ll relevant documentation is entered into a medical record 19 at the time of service,” but CMS recognizes “there may be times that a provider 20 will need to amend, correct, or enter documentation related to an encounter. CMS 21 7 The United States’ Complaint provides further context on refreshing. See, e.g., U.S. Compl. ¶ 7 22 (“Kaiser also employed a related data-mining program called ‘refresh,’ where Kaiser would mine patient medical files to find old diagnoses that had not yet been diagnosed in the current service 23 year. If a physician failed to address any of these old diagnoses at a patient visit, the physician would be provided a list of these ‘missed opportunities’ – i.e., opportunities for risk-adjustment 24 payment – to create an addendum to retrospectively add these diagnoses to the medical record.”); U.S. Compl. ¶ 151 (“Another category of Kaiser’s data-mining efforts focused on capturing 25 diagnoses that had been made in a prior year. Kaiser referred to this program as ‘refresh’ and to conditions that needed to be captured as ‘unrefreshed diagnoses.’ Kaiser created algorithms that 26 mined patients’ electronic medical records for any diagnoses that had been made in any setting during the past several (typically three) years. As detailed below, Kaiser meticulously monitored 27 and tracked these diagnoses, and if a physician failed to re-diagnose these conditions at a patient 1 expects supplemental documentation to be occasional and that delayed or amended 2 entries will be entered within a reasonable time frame. CMS will consider delayed 3 or amended explanations for diagnoses so long as the explanations are for 4 clarification and not for substantiating retroactive diagnoses.” Osinek Compl. ¶ 20 5 (emphasis added). Kaiser Permanente had its doctors use addenda to retroactively 6 diagnose – e.g., long after a patient visit, for a condition for which the patient was 7 not treated at the time of the face-to-face visit, based on tests run after the face-to- 8 face visit, to change a diagnosis to a higher value and more complicated form of 9 disease, without proper support/documentation, and/or using boilerplate language. 10 See, e.g., Osinek Compl. ¶¶ 28-32. 11 • Pressuring doctors. “Kaiser pressures its physicians to addend diagnoses and 12 capture the high value HCCs” – e.g., there is “an escalation process for physicians 13 who do not agree with the data mining prompts”; “[p]hysicians will have to meet 14 one-on-one with Data Quality Trainers if they refused to make diagnoses changes 15 that are presented by data mining”; “physicians have personal report cards based on 16 how they perform in certain areas [including response to refreshing and data 17 mining prompts], which are tied to their compensation”; and there are “mandatory 18 meetings called ‘coding parties,’ where physicians are gathered in a single room 19 with computers and asked to review past progress notes for addenda related to 20 revised medical diagnoses.” Osinek Compl. ¶¶ 33-35. 21 II. LEGAL ISSUES 22 Before the Court compares Osinek and the later-filed cases, it first takes into consideration 23 four legal issues related to the first-to-file bar, each of which will have an impact on the Court’s 24 comparison of the cases. 25 (1) Is the first-to-file bar jurisdictional in nature? 26 (2) In comparing the first-filed and later-filed actions, should a court look at the 27 original complaints or any amended complaints instead (assuming amended 1 (3) In comparing the first-filed and later-filed actions, must the facts in the actions be 2 identical in order for a court to apply the first-to-file bar? 3 (4) In comparing the first-filed and later-filed actions, how should a court proceed 4 where there are different defendants? 5 A. Jurisdiction 6 Although not all courts agree, the Ninth Circuit has expressly held that the first-to-file 7 provision (§ 3730(b)(5)) is jurisdictional in nature. See, e.g., United States ex rel. Hartpence v. 8 Kinetic Concepts, Inc., 792 F.3d 1121, 1130 (9th Cir. 2015) (stating that “[w]e treat the first-to-file 9 bar as jurisdictional”); see also United States ex rel. Marshall v. Univ. of TN Med. Ctr. Home Care 10 Servs., LLC, No. 3:17-CV-96, 2021 U.S. Dist. LEXIS 159167, at *41 & n.4 (E.D. Tenn. Aug. 23, 11 2021) (noting that the Fourth, Fifth, Sixth, Ninth, and Tenth Circuits have held that § 3730(b)(5) is 12 jurisdictional but that the D.C., First, Second, and Third Circuits have held that it is not; citing 13 cases). 14 The Ninth Circuit’s view of § 3730(b)(5) as jurisdictional is important because it impacts 15 which complaints should be considered when a court compares the first-filed and later-filed 16 actions. See Reply at 4 (making this same point). That issue is addressed below. 17 In their papers, the Stein plaintiffs argue that the first-to-file bar is not jurisdictional. In 18 support, they rely on Gonzalez v. Thaler, 565 U.S. 134 (2012), where the Supreme Court stated as 19 follows: “A rule is jurisdictional ‘[i]f the Legislature clearly states that a threshold limitation on a 20 statute’s scope shall count as jurisdictional.’” Id. at 141. The problem for the Stein plaintiffs is 21 that, post-Gonzalez, the Ninth Circuit issued Hartpence which clearly held that the first-to-file bar 22 is jurisdictional. See Hartpence, 792 F.3d at 1130. The Stein plaintiffs acknowledge Hartpence 23 but contend that the decision should not be given any weight as it relied solely on Lujan, 243 F.3d 24 at 1181, a pre-Gonzalez decision. The Court rejects the Stein Plaintiffs’ attempt to avoid 25 Hartpence. Hartpence is binding authority that is clearly on point and was issued post-Gonzalez. 26 It is not up to this Court to decide whether Hartpence was wrongly decided because the Ninth 27 Circuit did not explicitly address Gonzalez. 1 B. Original v. Amended Complaint 2 The next issue for the Court to consider is which complaints should be evaluated in 3 determining whether the first-filed and later-filed actions are related: the original complaints or the 4 amended complaints? All of the cases before the Court – including Osinek – have amended 5 complaints except for Arefi. (As noted above, the Arefi plaintiffs do not oppose Defendants’ 6 motion to dismiss their suit.) Below is a timeline with respect to the filing of the complaints. 7 Osinek Taylor Arefi (2015) Stein (2016) Bryant Bicocca 8 (2013) (2014) (2018) (2020) 8/22/2013 9 (original complaint) 10 10/22/2014 (original 11 complaint) 11/3/2014 12 (FAC) 9/4/2015 13 5/16/2016 (original 14 complaint) 11/3/2016 15 (FAC) 3/1/2018 16 (original complaint) 17 2/10/2020 (original 18 complaint) 19 10/9/2020 (FAC) 20 7/27/2021 – U.S. notice of election to intervene in part 7/29/2021 – Court order granting U.S. request to unseal complaints 21 10/7/2021 (FAC) 22 10/25/2021 – U.S. complaint in intervention 11/12/2021 23 (SAC) 11/15/2021 11/15/2021 24 (SAC) (FAC) 25 A number of courts have held that a court should compare (1) the original complaint in the 26 later-filed action with (2) whatever was the operative complaint in the first-filed action at the time 27 the later-filed action was filed (which in this case would be the original complaint in Osinek). See, 1 • Grynberg v. Koch Gateway Pipeline Co., 390 F.3d 1276, 1279 (10th Cir. 2004) 2 (stating that “[w]e judge whether § 3730(b)(5) barred Grynberg’s [later-filed] qui 3 tam action by looking at the facts as they existed at the time the action was 4 brought”; at the time the Grynberg suit was filed, “Precision’s 1992 amended 5 complaint [in the first-filed case] was pending in federal district court”). 6 • U.S. ex rel. Branch Consultants, L.L.C. v. Allstate Ins. Co., 782 F. Supp. 2d 248, 7 259 (E.D. La. 2011) (hereinafter “Branch II”) (indicating that “the Court should 8 look to the jurisdictional facts that existed at the time the action was filed, as 9 opposed to facts that existed when the relator later filed an amended complaint”). 10 • United States ex rel. Cestra v. Cephalon, Inc., No. 14-01842, 2014 U.S. Dist. 11 LEXIS 143745, at *7-8 (E.D. Pa. Oct. 9, 2014) (agreeing with Grynberg and 12 Branch). 13 • United States ex rel. Carter v. Halliburton, No. 1:11cv602 (JCC/JFA), 144 F. 14 Supp. 3d 869, 881 (E.D. Va. 2015) (noting, inter alia, that “[i]t is consistent with 15 the jurisdictional limitation to apply the first-to-file bar at the time the initial 16 complaint is filed, rather than when the complaint is amended”). 17 • United States ex rel. Marshall v. Univ. of TN Med. Ctr. Home Care Servs., LLC, 18 No. 3:17-CV-96, 2021 U.S. Dist. LEXIS 159167, at *24 (E.D. Tenn. Aug. 23, 19 2021) (also agreeing with Grynberg and Branch). 20 In the case at bar, Defendants advocate for this approach, and most of the plaintiffs agree – 21 but not all. See, e.g., Stein Opp’n at 3 n.2 (arguing that the Court should consider the SAC which 22 was filed in November 2021); Bicocca Opp’n at 5 (arguing that the Court should consider the 23 FAC which was filed in October 2020). 24 The district court in Branch II has provided the most extensive analysis as to why the 25 above approach should be followed. Branch II was the later-filed action. The first-filed action 26 was known as Rigsby. After Branch filed its original complaint in August 2006, it filed two 27 different amended complaints. The court gave several reasons why – for purposes of § 3730(b)(5) 1 the complaint in Rigsby. 2 First, the text of § 3730(b)(5) supports this approach.

3 The first-to-file bar [in § 3730(b)(5)] and the original source exception to the public disclosure bar [in § 3730(e)(4)] refer 4 specifically to jurisdictional facts that must exist when an "action," not a complaint, is filed. Under 31 U.S.C. § 3730(b)(5), a qui tam 5 plaintiff may not "bring a related action based on the facts underlying the pending action." As the Seventh Circuit has noted, 6 "[o]ne 'brings' an action by commencing suit." United States ex rel. Chovanec v. Apria Healthcare Group Inc., 606 F.3d 361, 362 (7th 7 Cir. 2010). Further, in order to be an original source under 31 U.S.C. § 3730(e)(4)(B), a relator must provide the information on 8 which the allegations are based to the government "before filing an action under this section which is based on the information." Both 9 provisions appear to contemplate that certain requirements must be met at the time a qui tam action is filed. The use of the term 10 "action" in both provisions indicates that the Court should look to the jurisdictional facts that existed at the time the action was filed, as 11 opposed to facts that existed when the relator later filed an amended complaint. 12 As the Third Circuit has noted, however the FCA is based on the 13 model of a single-count complaint, and it sometimes uses the term "action" when it likely means "claim." United States ex rel. Merena 14 v. SmithKline Beecham Corp., 205 F.3d 97, 101-02 (3d Cir. 2000). For example, under § 3730(b)(2) and (4), the government may 15 choose to "proceed with the action" or may "decline to take over the action," yet it is commonplace for the government to proceed with 16 only certain claims and not with others. Id. at 102. But even if "action" can mean "claim" in some contexts, it is perfectly natural to 17 read the first-to-file bar and the original source provision as imposing certain requirements that must be met at the time the suit 18 begins. 19 Branch II, 782 F. Supp. 2d at 259-60 (emphasis added). 20 Second, general jurisdictional principles also support the approach.

