In Re Pharmaceutical Industry Average Wholesale Price Litigation

491 F. Supp. 2d 20, 2007 U.S. Dist. LEXIS 44817, 2007 WL 1774644

Opinion

FINDINGS OF FACT AND CONCLUSIONS OF LAW

SARIS, District Judge.

TABLE OF CONTENTS

Page

INTRODUCTION AND SUMMARY. 29

I. Findings of Fact.32

A. The Origins of Average Wholesale Price.32

B. Medicare Part B.33

C. Manipulating and Marketing the Spread.34

D. Cross-Subsidization.37

E. Patient, The Vulnerable Victim.38

F. Self-Administered Drugs.39

G. Knowledge in the Industry. 39

H. Mega-Spreads.40

I. The Government Pit Bull.41

J. The Demise of AWP as Government Pricing Benchmark.44

K. Stuck.45

L. The Plaintiffs/TPPs.46

1. Blue Cross/Blue Shield Class 2 and Class 3 Representative.46

2. Pipefitters: Class 3 Representative .49

3. Sheet Metal Workers: Class 2 Representatives.50

M. Defendants.50

1. AstraZeneca.50

2. The Johnson & Johnson Group.54

a. Procrit. 54

b. Remicade.57

3. The Bristol-Myers Squibb Group.59

a. Single Source Drugs.1.62

i. Paraplatin .62

ii. Etopophos .63

b. Single-Source Drugs Later Subject to Generic Competition.64

i. Taxol ..64

ii. Vepesid.66

iii. Cytoxan.67

iv. Blenoxane.68

c. Multi-Source Drugs.69

i. Rubex .69

4. The Schering-Plough Group .70

a. Temodar.72

b. Intron-A.72

c. Proventil .73

d. Generic Albuterol Sulfate.74

II. CONCLUSIONS OF LAW.75

A. Statute of Limitations.75

B. Liability Under Section 9 or 11 of Chapter 93A.80

C. Per Se Unfair or Deceptive Conduct under Chapter 93A.82

D. The Daubert Challenge.85

1. The Hartman Speed Limit .86

2. Defendants’ Critique .89

a. Payors’ Expectations .89

b. Spreads of thirty percent.91

c. Changes in reimbursement.92

d. Living in the “but for” world. 92

E. The Merits: Chapter 93A Unfair or Deceptive Acts.93

1. The Standard.93

2. The Inflation of AWP.94

3. Causation .96

4. Class 2 Liability and Damages. 97

5. Multi-source drugs.97

a. Causation.98

b. Apportionment.100

6. Drug-by-Drug.101

a. AstraZeneca .102

b. Johnson & Johnson.103

1. Procrit.103

2. Remicade.104

c. Bristol-Myers Squibb.104

1. Etopophos .106

2. Paraplatin .106

3. Taxol.106

4. Vepesid.106

5. Cytoxan.107

6. Blenoxane.107

7. Rubex .108

d. Schering-Plough.108

1. Intron-A and Temodar.108

2. Proventil.108

3. Warrick’s Albuterol Sulfate.108

F. Class 2 Damages.109

III. ORDER.109

INTRODUCTION AND SUMMARY

This massive nationwide multi-district class action involves the pricing of pharmaceutical drugs reimbursed by Medicare, private insurers, and patients making coinsurance payments based on average wholesale price (“AWP”) 1 between 1991 and 2003. For the most part, the drugs at issue are administered by doctors for the treatment of cancer and other serious ailments.

Class plaintiffs have alleged that four pharmaceutical companies, AstraZeneca, Schering-Plough, Bristol-Myers Squibb (BMS) and Johnson and Johnson (J & J), 2 have engaged in unfair and deceptive trade practices in violation of Mass. Gen. Laws ch. 93A by grossly inflating the AWPs of certain specified drugs, which are published in commercial publications (Red Book, Medispan, First DataBank), and that these inflated prices have caused damages to Medicare, third-party payors, and patients making percentage co-payments.

The physician-administered drugs at issue in this litigation are typically quite expensive. For example, during the class period, Zoladex, manufactured by AstraZ-eneca to treat prostate cancer, had an AWP ranging from $320 to $450 for a one month dose; a typical dose of Taxol, manufactured by BMS to treat breast and ovarian cancer, had an AWP of over $1800; Remicade, a J & J product used to treat Crohn’s Disease and rheumatoid arthritis, cost over $1000 per dose; and Intron A, manufactured by Schering-Plough and used to treat melanoma, leukemia, and hepatitis, cost nearly $500 per week for a typical recommended dosage. 3 (Rosenthal Dir. ¶ 14.) Certain drugs that are self-administered with durable medical equipment are compensated under Medicare Part B and are therefore also included in the class action. The primary drug in this category is albuterol sulfate, a self-administered drug commonly administered by a nebulizer for asthma, and manufactured by Warrick, a subsidiary of Schering-Plough.

Plaintiffs’ core claim is that the published AWPs for defendants’ drugs are fictitious because they do not reflect the true average sales price (“ASP”) to providers, like doctors and pharmacists. Because AWP is the predominant benchmark for reimbursement by the government and third-party payors, plaintiffs contend that manufacturers grossly inflate each drug’s AWP to create a “spread” between the doctor’s real acquisition cost and the fictitious published AWP, and that drug manufacturers then “market the spread” in order to obtain market share over a competitor’s drug. Indeed, some doctors began to refer to “AWP” as “ain’t what’s paid.” Some of the representative “markups” 4 at issue in this litigation are reflected in the chart below.

Percentage Markup

Defendant_Drug Name_Spread (Year)_Spread (Year)

AstraZeneca_Zoladex_40.7% (1995)_149.7% (2001)

Bristol-Myers Squibb_Blenoxane_72.8% (1998)_85.9% (2002)

Bristol-Myers Squibb Taxol_27.0% (1997)_128.7% (2002)

Bristol-Myers Squibb Cytoxan_257.7% (1997)_676.8% (1999)

Bristol-Myers Squibb_Rubex_180.7% (1995)_66.2% (2001)

Bristol-Myers Squibb_Vepesid_70.7% (1995)_1131.7% (1999)

Johnson & Johnson_Remicade_32.1% (1999)_31.9% (2001)

Schering-Plough_Proventil_53.4% (1993)_28.6% (2001)

Warrick (Sphering)_albuterol sulfate_186.8% (1995)_651.4% (2002)

(PX 4030 at ¶ 38, Table 1.) 5

This bench trial involved two Massachusetts classes. One class, Class 2, 6 consists of third-party payors (“TPPs”) in Massachusetts that reimburse Medicare beneficiaries for their statutory twenty percent coinsurance obligations under Medicare, known as Medigap insurance or supplemental insurance. The other class of plaintiffs, Class 3, 7 consists of all third party payors, end-payors, consumers who make coinsurance payments, and consumers who have no insurance for these drugs in Massachusetts and who pay for drugs based on AWP. 8

The bench trial spanned twenty days, included nearly forty witnesses, and involved hundreds of documents and deposition transcripts. In essence, the evidence established that the Medicare system ere- ated perverse incentives by pegging the nationwide reimbursement for billions of drug transactions a year to a price reported by the pharmaceutical industry without any oversight. Many pharmaceutical companies unscrupulously took advantage of that flawed AWP system by establishing secret mega-spreads between the fictitious reimbursement price they reported and the actual acquisition costs of doctors and pharmacies. These spreads grossly exceeded the standard industry markup. The publication of false, inflated AWPs caused real injuries to the government, insurers, and patients who were paying grossly inflated coinsurance payments for critically important, often life-sustaining, drugs. Once the mega-spreads became widely known, the conduct was still egregious under the unfairness prong of Chapter 93A because neither the third party payors nor the government could move quickly or effectively to fix the problem. In 2003, Congress finally fixed the problem by moving to a reimbursement system not based on AWP.

I make the following findings with respect to the individual defendants:

1. AstraZeneca acted unfairly and deceptively by causing the publication of false and inflated average wholesale prices for Zoladex which grossly exceeded actual physician acquisition costs by as much as 169% and then marketing these mega-spreads between the physician’s acquisition costs and the AWP reimbursement benchmark in order to induce doctors to buy its drug based on the drug’s profitability. The spread on Zoladex exceeded 100% from 1998 forward. The Court finds damages of $4,451,429 to Class 3. The Court needs additional information to calculate damages for Class 2.

2. Bristol-Myers Squibb acted unfairly and deceptively by causing the publication of false and inflated average wholesale prices for five drugs, which grossly exceeded actual physician acquisition costs and then marketing these mega-spreads between the physician’s acquisition costs and the AWP reimbursement benchmark in order to induce doctors and other providers to buy its drugs based on the drugs’ profitability. I find liability for Bristol-Myers Squibb’s drugs Taxol (with spreads as high as 500%), Vepesid injectable (with spreads as high as 1,000%), Cytoxan injectable (with spreads as high as 676%), Blenoxane (with spreads as high as 199%), and Rubex (with spreads as high as 438%). The Court finds damages of $183,454 to Class 3.The Court needs additional information to calculate damages for Class 2.

3. Schering-Plough''s subsidiary War-rick acted unfairly and deceptively by causing the publication of false and inflated average wholesale prices for its generic drug albuterol sulfate, which had mega-spreads between 100% and 800% throughout the class period. The Court needs additional information to calculate damages for Class 2.

4. While Johnson & Johnson’s conduct was at times troubling, it did not rise to the level of egregious misconduct actionable under the Massachusetts Chapter 93A, because its spreads never substantially exceeded the range of what was generally expected by the industry and government.

5. The statute of limitations bars all claims by Classes 2 and 3 before December 1997. When Congress passed the Balanced Budget Act of 1997, it put third party payors on inquiry notice that many AWPs were not true prices paid by physicians and pharmacies to acquire the pharmaceuticals. The class period ends in 2003 when Congress passed the Medicare statute setting new reimbursement benchmarks. Thus, Classes 2 and 3 will include payments from December 1997 to December 2003.

6. The Court rejects plaintiffs’ position with respect to the Medicare Class 2, that defendants acted unfairly and deceptively by having any spread between the published AWP and the true average of prices charged to providers, because the government and industry were well aware by the late 1990’s that there was a 20 to 25 percent spread. This discrepancy was tolerated, in part, because of the need to cross-subsidize physician administration costs. Thus, while the spread violated the plain meaning of the Medicare statute, defendants’ actions cannot be said to be unfair or deceptive within the meaning of Chapter 93A so long as the spread stayed generally within that expected range.

7. Damages to Class 2 cannot be determined from the current trial record. The Court needs a breakdown of the damages for each drug, using the 30% threshold, for each of the years from 1998 until 2003 for which liability has been found. Defendants may provide their market shares in Massachusetts so that the Court can apportion the damage amount on that basis. If necessary, the Court will hold a damages phase of the bench trial.

The findings of fact and conclusions of law follow.

I. FINDINGS OF FACT

A. The Origins of Average Wholesale Price

Since the late 1960’s, almost every brand and generic prescription drug sold in the United States has had an “average wholesale price,” which is published in commercial compendia like Red Book, First DataBank, and Medispan. AWP is used as the basis for drug reimbursement both for drugs administered in physicians’ offices (“physician-administered drugs” or “PADs”) and for self-administered drugs dispensed by pharmacies (“self-administered drugs” or “SADs”). “Average Wholesale Price” or “AWP” was the pricing benchmark used by the federal government for Medicare reimbursement throughout the class period (1991 to 2003) until the passage of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003. See Pub.L. No. 108-173, 117 Stat. 2066. Throughout this period (and until today), it has also been the pricing benchmark used by most TPPs in Massachusetts and the nation. In 2002, Dyckman & Associates conducted a survey of private health plans regarding their payments for physician-administered drugs and found that all of the plans used a percentage of AWP as a formula to reimburse physicians for these drugs; that most plans used an AWP pricing formula that was in the range of 90 to 100 percent of AWP; and that the average percentage was 98 percent. (Rosenthal Dir. ¶ 26.)

AWP provides a common standard to process millions of drug transactions. A common benchmark is useful because TPPs reimburse for thousands of drugs and services. As the independent court expert Professor Ernst Berndt, a healthcare economist from MIT, stated, AWP is “a convenient focal point metric for contractually specifying various reimbursements, and for efficiently adjudicating pharmacy transactions electronically.” (DX 1275, Berndt Report ¶ 23.)

The federal government’s Centers for Medicare and Medicaid Services (“CMS”) (and its predecessor the Health Care Finance Administration, or “HCFA”) do not regulate or set the AWPs, but have entrusted the pharmaceutical companies with the task of reporting the AWPs accurately to the publications. While CMS had the authority to conduct surveys to verify the acquisition costs of providers, it never did so. The TPPs also rely on the AWPs reported by the pharmaceutical companies to the publications.

Initially, AWP was the average price charged by wholesalers to providers, like doctors and pharmacies. It was derived from the markup charged by wholesalers over their actual acquisition cost, sometimes called the “Wholesale Acquisition Cost” or “WAC.” Historically, there was an industry-wide formulaic 20 or 25 percent markup between WAC and AWP. At some point, though, because of consolidation and competition among wholesalers, these standard markups on branded drugs no longer reflected actual wholesaler margins, which were reduced to 2 to 3 percent. Therefore, the actual average wholesale price charged by wholesalers to providers was much lower than the 20 or 25 percent markup over WAC.

Nonetheless, most manufacturers, including AstraZeneca, Schering-Plough, and J & J, continued to report AWPs to the publications based upon the historic formulaic 20 to 25 percent markup, rather than adjusting these prices to reflect the lower, true margins. These manufacturers knew that wholesalers were not actually charging these prices to providers, that the AWP was not a true average of prices charged by wholesalers, and that the “AWP” based on the formulaic 20 to 25 percent markup had become an anachronism. BMS emphasizes that it reported a Wholesale List Price (“WLP”) to the publications, rather than an AAYP, but it expected — and indeed directed 9 — that the publishing compendia would apply a standard markup to their WLPs to derive an AAYP. As such, BMS effectively controlled the AAYP published in the compendia. This formulaic markup has never been reduced to reflect actual market conditions.

B. Medicare Part B

Medicare is the largest insurer of physician-administered drugs. During the class period, there were approximately 450 covered drugs reimbursed by Medicare Part B. 10 Medicare Part B covers professional services, including those drugs that were “incident to” a physician’s services, drugs administered with durable medical equipment (“DME”), and drugs specifically covered by statute. These specialty drugs are typically administered by physicians in an office setting or in hospital outpatient departments, the latter being more expensive. Covered drugs also included some self-administered drugs.

For a Medicare Part B covered drug, 80 percent of the cost is paid for by the federal government, and 20 percent is paid for by whoever is responsible for the co-payment. See 42 U.S.C. § 1395Í. Many individual Medicare recipients have a private supplemental insurance policy that covers all or part of their 20 percent co-payment. In Massachusetts, these TPPs that provide supplemental insurance (sometimes called Medigap insurance) are members of Class 2.

