Sidibe v. Sutter Health

Filing 714

ORDER re 698 Order re Motion for Class Certification.The court issues this redacted version of its August 30, 2019 order granting the plaintiffs' motion to certify their proposed class under Federal Rule of Civil Procedure 23(b)(2) an d denying without prejudice their motion to certify their proposed class under Federal Rule of Civil Procedure 23(b)(3). (This version also makes non-substantive changes to re-insert an inadvertently omitted word at 47:18 and to update the citations to Senne v. Kansas City Royals Baseball Corp. at 51–52 from Westlaw citations to F.3d citations.) Signed by Judge Laurel Beeler. (lblc1S, COURT STAFF) (Filed on 10/18/2019)

1
2
3
4
5
6
7
8

UNITED STATES DISTRICT COURT

9

NORTHERN DISTRICT OF CALIFORNIA

10

San Francisco Division

United States District Court
Northern District of California

11

DJENEBA SIDIBE, et al.,
Plaintiffs,

12
v.

13
14

SUTTER HEALTH,
Defendant.

15

Case No. 12-cv-04854-LB
(REDACTED) ORDER GRANTING
MOTION TO CERTIFY CLASS
UNDER RULE 23(B)(2) AND DENYING
WITHOUT PREJUDICE MOTION TO
CERTIFY CLASS UNDER RULE
23(B)(3)
Re: ECF Nos. 348 (under seal) and 379
(redacted version)

16
17
18

TABLE OF CONTENTS

19

INTRODUCTION ............................................................................................................................. 3

20

STATEMENT ................................................................................................................................... 5

21

1. The Proposed Class ..................................................................................................................... 5

22

2. Background ................................................................................................................................. 5

23

3. The Plaintiffs’ Antitrust Allegations Regarding Sutter’s Anticompetitive Practices.................. 7

24

3.1

“Penalty Rate” Provisions ................................................................................................. 7

25
26
27
28

Sutter’s Systemwide Contracting and Its “All-or-Nothing,” “Anti-Steering,” and

3.2

The “But For” World ....................................................................................................... 12

4. The Plaintiffs’ Calculations of Antitrust Injury and Damages to Class Members .................... 13
4.1

Calculating Sutter’s Overcharges .................................................................................... 15

ORDER – No. 12-cv-04854-LB

1

4.2

Assuming Health Plans “Pass On” 100 Percent of Any Sutter Overcharges They

2

Have to Pay Through to the Premiums That They Charge Their Customers .................. 20

3

ANALYSIS ..................................................................................................................................... 27

4

1. Rule 23(a) Prerequisites ............................................................................................................ 28

5

1.1

Numerosity — Rule 23(a)(1)........................................................................................... 28

6

1.2

Commonality — Rule 23(a)(2) ....................................................................................... 29

7

1.3

Typicality — Rule 23(a)(3) ............................................................................................. 30

8

1.4

Adequacy — Rule 23(a)(4) ............................................................................................. 32

9

1.4.1

Prosecuting the action vigorously on behalf of the class....................................... 33

10

1.4.2

Conflicts of interest ............................................................................................... 33

United States District Court
Northern District of California

11
12

2. Rule 23(b) Prerequisites ............................................................................................................ 38
2.1

Predominance — Rule 23(b)(3) ...................................................................................... 38

13

2.1.1

Antitrust violations ................................................................................................ 39

14

2.1.2

Antitrust injury and calculating damages .............................................................. 39

15

2.1.2.1

Calculating Sutter’s overcharges ................................................................. 42

16

2.1.2.2

Assuming health plans “pass on” 100 percent of any Sutter overcharges

17

they have to pay through to the premiums that they charge their

18

customers ..................................................................................................... 44

19
20

2.2

Acting on Grounds That Apply Generally to the Class — Rule 23(b)(2) ....................... 50

CONCLUSION ............................................................................................................................... 53

21
22
23
24
25
26
27
28
ORDER – No. 12-cv-04854-LB

2

INTRODUCTION

1

In this putative class action, six plaintiffs (four individuals who enrolled in health-insurance

2
3

policies from the health plans Aetna, Anthem Blue Cross, and Blue Shield and two small

4

companies that paid for health insurance for their employees) are suing Sutter Health, which owns

5

and operates a network of hospitals and medical-service providers in Northern California, for

6

violations of the federal Sherman Antitrust Act, the California Cartwright Act, and the California

7

Unfair Competition Law.

8

The plaintiffs allege that Sutter has “market power” in seven specific “geographic markets”

9

(the “Tying Markets”) in Northern California, where Sutter’s hospitals are either the only hospital
(i.e., a monopoly) or the dominant hospital in the market.1 Health plans like Anthem and Blue

11

United States District Court
Northern District of California

10

Shield must include those Sutter hospitals in their provider networks to be able to assemble health-

12

insurance products that are commercially marketable.2 Sutter allegedly uses that leverage to

13

require that health plans enter into “systemwide contracts” that include “all-or-nothing” and “anti-

14

steering” provisions. Those provisions (1) require health plans to accept as in-network providers

15

all of Sutter’s hospitals, at the prices Sutter dictates, and (2) prevent health plans from

16

incentivizing their enrollees to go to lower-cost hospitals instead of Sutter’s higher-cost hospitals.3

17

In particular, the plaintiffs allege that Sutter (1) requires health plans to include its hospitals in

18

four other geographic markets (the “Tied Markets”), at the prices Sutter dictates, and (2) prevents

19

health plans from incentivizing their enrollees to go to non-Sutter hospitals in the Tied Markets.4

20

Unlike in the Tying Markets, where Sutter has market power, in the Tied Markets, there are more

21

hospitals and more hospital competition. This competition normally would drive Sutter’s prices

22
23
24
25

Fourth Amend. Compl. (“4AC”) – ECF No. 204 at 4 (¶ 4), 10 (¶¶ 30–31), 29 (¶¶ 86–87). The
plaintiffs originally alleged eight Tying Markets but stipulated that summary judgment should be
granted with respect to one of them. Citations refer to material in the Electronic Case File (“ECF”);
pinpoint citations are to the ECF-generated page numbers at the top of documents.
1

27

2

Id. at 4 (¶ 6), 11–12 (¶ 35).

3

Id. at 4–5 (¶¶ 5–7), 11–12 (¶¶ 33–36), 13–15 (¶¶ 40–45).

4

26

Id.

28
ORDER – No. 12-cv-04854-LB

3

1

down.5 But by tying its hospitals in the Tied Markets to its “must have” hospitals in the Tying

2

Markets, Sutter forecloses competition by other hospitals in the Tied Markets and thus is able to

3

charge and maintain supra-competitive prices at its hospitals.6 The plaintiffs allege that the health

4

plans have to pay Sutter supra-competitive prices and then, in turn, pass on those costs through to

5

their customers in the form of higher premiums.7 Consequently, it is the health plans’ customers

6

— individuals and employers that buy health insurance — that ultimately bear the burden of

7

paying Sutter’s supra-competitive prices.8

8

The plaintiffs seek (1) treble damages and restitution from Sutter to compensate them for the

9

overcharges they incurred from Sutter’s alleged anticompetitive behavior and (2) a declaration that
Sutter’s practices are anticompetitive and an injunction barring Sutter from continuing to engage

11

United States District Court
Northern District of California

10

in anticompetitive behavior, including its “tying,” “all-or-nothing,” and “anti-steering”

12

arrangements.9 They move to certify a class under Federal Rule of Civil Procedure 23(b)(2) and

13

(b)(3) of all individuals and entities located in nine specific California Rating Areas (“RAs”)10 that

14

paid premiums for fully insured health-insurance policies from the health plans Blue Shield,

15

Anthem, Aetna, Health Net, or UnitedHealthcare from September 28, 2008 to the present.11
Sutter opposes the plaintiffs’ motion. Sutter’s main arguments are that (1) there are intraclass

16
17

differences and conflicts that render the plaintiffs atypical and inadequate to represent the class

18

and (2) individual issues about whether class members suffered antitrust injury, and how class

19

members’ damages would be calculated, predominate over common issues.

20
21
22
5

See id. at 6 (¶ 9), 30 (¶ 94).

23

6

Id. at 4–5 (¶¶ 5–8), 11–12 (¶¶ 35–36), 30 (¶ 94), 33 (¶¶ 103–05).

24

7

Id. at 3 (¶ 2), 5 (¶ 8), 10 (¶ 28), 34–35 (¶¶ 109–12).

8

Id.

9

Id. at 43.

25
26
27
28

Under the Affordable Care Act, states are required to define geographic “rating areas” to be used by
health plans in setting premium prices for individual and small-group health-insurance policies. 42
U.S.C. § 300gg(a)(2).
10

11

Pls. Mot. for Class Certification (“MCC”) – ECF No. 379 at 3.

ORDER – No. 12-cv-04854-LB

4

The court held a hearing and now rules as follows. The court finds that the plaintiffs have not

1
2

made a showing that issues of antitrust injury and damages are subject to common proof such that

3

certification of a damages class under Rule 23(b)(3) is appropriate. But the court also finds that the

4

plaintiffs have met the requirements for certification of an injunctive- and declaratory-relief class

5

under Rule 23(b)(2). The court thus grants the plaintiffs’ motion to certify their proposed class

6

under Rule 23(b)(2) and denies without prejudice their motion to certify their proposed class under

7

Rule 23(b)(3).

8
STATEMENT

9
10

1. The Proposed Class
The plaintiffs seek to certify a proposed class of:

United States District Court
Northern District of California

11

All entities in California Rating area 1, 2, 3, 4, 5, 6, 8, 9 or 10 (the “Nine RAs”),
and all individuals that either live or work in one of the Nine RAs, that paid
premiums for a fully-insured health insurance policy from Blue Shield, Anthem
Blue Cross, Aetna, Health Net or United Healthcare from September 28, 2008 to
the present. This class definition includes Class Members that paid premiums for
individual health insurance policies that they purchased from these health plans and
Class Members that paid premiums, in whole or in part, for health insurance
policies provided to them as a benefit from an employer or other group purchaser
located in one of the Nine RAs.12

12
13
14
15
16
17
18
19

2. Background
“‘The market for hospital services and medical care is complex.’” Sidibe v. Sutter Health, No.

20
21

12-cv-04854-LB, 2019 WL 2078788, at *4 (N.D. Cal. Apr. 12, 2019) (quoting Cascade Health

22

Sols. v. PeaceHealth, 515 F.3d 883, 891 (9th Cir. 2008)). “There are at least three transactions

23

involved in providing hospital services and health care in connection with health insurance.” Id.
“First, hospitals [such as Sutter] sell hospital services to health-insurance plans [such as Blue

24
25

Shield, Anthem, Aetna, Health Net, or UnitedHealthcare]. Hospitals and health plans negotiate

26

whether a given hospital will be included in the health plan’s network and negotiate the rates that

27
28

12

Pls. MCC – ECF No. 379 at 13.

ORDER – No. 12-cv-04854-LB

5

1

the health plan will pay the hospital for its hospital services.” Id. (citing Saint Alphonsus Med.

2

Ctr.-Nampa Inc. v. St. Luke’s Health Sys., Ltd., 778 F.3d 775, 784 & n.10 (9th Cir. 2015) (citing

3

Gregory Vistnes, Hospitals, Mergers, and Two-Stage Competition, 67 Antitrust L.J. 671, 672, 674

4

(2000))). “These negotiations are highly price-sensitive.” Id. (citing FTC v. Advocate Health Care

5

Network, 841 F.3d 460, 465 (7th Cir. 2016) (citing Vistnes, 67 Antitrust L.J. at 674–75)). “All else

6

being equal, hospitals prefer higher rates and health plans prefer lower rates.” Id. (citing Cascade

7

Health, 515 F.3d at 892).13
“Second, health plans sell health insurance to consumers. The consumers are individuals (who

8
9

directly purchase health insurance for themselves or their families) and employers (which
purchase health insurance for their employees).” Id. at *5 (citing Cascade Health, 515 F.3d at 892;

11

United States District Court
Northern District of California

10

St. Luke’s, 778 F.3d at 784). “An important way that health plans compete for consumers is their

12

provider networks: the hospitals, physicians, and ancillary providers that the health plan offers ‘in

13

network’ and that enrollees are encouraged to use.” Id. (citing Gregory S. Vistnes & Yianis

14

Sarafidis, Cross-Market Hospital Mergers: A Holistic Approach, 79 Antitrust L.J. 253, 267

15

(2013)). “All else being equal, a health plan with a more comprehensive provider network will be

16

more attractive to consumers.” Id. (citing Vistnes & Sarafidis, 79 Antitrust L.J. at 267). “At the

17

same time, health plans that have high-priced providers in their networks have higher costs.” Id.

18

(citing Vistnes & Sarafidis, 79 Antitrust L.J. at 267). “Thus, in choosing how inclusive their

19

provider network is, health plans balance the benefit of more comprehensive networks with the

20

costs of paying more to providers in their networks.” Id. (citing Vistnes & Sarafidis, 79 Antitrust

21

L.J. at 267).14
“Third, hospitals seek to attract health-plan enrollees who need hospital services to come to

22
23

them (as opposed to other hospitals).” Id. (citing St. Luke’s, 778 F.3d at 784 n.10 (citing Vistnes,

24

67 Antitrust L.J. at 681–82); Advocate Health, 841 F.3d at 471 (citing Vistnes, 67 Antitrust L.J. at

25

672); FTC v. Penn State Hershey Med. Ctr., 838 F.3d 327, 342 (3d Cir. 2016) (citing Vistnes,

26
27
28

Accord Chipty Decl. – ECF No. 379-1 at 17 (¶ 14); Chipty Dep. – ECF No. 415-1 (under seal) at
122 (p. 425).
13

14

Accord Chipty Dep. – ECF No. 415-1 (under seal) at 123–24 (pp. 426–31).

ORDER – No. 12-cv-04854-LB

6

1

Antitrust L.J. at 672)). “Unlike health plans, which are sensitive to the prices that hospitals charge

2

for their services, enrollees are ‘largely insensitive’ to price because the prices that hospitals

3

charge are largely borne by the enrollees’ health plans, not by the enrollees.” Id. (some internal

4

quotation marks omitted) (quoting St. Luke’s, 778 F.3d at 784 n.10 (citing Vistnes, 67 Antitrust

5

L.J. at 682) and citing Advocate Health, 841 F.3d at 471 (citing Vistnes, 67 Antitrust L.J. at 677,

6

680); Penn State Hershey, 838 F.3d at 342). “Instead of taking price into account, enrollees choose

7

hospitals based mostly on non-price factors, such as location or quality of services.” Id. (internal

8

quotation marks omitted) (quoting Penn State Hershey, 838 F.3d at 341 and citing St. Luke’s, 778

9

F.3d at 784 n.10 (citing Vistnes, 67 Antitrust L.J. at 682); Advocate Health, 841 F.3d at 465

10

(citing Vistnes, 67 Antitrust L.J. at 677, 682)).15

United States District Court
Northern District of California

11
12

3. The Plaintiffs’ Antitrust Allegations Regarding Sutter’s Anticompetitive Practices

13

3.1

Sutter’s Systemwide Contracting and Its “All-or-Nothing,” “Anti-Steering,” and
“Penalty Rate” Provisions

14
As of 2015, Sutter was the largest health system (other than Kaiser Permanente) in Northern

15
16

California, with 22 hospitals and approximately 4,000 patient beds.16 Currently, Sutter has 24

17

hospitals and approximately 5,200 patient beds.17

18

Before the early 2000s, health plans were able to negotiate with Sutter’s various hospitals

19

individually. Health plans could include some Sutter hospitals in their provider network while

20

excluding (or threatening to exclude) others. Under that structure, if a health plan determined that

21

a particular Sutter hospital was too expensive, it had the option to exclude that hospital from its

22

provider networks. Consequently, each Sutter hospital had an incentive to offer competitive prices

23
24
25
26

Accord Chipty Decl. – ECF No. 379-1 at 17 (¶ 14); Chipty Dep. – ECF No. 415-1 (under seal) at
122 (p. 425). Some enrollees may be more directly sensitive to price, e.g., enrollees in high-deductible
plans. Sidibe, 2019 WL 2078788, at *5 n.13 (citing Penn State Hershey, 838 F.3d at 342 n.6).
15

16

Chipty Decl. – ECF No. 379-1 at 20–22 (¶ 20).

17

27

Id. at 22–23 (¶¶ 21–22).