21 The notion that a court cannot proceed if it lacked jurisdiction at the time the original complaint was filed is consistent with the "time-of- 22 filing rule," under which "the jurisdiction of the Court depends upon the state of things at the time of the action brought[.]" Mollan v. 23 Torrance, 22 U.S. 537, 539 (1824) (diversity jurisdiction exists if the parties are diverse when the action was brought, even if 24 diversity is not maintained throughout the litigation) . . . . 25 Id. at 260 (emphasis added). 26 Third, the Supreme Court’s decision in Rockwell International Corp. v. United States, 549 27 U.S. 457 (2007), does not conflict with the above approach. provision [rather than the first-to-file provision], it also made 1 broader jurisdictional statements that are relevant to the FCA as a whole. In Rockwell, the relator brought a qui tam action relating to 2 toxic waste disposal at a nuclear weapons plant. The Supreme Court held that the relator was not an original source of new allegations in 3 the amended complaint just because he was an original source of the allegations in the original complaint. Id. at 473-74. The Court 4 concluded that the relator, "at a minimum," must be an original source of the claims in the amended complaint. Id. at 473. But the 5 Court did not suggest that the original complaint becomes irrelevant for jurisdictional purposes once an amended complaint is filed. To 6 the contrary, the Court stated that its holding was consistent with "[t]he rule that subject-matter jurisdiction 'depends on the state of 7 things at the time of the action brought.'" Id. (quoting Mollan v. Torrance, 22 U.S. 537, 539 (1824)). . . . [Cases cited in Rockwell] 8 indicate that a court cannot proceed if it lacked jurisdiction at the time the initial complaint was filed. 9 Rockwell goes on to state that jurisdiction is also defeated if a 10 plaintiff amends the complaint to withdraw the allegations upon which the court’s jurisdiction is based, “unless they are replaced by 11 others that establish jurisdiction.” . . . But Rockwell does not suggest that a plaintiff can establish jurisdiction by amendment when 12 jurisdiction did not previously exist. Indeed, such a conclusion would be directly contrary to the Court's statement that 13 "demonstration that the original allegations were false will defeat jurisdiction." 14 15 Id. at 261-62. 16 Finally, there are several practical/policy reasons to support the approach. For example,

17 the pre-filing disclosure requirement of § 3730(e)(4)(B) could not function if a court could acquire jurisdiction over a qui tam 18 complaint through amendment. If a court could gain jurisdiction over a qui tam action by amendment, then a relator could neglect to 19 inform the government of the information upon which the allegations are based before filing his or her action. Instead, the 20 relator could provide that information to the government at a later time, and then amend the complaint, even in a trivial fashion, to 21 ensure jurisdiction. Such a procedure would make the statutory language requiring disclosure to the government "before filing an 22 action" meaningless. 23 Id. at 263. 24 In addition,

25 while the first-to-file bar of 31 U.S.C. § 3730(b)(5) encourages relators to quickly report fraud about which they become aware, 26 problems arise when a relator files without yet having direct and independent knowledge of the information underlying the 27 allegations. As discussed infra, the Fifth Circuit has held that even Consultants v. Allstate Ins. Co., 560 F.3d 371, 379 (5th Cir. 2009). 1 It would be anomalous if a relator could secure a place in the jurisdictional queue with merely skeletal allegations, only to then 2 file an amended complaint after actually becoming an original source, and thereby trump any meritorious, related actions that were 3 filed in the meantime. Cf. United States ex rel. Ortega v. Columbia Healthcare, Inc., 240 F. Supp.2d 8, 14 (D.D.C. 2003) (amended 4 complaint could not "relate back" to the date the original complaint was filed in order to jump ahead in line). Such an approach would 5 shut out deserving relators while rewarding those who bring actions without having direct and independent knowledge of their publicly 6 disclosed allegations. A relator, under this scenario, could secure first-to-file status before actually conducting the investigation that 7 uncovers direct and independent information about the fraud. Requiring jurisdiction at the time the original complaint was filed 8 allows a court to dismiss such an attempt, regardless of later amendments. 9 10 Id. at 264.8

11 Finally, the time-of-filing rule has the advantage of simplicity. The benefits of a clear-cut rule are apparent in this case, which involves 12 multiple claims, complaints, and defendants, as well as other relators whose complaints have themselves been amended and involve 13 multiple defendants. See United States ex rel Rigsby v. State Farm Ins. Co., No. 06-433, 2006 U.S. Dist. LEXIS 98179 (S.D. Miss. 14 2006) (discussed infra); United States ex rel Denenea v. Allstate Ins. Co., No. 07-2795, 2011 U.S. Dist. LEXIS 6419 (E.D. La.); United 15 States ex rel Sonnier v. Allstate Ins. Co., No. 09-1038 (M.D. La.). The Court's jurisdiction may expand or shrink as amendments are 16 made to the complaint, but that jurisdiction must rest upon a solid foundation. 17 18 Id. 19 The analysis in Branch II is sound and persuasive. Furthermore, the analysis in Branch II 20 does not conflict with Ninth Circuit law, including Hartpence. Admittedly, in Hartpence, the 21 Ninth Circuit made the following comment in a footnote: “For purposes of determining 22 jurisdiction, we look to the allegations in the amended complaints. Rockwell Int’l Corp. v. United 23 States, 549 U.S. 457, 473-74 (2007).” Hartpence, 792 F.3d at 1125 n.2. But this footnote in 24 Hartpence does not mean that, for purposes of the first-to-file bar, that a court should look to an 25 amended pleading in a later-filed case. 26 8 Relatedly, if the rule were that an amended complaint in the later-filed action should be 27 considered (and not the original), that would give the relator in the later-filed action an incentive 1 First, Hartpence referred to two different FCA provisions, not only the first-to-file bar (in 2 § 3730(b)(5)) but also the public disclosure bar (in § 3730(e)(4)). Hartpence’s reference to 3 Rockwell in footnote 2 may well have related to the public disclosure bar, which would make 4 sense since Rockwell was a public disclosure case and not a first-to-file case 5 Second, as the Branch II court pointed out, Rockwell (the case that Hartpence cited) 6 acknowledged the “rule that subject-matter jurisdiction ‘depends on the state of things at the time 7 of the action brought.’” Rockwell, 549 U.S. at 473. 8 Third, Rockwell’s statement that “courts look to the amended complaint” must be 9 evaluated in context. The Rockwell Court noted that, if “original allegations [related to 10 jurisdiction] were false,” then jurisdiction is defeated. Id. “So also will the withdrawal of those 11 allegations unless they are replaced by others that establish jurisdiction. Thus, when a plaintiff 12 files a complaint in federal court and then voluntarily amends the complaint, courts look to the 13 amended complaint to determine jurisdiction.” Id. at 473-74. As indicated by the above text, the 14 Rockwell Court made the last statement in the context of a plaintiff withdrawing allegations that 15 gave rise to jurisdiction and pleading new allegations. 16 Finally, courts have recognized the context in which the Rockwell statement above was 17 made and thus taken note of the limits of Rockwell. For instance, the Fifth Circuit has stated:

18 The [Rockwell] Court did not hold . . . that the original complaint is irrelevant to jurisdiction or that a relator need not establish 19 jurisdiction from the moment he first files his action. Indeed, Rockwell did not speak to the question whether a relator can use an 20 amended complaint to establish jurisdiction when the original complaint is lacking. Consequently, we fall back on the 21 longstanding rule that the amendment process cannot “be used to create jurisdiction retroactively where it did not previously exist.” If 22 [the relator's] complaint did not establish jurisdiction, it should have been dismissed; his amendments cannot save it. 23 24 U.S. ex rel. Jamison v. McKesson Corp., 649 F.3d 322, 328 (5th Cir. 2011). Similarly, in Strudley 25 v. Santa Cruz County Bank, 747 F. App’x 617 (9th Cir. 2019), the Ninth Circuit held that the 26 plaintiffs could not amend as a matter of right to cure a jurisdictional defect in the original 27 complaint. “In line with Supreme Court precedent, this Circuit has adhered to the time-of-filing 1 commenced.’” Id. at 618. The court went on to reject the plaintiffs’ reliance on Rockwell.

2 Rockwell stands for the proposition that a plaintiff may voluntarily amend its original complaint to remove federal jurisdiction (except 3 when a case has been removed to federal court). Plaintiffs amended their complaint for the exact opposite purpose in this case [i.e., in 4 the attempt to create jurisdiction]. Therefore, the district court correctly looked to the original complaint in concluding that it 5 lacked subject matter jurisdiction over this case. 6 Id.; see also Black Hills Media, LLC v. Pioneer Corp., No. CV 13-05980 SJO (PJWx), 2014 U.S. 7 Dist. LEXIS 132030, at *9 (C.D. Cal. Jan. 14, 2014) (“The court in Rockwell had jurisdiction over 8 the original complaint when it was first filed, and the issue before the Supreme Court was whether 9 the amended complaint divested the court of that existing jurisdiction. The Supreme Court 10 determined that it did so. The opposite was true in [a Federal Circuit case], where the court had no 11 jurisdiction over the original complaint, and the amended complaint would therefore grant the 12 court jurisdiction that would not otherwise exist.”) (emphasis added). 13 Accordingly, the Court shall compare the original complaint in Osinek (i.e., the operative 14 complaint in the first-filed action at the time the later-filed action was filed) with the original 15 complaints in the later-filed actions. 16 C. “Identical Facts” Test v. “Material Facts” Test 17 Turning to the heart of the matter, the Court considers next what is the legal standard for 18 determining whether a first-filed suit and a later-filed suit are related for purposes of § 3730(b)(5). 19 Like other circuit courts, the Ninth Circuit has rejected the position that the first-filed and later- 20 filed actions must be based on “identical facts” in order to be deemed related. Instead of an 21 “identical facts” test, the Ninth Circuit applies a “material facts” test.