Initially, reimbursement for prescription drugs under Part B in the Medicare program was on a “reasonable charge” basis. 56 Fed.Reg. 25,792 (June 5, 1991). (See also Bell T1 Aff. ¶ 78; Hartman Decl. ¶ 12.) Prior to 1992, Medicare’s carriers used the customary or prevailing charge among physicians in a geographic area. (See Bell T1 Aff. ¶78.) In June 1991, HCFA proposed changing reimbursement to the lower of 85 percent of AWP or estimated acquisition cost (“EAC”), as determined by HCFA through surveys. Based on comments from doctors that they could not procure many drugs at that level of reimbursement and that there were shortfalls in chemotherapy administration payments, HCFA backed off this proposal. (Id. ¶ 80.) Instead, it adopted reimbursement for PADs under Part B at the lower of AWP or EAC (plus an allowance for other costs), effective January 1, 1992. (Id. ¶ 81.) The EAC could be determined based on a survey of physicians or actual invoice prices paid by physicians for the drug. 56 Fed.Reg. 59,502 (Nov. 25, 1991). 11

Unfortunately, Medicare carriers did not conduct surveys of actual invoices, and took the shortcut of reimbursing based on AWP. On one occasion, when one carrier attempted to base reimbursement on physician acquisition costs, HCFA actually directed the carrier not to collect invoices from physicians in order to implement an acquisition cost survey. (Bell T1 Aff. ¶ 84.)

Effective January 1, 1998, pursuant to a statutory change, the Medicare regulations were amended so that the allowed amount would be based on the lower of the billed charge or 95 percent of AWP. See 42 C.F.R. § 405.517 (1999) (Department of Health and Human Services (“DHHS”) Regulations); see also 42 U.S.C. § 1395u(o) (Medicare statute). Significantly, there is no statutory or regulatory definition of AWP.

Up until 1998, multi-source drugs 12 were reimbursed at the lower of the estimated acquisition cost or the “median price for all sources of the generic form of the drug.” 56 Fed.Reg. at 59,621 (DX 1049). Since 1998, multi-source reimbursement has been set at the lower of the billed charge or 95 percent of an average wholesale price, defined to be the lesser of the median generic AWP and the lowest brand name product AWP. 42 C.F.R. § 405.517 (2003) (DX 1852); see 42 U.S.C. § 1395u(o).

C. Manipulating and Marketing the Spread

The use of AWP as an embedded pricing benchmark used by the federal government, state governments, 13 and private insurers alike created perverse incentives for the drug manufacturers and the physicians. Typically, a single-source drug 14 without therapeutic competition bore a predictable relationship to acquisition costs. (Bell T1 Aff. ¶ 6.) When a branded drug faced competition from a therapeutic equivalent, though, the drug manufacturer could manipulate the spread — the difference between the actual selling price and the AWP based reimbursement — to make the drugs more attractive to a physician. The manufacturer could then “market the spread” to the physician to increase sales and market share.

To fully understand the strategy of manipulating and marketing the spread, one needs to understand that physicians purchase drugs in essentially three ways. The first route is a direct sale from the manufacturer to the provider of physician-administered drugs. AstraZeneca’s Zola-dex is one example of this direct distribution chain. In these instances, the doctor purchases the pharmaceutical from the manufacturer and bills the TPP, making a profit on the difference between the acquisition cost and the reimbursement amount. When there is therapeutic or generic competition, some providers may be “preferred purchasers” from a manufacturer’s perspective and be able to acquire the pharmaceutical at a lower price, increasing the spread.

The second route is an “indirect” path, which involves a sale by the manufacturer to an intermediary such as a wholesaler or specialty distributor that provides services, like refrigeration and overnight delivery, needed to deliver perishable biologies and pharmaceuticals. (Bell T1 Aff. ¶ 10.) As Dr. Bell describes it:

Due to the intermediary mark-up, in instances of indirect distribution, the provider often purchases the pharmaceutical at a wholesale price that is higher than the price paid to the manufacturer by the specialty distributor or wholesaler. Some of the ultimate physician or hospital purchasers, however, may be preferred providers from the manufacturers’ perspective. The manufacturers compete for their business by offering a lower price for the pharmaceutical. The preferred provider receives such a lower price either through a chargeback (the provider purchases the product at the lower price from the specialty distributor or wholesaler who then “charges back” the amount of the price concession to the manufacturer) or a rebate (a price concession provided by the manufacturer directly to the provider).

(Id. ¶ 11.)

A third route of distribution involves the sale by a drug manufacturer to a specialty pharmacy which takes title to the pharmaceutical. The physician bills the TPP for administering the drug and the specialty pharmacy bills for supplying it. This last method of distribution was rare during the class period.

Whether doctors purchased the drugs directly from manufacturers or indirectly through wholesalers or specialty distributors, they had to seek reimbursement for the drugs from the TPPs and Medicare Part B. During the class period, TPPs typically did not use pharmacy benefit managers (“PBMs”) or consultants to negotiate drug prices with doctors and did not use formularies to control drug costs. Typically, the government and private insurers reimbursed for whatever drugs the doctor prescribed because of the serious nature of the diseases, especially cancer— a target of many of the drugs in this case. When a drug was a single-source pharmaceutical with no therapeutic competition, the doctor had little leverage over pricing. However, when there was therapeutic competition with another branded drug or a multi-source PAD, the physician had huge leverage over price because she con trolled the prescription of the drug and could choose which drug to administer.

Knowing that the doctor played this key role, the drug manufacturers launched sales forces directly into the doctors’ offices to negotiate drug pricing. Significantly, for this case, the terms of the contracts were kept confidential. 15 Rather than marketing simply the therapeutic qualities of the drug, many in the pharmaceutical sales force nimbly marketed the “spread” (also called the “margin” or “return to practice”) between what the doctor paid for the drug and what she would be reimbursed.

A pharmaceutical company manipulated the spread in two ways. Sometimes, it would raise the AWP reported to the publishing company, which would increase the spread. The manufacturer would either report an AWP or report a wholesale acquisition price, also called a direct price or list price, with the expectation that the publishing company would apply the formulaic markup to determine the AWP. Thus, all else being equal, physicians would have an incentive to select a product with a larger spread, even if the acquisition cost of that drug exceeded that of a therapeutic substitute. This was also a cost-free approach from the manufacturer’s point of view, as raising the AWP did not diminish its profit margin on the drug. Sometimes, the pharmaceutical manufacturer would increase the spread by providing the doctors with rebates, chargebacks, discounts, or free samples, which would decrease the actual acquisition cost of the drug. This approach, of course, results in less income to the manufacturer. A helpful metaphor is a pair of scissors: the spread could be increased by raising the top blade (the AWP) or lowering the bottom (the acquisition cost), or both. This “spread” existed regardless of whether the drug was reimbursed by Medicare or TPPs which, as discussed above, typically predicated contractually-based reimbursements on AWP.

During the class period, many doctors (particularly oncologists and urologists) eagerly entered the fray by exacting discounts and rebates from manufacturers. Many doctors purchased the drugs based on their “return to practice,” which means the profitability of the drug to the practice. Some physicians had significant marketing leverage because of the nature of their specialties, geographic location, and reputation. The doctor would pay a discounted price for the drugs, and seek the much higher reimbursement amount from the government and TPPs. Medicare required that the doctors charge the Medicare patients their 20 percent co-payment based on AWP. Despite knowing that their acquisition cost was much lower than the published AWP, the doctors charged patients a co-payment based on this inflated AWP. Doctors, however, could not always collect the entire co-payment from those patients who were unable to pay, and therefore they had to absorb that loss of reimbursement. Also, some doctors did not charge Medicare beneficiaries who could not afford the coinsurance payment. Many third-party payors also required beneficiaries to make percentage coinsurance payments.

Plaintiffs’ expert Professor Meredith Rosenthal, a health care economist who teaches at the Harvard School of Public Health, explains why there was no competitive pressure on doctors to lower the prices of drugs they charged to TPPs and patients:

As professionals, physicians command a large amount of technical and clinical information that is not accessible to patients or payers who can only imperfectly judge whether a physician is making appropriate diagnoses or treatment choices. Such asymmetric information about the nature of the service being delivered poses problems for price competition because patients and payers are unable to make “apples to apples” comparisons of providers — that is, to compare prices for services of equal value. This asymmetry of information, along with the high stakes involved (health), also leads to the importance of trust in physician-patient interactions. Trust, in turn, limits the substitutability of physicians from a given patient’s perspective. This perceived differentiation of physicians on the basis of trust weakens price competition, particularly when the patients concerned are acutely or chronically ill as are most recipients of the physician-administered drugs now at issue.

(Rosenthal Dir. ¶ 30.) Because there was little or no payor oversight of the physician’s choice of drug, doctors had no incentives to lower prices. (Hartman Decl. ¶¶ 108-09.)

Professor Rosenthal also described the economic incentives for pharmaceutical manufacturers:

For class drugs, the relevant measure of the financial consequence of choosing a particular physician-administered drug is the difference between the reimbursement for the drug, which is a function of AWP, and the acquisition cost of the drug to the physician or clinic. This means, as is true in other markets, that manufacturers can increase their market share by reducing the cost of their product to physicians through discounts or rebates. But the unique and perverse feature of this market is that pharmaceutical manufacturers can also increase market share through raising their AWP, since this list price is the basis for third-party reimbursement. Unlike offering big discounts to physicians, raising the AWP relative to the acquisition cost to the physician does not reduce profit margins on the drug in question.

(Rosenthal Dir. ¶ 33.)

The paradigm case of “marketing the spread” involves the marketing battle between Zoladex and Lupron, both used to treat prostate cancer. An AstraZeneca document sums up this motivation with respect to the sale of Zoladex: “As we have come to understand in our experience with Zoladex, urologists are motivated by economics.... Zeneca has learned that in order to compete in [a] market dominated by Medicare, there needs to be a compelling argument based on ‘total return to practice.’ ” (PX 14 at 7143.)

D. Cross-Subsidization

One oft-cited justification for inflating the AWP above true market costs is that reimbursement for the physician services rendered in administering the drugs often fell short of the costs of administration incurred by the physicians. (Bell T1 Aff. ¶ 7.) For example, CMS acknowledged that “Medicare payments related to the provision of chemotherapy drugs and clotting factors used to treat hemophilia and similar disorders are inadequate.” (DX 1090 at 0059.) Accordingly, doctors used the profit margin on the drugs to cross-subsidize administration fees and other risks (like spoilage) associated with physician-administered drugs. (Bell T1 Aff. ¶ 75.) At trial, there was no evidence about the extent of a shortfall in the costs of administration of the drugs in question in this litigation. Moreover, there was no evidence that any margin over the 20 to 25 percent industry-wide formula was needed to compensate doctors for their costs of administration for these drugs and risks like spoilage. Significantly, the pharmaceutical companies marketed the spread by demonstrating to the doctor that he would make a profit on the drug, not by demonstrating that the drug would cover costs of administration or other risks.

E. Patient, The Vulnerable Victim

Disturbingly, the patient was a vulnerable victim of this strategy of “marketing the spread” because when the AWP was raised, the Medicare patient was required to make a co-payment of 20 percent of the inflated AWP (or AWP-5% after 1998). The manufacturers understood well the harmful impact that publishing inflated AWPs had on the elderly cancer patient. For example, Mr. Buckanavage of AstraZ-eneca testified as follows:

THE COURT: Excuse me. Did you understand that Medicare beneficiaries paid 20 percent of AWP?

THE WITNESS: Yes. They paid 20 percent out of pocket.

THE COURT: So you understood that every time you raised AWP, they had to pay 20 percent of the increase?

THE WITNESS: Yes. Whenever we took a price increase, it would raise the copay and also raise the reimbursement.

(11/14/06 Tr. 13:2-10 (Buckanavage).)

Beneficiaries of private insurers also had to make higher percentage co-payments. Not surprisingly, none of the cancer patients who testified had ever heard of AWP, and they trusted their doctors to pick drugs for them based on effective treatment criteria, not profitability. {See, e.g., 11/07/06 Tr. 74:13-15, 75:20-76:1 (Choice); 11/07/06 Tr. 101:9-16, 107:23-25 (Hopkins).)

The pharmaceutical companies were aware of the political ramifications if the impact of raising AWPs on patients became publicly known. For example, on January 2, 1998, in response to an inquiry about a government report on drug reimbursement, Cathleen Dooley of J & J’s subsidiary OBI acknowledged in an email:

By law, the physician must bill the patient the remaining 20% copay.... This will be a sensitive issue because the physician is able to bill Medicare and the patient off of AWP; the patient’s 20% copay is higher than it would be if it was billed off of acquisition cost {public relations issue).

(PX 259 (emphasis added).) In addition, pharmaceutical manufacturers understood that this strategy had the effect of inducing doctors to prescribe a drug at least partly based on return to practice rather than just on the therapeutic quality of the drug.

The pharmaceutical companies made some attempts to deal with the public relations issue. Many manufacturers instituted programs to help patients make the co-payment. 16 Two companies eventually instituted internal ethical guidelines to ban marketing the spread. In January of 2001, BMS sent a memo to all U.S. Sales & Marketing Personnel advising that, “in accordance with its Code of Conduct, ... the spread should not be used as a promotional or marketing tool.” (PX 223.) Later that year, Ortho Biotech, a subsidiary of J & J, sent a memo to its sales force stating: “It is absolutely inappropriate to sell product based upon the difference between AWP and acquisition cost.” (DX 2767.) What was remarkable, though, was how few of the pharmaceutical witnesses at trial were concerned about the impact of an inflated AWP on old and sick people making co-payments based on a percentage of AWP. Indeed, from the vantage point of AstraZeneca’s sales team, they were actually assisting patients because Zoladex was cheaper than Lupron in treating prostate cancer.

F. Self-Administered Drugs

Some of the Medicare Part B drugs are self-administered and primarily dispensed by pharmacies. Examples are Temodar, a single-source drag manufactured by Scher-ing-Plough, and albuterol, a multi-source generic manufactured by Schering-Plough’s subsidiary Warrick. Retail pharmacies have little ability to determine which single-source drugs will be dispensed to a patient, because they must dispense whichever drug is prescribed by the physician. Pharmacies therefore receive few price concessions on single-source drugs.

With respect to generic drugs, though, pharmacists determine which manufacturer’s version of a multi-source drag will be sold. Generic manufacturers thus compete on price so that a pharmacy or pharmacy chain will stock their version of a generic drug. However, Medicare generally reimburses multi-source drugs at 95 percent of the median of the generic AWPs. In private contracts, TPPs typically impose a maximum allowable cost (“MAC”) or other limit to curtail costs. 17 Thus, in Class 3, TPP reimbursement for multi-source drugs is generally not calculated based on the drug’s AWP. Accordingly, no damages have been calculated for multi-source drugs in Class 3. 18

G. Knowledge in the Industry

At least since the start of the class period, the most knowledgeable industry insiders, like the larger TPPs, including Blue Cross/Blue Shield of Massachusetts (“BCBSMA”), the named plaintiff, came to understand that with respect to self-administered drugs, like pills, the AWP of the pill did not reflect the actual average price charged by wholesalers to retail pharmacists. Knowledge about the AWP of SADs was available in the industry largely because of the role of the PBMs, which represent TPPs in negotiating drug prices with pharmaceutical manufacturers to get discounts and rebates on SADs sold by pharmacies. 19 There were also commercial data services like IMS Health which published marketing data. With these SADs, institutions like TPPs exercised control over physician prescribing patterns through formularies, and secured discounts from manufacturers selling competing single-source and multi-source drugs. (Bell T1 Aff. ¶ 39.) Moreover, as Dr. Bell points out, some TPPs were vertically integrated, running staff model health maintenance organizations (“HMOs”) which purchased SADs. In this way, they learned that the AWP of the drug was not the price of acquisition.