28
ORDER – No. 12-cv-04854-LB

7

1

to incentivize health plans to include the hospital in their provider networks (which in turn would

2

result in the hospital getting more patient volume).18
The plaintiffs maintain that beginning in the early 2000s, Sutter began requiring health plans to

3
4

enter into “systemwide” contracts that included “all-or-nothing” requirements. Sutter’s

5

systemwide contracts require, or effectively require, a health plan that wants to include one Sutter

6

hospital in its provider network to include all Sutter hospitals. Additionally, Sutter’s systemwide

7

contracts contain “anti-steering” provisions that effectively bar health plans from creating “tiered”

8

insurance products to incentivize their enrollees to use lower-cost non-Sutter hospitals rather than

9

more expensive Sutter hospitals (i.e., products where enrollees pay lower co-pays to use providers
in higher tiers and higher co-pays to use providers in lower tiers where Sutter hospitals are in a

11

United States District Court
Northern District of California

10

lower tier).19

12

More specifically, Sutter’s systemwide contracts include provisions under which health plans

13

must pay Sutter what they have characterized as “penalty rates”20 for any Sutter hospital that they

14
15
Joyner Decl. – ECF No. 497 at 4–5 (¶ 7) (“Prior to 2002, Blue Shield negotiated with Sutter’s
various hospitals, physician groups, and other providers individually to assemble provider networks.
Blue Shield had the ability to decide whether and under what conditions each Sutter provider could
participate in Blue Shield’s provider networks. During that time, if Blue Shield found that a particular
Sutter provider was too expensive, it had the option to exclude that provider from its networks.
Therefore, each Sutter provider had an incentive to offer competitive prices if it wanted to be included
in Blue Shield’s networks.”); Melody Decl. – ECF No. 313-1 (under seal) at 5 (¶ 9); Lacroix-Milani
Decl. – ECF No. 314 (under seal) at 6 (¶ 12); Welsh Decl. – ECF No. 335 (under seal) at 5 (¶ 11).
18

16
17
18
19
20
21
22
23
24
25
26
27

Joyner Decl. – ECF No. 497 at 5 (¶ 8) (“This dynamic changed dramatically when Sutter announced
that all of its providers would be negotiating as a bloc on a ‘systemwide’ basis. Beginning with the
negotiations leading to the 2002 Systemwide Amendment between Sutter and Blue Shield, it became
clear that selecting anything less than all Sutter hospitals and physician groups would have resulted in
a requirement that Blue Shield and its self-funded health plan customers pay unsustainably high prices
for the healthcare services of Sutter’s providers. Thus, as a practical matter, inclusion of one Sutter
hospital suddenly required inclusion of all Sutter hospitals and physician groups, and one systemwide
agreement would govern all of the relationships between Blue Shield and each Sutter provider. This
was Sutter’s ‘systemwide’ negotiation strategy. It required Blue Shield to include ‘all’ of Sutter’s
hospitals and physician groups in any new networks for the health plans to be competitive.”), 6 (¶¶ 15–
16), 10 (¶ 30); Melody Decl. – ECF No. 313-1 (under seal) at 5–6 (¶¶ 9–13), 7 (¶ 18), 9 (¶ 26); de la
Torre Decl. – ECF No. 312-4 (under seal) at 5 (¶ 12); Lacroix-Milani Decl. – ECF No. 314 (under
seal) at 5–6 (¶¶ 9, 12–14); Lundbye Decl. – ECF No. 321 (under seal) at 4 (¶ 8), 8 (¶ 16); Welsh Decl.
– ECF No. 335 (under seal) at 4 (¶¶ 9–10), 5 (¶¶ 12–14).
19

Joyner Decl. – ECF No. 497 at 11–12 (¶¶ 34–35); Melody Decl. – ECF No. 313-1 (under seal) at 8
(¶ 23); de la Torre Decl. – ECF No. 312-4 (under seal) at 6 (¶ 14), 7 (¶¶ 17–18).
20

28
ORDER – No. 12-cv-04854-LB

8

1

place out-of-network or in a lower tier. Typically, when health-plan enrollees use a hospital that is

2

not in-network, the hospital can charge the health plan only the “reasonable and customary value”

3

of the hospital services provided, as mandated by state law. Sutter, however, requires health plans

4

to sign contracts that supersede the “reasonable and customary value” limit on charges and instead

5

require health plans to pay 95 percent of Sutter’s “full billed charges” — a substantially higher

6

amount — if their enrollees use Sutter hospitals that they have placed out-of-network or in a lower

7

tier. No matter how a health plan might structure its health-insurance products, it is inevitable that

8

some enrollees will have to use Sutter hospitals (e.g., if they need emergency care and a Sutter

9

hospital is the nearest hospital), in which case Sutter’s “penalty rates” would apply. These higher
“penalty rates” reduce or eliminate any savings the health plans could achieve by excluding or

11

United States District Court
Northern District of California

10

tiering Sutter hospitals in the first place. Consequently, even where health plans technically are

12

permitted under their contracts to exclude or tier Sutter hospitals, they are effectively prevented

13

from doing so in a way that would give Sutter any incentive to lower its prices.21

14
15
16
17
18
19
20
21
22
23
24
25

21

As a former Senior Vice President from Blue Shield explained:
Sutter’s unique out-of-network pricing terms worked as follows: In California, health
plans must pay the appropriate charges for emergency room services even when their
members seek emergency room treatment at out-of-network hospitals. Typically, when a
Blue Shield member uses a hospital outside of Blue Shield’s provider network (e.g., in an
emergency situation), the amount that Blue Shield (or its self-funded customers) must pay
may be limited to the “reasonable and customary value for the health care services
rendered,” as mandated by state law for out-of-network providers. (Cal. Code Regs., title
28 §1300.71(a)(3)(B).) The “reasonable and customary” rates that Blue Shield pays for
services at an out-of-network hospital are often somewhat higher than the rates it would
pay to an in-network hospital but they typically are substantially less than the “full billed
charges” that hospitals often list on their hospital “chargemasters” for their services.
A substantial percentage of hospital healthcare is provided through hospital emergency
rooms and large numbers of CalPERS [California Public Employees’ Retirement System]
health plan enrollees inevitably would end up in the emergency rooms of Sutter’s 26 acute
care hospitals in Northern California. Accordingly, Sutter demanded that CalPERS agree
in advance to pay 95% of full billed charges for any services provided at an out-of-network
(“non-participating”) Sutter hospital. As a result, CalPERS had to agree in advance to pay
“non-par” rates that substantially exceeded the “reasonable and customary” out-of-network
rates otherwise mandated by state law.
....

26
27

In 2005, Sutter introduced the concept of the Non-Par rate, or 95% of billed charges,
into its Systemwide Amendment with Blue Shield. . . .
....

28
ORDER – No. 12-cv-04854-LB

9

1

The plaintiffs maintain that Sutter can impose these systemwide contracts on health plans

2

because it has “market power” in seven specific geographic markets — the Tying Markets — in

3

Northern California. Sutter hospitals are either the only hospital or the predominant hospital in

4
5
6
7
8
9
10

United States District Court
Northern District of California

11
12
13
14
15
16

The power of the Non-Par (95% of billed charges) rate cannot be overstated. Blue
Shield has performed financial analyses that have concluded that payment of 95% of
Sutter’s full billed charges erases any possible benefit of excluding some higher-priced
Sutter providers from a network. This is because it is impossible to prevent health plan
enrollees from using Sutter hospitals in all instances. If a member needs emergency care
and uses a Sutter hospital outside of his plan (a common scenario we refer to as “leakage”),
then Blue Shield (and its self-funded payor customers) would have to pay 95% of billed
charges to Sutter. These prices are so high that having to pay them would defeat any
significant benefit of a narrow network.
....
For many years Blue Shield has attempted to create tiered networks to foster price and
quality competition among the health care providers in its networks. In 2001, for example,
Blue Shield created a tiered network called Network Choice that included Sutter hospitals,
but Sutter required that all of its facilities be included as “Tier 1 (Choice) providers.”
Sutter threatened that Blue Shield and its self-funded health plan customers would have to
pay Non-Par rates equal to 95% of full-billed charges if all its facilities were not placed in
the preferred tier. By requiring that all of its facilities be included in the first tier, Sutter
prevented Blue Shield from offering customers financial incentives to utilize more costeffective hospitals. Thus, although Sutter technically was participating in a tiered product,
its hospitals were not placed in the lower tier that actually corresponded with its
significantly higher prices.
....

17
18
19
20
21
22
23
24
25
26
27

Sutter’s threat to charge Non-Par rates in response to Blue Shield’s proposals for tiered
products was typically the death knell for these products. If Blue Shield and its self-funded
health plan customers had to pay this exorbitant 95% of billed charges rate, an amount
much higher than Blue Shield had to pay any other out-of-network provider, these products
would be economically unfeasible. Any benefit to be gained by excluding high-priced
Sutter hospitals from Tier 1 would be erased once the inevitable emergency room
“leakage” to Non-Par Sutter hospitals occurred.
....
Sutter’s anti-tiering contractual restraints affected Blue Shield and each of Blue
Shield’s self-funded health plan customers in precisely the same manner. Because Sutter
contractually prohibited health plans from giving their members financial incentives to
select lower-priced alternatives, insisted upon 1st Tier status for Sutter providers that did
not qualify and imposed charges for non-participating Sutter providers at non-par rates of
95% of full billed charges, Sutter was able to destroy the ability of any “tiered” health plan
to moderate Sutter’s above-market pricing.
Joyner Decl. – ECF No. 497 at 10–18 (¶¶ 31–55); accord Melody Decl. – ECF No. 313-1 (under seal)
at 6–10 (¶¶ 14–31); de la Torre Decl. – ECF No. 312-4 (under seal) at 6–9 (¶¶ 13–28); Lacroix-Milani
Decl. – ECF No. 314 (under seal) at 6–15 (¶¶ 12–36); Lundbye Decl. – ECF No. 321 (under seal) at 4–
7 (¶¶ 9–13), 8–10 (¶¶ 18–21); Welsh Decl. – ECF No. 335 (under seal) at 5–11 (¶¶ 11–36), 12–13
(¶¶ 45–46).

28
ORDER – No. 12-cv-04854-LB

10

1

those markets. Because health plans are required to offer their enrollees at least one nearby in-

2

network hospital, they have no choice but to contract with Sutter so that they can include those

3

Tying Market hospitals in their provider networks. Sutter then uses the fact that health plans have

4

no choice but to contract with it to force health plans to accept its systemwide-contract terms,

5

including its all-or-nothing, anti-steering, and penalty-rate provisions.22
The plaintiffs maintain that Sutter’s systemwide contracts allow it to charge higher prices at its

7

hospitals in certain other geographic markets — the Tied Markets. Unlike in the Tying Markets, in

8

the Tied Markets, there are more hospitals and more competition among hospitals. Those other

9

hospitals normally would act as price constraints on Sutter hospitals. Health plans could threaten

10

to exclude Sutter hospitals in favor of those other hospitals or place Sutter hospitals in lower tiers

11

United States District Court
Northern District of California

6

— thereby reducing Sutter hospitals’ patient volume (and, thus, their revenues) — to negotiate

12

with Sutter to get it to lower its prices. But the systemwide contracts that Sutter imposes on health

13

plans effectively bar health plans from using these negotiating tactics. Health plans thus are

14

prevented from exposing Sutter to price competition. As a result, Sutter is free to charge supra-

15

competitive prices at its hospitals in the Tied Markets.23

16
17
18
19
20
21
22
23
24
25
26
27
28

Joyner Decl. – ECF No. 497 at 7 (¶ 20) (“Sutter’s new ‘systemwide’ approach was very effective in
promoting all-or-nothing contracting in the context of California’s healthcare laws regulating the
assembly of provider networks. For example, in assembling a provider network to be marketed to selffunded payors and insured clients, California law requires Blue Shield to create a network with access
to health care providers such that ‘all enrollees have a residence or workplace within 30 minutes or 15
miles of a contracting or plan-operated hospital . . . .’ (Cal. Code Regs. title 28, § 1300.51(d)(H)(ii)).
For example, if one of Blue Shield’s self-funded customers has an employee living in a rural area of
Northern California, and the only hospital within 30 minutes or 15 miles of his residence or workplace
was a Sutter Hospital, the law mandates that Sutter’s hospital be included in the network. Beginning
with Sutter’s 2002 Systemwide Amendment, Blue Shield could not feasibly contract with that one
Sutter hospital alone. It had to include all Sutter providers on a ‘systemwide’ basis.”) (ellipsis in
original), 9 (¶ 27); Melody Decl. – ECF No. 313-1 (under seal) at 5 (¶¶ 7–8), 6 (¶ 11); de la Torre
Decl. – ECF No. 312-4 (under seal) at 4 (¶¶ 8–9); Lacroix-Milani Decl. – ECF No. 314 (under seal) at
5–6 (¶ 9), 11 (¶ 28); Lundbye Decl. – ECF No. 321 (under seal) at 3–4 (¶¶ 5–6); Welsh Decl. – ECF
No. 335 (under seal) at 4 (¶ 8).
22

Joyner Decl. – ECF No. 497 at 8–9 (¶¶ 23–25) (“It would have been desirable for Blue Shield to
foster price competition by excluding selected Sutter hospitals in locations where there were
competing hospitals that had higher or comparable quality but charged significantly lower prices.
However, for all of the reasons described above, it was not economically feasible to assemble a
provider network in Northern California that excluded all of Sutter’s hospitals and medical practices
from Blue Shield’s provider networks. Therefore, Sutter’s ‘all or none’ contracting policy substantially
impaired Blue Shield’s ability to foster price competition. As a result, Sutter was able to quickly raise
23

ORDER – No. 12-cv-04854-LB

11

The plaintiffs maintain that Sutter engages in systemwide contracting with its all-or-nothing,

1
2

anti-steering, and penalty-rate provisions with each of the health plans Blue Shield, Anthem,

3

Aetna, Health Net, or UnitedHealthcare in the same general way. Sutter also charges these health

4

plans the same allegedly supra-competitive prices, across different insurance products offered by

5

any one health plan, and across different health plans.24
The “But For” World

6

3.2

7

The plaintiffs maintain that in the “but for” world — the counterfactual world where Sutter did
not tie its hospitals or engage in anticompetitive practices25 — health plans would be able to

9

negotiate lower rates from Sutter. While health plans might still have no choice but to contract

10

with Sutter hospitals in the Tying Markets, in the but-for world, health plans could threaten to

11

United States District Court
Northern District of California

8

exclude Sutter hospitals in the Tied Markets from their provider networks or place those Sutter

12

hospitals in a lower tier in lieu of other competitor hospitals. In the face of this competition, Sutter

13

would no longer be able to charge supra-competitive prices at its Tied Market hospitals and

14

instead would have to lower its prices to stay competitive. The plaintiffs maintain that this would

15

benefit all healthcare consumers and class members.26

16

For example, suppose a health plan currently offers a health-insurance policy that costs $100 a

17

month and includes Tied Market Sutter hospitals in the top tier of its in-network providers. Under

18
19
20
21
22
23

its reimbursement rates significantly above the rates that competing providers charged. . . . At or
around the time that Blue Shield entered into its ‘systemwide’ agreement with Sutter in 2002, Blue
Shield determined that Sutter’s pricing for acute care hospital services began increasing at a
dramatically faster pace than the prices of all or nearly all of its competitors.”) (emphasis in original),
9 (¶ 27), 12 (¶ 36), 18–20 (¶¶ 54–61); Melody Decl. – ECF No. 313-1 (under seal) at 8–9 (¶ 24), 10
(¶ 30); de la Torre Decl. – ECF No. 312-4 (under seal) at 3–4 (¶¶ 4–7); de la Torre Decl. – ECF No.
312-4 (under seal) at 5 (¶ 10), 7 (¶ 16); Lacroix-Milani Decl. – ECF No. 314 (under seal) at 4–5 (¶¶ 5–
7), 6 (¶ 10); Lundbye Decl. – ECF No. 321 (under seal) at 9–10 (¶ 20).
Joyner Decl. – ECF No. 497 at 5–6 (¶ 13), 20 (¶ 61); Melody Decl. – ECF No. 313-1 (under seal) at
4 (¶¶ 4, 6); de la Torre Decl. – ECF No. 312-4 (under seal) at 5 (¶ 11); Lacroix-Milani Decl. – ECF
No. 314 (under seal) at 15 (¶¶ 37–39); Welsh Decl. – ECF No. 335 (under seal) at 5 (¶ 12), 13 (¶¶ 47–
48).
24

24
25
26
27

See Chipty Dep. – ECF No. 415-1 (under seal) at 125 (p. 436) (“Q. And in this case — taking the
general to the specific, in this case, the but-for world would be Sutter negotiating contracts that did not
have the challenge[d] contract terms? A. That’s correct.”).
25

26

Pls. MCC Reply – ECF No. 559 at 14.

28
ORDER – No. 12-cv-04854-LB

12

1

Sutter’s systemwide contracts, the health plan effectively is barred from offering an alternative

2

health-insurance policy for, say, $80 a month,27 that excludes those Sutter hospitals or places them

3

in a lower tier.28 In the but-for world, however, the health plan could offer such a policy. Some

4

health-insurance buyers who prioritize having in-network access to Sutter hospitals might continue

5

to enroll in the first policy. But others who place less importance on access to Sutter hospitals or

6

more importance on lowering their insurance premiums might enroll in the new second policy

7

instead. Under the second policy, those latter buyers would use Sutter hospitals less often to avoid

8

the higher co-pays associated with using out-of-network or lower-tiered providers. This would

9

reduce Sutter’s patient volume and, consequently, its revenues. In response, the plaintiffs argue,
Sutter would reduce the prices it charges the health plan, so that, in turn, the health plan could

11

United States District Court
Northern District of California

10

reduce the premiums it has to charge for the first policy (perhaps down to, say, $90 a month

12

instead of $100), which in turn would entice more customers to sign up for the first policy again

13

instead of the second and use more Sutter hospitals, which in turn would restore some of Sutter’s

14

lost patient volume. As a result, even Sutter loyalists who would not enroll in a tiered health-

15

insurance policy (even if one were offered) nonetheless would be better off in the but-for world

16

because they would be able to enroll in the first non-tiered policy at a cheaper rate.29

17
18

4. The Plaintiffs’ Calculations of Antitrust Injury and Damages to Class Members

19

One of the central issues in the litigation and in the plaintiffs’ motion for class certification

20

(and Sutter’s opposition thereto) is how to assess the antitrust injury that class members allegedly

21

have sustained and how to calculate class members’ purported damages.

22
23
27

24
25
26
27

These numbers are arbitrary and are used solely for illustrative purposes.