22 Most of the few courts that have addressed § 3730(b)(5) have rejected an identical facts test. The cases' common principle is that 23 "section 3730(b)(5) precludes a subsequent relator's claim that alleges the defendant engaged in the same type of wrongdoing as 24 that claimed in a prior action even if the allegations cover a different time period or location within a company." United States ex rel. 25 Capella v. United Technologies Corp., 1999 U.S. Dist. LEXIS 10520, 1999 WL 464536, at *9 (D. Conn. June 3, 1999) 26 (summarizing the tests used by other courts). The Third Circuit, the only appellate court to discuss and apply § 3730(b)(5), rejected an 27 identical facts test. See LaCorte, 149 F.3d at 233-34. We find the Section 3730(b)(5)'s plain language refers to "related" not 1 "identical" actions. Therefore, we need not review the legislative history. See Hockings, 129 F.3d at 1071. Even if the language were 2 considered ambiguous, the single sentence from the legislative history does not compel a different result. Furthermore, an identical 3 facts test would defeat the congressional objectives for the 1986 amendments: "adequate incentives for whistle-blowing insiders with 4 genuinely valuable information and discouragement of opportunistic plaintiffs who have no significant information to contribute of their 5 own." United States ex rel. Springfield Terminal Ry. v. Quinn, 304 U.S. App. D.C. 347, 14 F.3d 645, 649 (D.C. Cir. 1994). Limiting § 6 3730(b)(5) to only bar actions with identical facts would be contrary to the plain language and legislative intent: (1) using a narrow 7 jurisdictional bar, such as an identical facts test, would decrease incentives to promptly bring qui tam actions; (2) multiple relators 8 would expect a recovery for the same conduct, thereby decreasing the total amount each relator would potentially receive and 9 incentives to bring the suit; and (3) a narrow identical facts bar would encourage piggyback claims, which would have no additional 10 benefit for the government," since once the government knows the essential facts of a fraudulent scheme, it has enough information to 11 discover related frauds." LaCorte, 149 F.3d at 234.

12 Therefore, we hold that § 3730(b)(5) bars later-filed actions alleging the same material elements of fraud described in an earlier suit, 13 regardless of whether the allegations incorporate somewhat different details. 14 15 Lujan, 243 F.3d at 1188-89; see also United States ex rel. St. John LaCorte v. SmithKline 16 Beecham Clinical Labs., Inc., 149 F.3d 227, 234 (3d Cir. 1998) (stating that, “once the 17 government knows the essential facts of a fraudulent scheme, it has enough information to 18 discover related frauds”). 19 As a practical matter, the material facts test often has a court consider “whether the [later- 20 filed] Complaint alleges a fraudulent scheme the government already would be equipped to 21 investigate based on the [first-filed] Complaint.” United States ex rel. Batiste v. SLM Corp., 659 22 F.3d 1204, 1209 (D.C. Cir. 2011) (noting, for example, “[i]f the government investigated the facts 23 alleged in [first-filed] complaint on a nationwide basis, it would discover continuing fraud in the 24 New Jersey offices [which was the focus of the later-filed complaint”); see also id. at 1210 (stating 25 that “[s]ection 3730(b) is designed to allow recovery when a qui tam relator puts the government 26 on notice of potential fraud being worked against the government, but to bar copycat actions that 27 provide no additional material information”); Hartpence, 792 F.3d at 1125, 1131-32 1 plaintiff in the later-filed suit “provided information about a different form of fraud, and without 2 that information the government might not have investigated beyond [defendant’s] fraudulent 3 coding practices”).9 4 As a frame of reference, below is a brief discussion of some cases where the material facts 5 test was not met and where the material facts test was met. 6 1. Material Facts Test Not Satisfied 7 In Hartpence, the Ninth Circuit held that the material facts test was not satisfied. The 8 defendant in Hartpence was KCI, a company that manufactured medical devices that speeded the 9 healing of wounds. One such device was a V.A.C. (vacuum assisted closure) device. See id. at 10 1124. “V.A.C. devices perform negative pressure wound therapy (‘NPWT’),” and Medicare has 11 covered NPWT devices as durable medical equipment. Id. 12 In the first-filed suit, the plaintiff Hartpence alleged that KCI had engaged in fraudulent 13 conduct by submitting claims to Medicare related to the V.A.C. devices. Specifically, Hartpence 14 asserted that KCI had submitted claims with a certain billing code, which indicated compliance 15 with certain requirements even though those requirements had not, in fact, been met in various 16 ways. See, e.g., id. at 1125 (noting that “Hartpence alleges that KCI improperly submitted claims 17 with the KX modifier [i.e., billing code]: (1) when there was no wound improvement in the 18

19 9 Dr. Taylor contends that this notice standard is no longer applicable after Kellogg Brown & Root Servs. v. United States ex rel. Carter, 575 U.S. 650 (2015). See Taylor Opp’n at 21-22 n.23. In 20 support, he cites an opinion from a Washington district court, United States ex rel. Savage v. CH2M Hill Plateau Remediation Co., No. 4:14-cv-5002-EFS, 2015 U.S. Dist. LEXIS 137979 21 (E.D. Wash. Oct. 1, 2015). There, the court stated: “The Supreme Court's ruling in Brown inherently limits this ‘notice’ analysis. Applying a broad ‘notice’ test does not serve the FCA's 22 purpose of providing private parties the opportunity to pursue actions alleging fraud against the government once the first-to-file bar lifts following the dismissal of the earlier action.” Id. at *22 23 (emphasis added).

24 As indicated by the language italicized above, Savage has no application here because the first-filed action – Osinek – has not been dismissed and therefore the first-to-file bar cannot have 25 not been lifted. See also Brown, 575 U.S. at 662 (indicating that “an earlier suit bars a later suit while the earlier suit remains undecided but ceases to bar that suit once it is dismissed”; rejecting 26 the argument that the first-filed action remains pending even after it has been dismissed).

27 Moreover, the Ninth Circuit’s decision in Hartpence – which applied the notice standard 1 previous month; (2) for the treatment of wounds for which V.A.C. therapy was neither reasonable 2 nor necessary; (3) when the required wound measurement documentation was absent; [etc.]”). 3 In the later-filed suit, the plaintiff Godecke also claimed that KCI had improperly used the 4 same billing code – albeit for a different reason. See id. (taking note of allegation that “KCI 5 violated the FCA by knowingly misusing the KX modifier in submitting claims for a full month of 6 V.A.C. therapy, even when the therapy . . . had been stopped and restarted within the same 7 month”). In addition, Godecke claimed that there was a FCA violation related to DWOs (detailed 8 written orders). Suppliers of durable medical requirement were required to obtain DWOs from a 9 patient’s treating physician before dispensing the supplied for which they sought reimbursement 10 from Medicare. See id. at 1125 n.4. According to Godecke, “KCI ignored the requirement to 11 receive correct and completed [DWOs] before delivering supplies and beginning therapy.” Id. at 12 1125. 13 The Ninth Circuit effectively acknowledged the similarity of the Hartpence and Godecke 14 complaints in that both implicated improper use of the same billing code (even though there were 15 different reasons why the billing code was not properly used). However,

16 Godecke's second claim involves different underlying facts. Whereas Hartpence's claims all allege knowing misuse of the KX 17 modifier [i.e., billing code], Godecke's second claim is based on facts which show KCI's violation of a different Medicare program 18 requirement – the requirement that a provider receive Detailed Written Orders for the V.A.C. device before beginning to treat 19 patients with the device. . . . [T]he claims are based on different material facts. The rules governing use of KX modifiers and DWOs 20 were disseminated at different times, in different publications, and are plainly treated as separate regulations under the program. 21 We further disagree that Godecke's action provided no additional 22 benefit to the government. Unaided by Godecke's complaint, the government may have never discovered that KCI, in addition 23 allegedly to misusing the KX coding system, was allegedly submitting V.A.C. claims before receiving DWOs. The two alleged 24 frauds are materially different: the KX fraud allegations are based on government payment for devices which were used, but unnecessary 25 for treatment, while the DWOs fraud allegations are based on the government paying for devices that were never used at all. The 26 alleged frauds, in short, exist completely independent of one another. 27 1 2. Material Facts Test Satisfied 2 In United States ex rel. Hampton v. Columbia/Hca Healthcare Corp., 318 F.3d 214 (D.C. 3 Cir. 2003), the D.C. Circuit concluded that the material facts test was satisfied. The plaintiff 4 Hampton’s suit was the later-filed suit. The first-filed suit was brought by Boston. According to 5 Hampton, the defendant companies and several employees had improperly billed the government 6 under the Medicare program for home health services – e.g.,

7 the companies billed for services that were miscoded; already paid for; performed by others; never administered; or supposedly 8 administered to Hampton's mother after she died in 1996. Hampton also claimed that [the companies] submitted bills for supplies and 9 medications that were unnecessary or never received; and that they billed for services to patients who did not qualify under the 10 Medicare guidelines, did not need treatment, or were not charged required copayments. The companies submitted false or inaccurate 11 documentation to the government and, so she alleged, shredded documents in order to destroy evidence of the fraud. 12 [The D.C. Circuit held that Hampton’s case and Boston’s case were 13 related because] Boston's allegations were along very much the same lines. He asserted that HCA home health subsidiaries billed 14 the government for services that did not meet the Medicare eligibility criteria, for undocumented services, and for services not 15 medically necessary. He also alleged that they submitted false or inaccurate Medicare documentation and destroyed documents. 16 17 Id. at 219. 18 Likewise, in Batiste, the D.C. Circuit also found the first- and later-filed cases related. The 19 plaintiff Batiste filed the later action. Zahara filed the earlier action. The court found that “[a] 20 side-by-side comparison has persuaded us that, although the complaints allege somewhat different 21 facts, Zahara's complaint suffices to put the U.S. government on notice of allegedly fraudulent 22 forbearance practices at [Sallie Mae] and its subsidiaries, and Batiste's complaint alleges the same 23 material elements of the same fraud.” Id. at 1209. In particular,

24 Zahara and Batiste broadly allege that the same fraudulent activities occurred at each of their offices, for the same reasons, and that 25 similar SLM corporate policies promoted the fraudulent behavior. They both allege SLM fraudulently increased its profits and 26 promoted its standing with the Department of Education by falsifying forbearances. And both allege that SLM's corporate 27 culture promoted increasing the dispensation of forbearances on the offering of forbearances to unqualified borrowers, the 1 allegations of the first complaint give the government grounds to investigate all that is in the second. 2 Under the . . . material facts test, these complaints allege essentially 3 the same corporation-wide scheme. The Zahara Complaint would suffice to equip the government to investigate SLM's allegedly 4 fraudulent forbearance practices nationwide. Batiste's additional details would not give rise to a different investigation or recovery. 5 6 Id. at 1209-10. 7 D. Different Defendants 8 Finally, the Court must consider whether it makes a difference in “material facts” when 9 different defendants are sued in the first-filed and later-filed actions. In Osinek, the original 10 complaint named only one defendant – Kaiser Permanente – which Ms. Osinek described as “a 11 private provider of Medicare Advantage insurance under Medicare Part C.” Osinek Compl. ¶ 2.