Information about the pricing of physician-administered drugs was far more opaque. The contracts required that all pricing terms be kept confidential. PBMs were not involved, and there was no standard published commercial transaction data for PADs available to TPPs. As such, industry experts, TPPs, academics, and the government typically did not have information as to the price paid by the doctors to acquire the drugs. While some TPPs had staff model HMOs which purchased PADs, knowledge about discounts given to bulk purchasers like HMOs did not provide a transparent picture of average prices charged to other classes of trade like physicians or physician groups. {See Bell TI Aff., App. C (describing vertical integration in drug purchasing); see, e.g., DX 1630 (1993 Los Angeles Times article reporting that Rite Aid complained in 1993 that HMOs and other classes of trade received better prices than drugstores).)

H. Mega-Spreads

To recap, throughout the class period, most knowledgeable insiders understood that AWP did not reflect the average sales price to providers, but that it bore a formulaic relationship to WAC of a 20 to 25 percent markup. In addition, payors were aware there was some discounting from WAC. However, I find that in the early 1990’s, payors typically did not understand that there were mega-spreads far in excess of the formulaic markup for physician-administered drugs when there was competition between therapeutic equivalents or multi-source drugs. Indeed, the named plaintiff TPPs had no knowledge or expectation as to the size of the spreads available to physicians.

Plaintiffs’ expert, Dr. Raymond S. Hartman, a healthcare economist, testified that the marketplace had an expectation that AWP did not exceed the average sales price by more than 30 percent. 20 (Hartman Decl. ¶ 77(a)-(c).) After reviewing studies produced by government offices, as well as academic and popular publications between 1992 and 2004 for PADs, he concluded that “the publicly available survey evidence generally informing the government, policy makers, and industry participants about spreads on single-source physician administered drugs over much of the Damage Period suggested that the spreads were not excessive.” (Hartman Decl. ¶ 77(c); see also Hartman Rebuttal ¶¶ 46-47.)

By the mid-1990’s, information about the existence of mega-spreads began to seep into the marketplace. For example, on June 10, 1996, Barron’s published an article titled Hooked on Drugs: Why Do Insurers Pay Such Outrageous Prices for Pharmaceuticals? describing AWP as “Ain’t What’s Paid.” {See DX 2641.) It stated that for “many drugs, especially the growing number coming off patent and going generic, the drug providers actually pay wholesale prices that are 60%-90% below the so-called average wholesale price, or AWP, used in reimbursement claims.” (Id. at 15.) In Dr. Hartman’s calculations, these are spreads of 150% to 900%. The Barron’s article reports on a doctor having a plaque reading “This is the house that leucovorin 21 built.” (Id.) The article also included a chart showing spreads for a number of the drugs in this litigation. The chart listed Doxorubicin (Rubex) as having a spread of 72% off AWP (Hartman spread of 271% above average sales price) and Etoposide (Vepesid) with a spread of 76% off of AWP (Hartman spread of 316% above average sales price). (Id.) There was a growing sense among doctors, TPPs, and others that AWP stood for “ain’t what’s paid.”

Less sophisticated participants, like Taft-Hartley Plans, which are union benefit funds, still did not understand that AWP was not a true market average because there was so much misinformation in the market. For example, in promoting its AWP price data, First DataBank, one of the major publishers, stated that AWP “is the average wholesale price. That is, AWP is the average of the prices charged by national drug wholesalers for a given product (NDC).” (DX 1275, Berndt Rep. ¶ 78.) This information was available on the website of the American Society of Consultant Pharmacists as late as 2005.(M)

By 2001, there was a perfect storm of information that reflected the size of the spreads, largely because of the compelling information collected by the HHS Office of Inspector General (“OIG”). In addition, the press began to report on the rampant abuse of the AWP system. 22

I. The Government Pit Bull

Initially, the government’s concern about the accuracy of AWPs focused on self-administered drugs. A 1984 OIG report involving self-administered drugs stated:

AWP cannot be the best — or even an adequate — estimate of the prices providers generally are paying for drugs. AWP represents a list price and does not reflect several types of discounts, such as prompt payment discounts, total order discounts, end-of-year discounts and any other trade discounts, rebates, or free goods that do not appear on the pharmacists’ invoices.

(DX 1039 at 10,206; see also DX 1985 at 20,255 (HCFA administrative decision in 1989 stating “AWP was not the price generally and currently paid by providers”).)

In 1992, the OIG began to focus on the shortcomings of AWP as a reimbursement benchmark for Medicare physician-administered drugs. The report stated: “Our review of invoices revealed that the 13 chemotherapy drugs can be purchased at amounts below AWP.” (DX 1053 at 5.) The OIG listed discounts off of AWP on a number of PADs, including, for example, discounts off of Doxorubicin (Rubex) of 56% to 59%, which are equivalent to spreads of 127% to 144%. (Id. at App. III.) Some of the drugs analyzed are at issue in this case, including Bleomycin (Blenoxane), Cyclophosphamide (Cytoxan), Doxorubicin (Rubex), and Etoposide (Vepesid). The OIG concluded: “AWP is not a reliable indicator of the cost of a drug to physicians.” (Id. at 11.) In 1996, it issued another report, examining the possibility of using the Medicaid Best Price rebating approach as a way to save money in the Medicare program. (See DX 1062 at 7.) Again, it flagged excessive pricing for Zoladex, Paraplatin, Taxol, Vepe-sid, Rubex, and Etopophos although it did not calculate a spread. (See id.)

Congressional committees also began to examine problems with the AWP system. In June 1997, prior to the passage of the Balanced Budget Act of 1997 (“BBA”), which inserted AWP into the Medicare statute, the Committee on the Budget of the House of Representatives issued a report stating:

The Inspector General for the Department of Health and Human Services has found evidence that over the past several years Medicare has paid significantly more for drugs and biologicals than physicians and pharmacists pay to acquire such pharmaceuticals. For example, the Office of Inspector General repoHs that Medicare reimbursement for the top 10 oncology drugs ranges from 20 percent to nearly 1000 percent per dosage more than acquisition costs.

(DX 1071 at 1354 (emphasis added).)

In December 1997, shortly after Congress decided to lower drug reimbursement to 95% of AWP, the OIG issued another report: “[Published AWPs ... bear little or no resemblance to actual wholesale prices that are available to the physician and supplier communities that bill for these drugs.... We believe that the 5 percent reduction [off of AWP] is not a large enough decrease.... [W]e’ve identified [spreads of] 11 to 900 percent....” (DX 1075 at ii-iii.)

At about the same time, President Clinton referred to AWP as a “sticker price” in his nationwide radio address: “Sometimes the waste and abuses aren’t even illegal; they’re just embedded in the practices of the system.... [T]hese overpayments occur because Medicare reimburses doctors according to the published average wholesale price, the so-called sticker price, for drugs.” (DX 1074 at 2033-34.)

In 1999, Donna E. Shalala, the Secretary of the Department of Health and Human Services reported to Congress:

For the past 13 years, the Office of Inspector General (OIG) has issued a series of reports that consistently show a finding that the Medicare program overpays for the drugs and biologicals it covers. This is because most drugs can be obtained at a much lower cost than the AWP. To address this problem, the President’s 1997 budget contained a legislative proposal that would have based payment on the lower of the billed charge or the actual acquisition cost (AAC) for the drug of the physician or supplier billing Medicare. However, as discussed above, in the BBA, Congress rejected this proposal in favor of the current rule, which is to pay based on the lower of the billed charge, or 95 percent of AWP.

(DX 1080 at 1-2.) She pointed out that “AWP is not a well-defined concept nor is it regulated in any way,” and concluded that AWP bore “no consistent or predictable relationship to the prices actually paid by physicians and suppliers to drug wholesalers in the marketplace.” (Id. at 2, 8.) A 1999 Medicare Bulletin flagged this awareness that AWP is “not a true discounted price and, therefore, does not reflect the cost to the physician or supplier rendering the drug to the Medicare beneficiary.” (DX 1166.)

Professor Ernst Berndt, the Court’s independent expert, explained that governmental inertia in fixing the problem of the inaccuracy of the AWP can be explained by the fact that drug expenditures were not a large portion of healthcare costs. (DX 1275, Berndt Rep. ¶ 187.) This inertia reflected the principle of “the importance of being unimportant.” (Id.)

Between 1998 and 2002, though, there was rapid growth in Medicare Part B expenditures, particularly with respect to amounts paid as drug expenses to oncologists and urologists due to drug product price increases at the manufacturer level and increases in utilization. According to a report cited by Professor Rosenthal,

The vast majority (77%) of the Medicare part B drug expense is paid to oncologists and urologists. Oncologist-based drug expenditures grew from $1.2 billion in 1998 to $3.8 billion in 2002 with the spending growth from 2001-2002 at 41 percent. The spending on drugs under Medicare Part B is highly concentrated with 7 of the approximately 450 drugs accounting for 49 percent of the spending ($4.0 billion out of the $8.4 billion).

(Rosenthal Dir. ¶ 22.)

In 2000, the Department of Justice (“DOJ”) compiled and reported actual average wholesale prices — the prices at which the wholesaler sells the drugs — for approximately 400 National Drug Codes (“NDCs”) 23 covered by Medicare. (See DX 1091 at 1.) The DOJ indicated that “these are more accurate wholesale prices for these drugs.” (Id.) Plaintiffs, in their complaint, calculated the spread between the DOJ’s actual AWP and the published AWP for many of the drugs in this litigation. (See Compl. ¶¶ 65, 112.) Several of those spreads exceed 100%. (See id.)

HCFA again attempted to administratively change reimbursement from an AWP basis to the cost-based prices calculated by the DOJ, (see DX 1091), but various members of Congress urged it to reconsider, primarily because of concerns that oncologists were being underpaid in administering their services, and that this underpayment needed to be corrected before reducing reimbursement. (See DX 1085 (letter to Secretary Shalala from Congress members); DX 1086 (same); DX 1090 (HCFA letter announcing the change).) The senators seemed particularly perturbed because “the Department’s unilateral declaration of a new definition of AWP, with no regulatory process, is inappropriate.” (DX 1086 at 2.)

After HCFA retracted its authorization regarding the use of these new AWPs in November 2000, Congress passed the Benefit Improvement and Protection Act of 2000, which prohibited the Secretary of HHS from implementing any payment reduction for drugs until the Government Accountability Office (“GAO”) prepared, and the Secretary reviewed, a report on revised payment methodologies for drugs. (Bell T1 Aff. ¶ 89.) In September 2001, the GAO released its report which found that: (1) the average discount from AWP for physician-administered drugs ranged from 13% to 34%, equating to “spreads” of 15% to 52%; (2) two physician-administered drugs had discounts of 65% and 86%, equating to “spreads” of 186% and 614%; and (3) two drugs used with durable medical equipment had discounts of 78% and 85%, equating to “spreads” of 355% and 567%. (Id.)

Like a pit bull, OIG pursued the AWP issue. In 2003 it issued a Compliance Program Guidance for pharmaceutical manufacturers, which admonished: If a pharmaceutical manufacturer purposefully manipulates the AWP to increase its customers’, profits by increasing the amount the federal health care programs reimburse its customers, the anti-kickback statute is implicated. Unlike bona fide discounts, which transfer remuneration from a seller to a buyer, manipulation of the AWP transfers remuneration to a seller’s immediate customer from a subsequent purchaser (the federal or state government). Under the anti-kickback statute, offering remuneration to a purchaser or referral source is improper if one purpose is to induce the purchase or referral of program business. In other words, it is illegal for a manufacturer knowingly to establish or inappropriately maintain a particular AWP if one purpose is to manipulate the “spread” to induce customers to purchase its product.

In the light of this risk, we recommend that manufacturers review their AWP reporting practices and methodology to confirm that marketing considerations do not influence the process. Furthermore, manufacturers should review their marketing practices. The conjunction of manipulation of the AWP to induce customers to purchase a product with active marketing of the spread is strong evidence of the unlawful intent necessary to trigger the anti-kickback statute. Active marketing of the spread includes, for example, sales representatives promoting the spread as a reason to purchase the product or guaranteeing a certain profit or spread in exchange for the purchase of a product. 68 Fed.Reg. 23,737 (May 5, 2003) (emphasis added) (PX 4016). This was the first written guidance from the government addressing marketing practices related to AWP.

J. The Demise of AWP as Government Pricing Benchmark

Finally, ten years after the OIG first reported the deficiencies in using unregulated AWPs as reported by the pharmaceutical industry as the benchmark for Medicare reimbursement, Congress took action with the passage of the Medicare Prescription Drug, Improvement and Modernization Act of 2003 (“MMA”) with an effective date of December 8, 2003. See Pub.L. No. 108-173, 117 Stat. 2066. The MMA provided for a shift from 95% of AWP to 85% of AWP in 2004 and then to 106% of “average sales price” 24 in 2005. See 42 U.S.C. § 1395u(o). The class period ends the day the MMA went into effect. On April 6, 2004, CMS issued a detailed interim rule on how manufacturers should calculate ASP data on Medicare Part B drugs. The final rule was issued on September 16, 2004. Reimbursement based on ASPs took three years to implement because the government not only had to determine the methodology for calculating the ASP, but also had to ascertain the amount needed to increase service fees for oncologists and other physicians administering drugs.

Even with the increase in administration fees paid to doctors, Medicare has had overall cost savings from the decrease in drug expenditures for Zoladex, Taxol, Remicade, Procrit, and albuterol. 25 The total reimbursement for a typical administration of Zoladex, including both product cost and administration fee, fell from $451.56 in 2002 to $226.48 in 2005 under the new ASP system. (PX 4069.) Looking at those same two years, the cost of a typical dose of Taxol dropped from $1785.16 to $428.07 (PX 4070), Remicade from $2,035.15 to $1,703.09 (PX 4071), Pro-crit from $150.65 to $131.96 (PX 4072), and albuterol from $109.74 to $71.63 (PX 4095). Medicare Reimbursement under the MMA

_2002_2005 Savings

Zoladex $ 451.56 $ 226.48 49.9%

Taxol $1785.16 $ 428.07 76.0%

Remicade $2035.15 $1703.09 16.3%

Procrit $ 150.65 $ 131.96 12.4%

albuterol $ 109.74 $ 71.68 34.7%

(See PX 4069; PX 4070; PX 4071; PX 4072; PX 4095.)

Dr. Rosenthal’s review of the hard data, as laid out above for several drugs in this case, shows that “the overall dollar reimbursement declined for each drug.” (Ro-senthal Rebuttal ¶ 15.)

Despite the Sturm und Drang from some medical providers, doctors have not generally shifted their patients to a more expensive hospital setting. According to a July 13, 2006 MedPAC report, the MMA “payment changes did not affect beneficiary access to chemotherapy services.” (PX 4019 at 5.) “Physicians provided more chemotherapy services and more Medicare beneficiaries received services in 2005 than in 2004.” (Id.)

K. Stuck

To date, TPPs have generally not shifted away from the AWP benchmark despite likely cost savings. On February 7, 2004, BCBSMA, the class plaintiff, did a study demonstrating that a shift would save them $6,010,576 even with the increase in administration fees. (DX 990 at 12.) As reasons for reform, BCBSMA states:

— Physicians benefit from the “spread” between AWP and acquisition cost creating an overpayment for drugs and costs for Medicare

— According to GAO and CMS, in 2001 Medicare overpaid Part B drugs by over $1 billion.