According to the plaintiffs, under Sutter’s systemwide contracts, the health plan either is actually
barred from offering such a policy or is effectively barred from doing so because Sutter would then
charge the health plan a “penalty rate” if any enrollee who held that policy used a Sutter hospital (e.g.,
in an emergency situation), a rate so onerous as to make the policy unsustainable. See supra notes 19,
21.
28

See Chipty Decl. – ECF No. 379-1 at 44–45 (¶¶ 61–62); Chipty Reply Decl. – ECF No. 559-1 at 20–
22 (¶¶ 19, 22).
29

28
ORDER – No. 12-cv-04854-LB

13

The plaintiffs offer the following formula, developed by their expert Dr. Tasneem Chipty, to

1
2

assess the purported antitrust injury and calculate damages for each class member. Broadly

3

speaking, the plaintiffs’ formula is as follows:
1. For each year between 2006 and 2015, calculate as a percentage how much of the

4
5

amount that each relevant Sutter hospital charged each health plan and its patients was

6

an overcharge (i.e., was beyond what the Sutter hospital would have charged in the but-

7

for world).30
2. Disaggregate the total amount each health plan paid to each Sutter hospital into

8

“cohorts” based on (1) the geographic Rating Area where health-plan enrollees live or

9

work, (2) the “group type” of the insurance buyer (individual buyers versus “small-

11

United States District Court
Northern District of California

10

group”31 employer buyers versus “large-group”32 employer buyers) and (3) year, and,

12

using the overcharge percentage from step one, calculate as a dollar figure each

13

cohort’s per-member-per-month (“PMPM”) share of the overcharge.33

14

3. Assume that (1) health plans pass on 100 percent of any Sutter overcharge they have to

15

pay through to the premiums they charge their customers and (2) there is a lag between

16

when a health plan is overcharged and when it passes that overcharge on to its

17

customers in the form of higher premiums, and, based on those assumptions, divide

18

each cohort’s PMPM overcharge for each year between 2006 and 2015 from step two

19
20
21

Chipty Decl. – ECF Nos. 379-1 at 89 (¶ 137); Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1
(redacted version) at 90–91 (¶ 138).
30

31

22

Until 2016, small-group employers were employers with 50 or fewer employees. From 2016
onward, small-group employers are employers with 100 or fewer employees. Chipty Decl. – ECF No.
348-3 at 25 n.53 (¶ 23 n.53); Axene Decl. – ECF No. 379-2 at 5 n.3 (¶ 5 n.3).

23

32

24
25
26
27

Until 2016, large-group employers were employers with over 50 employees. From 2016 onward,
large-group employers are employers over 100 employees. Axene Decl. – ECF No. 379-2 at 5 n.3 (¶ 5
n.3).
Chipty Decl. – ECF Nos. 379-1 at 89 (¶ 137); Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1
(redacted version) at 91–94 (¶¶ 139–41). “Member-months” are a measure of how many people were
enrolled in the health-insurance product and for how long. For example, if one enrollee was enrolled in
a health-insurance product from January 1 to June 30 of a given year and another was enrolled from
January 1 to December 31 of the same year, the two of them together would represent 18 membermonths. Chipty Decl. – ECF No. 379-1 at 93 n.276 (¶ 141 n.276).
33

28
ORDER – No. 12-cv-04854-LB

14

1

by the cohort’s total PMPM premium for each year between 2008 and 2017,

2

respectively (i.e., the PMPM premium two years later), to calculate as a percentage

3

how much of the premium was part of the Sutter overcharge.34

4

4. Multiply the individual premiums each class member paid by the premium-overcharge

5

percentage from step three to calculate that class member’s share of damages.35 Where

6

an employer and an employee both paid a portion of the employee’s health-insurance

7

premium, split the damages proportionately based on the amount that the employer and

8

employee respectively contributed toward the total premium.36
For the purposes of this order, the court focuses on two aspects of the plaintiffs’ formula:

10

(1) calculating Sutter’s overcharges and (2) the assumption that health plans pass on 100 percent

11

United States District Court
Northern District of California

9

of any Sutter overcharges through to the premiums they charge their customers.37
Calculating Sutter’s Overcharges

12

4.1

13

To calculate the amount that Sutter allegedly has overcharged health plans, Dr. Chipty

14

developed a regression-analysis model.38

15

Regression analysis is a statistical methodology for determining the relationship between a

16

dependent variable and a set of explanatory variables that can influence or drive the dependent

17

variable.39 In Dr. Chipty’s model, the dependent variable is hospital prices, or, more specifically,

18

“case-mixed adjusted hospital prices.”40 The explanatory variables are factors that Dr. Chipty

19
20

Id. at 89 (¶ 137); Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1 (redacted version) at 94–96
(¶¶ 142–44).
34

Chipty Decl. – ECF Nos. 379-1 at 89 (¶ 137); Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1
(redacted version) at 97–98 (¶ 146).

21

35

22

36

23

37

24
25

Chipty Reply Decl. – ECF No. 559-1 at 44 (¶ 60).

The parties additionally raise disputes about other aspects of the plaintiffs’ formula, including how
health plans distribute higher premiums caused by alleged Sutter overcharges across their customer
base and how higher premiums are apportioned between employers and employees when both pay a
portion of the premium for an employee’s health insurance. The court focuses in this order on the two
aspects of the plaintiffs’ formula listed above.

27
28

38

Chipty Decl. – ECF No. 379-1 at 56–76 (¶¶ 80–111).

39

26

Id. at 59–60 (¶ 88).

Id. at 59 (¶ 88). The “case-mixed adjusted hospital price” for a given hospital for a given year is the
total amount of money paid to the hospital (by either the health plan or the end patient) for all
40

ORDER – No. 12-cv-04854-LB

15

1

identified as being likely to drive hospital prices, such as the hospital’s operating costs, its

2

perceived quality, the size of its system, and how many competitors it has (or variables that are

3

measures or proxies of those factors).41 Dr. Chipty states that her regression model measures how

4

much Sutter hospital prices are driven by her explanatory variables (costs, perceived quality, size,

5

or number of competitors) versus how much they are driven by the hospital’s being in the Sutter

6

system — with the latter being a measure of how much the hospital is overcharging health plans

7

by virtue of being Sutter and benefiting from Sutter’s anticompetitive practices.42
Dr. Chipty examined Sutter hospitals in the Tied Markets and in one Tying Market (Berkeley-

8
9

Oakland) and group of “benchmark” hospitals that purportedly resemble Sutter hospitals except
that (unlike Sutter) they do not impose tying or other similar anticompetitive restrictions on health

11

United States District Court
Northern District of California

10

plans.43 Dr. Chipty initially used claims data from the health plan Anthem to calculate the case-

12

mix adjusted hospital prices charged to Anthem enrollees for each of these hospitals from 2006 to

13
14
15
16
17
18

individual claims divided by the “diagnosis related group” weights associated with those claims. Id. at
63–64 (¶ 97). Dr. Chipty explains that “[d]iagnosis-related groups (‘DRGs’) categorize inpatient stays
into groups based on the diagnosis and resources necessary to treat the condition. Each DRG is
assigned a weight based on the average resources used to treat the patient. Therefore, dividing the
allowed amount, or total eligible expenses including insurance payment and patient liability, by the
DRG weight is a method of adjusting for the acuity/complexity of the inpatient stay.” Id. at 54 n.161
(¶ 77 n.161).
Id. at 64 (¶¶ 98–99). Specifically, Dr. Chipty selected as explanatory variables (1) the hospital’s
wage index, as a measure of the cost of paying salary to its employees, (2) whether the hospital is a
major teaching hospital, (3) whether the hospital has trauma facilities, (4) how many competitors the
hospital has, (5) the hospital’s patient ratings, as a proxy for aspects of its quality, (6) the number of
inpatient beds the hospital has, as a measure of the size of its system, and (7) the year, to reflect the
general inflation in hospital prices over time. Id. at 64–67 (¶ 99).
41

19
20
21
22
23
24
25
26
27

Id. at 70–71 & n.215 (¶ 104 & n.215) (“The estimated coefficients on these indicator variables
capture the differences between the actual price at each Sutter Damage Hospital in each year and the
predicted but-for price that the Sutter Damage Hospital would have negotiated for that year had it not
been in the Sutter system, but otherwise maintained all of its other attributes (e.g. wage index, teaching
status, trauma status, system size, and competitive landscape). Thus, under the model assumptions, the
estimated coefficients on the Sutter Damage Hospital x Year indicators capture the overcharges
resulting from Sutter’s challenged conduct. If, for example, the coefficient on a Sutter Damage
Hospital x Year indicator takes the value of 0.35, one would say that the Sutter hospital price in that
year is 42 percent higher as a result of the challenged conduct, accounting for the log transformation
[100 × (e0.35 – 1), where e is the base of the natural logarithm, or approximately 2.71828].”), 72 (¶ 107)
(same).
42

43

Id. at 56 (¶ 80).

28
ORDER – No. 12-cv-04854-LB

16

1

2015.44 Dr. Chipty ran her regression model twice on the Anthem-based case-mixed adjusted

2

hospital prices, (1) once on only the benchmark hospitals to calculate an “out-of-sample

3

prediction” and (2) once on a sample of both Sutter and benchmark hospitals to calculate an “in-

4

sample prediction.”45 (Dr. Chipty explained that, among other things, the out-of-sample regression

5

generates year-by-year estimates of Sutter hospital overcharges, whereas the in-sample regression

6

generates only a single average estimate of Sutter hospital overcharges.46) Based on her regression

7

model, Dr. Chipty estimated Sutter’s overcharges of Anthem by year and hospital as set forth in

8

the chart below.

9
[Opening Declaration] Exhibit 14A47
Hospital Case-Mix Adjusted Price Overcharge Percentages, Anthem
Baseline Regression 2006-2015

10

United States District Court
Northern District of California

11
12
13
14
15

Hospital
Alta Bates-Main
Alta Bates-Summit
Sutter Sacramento
CPMC-Main
CPMC-St. Luke’s
Sutter Santa Rosa
Sutter Modesto

2006

2007

2008

2009

Out-of-Sample
2010
2011

2012

2013

2014

2015

In-Sample
2006-2015

16
17

In response to criticisms by Sutter’s expert Dr. Robert Willig that Dr. Chipty’s overcharge

18

estimates were based on data only from Anthem,48 Dr. Chipty applied her regression model to

19

claims data from the health plan Blue Shield.49 After processing the data, Dr. Chipty ran her

20
21
22
23

44

The claims data contained information, at the claim level, on the inpatient hospital services received
by each of Anthem’s health-plan enrollees, including the allowed amount (the amount paid to the
hospital) for that claim. Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1 (redacted version) at 11–
14 (¶ 10), 58 (¶¶ 84–85); Chipty Decl. App’x C – ECF No. 348-3 (under seal), 379-1 (redacted
version) at 129–37.
45

Chipty Decl. – ECF No. 379-1 at 70–72 (¶¶ 102–07).

46

Id. at 70 (¶ 103), 71–72 (¶ 106).

25

47

Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1 (redacted version) at 75 (¶ 110).

26

48

Willig Decl. – ECF Nos. 446-2 (under seal), 557 (redacted version) at 124–25 (¶¶ 216–17).

24

27
28

Chipty Reply Decl. – ECF No. 559-1 at 94 (¶ 148). Because the Blue Shield claims data did not
report the total amount of money paid to hospitals in the same way as the Anthem data did and did not
contain DRG codes at all — the two components that go into Dr. Chipty’s calculation of case-mix
49

ORDER – No. 12-cv-04854-LB

17

1

regression model twice on the Blue Shield-based case-mixed adjusted hospital prices to calculate

2

an “out-of-sample” and an “in-sample” prediction.50 Based on the Blue Shield claims data and her

3

regression model, Dr. Chipty estimated Sutter’s overcharges at its hospitals by year and hospital,

4

as set forth in the chart below.

5
[Reply Declaration] Exhibit 1751
Hospital Case-Mix Adjusted Price Overcharge Percentages, Blue Shield
Baseline Regression 2006-2015

6
7
8
9
10

United States District Court
Northern District of California

11

Out-of-Sample
Hospital
2006
2007
2008
2009
2010
2011
2012
2013
2014
Alta Bates-Main
53% **
44% *
58% **
42% *
40% *
21%
42% *
17%
144% ***
Alta Bates-Summit 135% ** 127% ** 158% *** 165% *** 172% *** 142% *** 140% *** 126% **
22% **
Sutter Sacramento 95% *** 37%
95% *** 114% *** 83% *** 99% *** 84% *** 65% **
39% *
CPMC-Main
25% **
11%
60% *** 44% *** 76% *** 78% *** 69% *** 57% *** 53% **
CPMC-St. Luke’s
33%
16%
70% *** 40% **
45% ** 109% *** 31%
30%
44% **
Sutter Santa Rosa
1%
-22% **
21% *
7%
-6%
-15%
-8%
6%
-4% **
Sutter Modesto
38% *** 57% *** 50% *** 29% **
24% *
26% **
28% **
31% **
19% **

In-Sample
2015
2006-2015
88% *
50% **
3%
96% ***
50% **
75% ***
40% **
50% ***
39% *
41% **
-6%
-3%
34% **
33% ***

12
13

Dr. Chipty has not conducted similar regression analyses on claims data from the health plans

14

Aetna, Health Net, or UnitedHealthcare. She testified at her deposition that she has not offered an

15

overcharge model for class members who bought insurance through those health plans.52 She

16

stated that she had two approaches for addressing those three health plans: (1) if she had access to

17

the same types of data for each health plan, she could run her regression models on each health

18

plan separately, or (2) she could try to extrapolate her findings from some health plans to other

19

health plans.53 She testified, “I have not reached an opinion as to whether I have a preferred

20
21
22

adjusted hospital prices — Dr. Chipty had to process the Blue Shield data somewhat differently than
she processed the Anthem data. Chipty Reply Decl. – ECF Nos. 462-3 (under seal), 559-1 (redacted
version) at 94 (¶ 149).

23

50

Chipty Reply Decl. – ECF No. 559-1 at 95 (¶ 150).

24

51

Chipty Reply Decl. – ECF Nos. 462-3 (under seal), 559-1 (redacted version) at 96 (¶ 151).

25
26
27

Chipty Dep. – ECF No. 531-4 (under seal) at 64–65 (pp. 666–67) (“Q. Have you offered an
overcharge model for class members who access through Aetna? A. No, I have not. Q. Have you
offered an overcharge model for class members who access through United? A. No, I have not.
Q. Have you offered an overcharge model for class members who accessed through Health Net?
A. No, I have not.”).
52

53

Id. at 70–71 (pp. 672–73).

28
ORDER – No. 12-cv-04854-LB

18

1

method between those two as yet.”54 With respect to the first approach, she stated that she did not

2

fully know whether she had sufficient data from Aetna, Health Net, or UnitedHealthcare to run her

3

regression analyses and would not know until she tried to run the analyses.55 With respect to the

4

second approach, she stated that she has not reached an opinion as to whether it would be

5

reasonable to use extrapolation as an overcharge-damages model.56

6

Sutter noted that Dr. Chipty’s regression model, when applied to Blue Shield claims data,

7

shows an undercharge (as opposed to an overcharge) at one of Sutter’s hospitals in a Tied Market,

8

Sutter Santa Rosa. When asked about it during her deposition, Dr. Chipty stated that this reflects

9

that her regression model has “an inability to measure the overcharge for Sutter-Santa Rosa.”57 Dr.

10

United States District Court
Northern District of California

11
12
13
14
15
16
17

54

Id. at 71 (p. 673). She began by saying that she would apply her regression-analysis model to Aetna,
Health Net, and UnitedHealthcare but then shifted to saying that she might use extrapolation instead.
Id. at 69 (p. 671) (“Q. So is it your opinion, Dr. Chipty, that the methodology for estimating aggregate
overcharge amounts that you use for Anthem and Blue Shield, that’s the methodology that you’re
proposing to estimate aggregate damages amounts for Aetna, United and Health Net? A. I would — I
would imagine using the same model to estimate those damages. Except to the point if I were to do —
do my continued work on their claims data and their specific circumstances, it’s possible that I reached
the conclusion that it’s reasonable to extrapolate from the experiences of Anthem and Blue Shield, but
I haven’t completed my work on that front yet.”).
Id. at 71–74 (pp. 673–76) (“Q. And have you reached an opinion on whether you have sufficient
data from Aetna, United and Health Net to apply the overcharge methodology that you use for Anthem
and Blue Shield for those health plans? A. Okay. So I may mess this up, in which case I’ll come back
and tell you. But I believe for United, we have the data that we need to move forward. By that I mean
we have adequate claims data and adequate premium data. I believe that’s true, but I would have to
talk to the team to make sure I’ve got that right. And for Aetna, I believe we’re still waiting on some of
the data. I believe we’re waiting on their premium data. On Health Net, there’s been a large production
of data from Health Net. It takes a different form, and I don’t think my team has completed its
assessment of the adequacy as yet of the Health Net data, but we’re working on that. . . . Q. . . . . Your
understanding, subject to verification from your staff, is that you and your staff have sufficient data
from United to apply the same overcharge methodology that you used for Anthem and Blue Shield; is
that correct? A. That’s correct. That’s my understanding. But I should also add that until we undertake
the exercise, we won’t know fully whether we do or we don’t. We actually — one never knows until
one gets deep into a problem. But, yes, my expectation is that we do and that we will look at that.
Q. And then subject to verification with respect to Aetna and Health Net, you and your staff are still
examining the data that has been received to determine whether there is sufficient data to use the same
methodology, and you noted that Aetna may not have the premium data to date? A. That’s correct. I
think that’s correct.”).
55

18
19
20
21
22
23
24
25
26

56

Id. at 71 (p. 673).

57

27

Id. at 111 (p. 713).