12 Kaiser Permanente is a California corporation with its principal place of business [in] Oakland, California 94612. Kaiser is one of 13 the largest Medicare Advantage organizations in the country and has more enrollees in its Medicare Advantage Plans than any other 14 organization in California. At all times relevant, Kaiser conducted business in California, including but not limited to providing 15 healthcare services through Medicare Advantage plans and to the general public in California. 16 17 Osinek Compl. ¶ 6. The later-filed actions named a variety of different Kaiser entities, including 18 but not limited to Kaiser Permanente. 19 COMPARISON OF DEFENDANTS NAMED IN ORIGINAL COMPLAINTS 20 Osinek Taylor Arefi (2015) Stein (2016) Bryant Bicocca 21 (2013) (2014) (2018) (2020)

22 Kaiser Kaiser Kaiser Kaiser Permanente Permanente Permanente Permanente 23

24 Kaiser Kaiser Kaiser Kaiser Foundation Foundation Foundation Foundation 25 Health Plan, Health Plan, Health Plan, Health Plan, Inc. Inc. Inc. Inc. 26 Kaiser Kaiser Kaiser Foundation Foundation Foundation 27 Health Plan Health Plan Health Plan COMPARISON OF DEFENDANTS NAMED IN ORIGINAL COMPLAINTS 1 Osinek Taylor Arefi (2015) Stein (2016) Bryant Bicocca 2 (2013) (2014) (2018) (2020)

3 Kaiser Kaiser Kaiser Foundation Foundation Foundation 4 Health Plan Health Plan Health Plan of Georgia of Georgia, of Georgia, 5 Inc. Inc. Kaiser Kaiser Kaiser 6 Foundation Foundation Foundation Health Plan Health Plan Health Plan 7 of the of the of the Northwest Northwest Northwest 8 Kaiser Kaiser Kaiser Foundation Foundation Foundation 9 Hospitals Hospitals Hospitals Southern Southern Southern Southern 10 California California California California Permanente Permanente Permanente Permanente 11 Medical Medical Medical Medical Group Group Group Group, Inc. 12 The The The Permanente Permanente Permanente Permanente Medical 13 Medical Medical Medical Group, Inc. Group Group Group 14 Colorado Colorado Colorado Permanente Permanente Permanente 15 Medical Medical Medical Group, P.C. Group Group P.C. 16 The The Southeast Southeast Southeast Permanente 17 Permanente Permanente Medical Medical Medical Group 18 Group Group 19 Hawaii Hawaii Hawaii Permanente Permanente Permanente 20 Medical Medical Medical Group Group Group 21 Northwest Northwest Permanente, Permanente, 22 P.C. P.C. Kaiser 23 Foundation Health Plan 24 of the Mid- Atlantic 25 States, Inc. Kaiser 26 Foundation Health Plan 27 of Washington COMPARISON OF DEFENDANTS NAMED IN ORIGINAL COMPLAINTS 1 Osinek Taylor Arefi (2015) Stein (2016) Bryant Bicocca 2 (2013) (2014) (2018) (2020)

3 Mid-Atlantic Mid-Atlantic Permanente Permanente 4 Medical Medical Group Group, PC 5 Group Health Permanente 6 The The Permanente Permanente 7 Federation, Federation, LLC LLC 8 Northwest Permanente 9 Physicians & Surgeons, 10 P.C. Washington 11 Permanente Medical 12 Group

14 The Ninth Circuit has not expressly addressed the issue of different defendants, but other 15 circuit courts have. Most have indicated that “adding a new defendant to the mix does not 16 necessarily allow a later-filed action to evade the first-to-file bar.” Cho v. Surgery Partners, Inc., 17 No. 20-14109, 2022 U.S. App. LEXIS 8774, at *15 (11th Cir. Apr. 1, 2022) (emphasis in 18 original). This is particularly true where the new defendant(s) named in the later-filed action is a 19 subsidiary or affiliate of the defendant(s) named in the first-filed action. See Branch I, 560 F.3d at 20 379 (noting that “allegations of fraud against a corporation may bar subsequent allegations of 21 fraud against the corporation’s subsidiaries”). 22 That being said, the fact that the new defendant(s) in the later-filed action is a subsidiary or 23 affiliate of the defendant(s) in the first-filed action does not automatically mean that the first-filed 24 and later-filed actions are related either. Ultimately, resolution depends on how the first-filed 25 action defines the scope of the misconduct. If there are, e.g., allegations in the first-filed suit that 26 there was a nationwide problem or a corporate-wide problem, then, most likely, the fact that new 27 subsidiaries or affiliates are named in the later-filed action will not make that action unrelated for 1 same corporate-wide scheme”). On the other hand, if the first-filed action focuses on a local 2 problem, then, most likely, a broader-in-scope later-filed action will not be related, even if the new 3 defendant(s) in the latter action is an affiliate of the defendant named in the earlier action. Cf. 4 United States ex rel. Chovanec v. Apria Healthcare Group, Inc., 606 F.3d 361, 364 (7th Cir. 2010) 5 (stating that, “to understand whether the suits materially overlap we must know whether the initial 6 suits alleged frauds by rogue personnel at scattered offices or instead alleged a scheme 7 orchestrated by Apria's national management”). 8 “Two cases from the D.C. Circuit, Hampton and Heath, serve as useful bookends for this 9 analysis.” Cho, 2022 U.S. App. LEXIS 8774, at *15. Hampton found the first- and later-filed 10 actions at issue related; Heath found the first- and later-filed actions at issue unrelated. 11 1. Hampton 12 In Hampton, the plaintiff Hampton named the following defendants in her later-filed 13 action: HCA; Clinical Arts (Georgia subsidiary of HCA); and several Clinical Arts employees. 14 She alleged that the defendants “had improperly billed the government under the Medicare 15 program for home health services.” Hampton, 318 F.3d at 218. The court was asked to decide 16 whether Hampton’s action was barred by an earlier lawsuit, filed by Boston. The D.C. Circuit 17 noted that

18 Hampton thinks her complaint differs significantly from Boston's because it named different defendants. Boston sued only HCA. 19 Hampton sued not only HCA but also HCA's subsidiary Clinical Arts and several Clinical Arts employees. As Hampton sees it, 20 Boston's complaint cannot possibly have covered fraud by Clinical Arts and its employees because it (1) fails to name Clinical Arts or 21 its employees as defendants and (2) specifically mentions fraud at HCA home health care subsidiaries in six states that do not include 22 Georgia. 23 Id. The court, however, found that

24 these are not differences in the material elements of the fraud. Boston was a senior manager in HCA's home care group. He 25 alleged a corporate-wide problem, revealed through internal audits, in which HCA perpetrated fraud in providing home health care 26 services through numerous subsidiaries. It is true that Boston's complaint mentioned instances of fraud at particular home health 27 agencies in only six specific states, not including Georgia. But by Columbia/HCA's 550 home health locations in 37 states.” Given 1 Boston's broad allegations based on his position as an HCA insider, Hampton's naming Clinical Arts – a specific HCA subsidiary – and 2 naming individual employees of Clinical Arts were merely variations on the fraud Boston's complaint described. 3 4 Id. (emphasis added); see also Batiste, 659 F.3d at 1209 (acknowledging that plaintiff in first-filed 5 suit “discusses activities at [a Sallie Mae] subsidiary office in Nevada, but [still] alleges a 6 nationwide scheme attributable not only to the subsidiary, but also to [Sallie Mae]”; thus, “[i]f the 7 government investigated the facts alleged in [that plaintiff’s] complaint on a nationwide basis, it 8 would discover continuing fraud in the New Jersey offices”); Chovanec, 606 F.3d at 364 (taking 9 note of allegations that fraud was enabled by changes made to a computer system used in all of a 10 company’s offices; also taking note of allegations that national headquarters provided guidance 11 that enabled fraud); United States ex rel. Marion v. Heald Coll., LLC, No. 5:12-cv-02067-PSG, 12 2015 U.S. Dist. LEXIS 97767, at *11 (N.D. Cal. July 24, 2015) (stating that “[a]llowing plaintiffs 13 to escape the first-to-file bar by naming specific employees who carried out a previously-alleged 14 corporate fraud contravenes the purpose of Section 3730(b)(5) – to prevent piggyback claims[;] 15 [h]ere, the previously-filed complaints against Corinthian [Colleges] allege that fraudulent conduct 16 extended far beyond individual campuses and pervaded the entire company”). 17 2. Heath 18 In United States ex rel. Todd Heath v. AT&T, Inc., 791 F.3d 112 (D.C. Cir. 2015), the 19 plaintiff Heath filed the later-filed action against AT&T and nineteen of its subsidiaries. The 20 lawsuit was related to a federal program known as the Universal Service Fund. See id. at 117. 21 Under federal law, “every interstate telecommunications carrier must contribute a portion of its 22 quarterly interstate and international telecommunications revenue to the . . . Fund.” Id. at 116. 23 One of the programs administered through the Fund is “E-Rate,” which “entitles qualifying 24 schools and libraries to receive Internet and telephone services at discounted rates.” Id. at 116-17. 25 According to Heath,

26 AT&T orchestrated and implemented through its subsidiaries a corporate-wide scheme to have false claims submitted to the 27 Universal Service Fund by depriving schools and libraries in the E- those inflated costs on to the federal government for reimbursement 1 through the Universal Service Fund. 2 Id. at 117. Of particular note, Heath asserted that AT&T deliberately or recklessly chose not to 3 train its employees in the lowest-corresponding-price requirement. See id. (taking note of 4 allegation that AT&T was a recidivist violator of the E-Rate Program). 5 Heath had also filed an earlier lawsuit but only against Wisconsin Bell, which was a 6 wholly owned subsidiary of AT&T. See id. at 118. In this suit, Heath asserted that “Wisconsin 7 Bell charged some E-Rate eligible schools more than others, and that Wisconsin Bell generally 8 failed to provide school districts with the benefit of the favorable pricing it offered to state 9 departments, agencies, and universities.” Id. Furthermore, “[w]hen informed of this pricing 10 discrepancy, Wisconsin Bell’s sales representatives ‘regularly denied the existence of the 11 agreements’ between Wisconsin Bell and other Wisconsin agencies.” Id. 12 One issue before the D.C. Circuit was whether Heath’s first-filed suit was a bar to his later- 13 filed action. The court held that the first-filed suit was not a bar because the

14 two complaints target factually distinct types of frauds. The Wisconsin Bell Complaint alerted the federal government only to a 15 limited scheme by Wisconsin Bell to defraud the E-Rate program within Wisconsin. That alleged fraud was accomplished, in part, 16 through affirmative misrepresentations by Wisconsin Bell employees to schools and libraries within Wisconsin, in which those 17 employees openly denied the existence of a state contract with a lower corresponding price. 18 In contrast, the AT&T Nationwide Complaint alleges a different and 19 more far-reaching scheme to defraud the federal government through service contracts entered into across the Nation, and then to 20 cover up that fraud. Critically, the alleged fraud was accomplished not through affirmative misrepresentations about the lowest 21 corresponding price, but through institutionalized disregard of the lowest-corresponding-price requirement altogether in AT&T's 22 employee-training and billing procedures. According to the AT&T Nationwide Complaint, AT&T and its subsidiaries deliberately 23 failed to enforce that lowest-price mandate by refusing to train or even tell employees about that limitation on charges, and by failing 24 to incorporate that limitation into its billing practices. 25 Id. at 121. 26 The court continued:

27 On its face, the Wisconsin Bell complaint discloses nothing more the complaint would have alerted the United States government to a 1 nationwide scheme centered in AT&T's corporate headquarters of mischarging the E-Rate program and subsequently concealing those 2 overpayments. Nor, given the affirmative misrepresentations at issue, would the Wisconsin Bell Complaint have pointed the federal 3 government to AT&T's systematic refusal to institutionalize compliance by employees with the lowest-corresponding-price 4 requirement.