— In 2002 oncologists collected approximately $600 million in overpay-ments.

— Patients who pay a coinsurance are adversely affected by the inflated AWP.

(Id.) But BCBSMA was afraid that its network of doctors would rebel at lower reimbursement rates: even if service fees went up, doctor profitability would go down. Cautious and fearful, BCBSMA decided not to follow Medicare and to continue using AWP as the pricing benchmark. Indeed, it recently decided to employ AWP in its fee schedule with hospitals, a different class of trade. As of the date of the trial, only a few TPPs have shifted to an ASP system. United Healthcare moved to ASP-based pricing, but chose to reimburse at ASP plus 12 percent for oncologists only. (Bell T1 Aff. ¶ 37.)

Defendants highlight the failure of TPPs to react when the mega-spreads became well known. It is hard to understand why the TPPs did not decrease the percentage off AWP during the five years after 2001 when knowledgeable TPPs typically understood that there were mega-spreads between cost and reimbursement prices, in excess of any reasonable compensation for service fees or risks (like spoilage or shelf life). While TPPs likely shut their eyes to the 20 to 25 percent spread to permit cross-subsidization of physician costs, the new Medicare structure provided an alternative reimbursement scheme. (See Bell T1 Aff. ¶¶ 77-92 (explaining Medicare’s assumption that drug reimbursement would cross-subsidize other costs).) This diffidence can no longer plausibly be explained by the “importance of being unimportant,” since drug costs had increased substantially over the class period.

Remarkably, BCBSMA, the behemoth insurer in the Massachusetts market, and other large TPPs, were not proactive in adjusting to cost data once Medicare did the legwork for them in devising more reasonable drug pricing and service fees. Medicare provided the TPPs with cover, by insulating them from protests by the network of providers. Dr. Rosenthal explains this inertia as “stickiness,” which is the economists’ label for the common sense phenomenon that it is much harder to decrease reimbursement rates than to increase them. TPPs were worried that they risked either losing network providers or pushing patients into the more expensive hospital setting if they pressed for lower AWPs on individual specialty drugs.

L. The Plaintiffs/TPPs

1. Blue Cross/Blue Shield: Class 2 and Class 3 Representative

Plaintiff BCBSMA, a class representative for both Class 2 and Class 3, provides health coverage for approximately 2.5 million lives in the State of Massachusetts. (11/7/06 Tr. 147:2-3 (DeVaux).) BCBSMA is currently the state’s largest health insurance company with over 4,000 employees and covering approximately 46% of the covered lives in Massachusetts. In 2005, BCBSMA paid $9.4 billion in claims. It has made payments for single-source drugs manufactured by each defendant in both Class 2 and Class 3, and in the case of multi-source drugs, has purchased a drug with a J-code 26 that matches a J-code of a drug manufactured by a defendant. (See PX 4012; Mulrey Aff. ¶¶ 20-22.)

Currently, BCBSMA primarily uses a fee-for-service arrangement with its physicians and physician groups, though it has also used capitated 27 and other risk sharing arrangements. (Mulrey Aff. ¶ 5.) Under the fee-for-service arrangement, BCBSMA establishes a fee schedule that governs provider reimbursement for the purchase and administration of drugs. (Devaux Aff. ¶ 7.) These fee schedules are contained in the contracts between BCBSMA and the physician or physicians group. (DeVaux Trial Aff. ¶ 8.)

From 1991 until 1995, the reimbursement amounts in the fee schedules were based on the usual and customary charge for the particular drug. (Mulrey Aff. ¶ 10.) Therefore, BCBSMA has no claim for Class 3 damages prior to 1995. In 1995, BCBSMA first began using AWP as a basis for reimbursement to physicians for PADs. (Id. ¶ 11.) From 1995 to 1998, BCBSMA used 100% of AWP as the basis for reimbursement of PADs, and in 1998, BCBSMA moved to using 95% of AWP. (Id.) Until 2005, BCBSMA obtained the AWP it used for these fee schedules from Medicare. (Id. ¶ 12.) Fee updates would be retrieved from Medicare/NHIC websites or the Medicare B Resource guide. (Id.)

In the Medicare context, BCBSMA offers Medex plans which are “MediGap” plans that cover a Medicare Part B beneficiary’s 20 percent co-payment. (Arruda Aff. ¶ 3.) BCBSMA has approximately 160,000 individuals who purchase these Medex plans directly, and approximately 85,000 individuals who receive the coverage through a group sponsored plan. (Id. ¶ 4.)

At trial, Kenneth Arruda, a BCBSMA marketing executive, explained that Medex premiums were set based on the prior two years’ claims experience by calculating projected benefit costs, expected administrative expenses, and a contribution to reserves. (11/08/06 Tr. 133:6-22, 154:25-155:3 (Arruda).) The contribution to reserves is an additional 2.5% of the premium that is added to cover shortfalls from miscalculation, increased utilization due to mass illness, previously unreported claims, and other unforeseen needs. (Id. 135:18-136:19, 137:11-14.) Mr. Arruda explained the need for the contribution to reserves: “This is generally a risky business because [we] are covering people who are over age 65 who have severe and in some cases catastrophic health care needs.” (Id. 136:17-19.)

Up until 1996, BCBSMA owned a staff model HMO, Medical West, Inc. (Coneys Aff. ¶ 13.) Medical West, Inc., comprised of Medical East and Medical West Health Plans, had several clinics located throughout Massachusetts. Medical West, Inc. operated an in-house pharmacy that provided PADs to physicians. (Id. ¶ 8.) Medical West, Inc. negotiated directly with drug manufacturers to purchase drugs for this pharmacy. (Curran Aff. ¶ 15.) According to the testimony of defense expert Eric Gaier, the staff model HMO purchased drugs at discounts as high as 92.6% below AWP, which under Dr. Hartman’s calculation is a spread of over 1200%. (Gaier Aff. ¶ 34; see DX 1389-DX 1403.) 28

One fact dispute is when and whether BCBSMA, the parent, knew about the spreads in PADs and other drugs reimbursed through Medicare Part B. The timing of this knowledge is significant to the statute of limitations and other issues. The level of knowledge among BCBSMA employees was uneven. Remember, BCBSMA did not reimburse physicians based on AWP until 1995. Michael T. Mulrey began working in 1987 as a Senior Financial Analyst for Medical East and Medical West, and worked from 1994 to 1998 in the Provider Contracting area as a Senior Contract Analyst. (Mulrey Aff. ¶ 3.) He believed until 2004 that AWP was the price at which, on average, physicians were paying to purchase PADs. (Id. ¶ 14.) Deborah Devaux, who was Senior Vice President for Health Care Contract Management at BCBSMA, had a long history in the area of health care reimbursement. She said:

I have personally been involved in negotiating such contracts, and in more recent years in supervising staff who negotiate such contracts. It is my ex perience that reimbursement for physician administered drugs is typically not a point of negotiation between BCBSMA and physicians. BCBSMA establishes, and periodically updates, fee schedules that govern the amount that any physician or physician organization in the BCBSMA network will be reimbursed both for the physician administered drugs and for the administration fee associated with administration of those drugs to insureds.

(DeVaux Aff. ¶ 7). While she was aware that some physicians or physician groups with large practices who could buy in greater quantities may have paid less, and other doctors may have paid more to acquire these drugs, she was not aware of spreads of more than 30% and did not know that pharmaceutical companies were inflating the AWP to increase their market share for these drugs at the expense of payors. (Id. ¶ 16.) She also didn’t know which manufacturers were involved in marketing the spread. (Id.)

Edward S. Curran, Jr., a key defense witness, worked at BCBSMA from 1988 to 1992 as the Director of Pharmacy. (Cur-ran Aff. ¶ 1.) His primary responsibilities included negotiating with drug manufacturers for rebates payable to BCBSMA in consideration for formulary status in the area of SADs. (Id. ¶ 14.) He also negotiated with drug manufacturers for rebates and discounts for use at the staff model HMO sites. (Id. ¶ 15.) He was a signatory on these contracts as finalized. He worked closely with the staff model HMO Medical East/West sites and knew that rebates varied widely from drug to drug. (Id. ¶ 16.) However, Curran testified he had “no clue” about how physician-administered drugs were purchased by doctors. (11/13/06 Tr. 36:24-37:5 (Curran).) He did not know that pharmaceutical manufacturers were marketing the spread so that doctors could make a profit.

Curran did say he had a role in purchasing PADs for the HMO but has no memory of any specific drugs. According to Maureen Coneys, the Medical East/West HMOs independently contracted with manufacturers to purchase PADs. (Coneys Aff. ¶ 10.) Even though Curran likely had some knowledge of negotiations involving rebates and discounts for PADs at the HMO, which would be attributable to BCBSMA, it was likely of quite limited significance to the AWP issues in this litigation because BCBSMA did not shift to AWP until 1995 after he left as Director of Pharmacy in 1992; thus, any knowledge he gained about the rebates available to the HMO sites did not give BCBSMA material information about the spreads between AWP and ASP available to private physicians in the network, or about marketing the spread to individual physicians or physician groups.

The HMOs were sold in 1997, two years after BCBSMA instituted AWP pricing. Therefore, any knowledge about the spread involving AWP gleaned by ongoing communications between the parent and the subsidiary likely existed only for a two year time period. While there is evidence that there were discussions with the parent, there is little evidence that detailed information about spreads was conveyed to the parent.

It is true that other employees (Gary Shramek and John Killion who worked at BCBSMA) knew that AWPs did not reflect acquisition costs, but they did not have detailed information about the size of the spreads until the late 1990’s. Gary Shra-mek, who was employed by BCBSMA as a Pharmacy Program Director from September 1999 to October 2002, became personally familiar with acquisition costs of 60 percent off AWP on PADs because of rebates and discounts from manufacturers and discussed this information with other BCBSMA employees. (Shramek Aff. ¶ 10.) In June 2002, he learned that U.S. Oncology (a large buyer) could purchase drugs for 18%-40% off AWP which is equivalent to spreads of 22%-67%. (See DX 1148 at 17376.) Several other BCBSMA employees also testified that sometime in the mid-to-late 1990’s they became aware that AWP was not a true average wholesale price. (See, e.g., Fox Dep. 126:16-127:10 (explaining his understanding that AWP was a “sticker price”); Fanale Dep. 84:13-86:13 (acknowledging that he knew doctors earned a profit on the drugs); Killion Dep. 119:9-22 (stating his knowledge that AWP was an “artificial price”).)

BCBSMA had limited knowledge of the spreads by the mid-1990’s. (See, e.g., DX 1979 at 30343 (minutes from the 1994 BCBSMA tri-regional carrier Medical Directors meeting stating in regard to a particular drug that BCBSMA “can’t rely on Red Book b/c physician’s are generating huge profit”); DX 1980 at 30378 (minutes from the May 1996 BCBS Technology Advisory Committee meeting stating that AWP is “grossly inflated”).) Beginning in 1999, BCBSMA employees understood that oncologists were generally making big money off of chemotherapy drugs, but BCBSMA was not able to determine the exact discount off of AWP that these oncologists were receiving. (See DX 1020 at 0066 (“We are not able to determine the exact discount off of AWP that MASCO is receiving, however, our contacts in the pharmacy business indicate drug companies offer substantial discounts to increase their market share. It appears that the physicians at MASCO are making money off of the drugs (we pay 95% of AWP, they buy the drugs for less), and are threatening to stop administering drugs in their office in order to keep reimbursement up.”).) Concerned about holding the “patient hostage,” BCBSMA explored different alternatives, like looking at the data as to whether the hospital setting was different and whether another method of purchasing the drugs and shipping them to the doctor should be explored.

I find that at least by 1999, employees at BCBSMA actually understood that AWP was not a real average and that doctors were receiving large discounts. However, for the most part, they still did not have any detailed knowledge on a drug-by-drug basis of the extent of the spreads.

2. Pipefitters: Class 3 Representative

Plaintiff Pipefitters Local 537 Trust Fund (“Pipefitters”) is a Class 3 representative. Pipefitters is a Taft-Hartley 29 multi-employer trust fund that provides health and welfare coverage for the local members of the Pipefitters union. Members of the Pipefitters union are tradesmen and tradeswomen who work on building systems. (Hannaford Aff. ¶ 9.) The small staff of the Pipefitters Trust Fund consists of six employees, including the Fund Administrator. (Id. ¶ 6.)

The Fund provides major medical benefits, including prescription drug benefits, to all union members and their eligible dependents. (Id. ¶ 7.) Currently there are approximately 4,600 individuals, both union members and their dependents, who receive major medical benefits, including prescription drug benefits, through the Pi-pefitters Fund. (Id. ¶ 8.) In general, the Pipefitters Fund covers 90 percent of all costs associated with treatment of its members, including the cost of pharmaceuticals. (Id. ¶ 10.)

Since approximately 1979, Pipefitters has contracted with BCBSMA to administer major medical benefits, including coverage for all prescription benefits, provided to members. (Hannaford Aff. ¶ 11.) Pipefitters uses this arrangement because it allows the Fund to obtain the “bargaining power” that BCBSMA has with doctors. (11/6/06 Tr. 177:2-5 (Hannaford).) Pipefitters has no ability to negotiate directly with providers and therefore is dependent on BCBSMA for its information on issues relating to prescription drug coverage. (Id. 184:5-7.) Pipefitters was aware that BCBSMA contracted to pay providers 95% of AWP for physician administered drugs. (Id. 167:11-168:2.)

The financial arrangement between Pi-pefitters and BCBSMA is “cost plus,” meaning that BCBSMA charges Pipefit-ters whatever it pays for the particular service or pharmaceutical plus an administrative fee. (Hannaford Aff. ¶ 11.) In this way, the Pipefitters Fund is fully responsible for all costs associated with benefits provided to its members. Based on claims data provided from BCBSMA, Pipefitters has paid for drugs manufactured by As-traZeneca, BMS, and J & J. (Id. ¶ 12; see PX 4012.) In the case of Schering-Plough’s multi-source albuterol, Pipefitters has purchased a drug with a J-code matching that of Schering-Plough’s products. (See PX 4012.) According to Charles Han-naford, the Fund Administrator, Pipefit-ters was not aware that AWP was not an average price, had no knowledge of any government studies, and did not know about the practice of marketing the spread. (Hannaford Aff. ¶¶ 13,16).

3. Sheet Metal Workers: Class 2 Representative

Plaintiff Sheet Metal Workers National Health Fund (“Sheet Metal Workers”), a Taft-Hartley multi-employer fund, is a Class 2 representative. Sheet Metal Workers offers a Supplemental Medicare Wraparound Plus program for over 15,000 retirees and covered beneficiaries. In this program, its payments are directly tied to what Medicare pays, covering 20 percent of Medicare’s allowable amount. (Randle Rev. Aff. ¶ 4.) Sheet Metal Workers believed AWP was an actual average of prices and did not know of spread marketing or any government studies about the spread. (Faulkner Rev. Aff. ¶¶ 6-14.)

Sheet Metal Workers employs a third-party administrator, Southern Benefits Administrators (“SBA”), to handle claims and to act both as a third-party administrator and as a consultant to advise the Fund on issues relating to healthcare. (11/6/06 Tr. 204:19-22 (Randle).) Sheet Metal Workers has employed SBA since 1996 to negotiate, contract for, and administer health benefits for its active and retired workers. (Id. 205:16-206:8.) Sheet Metal Workers relies on SBA to advise it on providing benefits to its members at the best possible price. (Id. 207:16-208:3.) Sheet Metal Workers has paid reimbursements for at least one of each of defendants’ drugs. 30 (See PX 4012.) Plaintiffs have not identified any individual Class 3 members.