28
ORDER – No. 12-cv-04854-LB

19

1

Chipty did not propose an alternative model for estimating the overcharge percentage for Sutter

2

Santa Rosa for Blue Shield class members.58

3

4.2

Assuming Health Plans “Pass On” 100 Percent of Any Sutter Overcharges They
Have to Pay Through to the Premiums That They Charge Their Customers

4
Dr. Chipty’s formula for converting the amount that Sutter allegedly overcharged health plans

5

to the amount that class members were overcharged for their insurance premiums relies on an

7

assumption that health plans “pass on” 100 percent of any Sutter overcharges they have to pay

8

through to the premiums that they charge their customers.59 In other words, her formula assumes

9

that if Sutter charged, say, Anthem $100,000 more than it would have been able to charge in the

10

but-for world, Anthem charged its customers (i.e., class members) $100,000 more in premiums

11

United States District Court
Northern District of California

6

than it would have charged in the but-for world as well.
Dr. Chipty stated in her report that health plans may actually pass on less than 100 percent of

12
13

hospital overcharges through to increased premiums to their customers. She explained in her

14

report:
Basic economic principles indicate that health plans will pass through at least some
portion of medical cost increases on to their consumers in the form of higher
premiums. How much is passed through depends on market conditions. Given
competitive conditions, a health plan may decide to absorb some portion of a
medical cost increase, by cutting into [its] margins, to remain price competitive
with other health plans who do not experience similar cost increases.60

15
16
17
18
19

As she testified in her deposition, “the different health plans may have different administrative

20

costs, the different health plans may have different medical loss ratios or profit margins that

21

they’re targeting.”61
Among other things, Dr. Chipty acknowledges that health plans like Anthem and Blue Shield

22
23

might decide not to pass on 100 percent of hospital overcharges because of the competitive

24
25

58

Id.

See Chipty Decl. – ECF No. 379-1 at 96 (¶ 144) (assuming for the purposes of her formula that the
pass-through ratio is equal to one, i.e., 100 percent).
59

26

60

Id. at 77 (¶ 112).

61

27

Chipty Dep. – ECF No. 415-1 (under seal) at 113 (p. 386).

28
ORDER – No. 12-cv-04854-LB

20

1

pressures they face from Kaiser Permanente.62 Kaiser is the largest health system in Northern

2

California (and California generally).63 Kaiser is a “closed” health system consisting of a Kaiser

3

health plan and Kaiser hospitals and other medical providers, where Kaiser health-plan enrollees

4

can receive healthcare only from Kaiser hospitals and medical providers (other than in

5

emergencies).64 Because Kaiser is a closed system that maintains its own network of hospitals,

6

Kaiser’s health plan does not contract with Sutter to include Sutter hospitals within its network

7

and thus is not subject to Sutter’s alleged systemwide contracting with its all-or-nothing, anti-

8

steering, and penalty-rate provisions.65 Kaiser’s health plan competes with health plans like

9

Anthem and Blue Shield in the sale of commercial health insurance to individuals and
employers.66 Dr. Chipty stated in her report that “[t]o the extent health plans like Anthem and Blue

11

United States District Court
Northern District of California

10

Shield would absorb cost increases to compete with Kaiser, pass-through may be less than 100

12

percent.”67 She explained in her deposition:
A. . . . . So the way I understand the premiums work is, at least actuarially, they’re
built to cover health care expenses. And actuarially, if health care expenses go u[p],
premiums will go up, but there’s also some room to adjust profit margin. So [my
report] is recognizing that a health plan might decide to absorb a cost increase to
better compete with Kaiser. . . .

13
14
15
16

Q. Okay. And one way that you’ve described is to — rather than passing through
100 percent, is pass through a portion of it and then either lower the profit margin
or perhaps cut expenses and maintain their profit margin?

17
18

A. Perhaps.68

19
20
21
22
62

Chipty Decl. – ECF No. 379-1 at 77 (¶ 112).

23

63

Id. at 5 (¶ 4), 18 (¶ 16).

24

64

Id. at 18 (¶ 16).

25

See id. (¶ 17) (“[N]o patient would have both Kaiser and Sutter hospitals in his or her provider
network.”).
65

27

66

Id. at 19 (¶ 17); Chipty Dep. – ECF No. 415-1 (under seal) at 129 (p. 450).

67

Chipty Decl. – ECF No. 379-1 at 77 (¶ 112).

68

26

Chipty Dep. – ECF No. 415-1 (under seal) at 134 (pp. 472–73).

28
ORDER – No. 12-cv-04854-LB

21

Sutter agrees and submits evidence (unrebutted by the plaintiffs) that supports Dr. Chipty’s

1
2

assessment that health plans might decide to not pass on their costs (including hospital costs)

3

through to the premiums they charge their customers, in order to remain competitive vis-à-vis rival

4

health plans. To take one example, in 2011,

5

one of its customers, Woodruff Sayer, and originally proposed

was discussing insurance rates for 2012 for

.69 Woodruff Sayer responded that

6

.70 In response,

7
8

71

9

To take

another example, in 2014, Blue Shield was discussing insurance rates for 2015 for one of its

11
United States District Court
Northern District of California

10

customers, the City and County of San Francisco.72 Blue Shield acknowledged that it was in a

12

“difficult position” given that competitor health plans Kaiser and UnitedHealthcare had both

13

proposed rate decreases.73 In response, Blue Shield offered the City a “rate pass” (a 0 percent

14

increase in premium rates) for 2015.74
Dr. Chipty’s damages-calculation formula nonetheless assumes that the rate at which health

15
16

plans pass on Sutter’s alleged overcharges through to their customers’ premiums will be 100

17

percent.75 Dr. Chipty stated that one of the things she relies on to assume that (1) health plans pass

18

on 100 percent of any Sutter overcharges and (2) the passthrough rate is the same across all health

19

plans and class members, is a slide from a PowerPoint presentation drafted by Sutter’s former

20
21
22
email chain – ECF Nos. 415-3 (under seal), 445-6 (redacted version) at 10
(WSAW003956).
69

23

70

Id. at 6 (WSAW003952).

71

Id. at 5 (WSAW003951).

25

72

Blue Shield email chain – ECF No. 445-10 at 274–75 (AON0009800–01).

26

73

Id. at 274 (AON0009800).

74

Id.

75

Chipty Decl. – ECF No. 379-1 at 96 (¶ 144).

24

27
28

ORDER – No. 12-cv-04854-LB

22

1

CFO Bob Reed.76

2
77

3
4

presentation discussed a hypothetical scenario where Sutter reduced its prices by

5

The

then examined how that might cause health plans to reduce their premiums
, assuming the health plans passed on 100 percent of Sutter’s price reductions to their

6
7

and

customers.78

8
9
10
79

United States District Court
Northern District of California

11

Dr. Chipty also cited other documents and analyses that she

12

said indicates that health plans try to set premiums at a level that covers all their expenses and that

13

when their costs increase, their premiums do as well.80
Dr. Chipty also conducted a regression analysis that she claims supports her formula’s

14
15

assumption that health plans pass on 100 percent of alleged overcharges through to their

16

customers. For her regression, Dr. Chipty used health plans’ PMPM health-care costs and

17

premiums between 2012 and 2015 for small-group-employer insurance buyers, as disclosed by the

18

plans in their annual Uniform Rate Review Template filings.81 (She did not include data from

19
20
21
22

Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1 (redacted) at 79–81 (¶¶ 118–19); Chipty Reply
Decl. – ECF No. 559-1 at 55 (¶ 80); Chipty Dep. – ECF No. 415-1 (under seal) at 115 (pp. 395–97),
131 (p. 459).
76

Sutter Strategy Session PowerPoint Presentation – ECF Nos. 349 at 130 (under seal), 379-4 at 131
(redacted version) (DEF001993774).
77

23

78

Id.

79

Id.

80

Chipty Decl. – ECF Nos. 348-3 (under seal), 379-1 (redacted version) at 77–82 (¶¶ 114–22).

81

24

Chipty Decl. – ECF No. 379-1 at 83 (¶ 123).

25
26
27
28
ORDER – No. 12-cv-04854-LB

23

1

individual or large-group-employer buyers.82) In her analysis, her dependent variable was the

2

natural logarithm of PMPM premium in a given year, and her sole explanatory variable was the

3

natural logarithm of PMPM cost in that year.83 She acknowledged that this was “a very simple

4

regression model.”84 Her regression analysis calculated an estimated passthrough coefficient for

5

small-group-employer insurance of approximately 0.9 (where a coefficient of 1 indicates 100-

6

percent passthrough85), as set forth in the chart below.

7
[Opening Declaration] Exhibit 1786
Pass-Through Analysis for Small Groups
Dependent Variable = Log ($ Premium)

8
9
10

Explanatory Variable

Estimates

Log(Cost)

0.914***
(0.0874)

12

Includes Health Plan
Fixed Effects

Yes

13

Observations

41

R-squared

0.965

United States District Court
Northern District of California

11

14

Notes:
1. Standard errors shown in parentheses below coefficient estimate.
The estimate is statistically significant at 1 percent significance level.
2. Coefficient estimated using ordinary least squares.

15
16

In response to criticisms by Sutter’s expert Dr. Willig that Dr. Chipty’s analysis did not

17
18

examine whether passthrough rates vary across different health plans, across health-plan business

19

lines (individual buyers versus small-group employer buyers versus large-group employer buyers),

20

or across different health-insurance products, or whether passthrough rates vary based on how

21

much competition a health plan faces (including competition from Kaiser),87 Dr. Chipty ran

22

additional regression analyses, as set forth in the charts below.

23
24

Id.; Chipty Dep. – ECF No. 415-1 at 116 (p. 399) (“[R]emember this is just for small group. It’s
exactly how Exhibit 17 is laid out.”).
82

83

Chipty Decl. – ECF No. 379-1 at 83 (¶ 125).

84

Chipty Dep. – ECF No. 415-1 at 114 (p. 391).

85

Chipty Decl. – ECF No. 379-1 at 83 (¶ 125).

27

86

Id. at 84 (¶ 125).

28

87

Willig Decl. – ECF No. 557 at 46 (¶ 67), 59–60 (¶ 98).

25
26

ORDER – No. 12-cv-04854-LB

24

1

[Reply Declaration] Exhibit 788
Dr. Willig’s Table 1, Showing More Information

2
Descriptions

Aetna

3

Coefficient

0.302

4

Standard Error

(0.228)

95% CI

-0.164, 0.768

5
6
7
8

(0.233)

Blue Shield

Health Net

1.042***

1.054***

(0.091)

(0.164)

0.857, 1.227

***

R-squared

UHC

0.719, 1.389

***
(0.216)

0.976

Plan/Year Observation

9

8

7

8

9

Notes:
1. Asterisks *** indicates statistical significance at the one percent level.
2. The data upon which the model relies are aggregated at the year-level for each health plan.

9

[Reply Declaration] Exhibit 889
Summary of Evidence Indicating Pass-Through Is at or Near 100 Percent

10
Health Plan
[1]

12
13

Source
[2]

Costs Included
[3]

Product
Included
[4]

Line-of-Business Included
[5]

Kaiser
Non-Kaiser
[6]

Coefficient
[7]

95 Percent
Confidence Interval
[8]

Observations
[9]

Aetna
Blue Cross
Blue Shield
Health Net
UHC

11
United States District Court
Northern District of California

Blue Cross

[1]

All Medical Costs

All

Small Group

Combined

0 914***

0 748, 1 079

41

-0 164, 0 768

Small Group

Combined

0 302
***
1 042***
1 054***
***

9
8
7
8
9

14

Aetna
Blue Cross
Blue Shield
Health Net
UHC

[1]

All Medical Costs

All

15

Blue Shield

[2]

All Medical Costs

All

Large Group

Combined

0 954***

0 800, 1 109

9

Combined

0 993***

0 816, 1 171

9

0 857, 1 227
0 719, 1 389

Blue Shield

All Medical Costs

All

Large Group:
100 - 500 Insured Employees

Combined

0 898***

0 723, 1 073

9

[2]

All Medical Costs

All

Large Group:
501 ≤ Insured Employees

Combined

0 988***

0 810, 1 165

9

[2]

All Medical Costs

All

Large Group:
2,000+ Insured Employees

Combined

1 002***

0 771, 1 233

9

[3]

Inpatient Spend

All

Individual

Combined

***

10

Anthem

[3]

Inpatient Spend

All

Large Group

Combined

***

10

Anthem

[3]

Inpatient Spend

All

Large Group

Non-Kaiser

***

10

Anthem

20

[2]

Anthem

19

All

Blue Shield

18

All Medical Costs

Blue Shield

17

[2]

Blue Shield

16

Large Group:
101 ≤ Insured Employees

[3]

Inpatient Spend

All

Large Group

Kaiser

***

10

***
***

20

Anthem

Inpatient Spend

All

Large Group

[3]

Inpatient Spend

All

Individual

Non-Kaiser

***

10

Anthem

22

[3]

Anthem

21

Non-Kaiser
Kaiser

[3]

Inpatient Spend

All

Individual

Kaiser

***

10

All

Individual

Non-Kaiser
Kaiser

***
***

20

Anthem

Inpatient Spend

[3]

Inpatient Spend

PPO

Large Group

Combined

***

10

Anthem

24

[3]

Anthem

23

[3]

Inpatient Spend

HMO

Large Group

Combined

***

10

Note: Asterisks *** indicates statistical significance at the one percent level.

25
26
27
28

88

Chipty Reply Decl. – ECF Nos. 462-3 (under seal), 559-1 (redacted version) at 49 (¶ 69).

Id. at 51–52 (¶ 76) (“Exhibit 8 summarizes pass-through analyses separately for Anthem and Blue
Shield data, by health plan, by line-of-business, by Kaiser’s presence, and by product. Column [1]
identifies the health plan whose data was analyzed. Column [4] identifies the product (e.g., all, HMO,
89

ORDER – No. 12-cv-04854-LB

25

1

Dr. Chipty maintains that these analyses support her decision to assume a uniform 100-

2

percent-passthrough rate across all health plans, health products, and class members in her

3

damages calculations. She stated that the health-plan-by-health-plan analysis (Exhibit 7, above),

4

which calculated passthrough coefficients of 0.32,

5

health plans at issue in this case, “supports the view, using Dr. Willig’s own results, that pass-

6

through rates for four of the five health plans are similar to each other and equal to 100 percent”90

7

and that the fifth health plan’s coefficient rate — “0.3 (or a 30 percent pass-through rate)” — is

8

“not statistically different from a pass-through rate as high as 77 percent.”91 She similarly said that

9

her analyses disaggregating health plans, business lines, and products (Exhibit 8, above), which

, 1.042, 1.054, and

for the five

generated coefficients ranging from 0.302 to

11
United States District Court
Northern District of California

10

, support her using a uniform 100-percent-

passthrough rate in her damages calculations.92 She acknowledged that for many of her analyses,

12

the upper bound of the 95-percent confidence interval around the passthrough-rate estimate93 fell

13

below 1 (i.e., below a 100-percent-passthrough rate) interval.94 She nonetheless maintains that

14
15
16
17
18
19

PPO) considered. Column [5] identifies the line-of-business (e.g., individual, small group, and large
group). Column [6] identifies whether pass-through rates in Kaiser and non-Kaiser areas were
estimated separately or jointly. Column [7] shows the estimated pass-through coefficient. Column [8]
describes the 95 percent confidence interval of the estimated passthrough rates, and Column [9] shows
the regression observations, where each observation reflects the premium and claims experience of
many. The first row shows results of the model presented in my Class Declaration. The second row
shows the results of Dr. Willig’s model. Statistically, the estimated pass-through rates are at or near
100 percent, for virtually all models. The lowest estimated pass-through rate comes from Dr. Willig’s
small group analysis, for Aetna.”).

21
22
23
24
25
26
27

90

Chipty Reply Decl. – ECF No. 559-1 at 48 (¶ 69).

91

Id. at 46 (¶ 66).

92

20

Chipty Reply Decl. – ECF Nos. 462-3 (under seal), 559-1 (redacted) at 50–54 (¶¶ 72–78).

Dr. Chipty explains that “a ‘confidence interval’ provides a range of values within which the true
population value is likely to lie.” Chipty Decl. – ECF No. 379-1 at 62 (¶ 94). “A 95 percent two-sided
confidence interval is an interval around the sample estimate, constructed based on the standard error,
that would contain the true population value 95 percent of the time if the sampling procedure were to
be repeated infinitely.” Id. at 62 n.191 (¶ 94 n.191).
93

Chipty Reply Decl. – ECF Nos. 462-3 (under seal), 559-1 (redacted version) at 50–51 nn.167–69
(¶¶ 72 n.167, 74 n.168–69) (noting that in row 8 of Exhibit 8, the upper bound of the 95-percent
confidence interval around the passthrough estimate is 0.9; in row 10, the upper bound of the 95percent confidence interval around the passthrough estimate is 0.87; in row 13, the upper bound of the
95-percent confidence interval around the passthrough estimate is 0.95; and in row 14, the upper
bound of the 95-percent confidence interval around the passthrough estimate is 0.92).
94

28
ORDER – No. 12-cv-04854-LB

26

1

these results are “near one [100 percent]” or “close to one [100 percent].”95 She does not cite any

2

sources to explain what being “near” or “close to” 1 means or how that supports her decision to

3

assume a passthrough rate of 100 percent.96

4

Dr. Chipty acknowledged in her deposition that she has never conducted in any other contexts

5

the passthrough regression analyses she is running here.97 She has not studied passthrough rates in

6

the context of medical premiums and has never offered an expert opinion on passthrough rates

7

outside of the context of this case.98

8
ANALYSIS

9

Class actions are governed by Federal Rule of Civil Procedure 23. A party seeking to certify a

10

United States District Court
Northern District of California

11

class must prove that all the prerequisites of Rule 23(a) are met, as well as those of at least one

12

subsection of Rule 23(b) (the relevant subsections here are (b)(2) and (b)(3)).
The following are the prerequisites of Rule 23(a):

13
14

1. the class is so numerous that joinder of all members is impracticable;

15

2. there are questions of law or fact common to the class;

16

3. the claims or defenses of the representative parties are typical of the claims or defenses
of the class; and

17

4. the representative parties will fairly and adequately protect the interests of the class.