5 The fraud thus manifested itself in sufficiently distinct ways in the two cases that the material elements of the fraud differ. As the 6 Seventh Circuit has recognized, "to understand whether the suits materially overlap we must know whether the initial suit[] alleged 7 frauds by rogue personnel at scattered offices or instead alleged a scheme orchestrated by * * * national management." Because the 8 Wisconsin Bell Complaint alleged only the former, it did not disclose the nationwide fraud grounded in institutionalized training 9 and enforcement failures, and compounded by efforts at concealment, that is the focus of Heath's later complaint. 10 11 Id. at 121-22 (citation omitted). The D.C. Circuit distinguished, inter alia, Hampton because, 12 there, “the first complaint alleged a broad fraudulent scheme orchestrated by a national or parent 13 company, and the second complaint merely added additional facts or widened the circle of victims 14 of the same fraudulent conduct.” Id.

15 Those cases stand for the simple proposition that the greater fraud often includes the lesser. The problem for AT&T is that the lesser 16 fraud does not, without more, include the greater. The Wisconsin Bell Complaint did not allege that AT&T encouraged Wisconsin 17 Bell's fraud or affirmative misrepresentations, or even knew anything about them. Nor did the Wisconsin Bell Complaint 18 suggest that AT&T and its subsidiaries engaged in "uniform billing practices" across the United States. There simply is no hint in the 19 Wisconsin Bell Complaint of a country-wide, institutionalized corporate practice of disregarding the lowest-price requirement or of 20 a calculated refusal to educate or train employees. 21 Id. at 122-23. 22 The court acknowledged AT&T’s point that the E-Rate program is a national program but, 23 it pointed out, so too “is virtually every law policed by the federal False Claims Act.” Id. at 123 24 (emphasis in original). The first-to-file bar could not be triggered “every time an initial complaint 25 alleges that a subsidiary of a national company violated a national law” or “a broad swath of False 26 Claims Act coverage” would be “erase[d].” Id. The court underscored that the purpose of the 27 first-to-file bar was “to prevent copycat litigation, which tells the government nothing it does not 1 see also United States ex rel. Branch Consultants v. Allstate Ins. Co., 560 F.3d 371, 379 (5th Cir. 2 2009) (stating that the first-filed action “Rigsby does not allege a true industry-wide fraud or 3 concerted action among a narrow group of participants[;] [r]ather, looking only at the facts 4 pleaded . . . , Rigsby implicates, at most, four specific [write-your-own] insurers among the 5 approximately ninety-five WYI insurers conducting business in the Louisiana and Mississippi 6 areas during Hurricane Katrina,” and, therefore, “Rigsby tells the government nothing about which 7 of the ninety-one other WYO insurers . . . , if any, actually engaged in any fraud”). 8 III. COMPARING RELATOR ACTIONS 9 Having reviewed the major legal issues, the Court may now turn to a comparison of the 10 first-filed action, Osinek, with the various later-filed actions. The Court makes the comparisons in 11 the order listed below because, e.g., if there are material differences between Osinek and Taylor, 12 that would effectively make Taylor the first-filed action for the new material facts. In other words, 13 the remaining cases would then need to be compared to both Osinek and Taylor. 14 • Case No. C-21-3894 EMC Taylor (filed in the District of Colorado in 2014 and 15 transferred to this District in 2021). 16 • Case No. C-16-1558 EMC Arefi (filed in the Central District of California in 2015 17 and transferred to this District in 2016). 18 • Case No. C-16-5337 EMC Stein (filed in the Central District of California in 2016 19 and transferred this District in 2016). 20 • Case No. C-18-1347 EMC Bryant. 21 • Case No. C-21-3124 EMC Bicocca (filed in the Eastern District of California in 22 2020 and transferred to this District in 2021). 23 A. Osinek and Taylor10 24 Dr. Taylor argues that his case is materially different from Osinek in two ways: (1) 25 different defendants were sued in each case and (2) different frauds were implicated in each case. 26 10 Dr. Taylor argues his FAC is the operative complaint for purposes of comparing his case to 27 cases filed after his. See Taylor Opp’n at 6 n.4. But for purposes of comparing Osinek and 1 1. Different Defendants 2 Ms. Osinek sued only one entity: Kaiser Permanente. In contrast, Dr. Taylor sued multiple 3 Kaiser entities: Kaiser Permanente; Kaiser Foundation Health Plan, Inc.; Kaiser Foundation 4 Health Plan of Colorado; Kaiser Foundation Health Plan of Georgia; and Kaiser Foundation 5 Health Plan of the Northwest. 6 Dr. Taylor argues that the difference in defendants is meaningful. Defendants argue to the 7 contrary. They take the position that, by suing “Kaiser Permanente,” Ms. Osinek implicitly sued a 8 national defendant which therefore covered all regional or local subsidiaries or affiliates. 9 Defendants further argue that allegations made in the Osinek Complaint show that Ms. Osinek was 10 implicating a nationwide fraud. See Mot. at 20. The Court finds that the Osinek complaint does 11 not allege a nationwide fraud. 12 As a starting point, the Court takes note that there does not appear to be any legal entity 13 with the name “Kaiser Permanente.” Rather, “Kaiser Permanente” seems to be a trade name used 14 by various Kaiser entities. See Mot. at 1 (asserting that there are “various healthcare organizations 15 operating under the Kaiser Permanente trade name”); cf. U.S. Compl. ¶ 28 (noting Kaiser’s Health 16 Plans, Permanente Medical Groups, and hospitals publicly hold themselves out and do business 17 collectively as an integrated healthcare providers called ‘Kaiser Permanente’”); Taylor Compl. ¶ 18 16 (alleging that “Kaiser Permanente is a non-profit care consortium” that “includes three main 19 groups: (1) the Kaiser Foundation Health Plan, Inc. and its subsidiaries; (2) the Kaiser Foundation 20 Hospitals and their subsidiaries; and (3) the Permanente Medical Groups”). Thus, the fact that 21 Ms. Osinek (not to mention Dr. Taylor and others) sued “Kaiser Permanente” is not particularly 22 telling one way or the other. 23 What is more significant are the allegations in Ms. Osinek’s complaint – specifically, the 24 geographic scope of the allegations. The complaint implicates California only. For example, Ms. 25 Osinek describes “Kaiser Permanente” as follows:

26 a California corporation with its principal place of business [in] Oakland . . . . Kaiser is one of the largest Medicare Advantage 27 organizations in the country and has more enrollees in its Medicare but not limited to providing healthcare services through Medicare 1 Advantage plans and to the general public in California. 2 Osinek Compl. ¶ 6 (emphasis added). 3 One might argue that the above paragraph is California-centric because Ms. Osinek was 4 simply trying to establish that jurisdiction in California is proper. However, nowhere in the 5 complaint does Ms. Osinek allege that there is a problem outside of California. She does not 6 mention any other state. She does not allege a “corporate-wide problem” as in Hampton, 318 F.3d 7 at 218. She never even uses the term “nationwide” or “corporate-wide” or otherwise suggest a 8 “‘scheme orchestrated by . . . national management.’” Heath, 791 F.3d at 122 (quoting Chovanec, 9 606 F.3d at 364). Nor is it not clear that any practice identified by Ms. Osinek was necessarily 10 nationwide or corporate-wide in nature – e.g., there is no suggestion that data mining through the 11 use of algorithms was implemented through a nationwide computer system. 12 Defendants contend that just because the Osinek Complaint gave California examples does 13 not mean that the pleading is limited in scope to California. In principle, this is true. But there 14 must be some indication in the pleading that the problem extends outside of California.11 15 Compare Hampton, 318 F.3d at 218 (“It is true that Boston's complaint mentioned instances of 16 fraud at particular home health agencies in only six specific states, not including Georgia. But 17 Boston's complaint described these as ‘examples’ and ‘samplings’ of ‘a huge number of illegal 18 payments from Medicare . . . received by Columbia/HCA's 550 home health locations in 37 19 states.’”). Indeed, if the rule were to the contrary, then there would seem to be serious policy 20 concerns. The mere fact that a defendant company is part of a larger network of affiliated 21 companies would be enough to deem the government on notice of a nationwide or corporate-wide 22 problem – which would then, under the first-to-file bar, cut off all other actions. But if the 23 government declined to investigate on a nationwide or corporate-wide scale (e.g., because of 24 limited resources), and the first filer stayed within the limited scope of its complaint as pled, then 25 no one else would be able to bring a FCA claim to address a potentially broader problem. This 26 11 Notably, where the Osinek Complaint does at one point refer to different regions, both of the 27 regions identified are still based in California. See Mot. at 20 (citing Osinek Compl. ¶ 37); Osinek 1 would run counter to the FCA which is meant to encourage the uncovering of fraud against the 2 government. The first-to-file bar would not only “prevent copycat litigation,” Heath, 791 F.3d at 3 123, but also a wide swath of litigation far broader than the first-filed suit. Here, Taylor is not 4 simply an “opportunistic successive suit,” Lujan, 243, F.3d at 1187; Taylor cannot be said to have 5 “provided no additional benefit to the government.” Hartpence, 792 F.3d at 1131. 6 Confronted with this obstacle, Defendants invoke the practical aspect of the material facts 7 test, which asks whether “[t]he first-filed claim provides the government notice of the essential 8 facts of an alleged fraud.” Lujan, 243 F.3d at 1187. Defendants note that, a few months after the 9 Osinek complaint was filed in 2013, the government issued four subpoenas to (1) the Kaiser 10 Foundation Health Plan; (2) the Permanente Medical Group; (3) Southern California Permanente 11 Medical Group; and (4) Kaiser Foundational Hospitals. The subpoenas covered not only the 12 named entities but also all subsidiaries and affiliates.12 See RJN, Exs. A-D (subpoenas). Thus, 13 Defendants argue, the government implicitly understood that Osinek pointed to a nationwide or 14 corporate-wide problem. Defendants add that, in 2017, the government sent a letter to Defendants 15 seeking the production of documents from the Colorado regions specifically – and cited to the 16 2013 subpoenas in support of that request. See RJN, Ex. E. 17 Defendants’ contention is not convincing for several reasons. First, Dr. Taylor raises a 18 legitimate argument that only the complaints should be considered in determining whether the 19 first-to-file bar applies – not evidence outside of the complaints. Cf. In re Natural Gas Royalties 20 ex rel. United States, 562 F.3d 1023, 1031 (10th Cir. 2009) (stating that “[t]he first-to-file bar is 21 designed to be quickly and easily determinable, simply requiring a side-by-side comparison of the 22 complaints”); Batiste, 659 F.3d at 1209 (making a “side-by-side comparison” of the complaints in 23 the first-filed and later-filed suits). Dr. Taylor has not been privy to all of the communications 24 12 See, e.g., RJN, Ex. A (Subpoena at 2) (“The term ‘KAISER FOUNDATION HEALTH PLAN, 25 INC.’ refers to the person or entity with its primary offices located at One Kaiser Plaza, Oakland, California, and also includes all current and former: directors, officers, principals, partners, 26 managers, and employees; independent contractors, attorneys, consultants, experts, investigators, agents and/or other persons or other representatives acting on your behalf, even if their actions 27 were not authorized by you or were outside the proper scope of their authority; corporate parents, 1 between the United States and Defendants. It is entirely possible that some communications might 2 support his position here rather than Defendants.13 And if the Court were to permit Dr. Taylor to, 3 in effect, conduct discovery into the government’s communications – or more generally, into the 4 government’s understanding of the Osinek complaint – then there would effectively be a mini trial 5 on a secondary matter that would only delay the process of moving forward with the case. The 6 task before the Court at this juncture is to compare the complaints, not conduct a mini-trial based 7 on facts. 8 Accordingly, the Court does not dismiss Taylor based on the first-to-file bar because 9 Taylor is broader in scope than Osinek in terms of defendants. See Heath, 791 F.3d at 122 10 (“Those cases [such as Hampton] stand for the simple proposition that the greater fraud often 11 includes the lesser. The problem for AT&T is that the lesser fraud does not, without more, include 12 the greater.”). 13 2. Different Frauds 14 The closer question is whether Taylor implicates different frauds than does Osinek. 15 As noted above, Osinek is about mining records to look for places to upcode, particularly 16 for high-value conditions. Upcoding was ultimately improper because it was based on, e.g., 17 exaggerating a patient’s condition, diagnosing a patient based on a test that took place after the 18 patient visit, diagnosing a patient for a condition for which the patient was not treated, diagnosing 19 without the proper support/documentation, and the like. 20 Taylor is essentially the flip side. It asserts the theory that, as a result of regular internal 21 audits, Kaiser knew there were high error rates in risk adjustment claims in certain areas but failed 22 to take action to find the false claims retroactively – and thus improperly retained the 23 24 13 Even if the Court were to consider the 2013 subpoenas issued by the government and its 2017 25 letter to Defendants, Defendants do not fare any better. Although the subpoenas did refer to subsidiaries and affiliates, that language is boilerplate in nature. And fact that the government 26 only issued subpoenas to four specific Kaiser entities (all of which appear to be based in California) points to a more limited scope of inquiry. As for the letter, the fact that the 27 government relied on the 2013 subpoenas to justify an inquiry into Colorado in 2017 largely 1 government’s overpayment for those false claims (essentially, a reverse false claim theory).14 See, 2 e.g., Taylor Compl. ¶¶ 81-82 (citing “[e]xamples of risk adjustment claims that the Kaiser audits 3 have identified as routinely false” and alleging that, “despite its knowledge that the categories of 4 risk adjustment claims . . . are false a significant percentage of the time, Kaiser routinely fails to 5 take reasonable steps to identify which of these claims are false . . . and then to prevent their 6 submission in the first place or to delete them after submission”). Taylor contrasts the lack of 7 effort by the Kaiser entities to take action to address the errors with their zealous pursuit of 8 reviewing records to find instances where diagnoses could be added (i.e., upcoding). See Taylor 9 Compl. ¶ 62 (alleging that Kaiser’s “lack of diligence contrasts starkly with [its] considerable 10 efforts and substantial commitment of resources to audit current and past claims to identify new 11 diagnoses that it could use to submit additional risk adjustment claims and thereby increase the 12 amount of the risk adjustment payments it receives from CMS.”). 13 Dr. Taylor argues that the flip side makes his case materially different from Osinek. 14 Specifically, he argues that his case, unlike Osinek, focuses on (1) Kaiser’s failure to act even after 15 audits revealed high error rates for certain HCCs or diagnoses; (2) Kaiser’s failure to act even after 16 audits revealed high error rates for diagnoses made by external providers; and (3) Kaiser’s failure 17 to act even after audits revealed high error rates for “True Positive” results associated with 18 Kaiser’s Natural Language Processing program. Each of these claimed differences is discussed 19 below. 20 14 In their papers, Defendants suggest that Dr. Taylor does not have a viable claim here: “Taylor’s 21 allegations about error rates identified in audits of diagnosis-code data do not identify a unique fraud scheme. Rather they purport to show that Defendants had knowledge of the upcoding 22 scheme that Osinek already alleged.” Reply at 3 (emphasis added). Defendants likely make this argument because, in its complaint, the United States alleges that “Kaiser’s internal audits put [it] 23 on . . . notice of fraudulent diagnoses.” U.S. Compl. at 68; see also U.S. Compl. ¶ 304 (alleging that “[a] variety of internal audits provided further notice that Kaiser’s addenda and query 24 practices were resulting in false claims to CMS”).