M. Defendants

1. AstraZeneca

AstraZeneca 31 manufactures and sells Zoladex, an injectable physieian-adminis- tered drug primarily used to treat prostate cancer. (Black Decl. ¶ 9.) Zoladex, the only AstraZeneca drug at issue in this trial, is a medical alternative to surgical castration. Typically, Zoladex is administered by a medical professional in the abdomen once a month or once every three months.

Launched in January 1990, Zoladex has been a single-source drug throughout the class period. Since its launch, however, Zoladex has been in direct competition with Lupron, manufactured by TAP Pharmaceuticals. Lupron is also injected by a physician. Although the method of injection differs, many physicians view Lupron and Zoladex as therapeutically equivalent. (Freeberry Dep. 26:19-27:6.)

AstraZeneca provided a WAC 32 and a corresponding AWP for Zoladex to First DataBank and Redbook. AstraZeneca’s suggested AWPs for Zoladex were 25% higher than WAC. This relationship remained constant over the class period. (Gould Decl. ¶ 8; Black Deck ¶ 16.) As-traZeneca effectively controlled the AWPs for its drugs.

AstraZeneca’s pricing decisions for Zola-dex were driven by the competitive market for Lupron and Zoladex. At launch, As-traZeneca set the WAC for Zoladex at $255, approximately $75 less per injection than Lupron. (See Gould Deck ¶ 9, Fig. 2.) AstraZeneca periodically increased the WAC for Zoladex, although for some time the company had a policy to keep the average WAC price increase of its products below the rate of inflation. (DX 2119, at AZ0049325; Black Deck ¶ 5; 11/28/06 Tr. 9:13-21 (Milbauer).) The average annual price increase for Zoladex was 2.6%, whereas the average increase in Lupron was 4.1% over the same time period. (Gould Deck ¶ 10.) Thus, Lupron always had a higher WAC and AWP than Zola-dex. (See Gould Deck ¶ 9, Fig. 2; 12/04/06 Tr. 68:23-69:2 (Gould).) As a result, patients, the Medicare program, and private insurers paid less when Zoladex was administered instead of Lupron. Reimbursement was also less when Medicare Part B carriers used a least costly alternative (“LCA”) policy, whereby claims for Lu-pron and Zoladex were reimbursed based on the AWP for the less costly of the two. (DX 2075 at I.) Since Zoladex had the lower AWP, under the LCA the Zoladex AWP was used for reimbursement of both products. 33 By 1999, the majority of Medicare Part B carriers had implemented a LCA policy. (Rosenthal Deck ¶ 47.)

From 1990 to 1993, AstraZeneca sold Zoladex directly to physicians and other purchasers at WAC, offering only a standard 2% prompt pay discount. (Milbauer Deck 1127; DX 2078.) Despite having a lower cost, Zoladex was unable to gain market share from Lupron, the market leader, because the AWP-based reimbursement system created a financial incentive for physicians to choose higher priced products for their Medicare patients. (11/28/06 Tr. 14:1-10 (Milbauer); Milbauer Deck ¶¶ 29-31; Black Deck ¶¶ 17-18; 11/14/06 Tr. 11:1-12:8, 36:11-18 (Buckana-vage); PX 14 at AZ0237143; PX 119 at AZ0010297.) Physicians could earn more income on Lupron because while both drugs published an AWP that was a 25 percent markup over WAC, 25 percent of a higher price created a larger absolute dollar spread for physicians. AstraZeneca expressed frustration with this dynamic, noting that “[o]ur campaigns to grow ZO-LADEX sales based on product attributes and somewhat straightforward pricing strategies have continually been thwarted by TAP response 34 as well as the method used by Medicare to reimburse for LhRh agonists.” 35 (PX 14 at AZ0237143.)

AstraZeneca faced a difficult competitive situation: find a way to compete with TAP, or see sales of Zoladex continue to languish. The company believed that “in order to compete in [a] market dominated by Medicare, there needs to be a compelling argument based on ‘total return to practice.’” (PX 14 at AZ0237143.) A 1995 Pricing Strategy memo explained:

Return to Practice is enhanced by widening the margin between the published price and the acquisition cost. This can be accomplished through several pricing manipulations:

1) Increase the AWP

2) Decrease the acquisition cost relative to the AWP, or

3) Both 1 and 2.

In order to maximize the Return to Practice, and to maximize our competitive position, it is recommended that we exercise option # 3 from above....

(PX 133 at AZ0080409; see also PX 19 at AZ0021763 (recommending an increase in AWP and additional discounts).) Thus, AstraZeneca chose to begin offering discounts to its physicians, while continuing to make increases in the WAC price and the corresponding published AWP. (See PX 4030 at ¶ 40, Fig. 3.) AstraZeneca knew that its AWP was a fictitious and artificial number, (Freeberry Dep. 168:6-20, 172:19-173:8.), but felt no need to correct its reported price because it was standard industry practice to leave the AWP at 25 percent above WAC. (Black Decl. ¶ 16.)

Furthermore, AstraZeneca rationalized that the leveling of the playing field between Zoladex and Lupron resulted in lower costs to patients and the healthcare system when physicians switched to using the lower priced drug, Zoladex. (11/28/06 Tr. 17:13-19 (Milbauer); Gould Decl. ¶ 42-44; Black Decl. ¶ 24.) For example, in 1996 Zoladex was priced $112.60 less per dose than Lupron, saving patients and the healthcare system $22.52 and $90.08 per dose, respectively, if they used Zoladex rather than Lupron. (PX 19 at AZ0021764.) AstraZeneca trumpets this cost savings to Medicare, noting that their economist estimated that the shift in market share between Lupron and Zoladex from 1991 to 2002 saved $129 million in patient co-payments and $516 million in Medicare payments. (Gould Decl. ¶¶ 42-44, fig. 12; 12/04/06 Tr. 100:19-102:21 (Gould).)

The reported AWP for Zoladex, however, was drifting farther and farther away from the actual selling price of the drug. In 1995 the spread rose to over 40% and continued rising steadily to reach over 140% in 2002. (PX 4028.) During that year, the AWP for a 3.6 mg dose of Zola- dex was $469.99, while the ASP 36 was only $194.62. 37 (Id.; Hartman Decl. Attach. G.l.b.)

Despite understanding that patients and payors were paying for Zoladex based upon these inflated AWPs, AstraZeneca seemed unconcerned. Alan Milbauer, As-traZeneca’s VP of Public Affairs, acknowledged that “yes, the reimbursements went up, but it was overall less cost to the health care system and less cost to the patient. So I actually felt good about that.” (11/28/06 Tr. 26:16-25 (Milbauer).)

In conjunction with increasing the spread, AstraZeneca began marketing Zo-ladex based upon the return to practice that physicians could earn. (See Chen Dep. 126:17-21.) Sales representatives sent a letter to potential accounts encouraging them to switch to Zoladex based on the current AWPs, cost to physicians, and the resulting return to practice in relation to . Lupron. (See PX 38 at AZ105880.) The letter emphasized that switching to Zoladex “could significantly increase your profits.” (Id.) A section titled “DO THE MATH!” then explained exactly how to calculate the “Return to Practice.” (Id.) Zoladex sales representatives provided physicians’ offices with information showing “how much money the doctor or office would save purchasing one drug over the other.” (Bowman Dep. 57:1-10, 58:2-9.) Sales representatives were also using spreadsheets on their sales calls that demonstrated the spread and compared the “Annual Return to Practice” for Zoladex and Lupron. (See, e.g., PX 33 (email with attached spreadsheet); Bowman Dep. 53:4-14 (discussing charts used to compare return to practice).)

In the course of these actions, there was some concern at AstraZeneca that this spread marketing was crossing over ethical or legal boundaries. In 1996, when AstraZeneca was proposing increasing the WAC and AWP while increasing discounts, an internal memo warned that the “challenge in this instance is to come up with a scenario which ... minimizes any perceived risk from a regulatory/legal/public relations perspective.” (DX 2127 at AZ0024480.) Similarly, another pricing strategy memo cautioned that “there is a possible, however likely, risk of a reaction from Medicare. It is feasible that HCFA may see through this strategy and take offense.” A 1996 pricing memo even outlined a “justification” for the price increases in the event that there was outside scrutiny:

[T]he aggressive nature of this price increase may draw some attention, although this is deemed to be unlikely. In the event that our increase is called to task, the move is easily justified on the basis of: 1) increased manufacturing costs, 2) no increase in realized revenue per unit over the last two years, and 3) we are still maintaining our price at a level that is $112.50 less that [sic] our competitor.

(PX 19 at AZ0021764.) Despite these concerns, AstraZeneca continued to price and market Zoladex based on the return to practice for physicians. (Bowman Dep. 102:12-103:1.) Significantly, AstraZeneca sought to thwart the 1998 Medicare legislation, which reduced reimbursement to 95% of AWP, by increasing the price of Zoladex by 6.9% to “compensate[ ] the customer for this 5% plus provide[ ] an additional improvement in return to practice.” (PX 146.)

To its credit, outside of the Medicare system, AstraZeneca attempted to compete with TAP by setting up reimbursement programs that didn’t rely on AWP. First, AstraZeneca encouraged health care plans to adopt a maximum allowable cost (“MAC”) on Zoladex and Lupron equal to Zoladex’s WAC price. (12/04/06 Tr. 17:7-19:13 (Tracy); PX 982D.) The MAC-based reimbursement removed the financial incentive for physicians to purchase the higher priced product. (12/04/06 Tr. 19:2-10 (Tracy); PX 982D; DX 2105.) Second, in 1996 AstraZeneca launched a “Bill to/ Ship to” program, which was later renamed the Managed Acquisition Program (“MAP”). (DX 2110; 12/04/06 Tr. 19:20-22 (Tracy); Tracy Deck ¶ 13; Buckanavage Decl. ¶ 16.) Under the MAP program, managed care organizations would buy Zo-ladex directly from AstraZeneca at the discount prices, and AstraZeneca would ship Zoladex directly to the physician. (Tracy Deck ¶¶ 13-14; DX 2110.) This took the physician entirely out of the financial transaction, allowing the health plans to benefit from the discounted prices. In 1999 or 2000, however, AstraZ-eneca decided not to continue marketing the MAP program because it feared a backlash from physicians. (See PX 4024 at AZ04313740; PX 4025 at AZ0431325.)

2. The Johnson & Johnson Group

The “J & J” Defendants include Johnson & Johnson and two wholly-owned subsidiaries, Centoeor, Inc. and Ortho Biotech Products, L.P. J & J has two drugs at issue in this case, Procrit and Remicade.

a. Promt

Procrit is the brand name for epoetin alfa, which is used to treat severe anemia, including anemia in AIDS and cancer patients. (Dooley Deck ¶ 3.) Epoetin alfa is manufactured by Amgen, Inc. and licensed to J & J’s Ortho Biotech for sale as Pro-crit. (11/16/06 Tr. 51:1-9 (Dooley).) Am-gen also sells epoetin alfa under the brand name Epogen. Procrit and Epogen are identical, having exactly the same FDA-approved indications for use. (Id.) Under an unusual licensing agreement, Amgen has the exclusive right to market epoetin alfa for use in the treatment of anemia in dialysis patients while Ortho Biotech has the exclusive right to market epoetin alfa for non-dialysis uses. (Dooley Deck ¶ 4.) Physicians, however, are not subject to the terms of the licensing agreement and may lawfully administer either brand of epoetin alfa to any patients they choose. (Id. ¶ 5.) Consequently, Procrit and Epogen are sometimes in direct competition with each other.

Ortho Biotech introduced Procrit in January 1991, over a year after Amgen launched Epogen. (Id. ¶ 14.) Ortho Bio-tech set the WAC price and the AWP for most of the Procrit NDCs equal to those already established for Epogen. (Id.) The published AWP for Procrit, like that of Epogen, was set 20% higher than the WAC price. (11/16/06 Tr. 57:22-58:3 (Dooley).) After launching Procrit, Ortho Bio-tech offered discounts below the WAC price to non-dialysis providers in order to encourage physicians to use Procrit rather than Epogen. (Id. 58:13-59:2.) These discounts generally ranged from 5% to 10% off of the WAC price, although some high volume purchasing physicians could receive higher discounts. (Dooley Decl. ¶ 15.) Ortho Biotech also offered rebate programs that ranged between 6% and 12% off of WAC. (11/16/06 Tr. 14:11-15:21 (Dooley).) The WAC price and AWP price remained constant for the six years following Procrit’s launch.

J & J fully understood the Medicare reimbursement system and its impact on physician choices. A 1993 memo emphasized that the “goal is to keep the physician ‘whole’ i.e. whole on the 80% as there is a fear that they will not be reimbursed on the remaining 20%.” (PX 339 at 61807.) A 1999 examination of reimbursement scenarios showed that a physician’s profit per patient, for a twenty week course of Procrit, could range from a loss of $304 to a gain of $1,520 depending on the percentage of the copayment collected. (PX 346 at 60861.) A 1996 McKinsey & Company consulting report for Ortho Bio-tech quoted a doctor as stating that “[m]y practice makes $6-8,000 per month on Procrit.” (PX 334 at 6790.) The report advised that “[Ortho Biotech] must preserve positive economics for physicians.” (PX 334 at 6810.) Significantly, in 1997 when Medicare decided to change Part B reimbursement from 100% of AWP to 95% of AWP, Ortho Biotech responded by making its first price increase since the launch of Procrit. In February of 1997, Ortho Biotech increased the prices on the most popular unit of Procrit by 3.5% and then in January of 1998 increased the prices an additional 1.8%. (PX 237, 238.) The result was that physicians would receive essentially the same reimbursement amount for Procrit after Medicare reduced its reimbursement percentage of AWP.

While J & J worked to “preserve physician economics,” there was serious concern at the company that the government would find out about the spreads and take action to reduce the reimbursement amounts. (,See PX 339 at 61805.) In 1998 Cathleen Dooley, then the Senior Director for Reimbursement and Health Policy, sent an email about Medicare’s reimbursement policy for Procrit in which she stated, “[r]ight now they do not know what the cost [of Procrit and Epogen] is for different providers.” (PX 259 at 842.) She cautioned that the fact that patients were paying a copayment of a price much higher than the acquisition cost would be a “public relations issue.” (Id. at 843.) She further noted that the only way that Medicare could determine Procrit’s market price was “to require an invoice be submitted with each Medicare claim that is sent in. This would be very cumbersome.... ” (Id. at 842.) Similarly, when Ortho Biotech considered taking a price increase in 1997 and 1998 it was concerned that raising the Procrit AWP above the Epogen AWP could “raise red flags” and “trigger a price survey.” (PX 262.) Ortho Biotech recognized that if a survey were taken, “the reimbursement rate would be lowered,” which would decrease the profit to providers. (PX 339 at 61805.)

Despite these concerns, J & J actively encouraged their sales representatives to market the spread on Procrit to physicians. The materials for a Sales Training Workshop indicate that one of the training objectives was to “[k]now how to explain PROCRIT Profit to the Pharmacist.” (PX 270 at 62599.) Dr. Bell, one of the defendants’ experts in this case who previously provided consulting services to Ortho Bio-tech, advised that the “Procrit sales force must provide compelling evidence that continuing with Procrit provides economic benefits.” (PX 344.) He further encouraged Ortho Biotech to develop a spreadsheet that would model those economic benefits of Procrit. (Id.)