18

A court may certify a class under Rule 23(b)(3) if “the court finds that the questions of law or

19
20

fact common to class members predominate over any questions affecting only individual

21
22

95

Id. at 50–51 (¶¶ 72, 74).

Dr. Chipty describes some of her results as being “not statistically different from one.” Chipty Reply
Decl. – ECF No. 559-1 at 50 (¶¶ 72, 74). She explains what she means when she says results are “not
statistically different from one”: that a value of 1 (a 100-percent-passthrough rate) is within the 95percent confidence interval. See Chipty Decl. – ECF No. 379-1 at 62 (¶ 94). She does not explain what
a result’s being “near one” or “close to one” means. It necessarily must mean that it is “statistically
different” from 1 (a 100-percent-passthrough rate), given that 1 falls above the 95-percent confidence
interval.
96

23
24
25
26

97

27

98

28

Chipty Dep. – ECF No. 415-1 (under seal) at 110 (pp. 376–77).

Id. at 110–11 (pp. 377–80). In addition to Dr. Chipty, the plaintiffs proffered another expert, David
Axene, but Mr. Axene did not conduct any analysis to calculate passthrough rates of hospital charges
to health-plan premiums. Axene Dep. – ECF No. 415-1 (under seal) at 20 (pp. 68–69), 29 (p. 105).
ORDER – No. 12-cv-04854-LB

27

1

members, and that a class action is superior to other available methods for fairly and efficiently

2

adjudicating the controversy.” Fed. R. Civ. P. 23(b)(3). A court may certify a class under Rule

3

23(b)(2) for injunctive or declaratory relief (i.e., not for money damages) if “the party opposing

4

the class has acted or refused to act on grounds that apply generally to the class, so that final

5

injunctive relief or corresponding declaratory relief is appropriate respecting the class as a

6

whole[.]” Fed. R. Civ. P. 23(b)(2).

7

“[P]laintiffs wishing to proceed through a class action must actually prove — not simply plead

8

— that their proposed class satisfies each requirement of Rule 23, including (if applicable) the

9

predominance requirement of Rule 23(b)(3).” Halliburton Co. v. Erica P. John Fund, Inc., 573
U.S. 258, 275 (2014) (emphasis in original) (citing Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338,

11

United States District Court
Northern District of California

10

350–51 (2011); Comcast Corp. v. Behrend, 569 U.S. 27, 32–33 (2013)). “[C]ertification is proper

12

only if ‘the trial court is satisfied, after a rigorous analysis, that the prerequisites of Rule 23[] have

13

been satisfied.’” Comcast, 569 U.S. at 33 (quoting Wal-Mart, 564 U.S. at 350–51). “Such an

14

analysis will frequently entail ‘overlap with the merits of the plaintiff’s underlying claim.’” Id. at

15

33–34 (quoting Wal-Mart, 564 U.S. at 351). “That is so because the ‘class determination generally

16

involves considerations that are enmeshed in the factual and legal issues comprising the plaintiff’s

17

cause of action.’” Id. at 34 (quoting Wal-Mart, 564 U.S. at 351). Still, “Rule 23 grants courts no

18

license to engage in free-ranging merits inquiries at the certification stage.” Amgen Inc. v. Conn.

19

Ret. Plans and Tr. Funds, 568 U.S. 455, 466 (2013). “Merits questions may be considered to the

20

extent — but only to the extent — that they are relevant to determining whether the Rule 23

21

prerequisites for class certification are satisfied.” Id. (citing Wal-Mart, 564 U.S. at 351 n.6).

22
23

1. Rule 23(a) Prerequisites
Numerosity — Rule 23(a)(1)

24

1.1

25

Rule 23(a)(1) requires that “the class [be] so numerous that joinder of all members is

26

impracticable.” There is no absolute minimum class size for establishing numerosity, but courts

27

have held that classes as small as 40 satisfy the numerosity requirement. See, e.g., In re Qualcomm

28

Antitrust Litig., 328 F.R.D. 280, 294 (N.D. Cal. 2018) (citing Twegbe v. Pharmaca Integrative
ORDER – No. 12-cv-04854-LB

28

1

Pharmacy, Inc., No. CV 12-5080 CRB, 2013 WL 3802807, at *3 (N.D. Cal. July 17, 2013)),

2

appeal docketed sub nom. Stromberg v. Qualcomm Inc., No. 18-80135 (9th Cir. filed Oct. 11,

3

2018); In re Lidoderm Antitrust Litig., No. 14-md-02521-WHO, 2017 WL 679367, at *13 (N.D.

4

Cal. Feb. 21, 2017) (citing Daniel R. Coquillette et al., 5 Moore’s Federal Practice – Civil § 23.22

5

(2016)). The plaintiffs reasonably estimate that the proposed class contains at least hundreds of

6

thousands of members.99 Sutter does not dispute numerosity.100 The court finds that the plaintiffs

7

have satisfied Rule 23(a)(1)’s numerosity requirement.
Commonality — Rule 23(a)(2)

1.2

9

Rule 23(a)(2) requires that “there [be] questions of law or fact common to the class.” “What

10

matters to class certification is not the raising of common ‘questions’ — even in droves — but,

11

United States District Court
Northern District of California

8

rather the capacity of a classwide proceeding to generate common answers apt to drive the

12

resolution of the litigation.” Torres v. Mercer Canyons Inc., 835 F.3d 1125, 1133 (9th Cir. 2016)

13

(internal ellipsis and some internal quotation marks omitted) (quoting Wal-Mart, 564 U.S. at 350).

14

“To satisfy Rule 23(a)(2) commonality, ‘even a single common question will do.’” Id. (some

15

internal quotation marks omitted) (quoting Wal-Mart, 564 U.S. at 359). Courts have held that

16

“‘antitrust liability alone constitutes a common question.’” In re Qualcomm, 328 F.R.D. at 294

17

(internal brackets omitted) (quoting In re High-Tech Emp. Antitrust Litig., 985 F. Supp. 2d 1167,

18

1180 (N.D. Cal. 2013)); accord, e.g., In re Lithium Ion Batteries Antitrust Litig., No. 13-MD-2420

19

YGR, 2017 WL 1391491, at *3 (N.D. Cal. Apr. 12, 2017) (In re Lithium Ion Batteries I) (“‘[T]he

20

very nature of a conspiracy antitrust action compels a finding that common questions of law and

21

fact exist.’”) (quoting In re Dynamic Random Access Memory (DRAM) Antitrust Litig., No. M 02-

22

1486 PJH, 2006 WL 1530166, at *3 (N.D. Cal. June 5, 2006)). The plaintiffs have raised common

23

questions about whether Sutter’s systemwide contracting with its all-or-nothing, anti-steering, and

24

penalty-rate provisions is anticompetitive that would generate common answers apt to drive

25
26
27

Chipty Decl. – ECF No. 379-1 at 31 (¶ 30) (“[T]here could be between 430,000 and 680,000 Class
Members in any given year.”).
99

See Def. MCC Opp’n – ECF No. 445 at 9, 30–31 (disputing whether the plaintiffs have satisfied
Rule 23(a)’s typicality and adequacy requirements but not numerosity or commonality).
100

28
ORDER – No. 12-cv-04854-LB

29

1

resolution of the litigation. Sutter does not dispute commonality.101 The court finds that the

2

plaintiffs have satisfied Rule 23(a)(2)’s commonality requirement.
Typicality — Rule 23(a)(3)

3

1.3

4

Rule 23(a)(3) requires that “the claims or defenses of the representative parties [be] typical of

5

the claims or defenses of the class.” “The test of typicality serves to ensure that ‘the interest of the

6

named representative aligns with the interests of the class.’” Torres, 835 F.3d at 1141 (quoting

7

Hanon v. Dataproducts Corp., 976 F.2d 497, 508 (9th Cir. 1992)). “‘Under the Rule’s permissive

8

standards, representative claims are ‘typical’ if they are reasonably coextensive with those of

9

absent class members; they need not be substantially identical.’” Id. (quoting Parsons v. Ryan, 754
F.3d 657, 685 (9th Cir. 2014)). “In this context, ‘typicality refers to the nature of the claim or

11

United States District Court
Northern District of California

10

defense and not to the specific facts from which it arose or the relief sought.’” Id. (internal ellipsis

12

omitted) (quoting Parsons, 754 F.3d at 685). “Measures of typicality include ‘whether other

13

members have the same or similar injury, whether the action is based on conduct which is not

14

unique to the named plaintiffs, and whether other class members have been injured by the same

15

course of conduct.’” Id. (quoting Hanon, 976 F.2d at 508). Put another way, “[t]ypicality is

16

present ‘when each class member’s claim arises from the same course of events, and each class

17

member makes similar legal arguments to prove the defendants’ liability.’” In re Qualcomm, 328

18

F.R.D. at 295 (quoting Rodriguez v. Hayes, 591 F.3d 1105, 1122 (9th Cir. 2010)). “Thus, ‘in

19

antitrust cases, typicality usually will be established by plaintiffs and all class members alleging

20

the same antitrust violations by defendants.’” Id. (internal brackets and some internal quotation

21

marks omitted) (quoting In re High-Tech Emp., 985 F. Supp. 2d at 1181).

22

The named plaintiffs’ claims are typical of the class’s claims. The conduct they challenge —

23

Sutter’s systemwide contracting with its all-or-nothing, anti-steering, and penalty-rate provisions

24

— is not unique to any plaintiff. Rather, Sutter engaged in this same conduct with the five health

25

plans at issue and allegedly charged all five health plans supra-competitive rates as a result, which

26
27
See Def. MCC Opp’n – ECF No. 445 at 9, 30–31 (disputing whether the plaintiffs have satisfied
Rule 23(a)’s typicality and adequacy requirements but not numerosity or commonality).
101

28

ORDER – No. 12-cv-04854-LB

30

1

the health plans then allegedly passed on to their customers (the class members). The alleged

2

injuries to the named plaintiffs and to the class arise from the same course of conduct.
Sutter argues that the named plaintiffs’ claims are not typical because the named plaintiffs are

3

two small employers and four individuals (two of whom worked for the same employer) and do

5

not include any large-group employers or any persons who bought an individual health-insurance

6

policy.102 Sutter argues that the plaintiffs do not have any personal stake or interest in establishing

7

that health plans passed on supra-competitive hospital costs through to large-group employers

8

despite the presence of individually negotiated premiums.103 But under Rule 23’s permissive

9

standards, the plaintiffs’ claims need only be “reasonably coextensive” with absent class

10

members’ claims; they need not be substantively identical. Torres, 835 F.3d at 1141. The

11

United States District Court
Northern District of California

4

overarching gravamen of the plaintiffs’ claims is Sutter’s alleged anticompetitive tying activity.

12

While differences in how health plans may have passed on Sutter’s alleged overcharges through to

13

their customers may go to Rule 23(b)(3)’s predominance requirement, it is not sufficient to defeat

14

typicality. Cf. In re Lithium Ion Batteries I, 2017 WL 1391491, at *7–8 (rejecting argument that

15

claims of individuals who bought battery products at non-negotiable prices were atypical of claims

16

of large institutional buyers who bought products in bulk and could negotiate prices because “the

17

overarching [antitrust] price-fixing scheme is the gravamen of the claim, regardless of the type of

18

product purchased, the quantity, the purchasing procedures, or the price paid”) (citing In re Static

19

Random Access Memory (SRAM) Antitrust Litig., 264 F.R.D. 603, 609 (N.D. Cal. 2009)); In re

20

Online DVD Antitrust Litig., No. M 09-2029 PJH, 2010 WL 5396064, at *4 (N.D. Cal. Dec. 23,

21

2010) (In re Online DVD I) (“The named plaintiffs’ claims are typical of the class because for all

22

claims, proof of the alleged violations in question will depend on proof of violation by defendants,

23

and not on the individual positioning of each plaintiff.”) (emphasis in original).104

24
25
26

Def. MCC Opp’n – ECF No. 445 at 31. Sutter also argues that the plaintiffs and class members
have conflicts of interest, id. at 30–31, which the court will address in the next section in connection
with Rule 23’s adequacy requirement.
102

103

27
28

Id. at 31.

It is worth noting that the “individually negotiated premiums” that Sutter discusses in its opposition
refers to large-group employers’ negotiations with health plans, not with Sutter. Sutter does not
104

ORDER – No. 12-cv-04854-LB

31

1

The court finds that the plaintiffs have satisfied Rule 23(a)(3)’s typicality requirement.105

2

1.4

3

Rule 23(a)(4) requires that “the representative parties [] fairly and adequately protect the

Adequacy — Rule 23(a)(4)

4

interests of the class.” “This adequacy requirement . . . ‘serves to uncover conflicts of interest

5

between named parties and the class they seek to represent’ as well as the ‘competency and

6

conflicts of class counsel.’” Espinosa v. Ahearn (In re Hyundai and Kia Fuel Econ. Litig.), 926

7

F.3d 539, 566 (9th Cir. 2019) (en banc) (quoting Amchem Prods., Inc. v. Windsor, 521 U.S. 591,

8

625, 626 n.20 (1997)). “To determine legal adequacy, [courts] resolve two questions: ‘(1) do the

9

named plaintiffs and their counsel have any conflicts of interest with other class members and

10

United States District Court
Northern District of California

11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

contend that large-group employers had any more ability to negotiate with Sutter over prices or
premiums than individuals or small-group employers did. The fact that large-group employers might
have engaged in negotiations with health plans does not render plaintiffs’ claims — which are against
Sutter, not the health plans — atypical of the claims of the class as a whole. Cf. In re Optical Disk
Drive Antitrust Litig., 303 F.R.D. 311, 317–18 (N.D. Cal. 2014) (in price-fixing antitrust case, finding
that typicality was not satisfied based on differences between large institutional class members and
small class members where the large institutional class members engaged in direct price negotiations
with the defendants — i.e., the parties that were actually doing the price-fixing — and small class
members did not) (citing In re Graphics Processing Units Antitrust Litig., 253 F.R.D. 478, 489–90
(N.D. Cal. 2008)).
105

Sutter cites Burkhead v. Louisville Gas & Electric Co., 250 F.R.D. 287 (W.D. Ky. 2008), and
Major v. Ocean Spray Cranberries, Inc., No. 5:12-CV-03067 EJD, 2013 WL 2558125 (N.D. Cal. June
10, 2013), to argue that the named plaintiffs’ claims are not typical of the class’s. Def. MCC Opp’n –
ECF No. 445 at 31. Neither case is apposite here. Burkhead was a class action brought against an
allegedly polluting power plant where the named plaintiffs were a “geographically concentrated group
living relatively close to Defendant’s plant.” Burkhead, 250 F.R.D. at 300. The plaintiffs sought to
represent a class of residents in a two-mile radius of the plant but offered no evidence that the plant
spread any pollution elsewhere within that radius. Id. at 292–93 & n.3. In light of the plaintiffs’
geographic concentration and their lack of evidence that other areas were affected, the court held that
their claims were not typical. Id. at 295–96. The court contrasted that situation from a situation where
“the harm suffered by the named plaintiffs may differ in degree from that suffered by other members
of the class so long as the harm suffered is of the same type,” which would support a finding of
typicality. Id. at 295 (emphasis in original, citation omitted). Major was a class action brought against
a food manufacturer where the named plaintiff alleged misleading product labeling. Major, 2013 WL
2558125, at *1. The plaintiff bought allegedly mislabeled fruit drinks but sought to represent classes of
persons who bought the defendant’s products more broadly. Id. at *2. The court held that the plaintiff
“fail[ed] to link any of those products to any alleged misbranding issue[.]” Id. at *4. For example, the
plaintiff alleged that a blueberry drink she purchased made misleading claims specifically about
blueberries, which the court held was not typical of claims related to other products that did not
contain blueberries. Id. (“[T]he content that purportedly gives rise to Plaintiffs claims is unique to the
specific and particular product she purchased and has no applicability to other products within the
same line.”). Neither of those cases is analogous to the case here.

28
ORDER – No. 12-cv-04854-LB

32

1

(2) will the named plaintiffs and their counsel prosecute the action vigorously on behalf of the

2

class?’” Id. (quoting Hanlon v. Chrysler Corp., 150 F.3d 1011, 1020 (9th Cir. 1998)).
1.4.1

3

Prosecuting the action vigorously on behalf of the class

4

Taking the second issue first, the plaintiffs and their counsel maintain that they will prosecute

5

this case vigorously on behalf of the class. The named plaintiffs have actively participated in this

6

case, cooperated in discovery, made themselves available for and sat for depositions, and followed

7

case developments.106 All are willing to prosecute the action vigorously and to sit for trial.107 They

8

have retained class counsel with experience in litigating antitrust class actions.108 Sutter does not

9

dispute that the named plaintiffs and their counsel will prosecute this case vigorously on behalf of

10

the class.109
The court finds that the plaintiffs have satisfied this component of Rule 23(a)(4)’s adequacy

United States District Court
Northern District of California

11
12

requirement.