25 Although Defendants fairly argue that the audits are relevant to Kaiser’s knowledge, it is not clear why Dr. Taylor would not also have a viable claim based on the theory that Kaiser failed 26 to take corrective action which resulted in its being able to keep government overpayments for false claims in violation of the FCA. Cf. United States v. United Healthcare Ins. Co., 848 F.3d 27 1161, 1173 (9th Cir. 2016) (indicating that plaintiff had a viable theory based on allegations that, 1 a. High Error Rates for Certain HCCs or Diagnoses 2 Dr. Taylor first claims a violation of the FCA because Kaiser failed to act even after audits 3 revealed high error rates in risk adjustment claims for certain HCCs or diagnoses. For example, 4 Dr. Taylor alleges as follows with respect to cancer: 5 • “Every year, Kaiser’s National Compliance Office (‘NCO’) conducts a nationwide 6 ‘Probe’ audit to test the accuracy of risk adjustment claims submitted the prior 7 year.” Taylor Compl. ¶ 69. 8 • The Probe Audits “have consistently identified cancer (HCCs 7-10) as the most 9 upcoded condition.” Taylor Compl. ¶ 102. 10 • “The most significant and consistent error is that Kaiser providers submit diagnosis 11 codes representing active, current treatment of cancer when, in fact, the patient’s 12 cancer is cured, in remission, or otherwise irrelevant to the services provided to the 13 patient.” Taylor Compl. ¶ 103. Notably, a diagnosis of “history of cancer” does 14 not result in a risk adjustment. See Taylor Compl. ¶ 105. 15 Cancer is not the only HCC/diagnosis called out as problematic. Other conditions that 16 have had high error rates include: 17 • Stroke. See, e.g., Taylor Compl. ¶ 118 (“Kaiser knew stroke was commonly coded 18 as an active event, when, in fact, the patient should have been classified as having a 19 history of stroke.”). 20 • Vascular disease. See, e.g., Taylor Compl. ¶ 127 (“[S]ome claims erroneously 21 claimed the patient had current vascular disease, when, in fact, they had only a 22 history of the condition.”); Taylor Compl. ¶ 129 (“[C]ertain claims were false 23 because of a ‘mismapping’ problem with HealthConnect, Kaiser’s EMR [electronic 24 medical records]. HealthConnect . . . allows physicians to choose a descriptive 25 diagnosis (as opposed to a specific ICD-9 code) when entering clinical information. 26 HealthConnect then ‘maps’ this descriptive diagnosis to a specific ICD-9 diagnosis 27 code, which is then inserted into the medical record documentation.”). 1 medical records] . . . pressured physicians to use the diagnosis for chronic 2 bronchitis (which risk adjusts) rather than acute bronchitis (which does not risk 3 adjust).”); Taylor Compl. ¶ 132 (“The probe audits regularly found COPD [chronic 4 obstructive pulmonary disease] claims erroneous based on lack of documentation in 5 the record, or because the doctor failed to document the patient’s condition with 6 sufficient specificity to determine if the patient actually had COPD.”). 7 • Malnutrition. See, e.g., Taylor Compl. ¶ 140 (“In some cases, the condition was 8 diagnosed as current when the patient actually only had a ‘history of’ the 9 condition.”). 10 • Renal insufficiency. See, e.g., Taylor Compl. ¶ 152 (“Chronic kidney disease 11 (‘CKD’) is a condition that is often miscoded . . . .”). 12 As indicated by the above, Dr. Taylor takes issue with Kaiser for not reacting to the high 13 error rates – i.e., had Kaiser done so then it would have seen that the diagnoses (of high-value 14 conditions) lacked documentation or proper support and/or that the diagnoses were irrelevant to 15 the treatment provided to the patient. Although Dr. Taylor is correct that his claim here is about 16 Kaiser ignoring an upcoding problem (as revealed by error rates) rather than actively creating 17 upcoding, the Court does not see this flip side as creating a material difference with respect to 18 Osinek. This is because both Taylor and Osinek are ultimately based on the same “underlying 19 facts,” Hartpence, 792 F.3d at 1131: that the high-level condition that was diagnosed did not have 20 documentation or proper support and/or did not affect patient care. 21 The practical aspect of the material facts test underscores that Taylor and Osinek are 22 related cases, at least with respect to the above. Here, the Court must ask whether the allegations 23 in Osinek “[gave] the government grounds to investigate all that is in” the Taylor Complaint. 24 Batiste, 659 F.3d at 1209; see also id. at 1209-10 (stating that the first-filed “[c]omplaint would 25 suffice to equip the government to investigate SLM's allegedly fraudulent forbearance practices 26 nationwide” and the “additional details” in the later-filed complaint “would not give rise to a 27 different investigation or recovery”). Based on Osinek, the government was put on notice that 1 was not treated for that condition at the time of service. Thus, in light of Osinek, the government 2 had grounds to investigate all that is in the Taylor Complaint which points to the same basic 3 problem. That, according to Dr. Taylor, the problem would have been revealed if Kaiser had 4 taken action in response to the high error rates, is somewhat beside the point. The error rates here 5 are not in themselves what is critical; rather, at bottom, Dr. Taylor’s broader claim is that high- 6 value conditions were diagnosed without following the practices required by Medicare regulations. 7 This is fundamentally the same charge that Ms. Osinek makes. 8 b. High Error Rates for Diagnoses Made by External Providers 9 Dr. Taylor also claims that his complaint is different from the Osinek Complaint because 10 he has alleged that Kaiser failed to act even after audits revealed high error rates with diagnoses 11 submitted by external providers. (Dr. Taylor refers to this as “one-way look chart review” in his 12 papers. See, e.g., Taylor Opp’n at 7.) The relevant allegations in support of this theory are as 13 follows: 14 • Several of Kaiser’s regions, including Colorado, Hawaii, and, until recently, 15 Georgia, “rely heavily on external providers (hospitals or other facilities who are 16 not owned by Kaiser) to provide inpatient care to Kaiser’s HMO members.” 17 Taylor Compl. ¶ 83. 18 • The external providers submit claims to Kaiser after they have provided services to 19 Kaiser members, and Kaiser uses the external providers’ diagnoses as the basis for 20 the risk adjustment claims that Kaiser submits to CMS. See Taylor Compl. ¶ 84. 21 • Kaiser’s audits “have identified significant error rates in risk adjustment claims 22 [that] Kaiser submitted to CMS based on diagnoses provided by external 23 providers.” Taylor Compl. ¶ 86. For example, for the Colorado region, the error 24 rates for external providers in some years was over 40% and 60%. See Taylor 25 Compl. ¶ 88; see also Taylor Compl. ¶ 89 (adding that “[t]he error rates for certain 26 large hospitals . . . are striking” – some more than 90%). 27 • “Despite knowing of the consistent errors in claims data from external providers, 1 claims submitted by external providers. This is particularly egregious because the 2 Colorado region does have a coder review each hospital stay at an external provider 3 to look for additional diagnoses present in the chart but not coded by the treating 4 physician.” Taylor Compl. ¶ 98 (emphasis added). 5 Here, the Court agrees with Dr. Taylor that this specific aspect of his case is not related to 6 Osinek. Problematic coding related to high-value conditions is different from problematic coding 7 by external providers. Notably, there is no indication that the problematic coding by external 8 providers was related to high-value conditions. Accordingly, this specific claim made by Dr. 9 Taylor involves “different underlying facts.” Hartpence, 792 F.3d at 1131. Furthermore, it cannot 10 be said that the government would likely have found the particular problem with external 11 providers based on its investigation into the kinds of internal upcoding practices identified in 12 Osinek. 13 c. Natural Language Processing Software 14 Finally, Dr. Taylor claims that his complaint differs from the Osinek Complaint because he 15 has made allegations about Kaiser’s Natural Language Processing (“NLP”) software. The main 16 allegations made in the Taylor Complaint with respect to the NLP software are as follows. 17 • “Broadly speaking, [a] NLP program uses an algorithm to search EMRs [electronic 18 medical records] to find words that, individually or in combination, indicate that a 19 patient has certain diagnoses.” Taylor Compl. ¶ 191. 20 • Kaiser developed its own NLP audit program “to try to find new diagnosis codes to 21 submit.” Taylor Compl. ¶ 191. 22 • “All face-to-face visits to a physician or hospital . . . are run through the NLP 23 software to identify new diagnoses that might be appropriate to use for submission 24 of additional risk adjustment claims. The results are grouped into four categories: 25 (a) True Positive: [meaning] diagnoses . . . have been confirmed by two Kaiser 26 coders; (b) More Information Needed: [meaning] diagnoses . . . may be present, but 27 further analysis is required to confirm; (c) Problem List Only: [meaning] diagnoses 1 no documentation of treatment; and (d) False Positives or Found Elsewhere.” 2 Taylor Compl. ¶ 196. 3 • “Kaiser allows the various regions to decide how to use [the above] information.” 4 Taylor Compl. ¶ 197. For some regions, if a result is True Positive, then a claim is 5 submitted to CMS for payment – without any further review. See Taylor Compl. ¶ 6 200. This is true even though audits have revealed that there is a high error rate for 7 True Positives. See Taylor Compl. ¶¶ 198, 200. 8 As indicated by the above, Dr. Taylor is not focusing here on the fact that Kaiser uses the 9 NLP program to mine records for instances where it can upcode. Had he done so, then his case 10 would clearly be related to Osinek. Rather, Dr. Taylor’s point is that there is a high error rate 11 associated with the NLP program’s True Positives, but Kaiser still submits claims based on True 12 Positives without any further inquiry. 13 The Court finds that the nature of wrongdoing claimed by Dr. Taylor here involves 14 different “material elements” from Osinek. Lujan, 243 F.3d at 1189. Dr. Taylor is charging 15 Kaiser with exploiting True Positives; this is different from Osinek which is focused on the 16 exploitation of high-value conditions. Similar to above, there is nothing that suggests True 17 Positives appear with high-value conditions only, or even primarily. Thus, Dr. Taylor here has 18 “‘significant information to contribute of [his] own.’” Id. The Court also notes that, under the 19 practical aspect of the material facts test, this part of Taylor should not be deemed related to 20 Osinek. Based on Osinek, the government would likely have looked at the NLP program given 21 that it was purportedly used to data mine; however, that would not lead the government to 22 question the True Positive results yielded by the NLP program. Rather, as a facial matter, the 23 more likely candidates for exploitation by Kaiser would be the categories of “More Information 24 Needed” and “Problem List Only,” not the True Positives. 25 d. Summary 26 The Taylor case is not dismissed in its entirety but only in part. Taylor differs materially 27 from Osinek in three ways: (1) Taylor points to a nationwide or corporate-wide problem whereas 1 related to external providers rather than high-value conditions; and (3) Taylor asserts a problem 2 with Kaiser failing to evaluate the True Positives results yielded by the NLP program. 3 B. Osinek and Arefi 4 As noted above, the Arefi Plaintiffs do not oppose the motion to dismiss their complaint. 5 Thus, the Court may move on to the next complaint filed after Arefi. 6 C. Osinek and Stein 7 As an initial matter, the Court takes note that the Stein plaintiffs contend that the Court 8 should compare the complaint in Osinek with their operative SAC, which was filed in November 9 2021. See Opp’n at 3 n.2. The Court rejects that argument for the reasons discussed above. 10 Based on their original complaint, the Stein plaintiffs focus on two specific conditions: 11 sepsis and malnutrition.15 The main allegations are as follows. 12 • Sepsis. The criteria for diagnosing sepsis is not straightforward. See, e.g., Stein 13 Compl. ¶¶ 42-43, 49. In 2003, the ICD-9 diagnostic codes and ICD Guidelines 14 were modified to, in essence, reflect that complexity. The Guidelines emphasize 15 that “coders will likely have to query physicians when documenting Sepsis to 16 trigger proper documentation that supports the Sepsis diagnosis due [to] the 17 complex nature of those diseases.” Stein Compl. ¶ 48. Defendants engaged in a 18 scheme “to up-code and falsely diagnose MA enrollees with sepsis and/or severe 19 sepsis, i.e., sepsis with acute organ failure[] (collectively referred to as ‘Sepsis’)[,] 20 when Sepsis was not present.” Stein Compl. ¶ 50. “[The] fraudulent scheme 21 [involved] the identification and treatment of Sepsis for . . . MA enrollees that 22 presented in the emergency room (ER) of [Kaiser] hospitals and was accomplished 23 by (a) [Defendants] implementing unwritten policies that prohibited coders 24 employed by [Defendants] from performing physician queries for Sepsis diagnoses 25 as required by the ICD-9 Guidelines, (b) implementing unwritten policies 26 15 The Stein plaintiffs also included in their original complaint allegations on Kaiser’s practice of 27 “refreshing.” However, in their opposition, they have conceded that this conduct was sufficiently 1 requiring Kaiser’s coders to code ICD-9 diagnosis codes for Sepsis based solely on 2 the physician’s instructions to code Sepsis instead of relying on the supporting 3 clinical findings documented in the medical record, (c) using an improper Sepsis 4 diagnostic standard that overstated the frequency of Sepsis diagnoses, (d) 5 aggressively diagnosing Sepsis as part of a strategy to lower the reported Sepsis 6 mortality rates at [Kaiser] hospitals throughout California, and (e) [Defendants], as 7 an express condition of receiving capitation payments from CMS, routinely and 8 annually falsely certifying that such ICD-9 diagnosis codes for Sepsis were 9 accurate, complete, and truthful . . . .” Stein Compl. ¶ 50. 10 • Malnutrition. Defendants “participated in a fraudulent scheme to up-code and 11 falsely diagnose malnutrition and severe malnutrition of their MA enrollees.” 12 Stein Compl. ¶ 70. The scheme “was conducted at all [Kaiser] Hospitals 13 throughout California and involved the diagnoses and coding of malnutrition and 14 severe malnutrition based upon assessment performed by dieticians . . . . The . . . 15 dietician used a rubber stamp on the MA enrollee’s medical record indicating that 16 in his/her opinion the MA enrollee suffered from malnutrition or severe 17 malnutrition. [The] physicians then countersigned the stamp in the MA enrollees’ 18 medical record. Based solely on the presence of the physician’s countersignature . 19 . . , Kaiser’s coders recorded the ICD-9 diagnosis codes for malnutrition or severe 20 malnutrition . . . .” Stein Compl. ¶ 70. There was no face-to-face encounter nor 21 were there clinical findings in support, as required by federal regulations. See 22 Stein Compl. ¶ 71. 23 These claims overlap with Osinek. Osinek asserts that upcoding was improper because it 24 was based on exploiting high-value conditions – e.g., exaggerating a patient’s condition, 25 diagnosing a patient based on a test that took place after the patient visit, diagnosing a patient for a 26 condition for which the patient was not treated, diagnosing without the necessary 27 support/documentation, and the like. Stein implicates the same kind of conduct; essentially, Stein 1 malnutrition) specifically. Notably, Osinek, like Stein, expressly identified malnutrition as one of 2 the high-value conditions that was being exploited.16 See Osinek Compl. ¶ 25. The fact that 3 Osinek did not also expressly identify sepsis as a high-value condition (as Stein did) is not 4 dispositive since it is but one example of the alleged upcoding. Cf. Hampton, 318 F.3d at 219 5 (concluding actions were related because both alleged that bills were submitted for ineligible and 6 undocumented Medicare services and for services not medically necessary). 7 The Stein plaintiffs suggest that their case is still materially different because the frauds 8 implicated in their complaint were committed when “patients . . . were admitted to a KFH hospital 9 or in the case of Sepsis, treated as a hospital outpatient through the emergency room.” Opp’n at 10 12-13. But nothing about Osinek excepts a hospital setting from the alleged upcoding. 11 The only place where Stein is materially different from Osinek is with respect to the scope 12 of the alleged misconduct. Stein suggests – by virtue of the Kaiser entities sued – that the alleged 13 misconduct goes beyond California. While this does make Stein different from Osinek (the latter 14 being California-centric in scope), Stein runs into a problem still because Taylor, the next case 15 filed after Osinek, implicates a nationwide or corporate-wide problem and is broad enough to 16 encompass the basic kind of upcoding practices alleged in Stein. 17 The Court therefore dismisses the Stein case based on the first-to-file bar – in its entirety. 18 The Stein plaintiffs have asked for leave to amend but that is a futile request since the Court’s 19 evaluation is limited to the original Stein Complaint. 20 D. Osinek and Bryant 21 As a preliminary matter, the Court takes note that Defendants do not seek to dismiss 22 Bryant in its entirety. It recognizes that Bryant has retaliation claims (based on the FCA and on 23 16 The Stein FAC (filed in May 2016, i.e., a few months after the original Stein Complaint) did 24 add in a new condition – i.e., aortic atherosclerosis (“AA”). See Stein FAC ¶ 83 (noting that AA “is a chronic condition that results in the build up of arterial plaque or fatty deposits in the 25 patient’s aorta”). According to the Stein Plaintiffs, “Kaiser’s coders coded . . . MA patients with an AA diagnosis based solely upon the physician’s notation of AA in the medical record, without 26 the medical record reflecting that the patient was treated for his/her AA condition.” Stein FAC ¶ 84. Even if the Court were to consider this new condition identified in the Stein FAC (as well as 27 the Stein SAC), there would still be overlap with the Osinek complaint. The nature of the conduct 1 other federal and state law) which are not subject to the first-to-file bar. See Mot at 1 n.3 (“This 2 Motion seeks dismissal of the Later-Filed Complaints . . . in their entirety except for the retaliation 3 causes of action in the operative Bryant complaint (Counts 5 through 8) and the California False 4 Claims Act causes of action in the operative Bicocca complaint (Counts 3 and 4) . . . .”).17 5 According to the Bryant plaintiffs, their case is materially different from Osinek in that,