In at least one region, the sales representatives were receiving specific instructions on ways “to tactfully discuss how an office can profit from providing Procrit in the office.” (PX 268 at 63656.) In a 1996 memo to his sales team, Sales Manager John Hess emphasized that the “office needs to understand that there is profit associated with Procrit.” (Id.) The memo then provides a chart showing a “return on equity for Procrit” and instructing the sales force to “ask for their real numbers” when “reviewing with a physician or office manager.” (Id.) The memo also specifically quantifies the profits per patient for Medicare and non-Medicare patients over various time periods. (Id. at 63657.) Mr. Hess also directed the sales representatives to be discreet in their use of the profit information, instructing them to “simply draw out the scenario on a piece of scratch paper asking for the office billing fee, injection fee, and acquisition fee based on medicare or non-medicare.” (Id.) The memo closes with an underlined directive: “Do not distribute this memo to your offices. This is for your information only!” (Id.)

The main Ortho Biotech office was also highlighting profit potential to physicians in a slide presentation created by an outside company. (See PX 331.) One slide asks, “Can you make money? ? ? ?,” and the next slide responds, “[djrugs have paid well under part B.” (Id. at 1833.) Another slide explains the Medicare reimbursement at 95% of AWP and quotes the current AWP for Procrit. (Id. at 1839.) The presentation concludes with the question “Should you give Procrit?” and the first reason supporting an affirmative answer is “Additional revenue.” (Id. at 1838.) Later in the class period, Ortho Biotech apparently instituted a policy prohibiting spread marketing. A November 2001 memo to the sales force states: “It is absolutely inappropriate to sell product based upon the difference between AWP and acquisition cost.” (DX 2767.)

Somewhat surprisingly, given J & J’s demonstrated focus on physician profit, the actual spread between the Procrit ASP and the published AWP never exceeded 30% during the class period. (See Hartman Deck, Attach. G.3.c and 1.3.) While it seems plausible that this would be a result of having only a 20 percent (rather than 25 percent) standard AWP markup over WAC, Dr. Hartman’s calculations show that 91 of the 114 spreads for Procrit were actually less than 25%. (Id.) Even Dr. Rosenthal concedes that using Hartman’s theory of market expectations, Procrit is one of the drugs that AWP seems to work well for because the AWP tracks the ASP. (11/27/06 Tr. 69:21-71:6. (Rosenthal).)

[[Image here]]

(J & J Post-Trial Mem. 6.)

b. Remicade

Centocor, Inc. launched Remicade in 1998 and Johnson & Johnson acquired Centocor in 1999. Remicade (infliximab) is used to treat rheumatoid arthritis, Crohn’s disease, and other conditions. (11/14/06 Tr. 53:10-16 (Hoffman).) Remi-cade is administered to patients via intravenous infusion, which frequently takes place in a physician’s office, but which may also take place in hospital out-patient departments. Remicade has been a single-source drug from its inception in 1998 and throughout the class period, although it faces therapeutic competition in the treatment of rheumatoid arthritis. (Id. 54:8-10.)

Unlike the standard 20 to 25 percent markups in the industry, Centocor set the AWP for Remicade at a 30 percent markup over its WAC price. John Hoffman, Vice President of the strategic customer franchise at Centocor, explained why the 30 percent markup was chosen: “It was a combination of looking at what the payors would bear in terms of the price of the product; and ... that it was going to be financially viable for [physicians] to be able to offer this service and not lose money.” (11/14/06 Tr. 56:12-58:25 (Hoffman).) He added that Centocor looked at the spreads between acquisition cost and AWP for other drugs in the same biological class and “picked something that we thought was at the reasonable, the low to middle range of that survey.” (Id. 58:18-25.) Throughout the class period, Centocor maintained this 30 percent difference between WAC and AWP. (Id. 55:20-23.)

Centocor was also unusual in that it did not offer discounts or rebates to physicians. (Id. 63, 88-89, 112-15; 11/27/06 77 (Rosenthal).) Centocor sold to specialty distributors, who in turn sold to physicians. The specialty distributors were entitled to prompt pay discounts of up to 2% and other small rebates, and thus upon resale the physicians could only purchase Remicade at or about the published WAC price. (11/14/06 Tr. 59-61 (Hoffman).) Consequently, the spreads for Remicade hovered very near to 30% throughout the class period. According to Dr. Hartman’s calculations, in only two years did the Remicade spreads exceed his 30% expectations yardstick: a spread of 32.1% in 1999 and a spread of 31.9% in 2001. (Hartman Decl., Attach. G.3.c.) J & J disputes these percentages, arguing that Dr. Hartman should have used a weighted average AWP rather than the June 30 AWP to determine the spread. Using their weighted averages, the spread is 30% or less for all years. (Dukes Deck ¶ 28.)

[[Image here]]

(Hartman Deck ¶ 60, Fig. 9.)

Nevertheless, Centocor pursued a strategy of marketing the spread to physicians. Centocor developed and implemented a Practice Management Program (“PMP”) to educate physicians on buying, infusing, and billing for Remicade. (Glassco Dep. 20:14-21:22; McHugh Dep. 252:15-253:14.) One of the PMP materials was a “Financial Impact Worksheet,” which listed the AWP and allowed the physician to fill in her acquisition cost, the percentage discount off AWP for reimbursement, her case load, and the number of vials per patient. (PX 252 at 3485.) The worksheet then showed the physician how to calculate an “Estimated margin per vial,” “Estimated revenue per patient,” and “Estimated monthly revenue from REMICADE.” (Id.) According to John Hoffman, a reimbursement specialist from Centocor would go over this worksheet with physicians and discuss the “financial ramifications” of using Remicade. (11/14/06 Tr. 67:4-7, 68:8-12 (Hoffman).)

Centocor also hosted PMP seminars, where sales representatives made presentations to groups of physicians explaining the profit potential of using Remicade given the AWPbased reimbursement. Senior Sales Executive Laura Glassco explained how she walked doctors through a PowerPoint presentation that illustrated the profitability:

Basically I would share with the physician ... that AWP was at that time the price that’s shown here, [and] that Medicare reimbursement was AWP less 5.... I then walked through with them the scenario which you see here of an example of a patient that might be a three-vial infused patient.... [I]f the cost of the drug was a certain amount, I show the cost of the drug to the physician and I compare that to what the reimbursement was from Medicare.... The last slide shows then the difference between what the physician paid for the drug and what the physician ... gets reimbursed from ... the Medicare carrier.

(Glassco Dep. 105:22-107:21.) The concluding slide showed that, assuming the drug is purchased at list price, the annual profit per patient on Remicade would be $2,293.41. (PX 254 at 90300.)

Laura Glassco also forwarded an email to her sales team, in which she praised one of the sales representatives for his “work in the field.” (PX 272 at 90283.) In the forwarded email, the sales representative writes about how he explained reimbursement to the physician and walked through a “Medicare AWP example” showing the potential reimbursement. (Id.) He notes that “Dr. Kassan seemed so excited about getting started....” (Id.)

3. The Bristol-Myers Squibb Group

The “BMS Group” of defendants is comprised of Bristol-Myers Squibb Co., Oncology Therapeutics Network Corp. (“OTN”), and Apothecon, Inc. 38 BMS is a major developer, manufacturer and marketer of “brand-name” prescription drugs. BMS has seven oncology drugs at issue in this case: Blenoxane, Cytoxan, Etopophos, Paraplatin, Rubex, Taxol, and Vepesid.

OTN is a specialty distributor that sells and distributes injectable drugs and supplies to medical providers who administer them in a hospital or office setting to patients. (Akscin Decl. ¶ 3.) OTN was a joint venture between BMS and another company until 1996 when BMS acquired OTN as a wholly-owned subsidiary. 39 (Id. ¶ 2.) OTN’s target customers are oncologists in private practice who administer chemotherapy to patients in their offices, rather than oncologists employed by a hospital or hospital out-patient clinic. (Peterson Decl. ¶ 5.)

As the sales agent for BMS oncology products and its wholly-owned subsidiary, OTN had a close relationship with BMS. (See Marré Dep. 26:8-20 (referring to the close cooperation using the phrase “One BMS”).) For example, OTN customers were able to obtain a four percent discount on BMS oncology products, a discount that was not offered through any other distributor. (12/8/06 Tr. 93:24-94:8 (Peterson).) BMS also established “floor” prices, or minimum prices, for BMS drugs sold by OTN. (Marré Aff. ¶ 6.) BMS’s Director of Marketing, Christof Marré, was in weekly contact with OTN to discuss the proper “floor” price and to coordinate joint marketing programs. (Marré Aff. ¶ 7; Marré Dep. 25:11-26:7.) OTN and BMS sales representatives communicated regularly, and OTN Territory Business Development Managers occasionally went on sales calls with their BMS counterparts as part of a strategy commonly referred to within the company as “BMS/OTN Synergy.” (Pe terson Dep. 104:2-105:22; see PX 843; PX 228 at 001483222.)

BMS claims to be unique among the defendants because it has never actually reported an AWP or a suggested AWP to the industry publications. (Kaszuba Aff. ¶ 6.) Rather, BMS only reports its wholesale list price, WLP. 40 (Rogers Aff. ¶¶ 1-4; Szabo Aff. ¶¶ 6-7.) The publications then routinely apply a markup factor of 20.5 percent or 25 percent to BMS’s WLP to calculate the published AWP. (11/13/06 Tr. 59, 120-21 (Kaszuba); DX 2611 at 6646, 6649.)

While BMS knew that its WLP would be marked up by 20 or 25 percent, BMS did not completely control the AWP percentage markup of its drugs. For example, in 1992, BMS wrote a letter instructing the publishers to change their practice and use a 25 percent markup factor for BMS oncology products. (See PX 183.) According to Ms. Kaszuba, this was because Bristol-Myers and Squibb had recently merged, and the publications were using different markup factors depending on whether the drug was a Bristol-Myers or Squibb drug. (Kaszuba Aff. ¶¶ 12-13.) She emphasized that this was the only time that BMS ever directly asked a publication to change the markup factor. (Id. ¶ 14.) The success of this request varied by the publisher. Red Book agreed to the change, while First DataBank and Medispan did not. (Kaszu-ba Aff. ¶ 13.; Szabo Aff. ¶ 6; DX 2554; DX 2650.)

BMS contends that this was an anomalous situation, and that BMS has never had any control over the publications. BMS points to several internal documents which repeatedly emphasize that “BMS does not set AWPs for its products. Third parties set AWP....” (DX 2545; see DX 2554; DX 2585 at 0398; DX 2595 at 9757; DX 2587 at 2095; DX 2588 at 9782; DX 2589 at 8211.) Furthermore, documents show that at least one time First DataBank independently changed the markup factor on BMS drugs. (DX 2588 at 9782; DX 2589 at 8206.)

Nevertheless, as a matter of industry practice, BMS knew, expected, and intended that when it reported a price, the publications would predictably calculate an AWP that was 20 to 25 percent higher than WLP. (Marré Aff. ¶ 10; 11/13/06 Tr. 55, 59 (Kaszuba); DX 2611 at 649; DX 2616; Szabo Aff. ¶ 6.) Frank Pasqualone, Senior VP of the Oncology Division, confirmed that the only possible issue was whether the WLP was going to be marked up by 20 or 25 percent. (12/06/06 Tr. 13:5— 7 (Pasqualone).) Internal BMS documents show first the list price that BMS was establishing for specific Apothecon drugs and then its “Anticipated AWP” based on the 25 percent markup factor being used at the time. (PX 209; PX 210; PX 211.)

BMS was actively involved with approving the AWP before publication. In a 1998 fax announcing a price change for certain BMS drugs, BMS wrote, “Please supply AWP’s for these products once the information has been processed through your database.” (PX 179 at 2173.) The publishers would then respond to BMS with a report showing the AWPs so that BMS could “review [the] AWP’s for reasonability” before publication. (PX 180 at 6649; see also PX 849 (Red Book product listing verification of BMS prices with BMS employee’s approval signature).) Denise Kaszuba, Associate Manager of Pricing Support, explained that a BMS employee would be “mathematically ... looking at the AWP to make sure that it is within [the publication’s] factor.” (11/13/06 Tr. 95:18-96:4 (Kaszuba).) If the AWP was different than expected by BMS, Ms. Kaszuba indicated that BMS would contact the publisher. (Id. 97:3-10.) All of this is sufficient to conclude that BMS could affect, and at times fully control, the AWP for its drugs.

BMS sells its oncology drugs to customers through intermediary wholesalers. BMS distributes the drugs to wholesalers, who pay WLP for the products. (PX 196 at 8200.) Many large providers contract with BMS to then purchase the drugs from the wholesalers. (Marré Aff. ¶ 6.) The wholesaler provides the product at the contract price and then issues a “charge-back” request to BMS for the difference between WLP and the contract price that the wholesaler collected from the purchaser. (11/14/06 Tr. 155-57 (Marré); see also PX 2591 at 6967 (graphic illustration of chargeback process).)

BMS used a similar business model in pricing all of the drugs at issue in this case. The pricing was dependent upon whether a drug was single-source with no competition, single-source with therapeutic competition, or multi-source facing generic competition. At launch, BMS set an initial list price, WLP, for sales to wholesalers. (Pasqualone. AfO 13.) Wholesalers were generally entitled to a possible 2% prompt pay discount. (Id.) BMS would sometimes provide a 5%-10% discount immediately after launch to help get the new product into the marketplace. (Id.) Otherwise, there were few discounts, rebates, or price concessions while a drug faced no therapeutic competition. (Id. ¶ 16.) During the patent period, BMS would take periodic list price increases “in recognition of prevailing market conditions.” (Bell BMS Aff. ¶ 26.) The AWP would rise in step with the WLP increases, so the spread would remain fairly constant throughout the period of patent protection.

Once competition was introduced, BMS would offer discounts and rebates in order to compete with the new alternatives. (Pasqualone ¶ 17.) Marré testified that “the average contract prices and floor prices for BMS drugs in the multi-source portfolio tended to trend down over the long term.” (Marré Aff. ¶ 9.) While these actual sales prices were falling, BMS kept the WLP the same as it was before the introduction of competition. (Pasqualone Aff. ¶ 18.) According to BMS employees, BMS did not decrease the list prices of drugs that became multi-source because there were still customers who were willing to pay that list price. (Id. ¶¶ 18-19.) Marré explained that some of these customers were just brand loyal, (id.), others lacked information about the discounts, (11/14/06 Tr. 130:9-131:12 (Marré)), and others were not entitled to discounts because they did not have contracts with BMS. (11/14/06 Tr. 159-164 (Marré).) As Dr. Bell notes, given these circumstances it would be economically irrational for BMS to lower its list price to wholesalers because “BMS would be losing revenues.” (Bell BMS Aff. ¶ 25.) Thus, the spreads increased over time as the drugs faced more and more competition, and simultaneously fewer and fewer sales were made at or near the list price.