13

1.4.2

Conflicts of interest

The Ninth Circuit has cautioned that “we do not ‘favor denial of class certification on the basis

14
15

of speculative conflicts.’” Resnick v. Frank (In re Online DVD-Rental Antitrust Litig.), 779 F.3d

16

934, 942 (9th Cir. 2015) (In re Online DVD II) (quoting Cummings v. Connell, 316 F.3d 886, 896

17

(9th Cir. 2003)). “Nor does a district court abuse its discretion [in certifying a class] when

18

conflicts are trivial.” Id. (citing Abbott v. Lockheed Martin Corp., 725 F.3d 803, 813 (7th Cir.

19

2013)). “‘Only conflicts that are fundamental to the suit and that go to the heart of the litigation

20

prevent a plaintiff from meeting the Rule 23(a)(4) adequacy requirement.’” Id. (quoting William

21
22
23
24

MacAusland Decl. – ECF No. 356-16; Feeney Decl. – ECF No. 356-17; Hansen Decl. – ECF No.
356-18; Herman Decl. – ECF No. 356-19; Stewart Decl. – ECF No. 356-20; Jankowski Decl. – ECF
No. 356-21; Sidibe Decl. – ECF No. 356-22; see, e.g., Feeney Dep. – ECF No. 558 at 3–33 (co-owner
of named employer plaintiff sitting for deposition); Herman Dep. – ECF No. 445-6 at 195–221 (named
individual plaintiff sitting for deposition).
106

25

MacAusland Decl. – ECF No. 356-16; Feeney Decl. – ECF No. 356-17; Hansen Decl. – ECF No.
356-18; Herman Decl. – ECF No. 356-19; Stewart Decl. – ECF No. 356-20; Jankowski Decl. – ECF
No. 356-21; Sidibe Decl. – ECF No. 356-22.

26

108

27

109

107

28

See Counsel Biographies – ECF No. 362-3.

See Def. MCC Opp’n – ECF No. 445 at 9, 30–31 (disputing whether the plaintiffs have satisfied
Rule 23(a)(4)’s adequacy requirements in light of purported conflicts but not disputing the plaintiffs’
or counsel’s prosecution of this case).
ORDER – No. 12-cv-04854-LB

33

1

B. Rubenstein et al., 1 Newberg on Class Actions § 3.58 (5th ed. 2011)). “A conflict is

2

fundamental when it goes to the specific issues in controversy.” Id. (quoting Newberg on Class

3

Actions § 3.58).

4

Sutter argues that there are two conflicts of interest that prevent class certification. Neither

5

purported conflict goes to the heart of the litigation so as to render the plaintiffs inadequate to

6

pursue their class claims.
First, Sutter argues that employers and employees have conflicts, because employers have an

8

incentive to argue that they bore the entire brunt of any premium increase and did not pass any of

9

it along to their employees, while employees have an incentive to argue that the entire alleged

10

increase was passed on to them.110 This is not a conflict that goes to the heart of the litigation.

11

United States District Court
Northern District of California

7

Employers and employees do not have a conflict regarding the central issues at controversy in this

12

case, namely, whether Sutter overcharged health plans for its hospital services, whether health

13

plans passed on those overcharges through to the premiums that they charged their customers,

14

whether they as the health plans’ customers paid more in premiums than they otherwise would

15

have but for Sutter’s anticompetitive conduct, and whether Sutter should pay damages to make

16

them whole. Only if the class collectively were to establish that Sutter is liable and should pay

17

damages would any purported conflicts arise between employers and employees regarding how

18

those damages should be allocated. This is not a conflict that is so fundamental to the suit or goes

19

so much to the heart of the litigation as to render the plaintiffs inadequate to pursue their class

20

claims. Cf. In re Lidoderm, 2017 WL 679367, at *26 (finding plaintiffs adequate in antitrust class

21

action alleging that defendants had inflated the price of lidocaine patches, rejecting arguments that

22

end purchasers had insuperable conflicts with employers or insurers that bore some of the costs,

23

and finding that “any theoretical disputes between, for example, an end payor consumer and her

24

health insurance plan over how their overcharge damages should be split. . . . does not create a

25

type of conflict that precludes certification”).

26
27
110

Def. MCC Opp’n – ECF No. 445 at 30.

28
ORDER – No. 12-cv-04854-LB

34

Second, Sutter argues that class members have conflicts because some class members received

1
2

a net benefit from its allegedly anticompetitive practices “and would have been harmed in the

3

world plaintiffs[] hypothesize absent Sutter’s challenged conduct.”111 Sutter and its expert Dr.

4

Willig maintain that some class members derive more benefit from having Sutter hospitals in their

5

health-insurance provider networks than any harm they suffer in increased health-insurance

6

premiums attributable to Sutter’s alleged overcharges.112 Sutter argues that these class members

7

thus have a conflict with other class members who do not derive such benefit and want to

8

challenge Sutter’s practices.
Sutter’s argument necessarily relies on an assumption that, in the but-for world where Sutter is

9

barred from engaging in allegedly anticompetitive practices, health plans would no longer offer

11

United States District Court
Northern District of California

10

health-insurance policies that have Sutter hospitals in-network. In other words, Sutter’s argument

12

assumes that in the but-for world, health plans would offer only narrow health-insurance policies

13

that exclude Sutter hospitals (or that place Sutter hospitals in a lower tier and increase the amount

14

enrollees had to pay to use those hospitals113), thereby harming class members who currently

15

benefit from having access to Sutter hospitals.
But that is not the plaintiffs’ case theory. The plaintiffs’ theory is that in the but-for world,

16
17

health plans would offer narrow policies that exclude some Sutter hospitals (e.g., the Sutter

18

hospital in the Tying Markets) in addition to — not to the exclusion of — the broader policies they

19

currently offer that include Sutter hospitals in-network.114 Class members who derive benefit from

20
21
22
23
24
25
26
27
28

111

Def. MCC Opp’n – ECF No. 445 at 30.

Def. MCC Sur-Reply – ECF No. 535 at 7–9; Willig Decl. – ECF Nos. 446-2 (under seal), 557
(redacted) at 27–35 (¶¶ 34–45).
112

113

The fact that health plans might place Sutter hospitals in a lower tier does not, on its own,
necessarily harm any class member. For example, if a health plan currently requires enrollees to make
a co-payment of, say, $40 for hospital services, and then were to create a new health-insurance policy
that placed Sutter hospitals in a lower tier and required enrollees to make a co-payment of $20 for
services from hospitals in the highest tier and $40 for services from hospitals in the lowest tier (with
all else being equal), no enrollee would be worse off.
As the plaintiffs explain, “Sutter’s ‘net harm’ argument distorts plaintiffs’ antitrust theory. Dr.
Chipty posited that, in the but for world: (a) Sutter would have faced price competition due to the
threat of exclusion, the launching of additional narrow networks and/or steering; (b) Sutter would have
dropped its rates in order to retain critical patient volume; and (c) consumers would thus have
114

ORDER – No. 12-cv-04854-LB

35

1

having Sutter hospitals in-network could remain in the original broader policies. Consequently,

2

class members would not be worse off. Class members would also have the additional option of

3

signing up for cheaper narrower policies that excluded Sutter hospitals (or placed them in a lower

4

tier). Not all class members would sign up for those cheaper narrower policies, but some would.

5

This would cause Sutter to lose patient volume. This loss in patient volume (and, thus, revenues)

6

would pressure Sutter to lower its supra-competitive prices, so that health plans could in turn

7

lower the premiums of the original Sutter-including broader policies to make them more appealing

8

to customers and more competitive vis-à-vis the narrower policies.115 Those lower prices and

9

premiums would benefit all class members. To take the example discussed above, no matter how
much a class member might benefit from having a health-insurance policy that includes Sutter

11

United States District Court
Northern District of California

10

hospitals in-network, that class member would benefit more (and not be harmed) if she were able

12

to enroll in that policy for, say, $90 a month instead of $100. That some class members benefit

13

from the inclusion of Sutter hospitals in-network does not mean that they benefit from Sutter’s

14

alleged overcharges and, consequently, does not present a conflict that is so fundamental to the

15

suit or goes so much to the heart of the litigation as to render the plaintiffs inadequate to pursue

16

their class claims.
Sutter’s expert Dr. Willig challenges the plaintiffs’ view and argues that in the but-for world

17
18

— where Sutter faces the prospect of losing patient volume due to increased competition from

19

cheaper narrower health-insurance policies that do not include Sutter hospitals in-network —

20

Sutter might respond by increasing its prices instead of reducing them.116 Sutter argues that if

21
22
23
24
25
26
27
28

continued access to the hospitals of their choice (including any in-demand Sutter hospitals) but at
lower prices. Consequently, in Dr. Chipty’s ‘but for’ world, all existing insurance products that include
Sutter hospitals in-network would continue to exist, but at substantially lower prices, making all Class
Members better off.” Pls. MCC Reply – ECF No. 559 at 14 (emphasis in original) (citing Chipty Reply
Decl. – ECF Nos. 462-3 (under seal), 559-1 (redacted) at 20–24 (¶¶ 19–27)); accord Def. MCC SurReply – ECF No. 535 at 6 (“Plaintiffs expressly allege that, in the but-for world, insurers ‘would have
launched a number of additional “tiered” or “limited” or “narrow” networks.’”) (emphasis in original)
(quoting 4AC – ECF No. 204 at 31 (¶ 97)).
115

Alternatively, health plans might be able simply to threaten to create new cheaper narrower policies
that excluded Sutter hospitals (without actually going through with it) because even the threat of
competing cheaper policies could exert pressure on Sutter to lower its prices. See infra note 117.
116

See, e.g., Willig Decl. – ECF No. 557 at 22 (¶ 22).

ORDER – No. 12-cv-04854-LB

36

1

health plans then passed those increased prices through to the premiums the health plans charged

2

their customers (i.e., class members), those class members who wanted to keep Sutter hospitals in-

3

network would be worse off. But whether, in the but-for world, Sutter would respond to increased

4

competition and the threatened loss of patient volume by reducing its prices or increasing them is

5

not for the court to resolve on this class-certification motion. Cf. In re Lidoderm, 2017 WL

6

679367, at *18 (“[W]hat the but-for price should have been is not appropriately resolved on this

7

[class-certification] motion.”); see also In re Aftermarket Automotive Lighting Prods. Antitrust

8

Litig., 276 F.R.D. 364, 373–74 (C.D. Cal. 2011) (“[I]n situations where a court is faced with two

9

opposing expert analyses or econometric models of what the ‘but for’ world would look like, the
Court is not supposed to decide at the certification stage which expert analysis or model is

11

United States District Court
Northern District of California

10

better.”) (citing In re TFT-LCD Antitrust Litig., 267 F.R.D. 291, 313 (N.D. Cal. 2010)). The

12

plaintiffs and Dr. Chipty sufficiently support their position that Sutter would respond to a

13

threatened loss of patient volume by lowering, not raising, its prices (and thus leaving no intraclass

14

conflict that goes to the heart of the litigation) to establish adequacy at this juncture.117

15
16
17
18
19
20
21
22
23
24
25
26
27
28

Among other things, the plaintiffs and Dr. Chipty point to Sutter’s own assertions before the court
in California v. Sutter Health, No. 3:99-cv-03803-MMC (N.D. Cal. filed Aug. 10, 1999), that health
plans’ ability to steer patients away from Sutter — and thereby threaten Sutter with a loss of patient
volume — would serve to “discipline” Sutter’s pricing.
117

The California Attorney General brought the California v. Sutter case in 1999, challenging as
anticompetitive a proposed merger between Sutter’s Alta Bates Medical Center in Berkeley, Alameda
County, and the then independently owned Summit Medical Center in Oakland, Alameda County. See
California v. Sutter Health Sys., 84 F. Supp. 2d 1057, 1059–60 (N.D. Cal. 2001). In defending why the
proposed merger would not be anticompetitive and would not result in higher prices for patients, Sutter
argued to the court that “hospitals are very sensitive to even small declines in [patient] volume” and
that “the loss of even a modest number of patients would be sufficient to discipline Alta Bates’ and
Summit’s prices. Defendants’ economist estimates that a shift of less than 8% (equal to about an
additional 1 patient per day in various competing East Bay hospitals) would be sufficient to prevent a
5% increase at Summit or Alta Bates.” California v. Sutter Defs. Proposed Findings of Fact and
Conclusions of Law – ECF No. 462-5 (under seal) at 15–17 (¶¶ 27, 30). Sutter argued that its
hospitals’ “viability depends on volume, and only a modest loss of patients would be sufficient to
discipline their pricing activity,” id. at 29 (¶ 55), that “the evidence establishes that relatively modest
redirection of patients would impose a critical loss sufficient to defeat a price increase” by Sutter, id. at
37 (¶ 70), that “[t]here are numerous mechanisms by which health plans can discipline the hospitals in
the event of an attempted price increase. The simplest, but rarely used, is to exclude hospitals from the
plans’ provider networks,” id. at 31 (¶¶ 58–59) (citations omitted), and thus that “if the combined Alta
Bates/Summit acted in an anticompetitive manner, the physician groups could, either on their own or
at the impetus of health plans, discipline the merged entity by admitting patients to other hospitals, by
threatening to do so, or by shifting referrals (particularly for high value specialty services) to physician
ORDER – No. 12-cv-04854-LB

37

1
2

The court finds that the plaintiffs have satisfied this component of Rule 23(a)(4)’s adequacy
requirement.

3
4

2. Rule 23(b) Prerequisites
Predominance — Rule 23(b)(3)

5

2.1

6

Among other things, Rule 23(b)(3) requires that “the questions of law or fact common to class

7

members predominate over any questions affecting only individual members.” “Considering

8

whether ‘questions of law or fact common to class members predominate’ begins . . . with the

9

elements of the underlying cause of action.” Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S.
804, 809 (2011). The plaintiffs here bring claims under Sections 1 and 2 of the federal Sherman

11

United States District Court
Northern District of California

10

Antitrust Act, the California Cartwright Act, and the California Unfair Competition Law

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

specialists at other hospitals,” id. at 36 (¶ 68). In 2001, the court denied the California Attorney
General’s request for an injunction blocking the merger and allowed the merger to go forward,
adopting Sutter’s arguments that “[w]hen faced with price increases, there are numerous mechanisms
through which health plans can discipline hospitals. The simplest, but rarely used, is to exclude
hospitals from the plans’ provider networks.” California v. Sutter, 84 F. Supp. 2d at 1078.
Sutter’s own assertions in the California v. Sutter case, together with the other analyses Dr. Chipty
discusses, Chipty Decl. – ECF No. 379-1 at 42–46 (¶¶ 57–63); Chipty Reply. Decl. – ECF Nos. 462-3
(under seal), 559-1 (redacted version) at 21–24 (¶¶ 22–27), provide support for the plaintiffs’ position
that in the but-for world, Sutter would lower its prices, not raise them, thereby benefiting all class
members and leaving no fundamental intraclass conflicts. Cf. In re Online DVD II, 779 F.3d at 942
(courts “do not ‘favor denial of class certification on the basis of speculative conflicts’”) (quoting
Cummings, 316 F.3d at 896). Sutter’s disagreement with the plaintiffs’ but-for world may be a merits
issue, but it does not defeat class certification in the first instance.
It is worth noting that around the time that the merger was allowed or shortly thereafter, Sutter
(according to the plaintiffs) began engaging in its systemwide-contracting practices, thereby allegedly
preventing health plans from using the patient-steering mechanisms that Sutter argued in California v.
Sutter would discipline it from imposing price increases. Following the merger, Sutter (according to
the plaintiffs) increased its prices at Summit by 29.0 to 72.0 percent, significantly more than other
hospitals. Chipty Reply Decl. – ECF No. 559-1 at 102–03 (¶ 164); see also Advocate Health, 841 F.3d
at 472 (“For example, in 2001 the Northern District of California refused to enjoin a hospital merger,
relying in part on patient movement data. In 2011, a follow-up study found that the cheaper of the two
hospitals raised its prices by 29 to 72 percent, much more than a control group had.”) (citing
California v. Sutter, 130 F. Supp. 2d at 1131–32, 1137; Steven Tenn, The Price Effects of Hospital
Mergers: A Case Study of the Sutter-Summit Transaction, 18 Int’l J. of the Econ. of Bus. 65, 75–76
(2011)). Whether Sutter in fact increased its prices to supra-competitive levels, and whether it was
because of its systemwide-contracting practices that it was able to do so, are issues that can be
addressed at the merits phase of the litigation.

28
ORDER – No. 12-cv-04854-LB

38

(“UCL”).118 “To establish a federal antitrust claim, ‘plaintiffs typically must prove (1) a violation

2

of antitrust laws, (2) an injury they suffered as a result of that violation, and (3) an estimated

3

measure of damages.’” In re Qualcomm, 328 F.R.D. at 296 (quoting In re High-Tech Emp., 985 F.

4

Supp. 2d at 1183). “With regard to Plaintiffs’ Cartwright Act claim, ‘the analysis mirrors the

5

analysis under federal law because the Cartwright Act was modeled after the Sherman Act.’” Cf.