6 unlike the Osinek complaint, the Bryant/Hernandez Complaint exposes Kaiser's upcoding fraud: (i) relating to a specific high-value 7 diagnosis code, mechanical ventilator dependence status, that first came to light after Osinek filed her suit[18]; (ii) not just on the 8 Medicare Advantage program, but on a different government program (the Affordable Care Act) run by a different government 9 agency entirely (the Department of Health and Human Services); and (iii) within Kaiser's insurance and physician-practice behemoths 10 in all regions, not just one region, and also within Kaiser's massive hospital operation across Kaiser's regions, which was not even part 11 of the Osinek suit. 12 Bryant Opp’n at 2. 13 For (iii), as noted above, it does not matter that the hospital setting was not expressly 14 implicated in Osinek. Nothing in Osinek suggests that it excludes the hospital setting. The Bryant 15 plaintiffs fairly argue that Osinek is California-centric; however, as noted above, Taylor – the case 16 next in line after Osinek – put the government on notice of a nationwide or corporate-wide 17 problem related to the upcoding of high-value conditions. 18 As for (i), see Bryant Compl. ¶¶ 85-86 (alleging that “a patient is vent dependent only if 19 the patient relies on the ventilation to live on a long-term basis and not for the short-term acute 20 phase of a condition” but Defendants fail to comply with that guidance) (emphasis omitted), the 21 Bryant plaintiffs are basically making the same kind of argument that the Stein plaintiffs did. See 22 Bryant Opp’n at 15 (“Given the sheer number of diagnosis codes, it would be impossible for the 23 government to identify fraudulent over-documenting and upcoding in particular diagnoses without 24 17 At one point, the Bryant plaintiffs also had California FCA claims but they dropped those 25 claims in their FAC.

26 18 In their complaint, the Bryant plaintiffs explicitly identified additional high-value HCCs or diagnoses but implicitly recognize that these conditions have already been expressly named by 27 plaintiffs in earlier-filed actions. See, e.g., Bryant Compl. ¶¶ 56, 107, 136, 144, 146-49 (referring 1 being pointed in the right direction.”). But as with Stein, this is lesser-included conduct and 2 sufficiently similar to the conduct put at issue in Osinek – i.e., Defendants were exploiting high- 3 value conditions by failing to provide support/documentation for the upcoding. Cf. Hampton, 318 4 F.3d at 219 (concluding actions were related because both alleged that bills were submitted for 5 ineligible and undocumented Medicare services and for services not medically necessary). The 6 fact that the practice allegedly was not discovered until after Osinek was filed does not mean that 7 Bryant cannot be a related case. That Defendants may have allegedly expanded their misconduct 8 to other high-value conditions does not negate the government being put on notice of the 9 fraudulent scheme in the first instance as a result of Osinek. 10 This leaves the Bryant plaintiffs with (ii). Here, the Bryant plaintiffs correctly point out 11 that their original complaint contains multiple allegations about payments Defendants receive 12 under the Affordable Care Act (and not just Medicare Advantage). For example, the Bryant 13 plaintiffs allege as follows: 14 • “The United States contributes to premiums that individuals pay to private health 15 insurance companies such as Kaiser under the Affordable Care Act. See Bryant 16 Compl. ¶ 7; see also Bryant Compl. ¶ 52 (alleging that the government contributes 17 through tax credits). 18 • “The Affordable Care Act sets up a program of risk adjustment in individual and 19 group markets to lessen or eliminate the influence of risk selection on the premiums 20 that plans charge. In the risk adjustment model utilized under the AA, which is 21 named the HHS-Hierarchical Condition Categories (‘HHS-HCC’) risk adjustment 22 model, HHS [i.e., the Department of Health and Human Services] utilizes criteria 23 and methods similar to those utilized under the Medicare Advantage Program, and 24 adapts Medicare Advantage HCCs for use in the HHS-HCC model.” Bryant 25 Compl. ¶ 51; see also Bryant Compl. ¶ 7. 26 • Thus, “as under the Medicare Advantage Program, the ACA risk adjustment model 27 creates powerful incentives for private health insurance companies like Kaiser to 1 their enrollees; the more codes that are reported, the higher premiums the 2 companies are permitted to charge, and the higher contributions will be made to 3 such premiums by the United States.” Bryant Compl. ¶ 7. 4 • “Defendants overdocument and upcode risk adjustment claims relevant to 5 individuals covered by the ACA in the same manner and pursuant to the same 6 schemes as relevant to the Medicare Advantage program . . . .” Bryant Compl. ¶ 7 11(b). 8 • The Bryant plaintiffs “seek . . . to recover damages and civil penalties arising from 9 the false or fraudulent records, statements and/or claims that the Defendants made 10 or caused to be made in connection with false and/or fraudulent claims for 11 Medicare Advantage risk adjustment payments [and] Affordable Care Act 12 insurance premiums . . . .” Bryant Compl. ¶ 18. 13 In response, Defendants point out that the actual causes of action asserted by the Bryant 14 plaintiffs at the end of their complaint refer to risk adjustment payments under Medicare only. 15 See, e.g., Bryant Compl. ¶¶ 209-10, 214-15, 220, 224-26. Nowhere do the causes of action refer 16 to risk adjustment payments under the Affordable Care Act. Although Defendants are correct, the 17 Bryant plaintiffs fairly point out that each cause of action does incorporate all paragraphs 18 previously pled. Accordingly, although the Bryant plaintiffs could have drafted a better and 19 clearer pleading, they have not pled themselves out of FCA claims based on the Affordable Care 20 Act as Defendants contend. 21 At the hearing, Defendants argued that, even if the Bryant plaintiffs have claims predicated 22 on the Affordable Care Act, they are nonetheless still barred by the first-to-file provision because 23 the government would have been put on notice of that alleged fraudulent scheme by virtue of 24 Osinek. The Court does not agree. The Affordable Care Act is an entirely different scheme, not 25 run by CMS specifically, and covering a broad range of individuals outside of the reach of 26 Medicare. The ACA claims thus state causes of action entirely different and distinct from the 27 Medicare Advantage claims. The government was put on notice of a problem with the Medicare 1 that the government’s investigation of Medicare Advantage would naturally lead to an 2 investigation of the Affordable Care Act. 3 Accordingly, the Court dismisses Bryant but only in part. The claims that survive are the 4 retaliation claims, as well as the claims based on the Affordable Care Act. 5 E. Osinek and Bicocca 6 Defendants do not seek to dismiss the entirety of Bicocca. Specifically, they recognize that 7 Dr. Bicocca has claims pursuant to the California FCA that are not subject to the first-to-file bar. 8 See Mot. at 1 n.3 (“This Motion seeks dismissal of the Later-Filed Complaints . . . in their entirety 9 except . . . the California False Claims Act causes of action in the operative Bicocca complaint 10 (Counts 3 and 4) . . . .”). 11 Second, the Court takes note that Dr. Bicocca admits his original complaint (filed in 12 February 2020) is barred by the first-to-file provision. He argues, however, that his FAC (filed in 13 October 2020) adds a theory that is not in Osinek. See Bicocca Opp’n at 2. Because, as discussed 14 above, it is Dr. Bicocca’s original complaint that must be compared to the Osinek Complaint, he is 15 out of luck. 16 However, even if the Court were to consider the FAC, Dr. Bicocca’s contention that his 17 case is materially different from Osinek is without merit. He states:

18 Relator Bicocca’s Amended Complaint describes “two sources” of diagnoses that Kaiser requires physicians to add. Bicocca Dkt. 16 ¶ 19 108. One of these sources is addenda that Kaiser gives physicians after a patient visit, which include additional diagnoses for the 20 physician to retroactively add to the patient’s chart. Id. at ¶¶ 106, 110. The other is a list of the patient’s past diagnoses, which Kaiser 21 gives to physicians before the physician meets with the patient, with the intention that the physician will re-diagnose each of the specific 22 diagnoses during the visit (“upfront list”). Id. at ¶ 109. 23 Bicocca Opp’n at 2 (emphasis in original); see also Bicocca FAC ¶ 109 (“The first source [for 24 diagnoses that physicians are required to add onto Medicare] are diagnoses that these patients had 25 already, confirmed in previous years by other physicians. While, as a matter of first impression, 26 this does not seem to be a violation of regulations, since the patients have already had the 27 diagnoses confirmed by others, having physicians re-confirm these diagnoses without spending 1 Medicare’s regulations on confirming diagnoses for the purpose of risk adjustment.”). The 2 problem for Dr. Bicocca is that the latter is essentially refreshing, which has already been put at 3 issue in Osinek. 4 Accordingly, the Court grants the motion to dismiss Dr. Bicocca’s claims, and the only 5 claims that survive are those based on the California FCA which Defendants have not contested 6 for purposes of the first-to-file bar. 7 IV. CONCLUSION 8 For the foregoing reasons, the Court grants in part and denies in part Defendants’ motion 9 to dismiss based on the first-to-file bar. Specifically: 10 • Arefi and Stein are dismissed in their entirety. 11 • Taylor is dismissed except to the extent that it pleads (1) a nationwide or corporate- 12 wide fraud; (2) a fraud based on improper coding by external providers; and (3) a 13 fraud based on True Positive results from the NLP program. 14 • Bryant is dismissed except to the extent that it pleads (1) retaliation claims and (2) 15 claims based on fraud in the Affordable Care Act program. 16 • Bicocca is dismissed except to the extent that it pleads claims based on the 17 California FCA. 18 This order disposes of Docket No. 141. 19 20 IT IS SO ORDERED. 21 22 Dated: May 5, 2022 23 24 ______________________________________ EDWARD M. CHEN 25 United States District Judge 26 27