BMS recognized that reimbursement was very important to physicians working in office-based oncology practices (“OBOs”). John Akscin, a Vice President at OTN, acknowledged that OBO revenue is highly Medicare driven because 50 to 55 percent of OBO patients are Medicare recipients. (Akscin Dep. 91-92.) He also noted that 64 percent of OBO revenues came from drug reimbursements. (Id. 93; see also PX 197 at 6634.) In a presentation to OTN and BMS sales representatives, Mr. Akscin displayed a slide which proclaimed that the “Top Three OBO Concerns” were “Reimbursement, Today,” “Reimbursement, Tomorrow,” and “Reimbursement!” (PX 197 at 6636.) BMS noted the impact of the spreads in a memo concerning the launch of Etopophos:

Currently, physician practices can take advantage of the growing disparity between Vepesid’s list price (and, subsequently, the Average Wholesale Price [AWP]) and the actual acquisition cost when obtaining reimbursement for eto-poside purchases. If the acquisition price of Etopophos is close to the list price, the physicians’ financial incentive for selecting the brand is largely diminished.

(PX 208 at 1221.)

With these financial incentives behind reimbursement, it is easy to see the temptation to market the spread to physicians. BMS, however, had a clear policy against such conduct. In January of 2001, BMS sent a memo to all U.S. Sales & Marketing Personnel advising that, “in accordance with its Code of Conduct, ... the spread should not be used as a promotional or marketing tool.” (PX 223.) When asked whether that policy was enforced at BMS, Frank Pasqualone, the Senior VP of the Oncology Division, responded, “Absolutely.” (12/06/06 Tr. 15:14-15 (Pasqualone).)

Nevertheless, plaintiffs presented substantial evidence suggesting that BMS was marketing the spread. While I will address drug-specific spread marketing below, there is one significant piece of spread marketing evidence that applies to all the BMS drugs at issue here. OTN offered customers an online “Cost Differential” report for BMS drugs. (See PX 219.) The site prompted the customer to input a variety of information, including their AWP reimbursement percentage. The site would then display, by regimen, the reimbursement rate, acquisition cost, and “AWP Cost Differential” (equivalent to the spread) for the requested drugs. (Id. at 134-36.)

BMS was well aware that AWP was used as a reimbursement mechanism both under Medicare Part B and through private reimbursement plans. (See Marré Aff. ¶ 12; 11/13/06 Tr. 62-64 (Kaszuba); Akscin Dep. 26-27; Peterson Dep. 114— 15.) BMS also knew that AWP was an “artificially inflated number.” (PX 195.) Yet despite these understandings, there was very little concern, if any, about pay-ors and cancer patients overpaying for their drugs. Sales Representative Douglas Soule best summed up the attitude of BMS when he said, “it’s just the system.” (12/08/06 Tr. 71:3 (Soule).) When asked if it ever bothered him that people were paying a percentage of a phony price, he finally responded, “No.” (Id. 71:10.)

In order to examine the selling and pricing of each drug, it is useful to group the BMS drugs into categories depending upon the type of competition that they faced. Two of the BMS drugs, Parapla-tin 41 and Etopophos, were patent-protected, single-source drugs for the entire class period. Four drugs, Taxol, Vepesid, Cy-toxan tablets, and Blenoxane, all began as single-source drugs and became subject to generic competition at some point during the class period. Finally, Rubex was a branded multi-source drug for the duration of the class period.

a. Single-source Drugs

i. Paraplatin

BMS launched Paraplatin (carboplatin) in 1989 as a second-generation product to first-generation Platinol (cisplatin). (Bell BMS Aff. ¶ 13.) Paraplatin is typically-used in the treatment of non-small-cell lung cancer (NSCLC), small-cell lung cancer (SCLC), and ovarian cancer. (I'd)

As expected with a single-source drug, there were few discounts given and thus the spreads were fairly close to Dr. Hartman’s 30% expectations yardstick. The majority of spreads were under 30%, though the spreads for a few NDCs rose as high as 40%-60% in the years 1997-2002. (See Hartman Decl., Attach. G.2.c.) Averaging across all NDCs, however, the overwhelming number of sales were made within 5% of the list price: for all NDCs across all years of the class period, 94.7% of sales were within 5% of the list price. (Bell BMS Aff. Exh. E.)

Paraplatin was often used in combination with Taxol, so BMS often marketed the two products together. Documents suggest that BMS marketed the spread on both drugs. Sales representatives received a presentation entitled “Practice Efficiencies & Quality Care Workshop” that provides revenue and expense information, including a display of the costs and reimbursement amounts for Taxol and Parapla-tin. (PX 222.) Each drug had a slide that conveniently listed its AWP, the Medicare allowable percentage, and the OTN cost to the physician. (Id. at 2315-16.) Although it was an internal presentation, BMS sales representative Greg Keighley testified that it “was a stand-alone presentation that we would verbally give on an account.” (Keighley Dep. 270:2-4.) Keighley used the information in this way on “one or two instances.” (Id. 270:18-20.) In several pages of call notes from 1998 through 2002, BMS sales representatives detailed their discussions with physicians about reimbursement for Paraplatin and Taxol. For example, in 1998 a sales representative noted that she “[w]ent over some numbers re reimbursement for Taxol/Car-bo vs VP/Cis for NSCLC. He agrees that the [Taxol] is better & you do make ....” (PX 229 at 4123.) In 1999 another representative noted that he had “gone over AWP numbers & fact that do make money on Taxol/Carbo.... ” (Id. at 8993.) In 2000, one wrote, “Jo is not aware of the ... value proposition on Paraplatin, so I covered all of this with her.” (Id. at 3009.) Similarly, in 2002, a representative wrote that he “talked about benefit for reimbursement for taxol -I- paraplatin regimen over non generic products.” (Id. at 2251.)

ii. Etopophos

Etopophos (etoposide phosphate) was launched in 1996 as the second generation of Yepesid, a product discussed below that had become subject to generic competition in 1994. (Bell BMS Aff. ¶ 14.) Etopophos is typically used in the treatment of SCLC and testicular cancer. (Id.) To treat these conditions, Etopophos is generally used in combination with one of the BMS platinum-based oncolytics, Platinol or Parapla-tin. The primary advantage of Etopophos over Vepesid is that Etopophos can be administered to the patient much more quickly. (Pasqualone Aff. ¶ 30.)

As with Paraplatin, BMS offered few price concessions for Etopophos. (Bell BMS Aff. ¶ 40.) At its launch in 1996, OTN developed a buy-in program to increase awareness and initial trial usage of Etopophos. (Id. ¶ 39.) Presumably, that is why the only Etopophos spread that exceeds 30% occurs in 1996, a 35.8% spread. (See Hartman Decl. Attach. G.2.c.) For all other years, the spreads were well below 30% and 100% of sales were made within 5% of the list price. (Id.; Bell BMS Aff. Exh. E.)

Plaintiffs presented no evidence that BMS specifically marketed the spread on Etopophos.

b. Single-Source Drugs Later Subject to Generic Competition

i. Taxol

Taxol (paclitaxel) was launched in 1992 and became subject to therapeutic competition in 2000 and generic competition in 2001. (Marré Aff. ¶ 5; Hartman Decl. ¶ 48.) Taxol was the first of a class of agents called taxanes that interrupt the cell cycle of a cancer cell growth stage and make the tumor more susceptible to the effects of radiation. Taxol is used alone or in combination with other products, most often for the treatment of breast cancer, NSCLC, and ovarian cancer. (Bell BMS Aff. ¶ 15.)

Unlike the other single-source drugs, BMS never increased the list price of Tax-ol. (Id. ¶ 42.) During the patent protected period, BMS offered few discounts and the spreads for Taxol were all under 30%. (See Hartman Decl. Attach. G.2.c.) Similarly, over 99% of sales were made within 5% of the WLP. (See Bell BMS Aff. Exh. E.) When generic entry loomed in 2000, however, BMS had to prepare a strategy to deal with the new low-priced competition. BMS decided to divide the market into three segments, each with its own marketing program: (1) accounts willing to pay a premium for Taxol, (2) accounts that preferred Taxol but were not willing to pay a premium, and (3) accounts that had switched to generic paclitaxel. (Bell BMS Aff. ¶¶ 45-46.) According to Dr. Bell, “[t]his segmentation allowed BMS and OTN to effectively charge a premium to customers who placed the highest value on Taxol and offer lower prices to more price-sensitive customers.” (Bell BMS Aff. ¶ 46.) Thus, actual sales prices began to plummet and the spread began to rise. In 2001, the ASP to providers for Taxol dropped by 25%-50%. (Hartman Decl. ¶ 48.) By 2002 the spreads for certain Taxol NDCs were over 500%. (See Hartman Decl. Attach G.2.c.) By the fourth quarter of 2002, BMS was routinely providing large discounts on Taxol to high volume customers, some as high as 80% off of WLP. (See PX 203 at 6988; PX 204 at 6293; PX 205; PX 206 at 9756.) The result was that in 2002, hardly anyone was paying the list price. Less than 0.5% of sales of Taxol were within 5% of WLP and over 46% of sales were made at a price less than half of WLP. (See Bell BMS Aff. Exh. E.)

[[Image here]]

(Hartman Decl. ¶ 49, Fig. 5.)

BMS carefully educated its sales force on the reimbursement system, the existence of the spread, and the subsequent profitability for a doctor administering Taxol. For example, BMS distributed to its sales force a document entitled “Taxane Economics.” (See PX 221.) The document presents in detail the costs, reimbursements, and spreads for Taxol and Aventis’s Taxotere for different administration periods. (Id.) The document indicates that it “should not be utilized in any sales presentations,” and there is no evidence that it ever was. (Id. at 423.) Sales representatives also received a presentation entitled “Practice Efficiencies & Quality Care Workshop” that provided revenue and expense information, including a display of the costs and reimbursement amounts for Taxol and Paraplatin. (PX 222.) Each drug had a slide that conveniently listed its AWP, the Medicare allowable percentage, and the OTN cost to the physician. (Id. at 2315-16.) As discussed above, this document was actually given to customers on at least a couple of occasions. (See Keighley Dep. 270:2-22.) Finally, BMS produced a series of sales documents that carefully calculate and illustrate the “profit to oncology practice” of using Taxol or a generic version. (See PX 225 at 8052.) Sales representatives were therefore fully prepared to discuss the spread and profitability.

There is substantial evidence that BMS marketed the spread on Taxol. As noted above, Taxol and Paraplatin were often marketed in combination. Thus, many of the sales representatives’ call notes cited in the section on Paraplatin also apply here. In addition, several other call notes focus specifically on Taxol. In 1998, a sales representative noted that he “[g]ot info on [Taxol] vs. [Taxotere] w. respect to AWC and AWPs. Also what medicare is reimbursing.” (PX 229 at 3600.) In some cases, it is clear that the sales representatives were responding to questions or concerns from the physicians. For example, a 1999 call note states, “message ... loud and clear. Bottom line, he wants us to raise our AWP or lower our price. I told him that our AWP is about 25% over acquisition cost, and that we are one of the best in terms of AWP.” (Id. at 6365.) Another call note reads: “Also said they are considering moving away from TAXOL due to cost issues and reimbursement. Talked about TAXOL going Generic and the advantages this will have for the office and reimbursement.” (Id. at 4566.) Many of the documents, however, simply show a focus on selling the economics of the drug. A 2000 call note candidly explains, “[w]e talked to him about Taxol and the profitability spread.” (Id. at 5895.) In other 2000 call notes, sales representatives were focusing specifically on explaining to physicians how Taxol would still be profitable after the entrance of generics in 2001. One sales representative wrote:

We discussed the financial impact of generic competition. I explained it as the greatest business opportunity for him in many years because for every dollar BMSO lost due to price reductions needed to stay competitive with generic competition, medical ocologists [sic] would make 95 cents due to the wide disparity of cost vs AWP reimbursement.

(Id. at 8646.) Another representative documented his “very good conversation on generic paclitaxel and AWP situations.” (Id. at 8664.)

Spread marketing continued in 2001 and 2002. In 2001, a sales representative noted that he gave a physician the taxol profitability sheet. (Id. at 6459.) It is likely that this referred to one of the presentations given to the sales representatives about reimbursement. (See, e.g., PX 222; PX 225.) In 2002, a sales representative documented his discussion of spread during a meeting at a physician’s office: “Discussed generic taxol. They do not want to switch. I told Melva that we are constantly lowering the cost of Taxol and that AWP is still strong. She will stay with OTN.” (PX 4048 at 8182.)

ii. Vepesid

Vepesid (etoposide) is produced in an injectable form and in a capsule form. (Bell BMS Aff. ¶ 49.) The injectable form was launched in 1983 and became subject to generic competition in 1994. (Id. ¶ 16.) Vepesid capsules were launched in 1987 and have been subject to generic competition since 2001.(Id) Vepesid is primarily used in combination with other agents for the treatment of testicular cancer and lung cancer. (Hartman Decl. ¶ 52.)

Injectable Vepesid and the capsule form had very different pricing experiences. From 1993 through 2001, BMS increased the WLP for Vepesid capsules and made further increases after the launch of generic competition in 2001. (Bell BMS Aff. ¶ 50.) For reasons which were never well explained at trial, 42 even with the advent of competition, BMS never increased price concessions more than 2 percent. (See Bell BMS Aff. ¶ 52.) Thus, over 90% of sales were made within 5% of list price for almost all years of the class period, including those after the entrance of generics. (See Bell DX 2524.) The spreads were similarly very low.

The story for the injectable form of Vepesid was much different. At the point generic competition entered the market in 1994, BMS halted all price increases and left WLP at its current level. (Bell BMS Aff. ¶ 50.) In order to protect market share, however, BMS began to offer substantial concessions to compete with the generics on price. Contract discounts to large purchasers were as high as 94% off of WLP. (See, e.g., PX 204 at 6293, PX 205; PX 206 at 9756, PX 207 at 4141.) For some NDCs the spread between ASP and AWP became astronomically high, exceeding 1000%. (See Hartman Decl. Attach. G.2.c.) Given those brand loyal and ignorant customers, however, BMS still made at least 10% of their Vepesid sales within 5% of the unchanged WLP. (See DX 2524.) Excluding the year 2000, however, virtually all the remaining sales were made at prices that were 50% or less of WLP. (See id.) Spreads thus reached over 1,000% percent. (See Hartman Decl. Attach. G.2.c.)

[[Image here]]

Hartman Decl. ¶ 52, Fig. 8.)

Aside from the “Cost Differential Report” available to customers online, (see PX 219), plaintiffs presented no further evidence that BMS proactively marketed the spread on either form of Vepesid. iii. Cytoxan

Cytoxan (cyclophosphamide) is also produced in two forms, injectable and tablet. The injectable form of Cytoxan was originally approved in 1959 and has been subject to generic competition since 1982, before the start of the class period. (Bell ¶ 17.) Cytoxan tablets were approved pri- or to 1982 and have been subject to generic competition since 2000.(M) Cytoxan is often used in the treatment of breast canly in combination with other oncolytics. (Id.) The pricing trajectory for the two versions of Cytoxan are similar to those of the two forms of Vepesid. cer and non-Hodgkin’s lymphoma, typical-

The Cytoxan tablets were relatively unaffected by generic competition. BMS increased the WLP for the Cytoxan tablets from 1993 up until the launch of generic competition in 2000. (Bell BMS Aff. ¶¶ 53-54.) From that point forward, BMS stabilized the WLP. (Id.) Despite the entry of generics, BMS offered less than 2% in price concessions, such that the spreads remained relatively low and the overwhelming majority of sales were made within 5% of WLP. (See Hartman Decl. Attach. G.2.c: DX2524.)