6

id. (quoting Cty. of Tuolumne v. Sonora Cmty. Hosp., 236 F.3d 1148, 1160 (9th Cir. 2001)). “Also,

7

Plaintiffs’ UCL claim is premised at least in part upon the Sherman and Cartwright Act

8

violations.” Cf. id. (citing Cel-Tech Commc’ns, Inc. v. L.A. Cellular Tel. Co., 20 Cal. 4th 163, 180

9

(1999)). “Neither party identifies any material difference between the federal and state claims

10

warranting separate treatment.” Cf. id. “Thus, the Court may treat the state law claims together

11

United States District Court
Northern District of California

1

with the federal claims in this case.” Cf. id.
2.1.1

12

Antitrust violations

13

As discussed above, the plaintiffs have introduced evidence that Sutter engaged in common

14

anticompetitive systemwide-contracting practices with respect to all five health plans at issue in

15

this case and that its practices allow it to charge all five health plans supra-competitive rates in the

16

same way.119 Sutter does not meaningfully dispute that the plaintiffs’ claims — that it violated the

17

antitrust laws — are subject to common proof.120 The court finds that common questions will

18

predominate with respect to the alleged antitrust violations.
2.1.2

19

Antitrust injury and calculating damages

“Antitrust ‘impact’ — also referred to as antitrust injury — is the ‘fact of damage’ that results

20
21

from a violation of the antitrust laws.” In re Qualcomm, 328 F.R.D. at 299 (some internal

22

quotation marks omitted) (quoting In re DRAM, 2006 WL 1530166, at *7). “‘It is the causal link

23

between the antitrust violation and the damages sought by plaintiffs.’” Id. (quoting Brown v. Am.

24
25
118

4AC – ECF No. 204 at 38–43 (¶¶ 124–70).

26

119

See supra notes 18–24 and accompanying text.

27

120

28

See Def. MCC Opp’n – ECF No. 445 at 16–29 (disputing whether antitrust injury or damages
calculations are subject to common proof but not disputing whether antitrust violations are subject to
common proof).
ORDER – No. 12-cv-04854-LB

39

1

Honda (In re New Motor Vehicles Canadian Antitrust Litig.), 522 F.3d 6, 19 n.18 (1st Cir. 2008)).

2

“Thus, Plaintiffs here ‘must be able to establish, predominantly with generalized evidence, that all

3

(or nearly all) members of the class suffered damage as a result of [defendant’s] alleged anti-

4

competitive conduct.’” Id. (quoting In re High-Tech Emp., 289 F.R.D. at 567).

5

On a motion for class certification, “plaintiffs [must] be able to show that their damages

6

stemmed from the defendant’s actions that created the legal liability” and “must show that

7

‘damages are capable of measurement on a classwide basis,’ in the sense that the whole class

8

suffered damages traceable to the same injurious course of conduct underlying the plaintiffs’ legal

9

theory.” Nguyen v. Nissan N. Am., Inc., 932 F.3d 811, 817 (9th Cir. 2019) (some internal quotation
marks omitted) (quoting Just Film, Inc. v. Buono, 847 F.3d 1108, 1120 (9th Cir. 2017)).

11

United States District Court
Northern District of California

10

“‘[U]ncertainty regarding class members’ damages does not prevent certification of a class as long

12

as a valid method has been proposed for calculating those damages.’” Id. (quoting Lambert v.

13

Nutraceutical Corp., 870 F.3d 1170, 1182 (9th Cir. 2017), rev’d on other grounds, 139 S. Ct. 710

14

(2019)). “Although uncertain damages calculations do not alone defeat certification, the Supreme

15

Court has emphasized that ‘at the class-certification stage (as at trial), any model supporting a

16

plaintiff’s damages case must be consistent with its liability case.’” Id. (emphasis in original, some

17

internal quotation marks omitted) (quoting Comcast, 569 U.S. at 35).

18

The plaintiffs’ liability theory is that Sutter charged the five health plans at issue in this case

19

— Blue Shield, Anthem, Aetna, Health Net, and UnitedHealthcare — supra-competitive rates. The

20

plaintiffs do not allege that Sutter directly charged class members supra-competitive rates. Instead,

21

the alleged antitrust injury is indirect: the class members’ harm comes only to the extent the health

22

plans passed on Sutter’s alleged overcharges through to class members in the form of higher

23

premiums than the health plans would have charged in the but-for world. As a result, the plaintiffs

24

have a two-fold burden: they must demonstrate that (1) the five health plans paid Sutter inflated

25
26
27
28
ORDER – No. 12-cv-04854-LB

40

1

prices for inpatient hospital services, and then (2) those overcharges were passed on to class

2

members in the form of inflated premiums. Cf., e.g., In re Qualcomm, 328 F.R.D. at 299.121

3

Courts have recognized that “‘antitrust plaintiffs have in recent years trended toward
presenting an econometric formula or other statistical analysis to show class-wide impact’ and that

5

such analysis has often been accepted at the certification stage.” In re Optical Disk Drive Antitrust

6

Litig., No. 3:10-md-2143 RS, 2016 WL 467444, at *7 (N.D. Cal. Feb. 8, 2016) (In re Optical Disk

7

Drive II) (quoting In re Graphics Processing Units Antitrust Litig., 253 F.R.D. 478, 491 (N.D.

8

Cal. 2008)). “‘[S]uch methods, where plausibly reliable, should be allowed as a means of common

9

proof. To rule otherwise would allow antitrust violators a free pass in many industries.’” Id.

10

(quoting In re Graphics Processing Units, 253 F.R.D. at 491). “Accordingly, it is clear that

11

United States District Court
Northern District of California

4

statistical and economic methodologies . . . may be employed to establish class-wide impact.” Id.

12

(emphasis in original). Courts are cautious about “engaging in a battle of expert testimony” at the

13

certification stage. Id. at *6 (citing In re DRAM, 2006 WL 1530166, at *9); accord, e.g., In re

14

Lidoderm, 2017 WL 679367, at *18. At the same time, “‘[c]ertification should not be automatic

15

every time counsel dazzle the courtroom with graphs and tables.’” In re Optical Disk Drive II,

16

2016 WL 467444, at *6 (internal brackets omitted) (quoting In re Graphics Processing Units, 253

17

F.R.D. at 491). “If the presumption were otherwise, ‘nearly all antitrust plaintiffs could survive

18

certification without fully complying with Rule 23.’” Id. (quoting In re Graphics Processing

19

Units, 253 F.R.D. at 492). “‘It is now clear that Rule 23 not only authorizes a hard look at the

20

soundness of statistical models that purport to show predominance — the rule commands it.’” Id.

21

(internal ellipsis omitted) (quoting In re Rail Freight Fuel Surcharge Antitrust Litig. 725 F.3d 244,

22

255 (D.C. Cir. 2013)). “Put another way, the inquiry must be to determine if the proffered expert

23

testimony has the requisite integrity to demonstrate class-wide impact.” Id.; accord Nguyen, 932

24

F.3d at 817 (plaintiff’s method for calculating damages must be “valid”).

25
26
27
28

121

This distinguishes this case from the UFCW & Employers Benefit Trust v. Sutter Health class
action pending against Sutter in state court. In contrast to the putative class here, the class in that case
consists of self-funded payors that paid Sutter directly for healthcare charges incurred by their
employee-members instead of buying health insurance. UFCW & Emps. Benefit Tr. v. Sutter Health,
No. CGC - 14-538451, 2017 WL 11405066, at *5 (Cal. Super. Ct. S.F. Cty. Aug. 14, 2017).
ORDER – No. 12-cv-04854-LB

41

The court has taken a “hard look” at the plaintiffs’ proposed methodology for assessing

1
2

antitrust injury and calculating antitrust damages. The court finds the plaintiffs’ proposed

3

methodology has at least two significant deficiencies that prevent the plaintiffs from relying on it

4

to demonstrate that they can prove antitrust injury or calculate antitrust damages on a class-wide

5

basis: (1) it does not include a reliable method for proving or calculating Sutter’s overcharges to

6

the five health plans, and (2) it does not include a reliable method for proving or calculating how

7

the overcharges were passed through to health-insurance premiums paid by class members.122
2.1.2.1

8

Calculating Sutter’s overcharges

As discussed above, the plaintiffs and Dr. Chipty used a regression-analysis model that

9

purports to calculate what portion of Sutter’s charges to the health plans Anthem and Blue Shield

11
United States District Court
Northern District of California

10

are attributable to Sutter’s alleged anticompetitive practices. Dr. Chipty has not offered an

12

overcharge model with respect to Aetna, Health Net, or UnitedHealthcare.123 She stated that she

13

had two approaches for addressing those three health plans — (1) trying to run her regression-

14

analysis model on them or (2) extrapolating her findings from some health plans to other health

15

plans — but she had not reached an opinion as to which approach she preferred, did not know if

16

she could run regression analyses on those three health plans, and did not have an opinion about

17

whether extrapolation would be reasonable.124 She does not explain how one of her two

18

alternatives — extrapolation — would actually work: despite the fact that her regression-analysis

19

model returned significantly different overcharge rates as between Anthem and Blue Shield,125 she

20

does not explain how she would reconcile those differences to use them to extrapolate rates for

21

Aetna, Health Net, or UnitedHealthcare. Assuming without deciding that Dr. Chipty’s regression-

22
23
24

The court is not engaging in a “battle of the experts” here. The court does not find that the
plaintiffs’ methodology is unreliable because Sutter’s experts or methodologies are more persuasive.
Rather, it finds that the plaintiffs’ methodology is unreliable due to deficiencies within the
methodology.
122

26
27

123

See supra note 52.

124

See supra notes 53–56.

125

25

For example, her model calculated the overcharge rate by Sutter’s Alta Bates Main hospital in 2006
for Anthem and 53 percent for Blue Shield,
. See supra notes 47,

as
51.

28
ORDER – No. 12-cv-04854-LB

42

1

analysis model for Anthem and Blue Shield is sound, the plaintiffs nonetheless have not

2

demonstrated that they can prove or calculate Sutter’s overcharges on a class-wide basis given that

3

they have not offered a method (or hedged between various methods) for calculating overcharges

4

for the remaining three health plans, customers of which Dr. Chipty estimates comprise 30 percent

5

of the class.126

6

The court in In re Lithium Ion Batteries was confronted with a similar situation. The plaintiffs

7

there alleged that battery manufacturers engaged in a price-fixing conspiracy to inflate the price of

8

batteries used in portable consumer electronics (e.g., laptops, smartphones, etc.). In re Lithium Ion

9

Batteries I, 2017 WL 1391491, at *1–2. The plaintiffs relied on an expert report and model to

10

prove and calculate damages on a class-wide basis. Id. at *16. The court found, however, that:
[The expert]’s analysis fails to provide a firm foundation for class certification
because he was unable to complete an analysis based on the actual cost data for any
products other than Toshiba laptops. Further, even in his analysis of the Toshiba
products, the data was limited and required some extrapolation. The data for all
other products was, by [the expert]’s own admission, insufficient to run an analysis
as to any of the other products types covered by the class definition. While it is
unclear where to lay the blame, the Court nevertheless cannot ignore the large gaps
in the evidence supporting the ability to demonstrate impact and damages on a
class-wide basis. . . . As it stands, the analysis of the Toshiba laptops alone does not
satisfy the Court that a showing of antitrust impact for that product can be
extrapolated as a measure of impact for the rest of the Cells, Batteries and Finished
Products in the class definition.

United States District Court
Northern District of California

11
12
13
14
15
16
17
18
19

Id. at *17–18. The court reaches a similar conclusion here. Dr. Chipty offers an overcharge model

20

only for Anthem and Blue Shield, which is insufficient to demonstrate a method for proving

21

antitrust impact or calculating damages on a class-wide basis across the entire class.127

22
23

126

Chipty Reply Decl. – ECF No. 559-1 at 16–17 (¶ 13).

The plaintiffs argue that “[c]lasses have . . . been certified where experts have proposed a damages
model, but have not yet actually calculated any damages for any class members.” Pls. Sur-Sur-Reply –
ECF No. 569 at 4 (emphasis in original). But it is not simply that the plaintiffs have not calculated
damages for Aetna, Health Net, or UnitedHealthcare. They have not proposed a damages model (or try
to hedge between two different models, in order to have it both ways). See supra notes 52, 54.
127

24
25
26
27
28

The court also has some questions about other aspects of Dr. Chipty’s overcharge model with respect
to Anthem and Blue Shield. For example, her model shows an undercharge — rather than an
overcharge — for Blue Shield by one of Sutter’s hospitals in a Tied Market, Sutter Santa Rosa. See
supra note 51. Dr. Chipty testified that this reflects that her model has “an inability to measure the
ORDER – No. 12-cv-04854-LB

43

1
2
3

2.1.2.2

Assuming health plans “pass on” 100 percent of any Sutter overcharges
they have to pay through to the premiums that they charge their
customers

“Where, as here, the class is composed of indirect purchasers, ‘proof of class-wide antitrust

4

impact is made more complex because plaintiffs must offer a model of impact and damages that

5

demonstrates the alleged overcharge was passed through to each successive link in the distribution

6

chain, and ultimately to the plaintiffs.’” In re Qualcomm, 328 F.R.D. at 301 (quoting In re Lithium

7

Ion Batteries Antitrust Litig., No. 13-MD-2420 YGR, 2018 WL 1156797, at *3 (N.D. Cal. Mar. 5,

8

2018) (In re Lithium Ion Batteries II)).

9

Courts in indirect-purchaser antitrust cases have recognized the use of averaging in calculating
passthrough rates where experts have shown the sound methodological steps through which they

11

United States District Court
Northern District of California

10

calculated their averages. See, e.g., id. at 315 (collecting cases). For example, In re Qualcomm —

12

an antitrust case against a manufacturer of “modem chips” used in cellphones brought by a class of

13

indirect purchasers who bought phones containing the chips — the plaintiffs’ expert calculated an

14

average passthrough rate from the chip manufacturer through cellphone manufacturers and

15

retailers to cellphone buyers through regression analyses based on sales data from (1) six major

16

cellphone manufacturers, including the five largest manufacturers in the U.S. market (Apple,

17

Samsung, Motorola, LG, and HTC), representing approximately 90 percent of total cellphone

18

sales, (2) six of the largest U.S. retailers (including Best Buy, Amazon, Wal-Mart, and Target),

19

representing approximately 84 percent of the retailer market, and (3) five wireless carriers,

20

including the four major U.S. carriers (AT&T, Sprint, T-Mobile, and Verizon) and one regional

21

carrier, representing approximately 97 percent of the wireless-carrier market. Id. at 302–03. The

22

expert used ten control variables in his regressions (the same ten control variables that the

23
24
25
26
27
28

overcharge for Sutter-Santa Rosa.” See supra note 57. It is unclear if the plaintiffs are taking the
position that Sutter did not overcharge Blue Shield at its Sutter Santa Rosa hospital or if they are
taking the position that it did overcharge and that they need a new model to determine how much it
overcharged (because their current model cannot measure that overcharge). The latter raises the
prospect of the plaintiffs proposing additional individualized overcharge models for specific hospitals
or health plans. In light of the court’s decision on other grounds that the plaintiffs have not provided a
sufficient model, it need not fully resolve these questions now, but it raises them as issues that may
need to be addressed on any further class-certification motion.
ORDER – No. 12-cv-04854-LB

44

1

defendant used in a submission to the FTC) and calculated separate passthrough rates for each of

2

18 sales channels, which he then weighted to calculate an overall sales-channel-weighted average

3

passthrough rate of 87.4 percent. Id. at 303–04. In finding that the expert’s average weighted

4

passthrough rate was sufficient to support class certification, the court found that:
[The plaintiffs’ expert] does not simply assume a uniform pass-through rate for
OEMs. Instead, he examines transactional data for six different OEMs — including
the five largest OEMs in the U.S. market (Apple, Samsung, Motorola, LG, and
HTC) — who “accounted for approximately 90% of total cell phone sales” during
the relevant period. [He] calculates individual pass-through rates for these six
OEMs in order to model a composite pass-through rate.

5
6
7
8
9

Id. at 308. By contrast, courts have rejected the use of averaging in calculating passthrough rates
where experts have not shown the sound methodological steps through which they calculated their

11

United States District Court
Northern District of California

10

averages. For example, in In re Lithium Ion Batteries — an antitrust case against manufacturers of

12

batteries used in consumer electronics — the plaintiffs’ expert assumed that the passthrough rate at

13

each level in the distribution chain from the battery manufacturers to device manufacturers to

14

retailers to the end consumers approached 100 percent without sufficiently supporting his analysis.

15

In re Lithium Ion Batteries I, 2017 WL 1391491, at *12. Among other things, “[the expert]

16

acknowledged that bundling, rebates, and discounts would affect the accuracy of cost data, but

17

apparently has offered no methodology to account for it in his analysis.” Id.128 The court there

18

“f[ound] the [expert’s] declarations insufficient to show that pass-through and damages can be

19

established by expert analysis on a class-wide basis” and denied certification. Id. On a renewed

20

motion a year later, the court found that the expert’s supplemental analysis still failed to account

21

for a central issue affecting passthrough rates (focal-point pricing; meaning, the practice of

22

retailers setting prices at certain “focal points,” such as prices ending with 9, and not adjusting

23

such prices based on small differences in costs) and thus “left too much uncertainty as to whether

24

pass-through can be estimated reliably at 100% as to retailers or distributors farther down the

25
26
By way of contrast, the expert in In re Qualcomm “perform[ed] separate pass-through rate
calculations for subsidized and unsubsidized phones and f[ound] statistically significant pass-through
rates for each wireless carrier for subsidized and unsubsidized phones.” In re Qualcomm, 328 F.R.D.
at 310.
128

27
28

ORDER – No. 12-cv-04854-LB

45

1

supply chain, and ultimately to the consumers who make up the proposed class,” and once again

2

denied certification. In re Lithium Ion Batteries II, 2018 WL 1156797, at *4.
The In re Optical Disk Drive antitrust litigation is instructive because the court there first

4

denied class certification due in part to deficiencies in the plaintiffs’ passthrough assumptions and

5

then, two years later, granted a renewed motion for class certification. The plaintiffs there alleged

6

that optical-disc-drive (“ODD”) manufacturers engaged in bid-rigging to prop up the price of

7

ODDs. In re Optical Disk Drive Antitrust Litig., 303 F.R.D. 311, 314 (N.D. Cal. 2014) (In re

8

Optical Disk Drive I).129 The plaintiffs sought to certify a class of direct-purchaser plaintiffs

9

(“DPPs”) that bought ODDs directly from the defendants, id. at 316, and a class of indirect-

10

purchaser plaintiffs (“IPPs”) that bought products that contained ODDs manufactured by the

11

United States District Court
Northern District of California

3

defendants (but which the IPPs did not buy directly from the defendants), id. at 323. On the

12

plaintiffs’ initial class-certification motion, the plaintiffs’ expert aggregated prices for all

13

purchasers who bought ODDs of particular types in given years before running a regression

14

analysis to calculate a passthrough rate. Id. at 324. The court found that this resulted in “class-

15

wide impact . . . being assumed by the models, rather than demonstrated by the results.” Id.

16

(emphasis in original). The expert “purport[ed] to test the validity of his models by looking to

17

specific examples in the data,” but the court rejected this approach, holding that “[i]dentifying

18

some instances where the empirical data appears to match the model does not transform the

19

analysis from one that assumes class-wide impact into one that proves it.” Id. The court ultimately

20

concluded that “the IPPs have not presented a persuasive explanation as to why it would be

21

reasonable to assume a uniform pass through rate” and denied certification. Id.
Two years later, the plaintiffs moved for certification again, this time for a narrower IPP class.

22
23

In re Optical Disk Drive II, 2016 WL 467444, at *3. The plaintiffs’ expert offered a modified

24

overcharge model that, among other things, “integrate[d] all ‘useable’ sales and costs data

25

produced — from 86 percent of the market” and “provide[d] further detail on the multivariable

26
27
28

“Optical discs” include CDs and DVDs. In re Optical Disk Drive I, 303 F.R.D. at 314. While the
action was captioned “Optical Disk Drive Antirust Litigation,” apparently “disc” (not “disk”) is the
preferred spelling for optical media. Id. at 314 n.1.
129

ORDER – No. 12-cv-04854-LB

46

1

regression analysis, which [the plaintiffs] contend shows that all factors other than [the antitrust]

2

conspiracy are being adequately controlled for in the overcharge model.” Id. at *6. The court

3

found that this new analysis, which measured passthrough rates for over 273 million ODD

4

products, id. at *9, was sufficient to support certification.
Dr. Chipty’s passthrough analyses here have more in common with the deficient models in In

5
6

re Lithium Ion Batteries and the first In re Optical Disk Drive decision than the models in In re

7

Qualcomm, the second In re Optical Disk Drive decision, or other decisions where courts have

8

accepted passthrough models. Dr. Chipty began by assuming what she needs to prove, namely,

9

that each of the five health plans pass on 100 percent of any Sutter overcharges through to the
premiums they charge their customers. Dr. Chipty based her 100-percent-passthrough assumption

11
United States District Court
Northern District of California

10

in large part on the Bob Reed PowerPoint presentation where Sutter supposedly assumed a 100-

12

percent-passthrough rate for the purpose of running hypotheticals

13
.130 Contrary to Dr. Chipty’s inference, the hypothetical in that presentation does not

14
15

establish that Sutter actually concluded that health plans pass on 100 percent of Sutter’s price

16

reductions (or increases) through to their customers’ premiums. Even if it did, the fact that Sutter

17

(which is not a health plan) might have drawn this conclusion does not establish that health plans

18

actually pass on 100 percent of price reductions (or increases) through to their customers’

19

premiums. Dr. Chipty further based her 100-percent-passthrough assumption on documents and

20

analyses that she said indicates that health plans try to set premiums at a level that covers all their

21

expenses and that when their costs increase, their premiums do as well.131 The fact that health

22

plans try to set premiums at a level that covers all their expenses as a whole, or that their

23

premiums generally increase when their costs increase, does not establish that health plans pass on

24

100 percent of any given cost increase or overcharge through to their customers. To the contrary,

25

Dr. Chipty herself acknowledged that “a health plan may decide to absorb some portion of a

26
130

See supra notes 77–79.

131

27

See supra note 80.

28
ORDER – No. 12-cv-04854-LB

47

1

medical cost increase, by cutting into [its] margins,” rather than passing on 100 percent of those

2

increases to its customers.132

3

Dr. Chipty tries to bolster her 100-percent-passthrough-rate assumption with a regression

4

analysis. Her original regression analysis is overly simplistic, using data only from small-group

5

employers (not individuals or large-group employers) and measuring the relationship between

6

PMPM premiums and only a single explanatory variable, PMPM costs. Dr. Chipty herself

7

acknowledged that this was “a very simple regression model” and that this model might not be

8

appropriate if the relationship between a health plan’s costs (including costs in paying hospitals

9

like Sutter) and the health plan’s premiums were complex.133 She said that she nonetheless elected
to use this simple model because she assumed that the relationship between costs and premiums

11
United States District Court
Northern District of California

10

was “formulaic.”134 In other words, Dr. Chipty assumed what she set out to prove — that the

12

method by which health plans pass on their costs through to their customers’ premiums is in fact

13

“formulaic” — and then developed a “simple” regression-analysis model to try to support that

14

assumption. Cf. In re Optical Disk Drive I, 303 F.R.D. at 324 (finding unreliable models that

15

assumed class-wide impact, rather than demonstrating class-wide impact).
Among the model’s other flaws is its failure to take into account how competition from rival

16
17

health plans (including Kaiser Permanente) affects a health plan’s decision to pass its costs

18
19
20
132

21
22
23
24
25
26
27

See supra note 60.

See Chipty Dep. – ECF No. 415-1 (under seal) at 114 (pp. 391–92) (“Q. Are you aware of any
generally accepted method of determining the pass-through rates? A. Oh, I believe this is a generally
accepted method. There are a range of methods ranging from simple to highly complicated that people
use and can use. And it just depends, I think, on the context and the available evidence to study as well
as the other corroborating evidence you have. So I view this as a component of the larger evidence that
I use to reach my conclusion. Now, would I do exactly the same thing in a different setting? Probably
not. I would have to adapt to the — to the circumstances and the facts of the case. Q. What would it
depend on? A. It would depend in part on the complexity of the relationship between the price and the
cost. Here are the reasons — clear reasons to believe, given the rate setting process, that there is a, if
you will, formulaic relationship between costs and premiums. So that’s part of the underpinning that
led me to believe that there would surely be a relationship, and the question was to what extent are
those movements correlated.”).
133

134

See id.

28
ORDER – No. 12-cv-04854-LB

48

1

through to its premiums.135 Dr. Chipty stated in her report that “[t]o the extent health plans like

2

Anthem and Blue Shield would absorb cost increases to compete with Kaiser, pass-through may

3

be less than 100 percent.”136 But she offers no methodology for taking health-plan competition

4

into account in her original regression-analysis model, undermining her model’s reliability. Cf.

5

Lithium Ion Batteries I, 2017 WL 1391491, at *12 (finding unreliable models where expert

6

acknowledged that certain factors would affect the accuracy of cost data but had no methodology

7

to account for those factors in his analysis).
Dr. Chipty’s additional analyses in her reply declaration fare no better. For one health plan,

8

Aetna, Dr. Chipty does not dispute Dr. Willig’s analysis estimating a passthrough rate of 30

10

percent, with a 95-percent confidence interval ranging from negative 16.4 percent and 76.8

11
United States District Court
Northern District of California

9

percent.137 Dr. Chipty describes this 30-percent-passthrough rate as being “not statistically

12

different from a pass-through rate as high as 77 percent”138 — ignoring that, by that same logic, it

13

also is not statistically different from a passthrough rate as low as negative 16.4 percent — and

14

then offers no analysis for how she went from a passthrough rate “as high as 77 percent” to

15

assuming a significantly higher 100-percent-passthrough rate for Aetna in her damages model.

16

Similarly, of the 23 passthrough calculations that she listed in Exhibit 8 of her reply declaration,139
of them (i.e., more than

17

of her results) had 95-percent-confidence intervals whose

18

upper bound fell below 1 — meaning that for those regressions, there was less than a five-percent

19

chance that the actual passthrough rate was as high as 100 percent. Other than saying that these

20

results nevertheless were “near” or “close to” 100 percent140 — terms she does not define with any

21

precision141 — she does not explain how these results or her analyses support her decision to

22
23

135

24

136

See supra note 67; see also supra notes 62, 68.

25

137

See supra note 88.

138

See supra note 91.

139

See supra note 89.

27

140

See supra note 95.

28

141

See supra note 96.

26

Cf. supra notes 69–74 and accompanying text (discussing examples where health plans did not pass
on costs through to their customers’ premiums due to competition from rival health plans).

ORDER – No. 12-cv-04854-LB

49

1

assume a uniform 100-percent-passthrough rate across all health plans, all business lines, all

2

health-insurance products, in all competitive situations.

3

The court finds that Dr. Chipty’s analyses are insufficient to show that antitrust injury and
damages can be established on a class-wide basis. And absent a sound methodology for proving

5

on a class-wide basis what Sutter’s overcharges were and how those overcharges were passed

6

through to the class, the plaintiffs have not met their burden of showing that common issues

7

predominate. Cf. In re Optical Disk Drive I, 303 F.R.D. at 324 (absent a class-wide methodology

8

for calculating overcharges and passthroughs, the case would result in “‘thousands of mini-trials,

9

rendering this case unmanageable and unsuitable for class action treatment’”). The court therefore

10

denies the plaintiffs’ motion to certify their proposed class under Rule 23(b)(3). Cf. In re Lithium

11

United States District Court
Northern District of California

4

Ion Batteries II, 2018 WL 1156797, at *5 (denying certification due to lack of valid class-wide

12

passthrough model); In re Lithium Ion Batteries I, 2017 WL 1391491, at *19 (same); In re Optical

13

Disk Drive I, 303 F.R.D. at 324–25 (same).

14

This denial is without prejudice to the plaintiffs moving to certify a Rule 23(b)(3) class if they

15

are able to make a fuller showing that they can prove antitrust injury and calculate damages on a

16

class-wide basis with a model that addresses the deficiencies the court identified here (and any

17

other deficiencies that may be present elsewhere in their models). Cf. In re Lithium Ion Batteries I,

18

2017 WL 1391491, at *12. This denial also is without prejudice to the plaintiffs moving to certify

19

a Rule 23(b)(3) class for settlement purposes, which does not raise the same manageability issues

20

that certifying a litigation class does. In re Hyundai and Kia, 926 F.3d at 556–57 (“The criteria for

21

class certification are applied differently in litigation classes and settlement classes. In deciding

22

whether to certify a litigation class, a district court must be concerned with manageability at trial.

23

However, such manageability is not a concern in certifying a settlement class where, by definition,

24

there will be no trial.”).
Acting on Grounds That Apply Generally to the Class — Rule 23(b)(2)

25

2.2

26

A court may certify a class under Rule 23(b)(2) if “the party opposing the class has acted or

27

refused to act on grounds that apply generally to the class, so that final injunctive relief or

28

corresponding declaratory relief is appropriate respecting the class as a whole[.]”
ORDER – No. 12-cv-04854-LB

50

1

“‘The key to the (b)(2) class is the ‘indivisible nature of the injunctive or declaratory remedy

2

warranted — the notion that the conduct is such that it can be enjoined or declared unlawful only

3

as to all of the class members or as to none of them.’” B.K. ex rel. Tinsley v. Snyder, 922 F.3d 957,

4

971 (9th Cir. 2019) (quoting Wal-Mart, 564 U.S. at 360). “‘In other words, Rule 23(b)(2) applies

5

only when a single injunction or declaratory judgment would provide relief to each member of the

6

class. It does not authorize class certification when each individual class member would be

7

entitled to a different injunction.’” Id. (quoting Wal-Mart, 564 U.S. at 360).

8
9

Rule 23(b)(2) “does not require [courts] to examine the viability or bases of class members’
claims for declaratory and injunctive relief, but only to look at whether class members seek
uniform relief from a practice applicable to all of them.” Rodriguez v. Hayes, 591 F.3d 1105, 1125

11
United States District Court
Northern District of California

10

(9th Cir. 2010). “‘[I]t is sufficient’ to meet the requirements of Rule 23(b)(2) that ‘class members

12

complain of a pattern or practice that is generally applicable to the class as a whole.’” Id. (quoting

13

Walters v. Reno, 145 F.3d 1032, 1047 (9th Cir. 1998)). “The fact that some class members may

14

have suffered no injury or different injuries from the challenged practice does not prevent the class

15

from meeting the requirements of Rule 23(b)(2).” Id. (citing Walters, 145 F.3d at 1047).

16

“Furthermore, unlike actions brought under one of the other 23(b) prongs, ‘questions of

17

manageability and judicial economy are irrelevant to 23(b)(2) class actions.’” Id. (internal ellipsis

18

omitted) (quoting Forbush v. J.C. Penney Co., Inc., 994 F.2d 1101, 1105 (5th Cir. 1993)).

19

The Ninth Circuit has recently instructed that courts should not impose a “cohesiveness”

20

requirement in assessing whether certification under Rule 23(b)(2) is appropriate. Senne v. Kan.

21

City Royals Baseball Corp., 934 F.3d 918, 937–38 (9th Cir. 2019) (internal brackets omitted)

22

(quoting Fed. R. Civ. P. 23(b)(2)). “Although common issues must predominate for class

23

certification under Rule 23(b)(3), no such requirement exists under 23(b)(2).” Id. at 938 (quoting

24

Walters, 145 F.3d at 1047). Instead, “Rule 23(b)(2) . . . requires only that ‘the party opposing the

25

class have acted or refused to act on grounds that apply generally to the class, so that final

26

injunctive relief or corresponding declaratory relief is appropriate respecting the class as a

27

whole.’” Id. at 928 (internal brackets omitted) (quoting Fed. R. Civ. P. 23(b)(2)). Due to Rule

28

23(b)(2)’s and Rule 23(b)(3)’s differing requirements, courts have granted motions to certify
ORDER – No. 12-cv-04854-LB

51

1

putative classes under Rule 23(b)(2) while denying motions to certify the classes under Rule

2

23(b)(3). See, e.g., Ang v. Bimbo Bakeries USA, Inc., No. 13-cv-01196-HSG, 2018 WL 4181896,

3

at *12, *17 (N.D. Cal. Aug. 31, 2018) (denying certification under Rule 23(b)(3) “[b]ecause

4

Plaintiffs have not shown that the economic harm they allegedly sustained . . . is capable of

5

measurement on a classwide basis” but granting certification under Rule 23(b)(2) because “a

6

single injunction would provide relief to each member of the class”) (internal ellipsis omitted);

7

Campbell v. Facebook Inc., 315 F.R.D. 250, 269–70 (N.D. Cal. 2016) (same).
The plaintiffs here seek a single injunction or declaratory judgment — namely, an injunction

8
9

barring Sutter from engaging in anticompetitive behavior, including its systemwide-contracting
practices with their “all-or-nothing,” “anti-steering,” and “penalty rate” provisions or a declaration

11
United States District Court
Northern District of California

10

that Sutter’s practices are anticompetitive and violate the antitrust laws — that would provide

12

relief to each member of the class.142 They therefore have satisfied the requirement for certifying a

13

class under Rule 23(b)(2). Cf. In re Qualcomm, 328 F.R.D. at 318–19 (certifying Rule 23(b)(2)

14

class in antitrust case); In re TFT-LCD, 267 F.R.D. at 595–97 (same).143

15
16
17
18
19
20
21

142

The plaintiffs do not need to specify at the class-certification stage the precise injunction they will
ultimately seek on the merits. B.K., 922 F.3d at 972.
Sutter argues that the putative class is not sufficiently “cohesive” to be certified under Rule 23(b)(2)
and argues that “[t]he need for cohesiveness under (b)(2) is similar to the requirements of
‘predominance’ and ‘superiority’ under Rule 23(b)(3),” citing cases from the Third Circuit. Def. MCC
Opp’n – ECF No. 445 at 31–32. This argument fails in light of the Ninth Circuit’s decision in Senne,
934 F.3d at 937–38. Sutter also argues that certification under Rule 23(b)(2) is improper because (so it
claims) many class members benefited from its practices. Def. MCC Opp’n – ECF No. 445 at 32. As
the court discussed above in the context of Rule 23(a)(4)’s adequacy requirement, Sutter’s argument
about ostensible “benefit” to some class members from its practices is misplaced. Additionally,
arguments that some hypothetical members of the class allegedly benefit from the practices that the
class seeks to enjoin and do not want those practices to end do not provide a basis for denying
certification under Rule 23(b)(2). Campbell, 315 F.R.D. at 269–70 (citing In re Yahoo Mail Litig., 308
F.R.D. 577, 601 (N.D. Cal. 2015)). “[C]lass certification is not a decision on the merits, and the
plaintiffs will only be entitled to injunctive relief if such relief is necessary to redress the . . . violations
they actually prove at trial.” B.K., 922 F.3d at 971.
143

22
23
24
25
26
27
28

ORDER – No. 12-cv-04854-LB

52

CONCLUSION

1
2
3

The court grants the plaintiffs’ motion to certify their proposed class under Rule 23(b)(2) and
denies without prejudice their motion to certify their proposed class under Rule 23(b)(3).

4
5

IT IS SO ORDERED.

6

Dated: August 30, 2019
______________________________________
LAUREL BEELER
United States Magistrate Judge

7
8
9
10

United States District Court
Northern District of California

11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
ORDER – No. 12-cv-04854-LB

53