Pricing for the injectable Cytoxan, however, was marked by substantial discounting and dramatic increases in the spread. While BMS kept the WLP relatively constant, contract discounts reached 65%-75% off of WLP. (See PX 204 at 6293; PX 205; PX 207.) This resulted in several spreads of over 100%, even reaching 500% in certain years. (See Hartman Deck Attach. G.2.c.) From 1995 on, the majority of sales were made at prices less than 50% of WLP, and in certain years, as little as 6% of Cytoxan sales were made within 5% of WLP. (DX 2524.) BMS did reduce discounting somewhat from 2000-2002 because generic manufacturers were having difficulty producing the drug, and were starting to exit the market. (Rosenthal Deck ¶ 55; Marré Dep. 88-90.) In fact, by 2003 all competitors had abandoned the market. (Marré Dep. 88:16-89:6.)

[[Image here]]

(Hartman Deck 150, Fig. 6.)

Aside from the “Cost Differential Report” available to customers online, (see PX 219), plaintiffs presented no further evidence that BMS marketed the spread on either form of Cytoxan.

iv. Blenoxane

Blenoxane (bleomycin) is a chemotherapy drug used to treat cancer including lymphomas and testicular cancers. (Hartman Deck ¶ 46.) Blenoxane was launched in 1973 and became subject to generic competition in 1996. (Bell BMS Aff. ¶ 18.)

Like most of its other drugs, BMS increased the WLP during the period of exclusivity and then held it constant once Blenoxane faced generic competition. (See Hartman Deck ¶ 47, Fig. 4.) Up until 1996, price concessions were small and the spread was therefore under 30%. (See id.; Bell BMS Aff. Exh. D.) In 1996, anticipating the entry of generics, BMS adjusted its pricing strategy. According to Dr. Bell, BMS attempted to get its top clients to commit to purchasing most of their bleomycin from BMS, and BMS would in return price Blenoxane competitively with any “bona fide offer for a generic.” (Bell BMS Aff. ¶ 56.) Thus, discounts quickly reached over 60% off of WLP, causing the ASP to drop and the spread to reach over 100% for certain NDCs. (See Hartman Decl. ¶ 47, Attach. G.2.c.) In the post-generic years, only 6% to 15% of Blenox-ane’s sales continued to be made within 5% of WLP. (See Bell BMS Aff. Exh. E.)

[[Image here]]

(Hartman Decl. ¶ 47, Fig. 4.)

Aside from the “Cost Differential Report” available to customers online, (see PX 219), plaintiffs presented no further evidence that BMS marketed the spread on Blenoxane.

c. Multi-Source Drugs

i. Rubex

Rubex (doxorubicin hydrochloride) is used to treat a broad variety of cancers, often in combination with other therapies. (Bell BMS Aff. ¶ 19.) Rubex, a multi-source drug for the entire class period, was launched by BMS in 1989 as a branded version of Adriamycin RDF. (Id.) During 1992 and 1993, Rubex was marketed by Immunex Corporation, but reverted back to BMS in 1994.(Id.) BMS phased out the drug in 2001 and discontinued production after 2002. (Marré Dep. 97:22-98:12.)

Since Rubex is a multi-source drug, BMS has always offered substantial concessions off of the WLP. While the WLP remained fairly constant, discounts have averaged as high as 94% off of WLP. (See Bell BMS Aff. Exh. D.) Thus, the spreads were large, peaking at over 400% toward the end of the 1990’s. (See Hartman Decl. Attach. G.2.c.) On average, 37% of the Rubex sales were made at list price, although that percentage ranged from 0% to 62% throughout the individual years. (DX 2524.)

[[Image here]]

(Hartman Decl. ¶ 51, Fig. 7.)

Aside from the “Cost Differential Report” available to customers online, (see PX 219), plaintiffs presented no further evidenbe that BMS marketed the spread on Rubex.

4. The Schering-Plough Group

The Schering-Plough Group includes Schering-Plough Corporation and Warrick Pharmaceuticals Corporation, its subsidiary.

The Schering products at issue include the branded drugs Temodar, Proventil, and Intron-A. The only Warrick product at issue is generic albuterol sulfate, the same chemical compound as Schering’s branded Proventil.

Schering-Plough refers to its list prices as “direct prices” or “net direct prices.” (Kane Dep. 34:7-35:21.) Schering-Plough reports AWPs for its branded drugs to the pricing compendia. (Zahn Dep. 173:1-9.) It derives its reported AWPs by marking up the direct prices by 20 percent. (Kane Dep. 34:7-35:21; PX 809 at 315085.)

Warrick also reports its AWPs to the pricing publications. (See Weintraub Decl. ¶ 58.) Warrick sets the AWPs for a new generic at a value 10 to 20 percent below the AWPs for the branded counterpart. (Weintraub Decl. ¶ 54; Aug. 25, 2005 Weintraub Dep. 31:2-17; Feb. 2003 Wein-traub Dep. 494:17-495:6; PX 425 at 6155.) When competitor generic products are already on the market, Warrick slots the AWP for its product “somewhere in the pack” of the competitor AWP values. (Sept. 2006 Weintraub Dep. 527:6-528:8.) According to Harvey Weintraub, a former Warrick sales and marketing consultant who was responsible for setting the AWP, this was done simply to save the “time and trouble” of calculating its own price. (Id. 527:15-23.)

Schering and Warrick never lowered their reported AWPs despite offering significant discounts that reduced the ASPs. (See Hartman Decl. Attach. G.4.b.) Schering-Plough and Warrick entered into contracts with pharmacies and other providers, which offered rebates for meeting certain market share targets. (See PX 453; PX 455; PX 612.) These rebates often reached 20%-25% of the direct price. (See id.) In addition, many customers were provided with free goods that further reduced their average acquisition costs. (See PX 505.) Some pharmacies, in order “to keep [their] pricing a secret” bought from a wholesaler at the wholesaler’s price and then made a “chargeback” to Warrick or Schering to make up the difference in the contracted price. (PX 433.) These various discounts all resulted in lower ASPs for Schering and Warrick drugs.

Schering and Warrick were well aware of the role that the spread played in driving purchasing decisions for their products. For example, in response to the government’s investigation into drug pricing under Medicare, a Schering document notes that “[t]he reduction or elimination of the ‘spread’ is likely to have a significant effect on choices of Medicare Part B drugs and on utilization in areas of treatment where generic and brand name drugs are available.” (PX 713 at 55875.) Similarly, Warrick produced a letter from a pharmacy buying group, indicating that Warrick had been chosen to be an “endorsed generic contract vendor” based on certain “Generic Product Selection Criteria,” one of which was “AWP Spread; MAC.” (PX 779 at 35102.)

The Schering and Warrick subject drugs are different from the other drugs in this litigation because they are primarily self-administered. Many of the drugs are administered through the use of a nebulizer, and patients are trained to self-administer the drug using a nebulizer at home. (See 11/15/06 Tr. 113:11-114:2 (Rosenthal).) These drugs are covered and reimbursed under the durable medical equipment (“DME”) provisions of Medicare Part B. (See 42 U.S.C. § 1395 et. seq.) Because the drugs are distributed through pharmacies, Schering sells principally to chain pharmacies, wholesalers, and other intermediaries, rather than physicians. (12/13/06 Tr. 42:19-43:8 (Kane).) Warrick does not market or sell albuterol sulfate to physicians at all. (Weintraub Deck ¶ 23.)

Schering-Plough and Warrick emphasize the differences in the market for SADs. First, many TPPs, including both BCBSMA and Sheet Metal Workers, use PBMs to manage their pharmacy dispensed drugs. (See 12/12/06 Tr. 88:21-22 (Kolassa); 11/7/06 Tr. 6:2-5 (Faulkner); 11/20/06 Tr. 154:23-25 (Shramek); 12/13/06 Tr. 15:1-15 (Dutch).) PBMs can serve a variety of functions in the administration of pharmacy benefits. (See PX 4002, Ro-senthal Tutorial 12-17, Exh. 13.) Manufacturers contract with the PBMs, often offering rebates and chargebacks for drug purchases. PBMs then contract with retail pharmacy networks that dispense the drugs to patients. (See id.; see also DX 1275, Berndt Report ¶ 15.) PBMs are generally large entities which consolidate market power to negotiate better drug prices for their customers. (See 11/28/06 Tr. 88:8-89:18 (Bell); 11/15/06 Tr. 21:25-22:2 (Rosenthal).) Schering and Warrick contend that the PBMs operate in a vigorously competitive market, which ensures that drug reimbursements are kept at a reasonable level.

Second, Schering-Plough argues that it had no incentive to manipulate or market AWPs. For a single-source self-administered drug, the prescribing physician is not being reimbursed for the drug, so there is no pecuniary reason to select drugs based upon the spread. (11/15/06 Tr. 115:1-14 (Rosenthal); see DX 1275, Berndt Report ¶ 188.) Furthermore, the pharmacies that are reimbursed for the drugs have “no control over the prescription” and must dispense whichever branded drug is prescribed by the physician. (11/15/06 Tr. 115:1-14 (Rosenthal); see Ad-danki Am. Decl. ¶ 28.) For a generic mul-ti-source drug, such as Warrick’s albuterol sulfate, pharmacies can choose which version of the drug that they will carry. (See Addanki Am. Decl. ¶ 29; 11/15/06 Tr. 115:15-116:5 (Rosenthal).) However, as discussed below, all versions of a generic drug are reimbursed based upon the same single measure, such as a median or MAC, so that there is no competitive gain from having a higher AWP. (See Addanki Am. Decl. ¶ 30.)

a. Temodar

Temodar is a self-administered pill used to treat brain cancer. (Kolassa Decl. ¶ 22.) Temodar was launched in 1999 and remained single-source throughout the class period. Throughout this time, 95% of all Temodar sales were within 5% of WAC. (DX 2935.) The spreads, as calculated by Dr. Hartman, 43 were all less than the 30% yardstick. (See PX 4109; DX 2968.) Furthermore, plaintiffs presented no evidence that Schering-Plough marketed the spread on Temodar.

b. Intron-A

Intron-A is used to treat hepatitis, leukemia, melanoma, follicular lymphoma, condyloma, and AIDS-related Kaposi’s Sarcoma. (Kolassa Decl. ¶ 21.) Intron-A is generally a pharmacy-dispensed drug, but certain larger dosage sizes are sometimes or always administered by physicians and thus can be reimbursed under Medicare Part B. (Id.) Dr. Hartman has identified six NDCs that are commonly physician-administered. (See Hartman Decl. ¶ 189 n. 221.) To be conservative, Dr. Hartman excluded all other Intron-A NDCs from his damage calculations. (See id.)

Plaintiffs have presented no direct evidence that Schering-Plough was marketing the spread on Intron-A. However, plaintiffs do offer a 1998 internal memorandum to the oncology sales representatives which emphasized the continuing existence of profit potential to physicians after Medicare’s move to reimbursing at 95% of AWP. The message exclaimed:

Treating bladder patients with Intron is still very profitable!! One patient on Intron can represent $16,956.36 of incremental sales and $2,373.84 of profit for our physicians just on the drug alone. These figures are based on having your physicians buy Intron-A at Net Direct pricing and treating on high dose of Intron (12 weeks of lOOrniu weekly then 50miu monthly for 1 year). As you know this dose is very tolerable when given intravesieally.

(PX 394.) It is unclear whether this information was used to market the spread to physicians, but as Schering-Plough points out, the spread that can be calculated from the numbers in this document is only 14%. Dr. Hartman’s spreads for 1998, the year of this memo, were also all under 30%. In fact, the spreads for the physician administered NDCs of Intron-A were nearly all under the 30% threshold. (See PX 4109; DX 2968.) In only four instances was the spread above 30%, and the largest of those spreads was merely 32.6%. 44 (See id.)

e. Proventil

Proventil is a branded form of albuterol sulfate used to treat the symptoms of asthma, chronic bronchitis, emphysema, and other lung diseases. (Kolassa Decl. ¶ 23.) In its solution form, which is the only form at issue in this case, Proventil is almost exclusively self-administered and dispensed by pharmacies. {See id. ¶¶ 20, 23; 11/15/06 Tr. 114:3-5 (Rosenthal).) Proven-til was subject to competition from brand or generic forms of albuterol sulfate since the beginning of the class period. (Hartman Decl. ¶ 62.)

Schering set the AWP for Proventil at 20 percent above its WAC. As a multi-source drug, however, Proventil’s (and generic albuterol sulfate’s) reimbursement under Medicare is not usually based upon the brand AWP. Instead, it is determined by the lower of the median of generic AWPs and the lowest branded version of the drug. {Id.) As a practical matter, though, the branded AWPs were generally much higher than the generic AWPs and thus the median generic was generally used for Medicare reimbursement. (Hartman Decl. ¶ 31 n. 47.) For liability purposes, Dr. Hartman still calculates the spread as the difference between the Pro-ventil brand AWP and Proventil’s ASP. {See 12/13/06 Tr. 88:7-19 (Addanki).) Dr. Hartman then uses the median generic AWP for his damage calculations. 45 {See id.) Dr. Addanki, however, notes that when calculating the spread using the median generic that is actually used for reimbursement, most of the spreads are below 30% and many are, in fact, negative because Proventil’s ASPs are higher than the median generic AWP. {See DX 2967.) Schering argues that there was no incentive for them to market or manipulate the spread on Proventil, because on average, “pharmacists would have lost money had they dispensed Proventil to a Medicare patient.” (Schering and Warrick’s Post-Trial Br. 16 (emphasis in original).) Consistent with this observation, plaintiffs produced no evidence that Schering was marketing the spread on Proventil.

Despite facing generic competition, Schering increased both the Proventil AWP and ASP throughout the period. It appears that this strategy was possible because of the introduction of Warrick’s generic albuterol sulfate which allowed Schering-Plough to segment the market; sophisticated, price sensitive customers could purchase Warrick’s generic albuter-ol, and the less powerful or brand loyal customers would continue to pay higher prices for the branded Proventil. {See PX 409; PX 418.) An internal Schering-Plough memorandum responds to a customer’s demand for lower prices on albu-terol by stating, “[rjather than lowering our Proventil contract prices, I recommend we offer a 2-year Warrick Solution and syrup market driven contract addendum to their existing GeriMed contract” because it would allow Schering to “maintain existing Proventil sales.” (PX 418 at 44924.) Given this segmentation, Schering-Plough maintained 83% of its sales within 5% of WAC. The spreads, as calculated by Dr. Hartman using the brand AWP and illustrated for one NDC in the chart below, exceeded 30% in every year from 1991 to 1997 and later for one NDC in 2002. {See Hartman Decl. Attach G.4.c.)

[[Image here]]

(Hartman Decl. ¶ 63, Fig. 10.)

d. Generic albuterol sulfate

Like the branded Proventil solution, Warrick’s generic albuterol sulfate solution is used to treat asthma, chronic bronchitis, emphysema, and other lung diseases. (Weintraub Decl. ¶22.) In fact, the two products, Proventil and generic albuterol, were identical and manufactured in the same facility, but had different NDCs. (Aug. 2005 Weintraub Dep. 63:9-64:16.) The generic albuterol sulfate is thus also primarily dispensed by pharmacies for patients to self-administer at home with a nebulizer. (Weintraub Decl. ¶ 23.) Both Warrick’s 0.5% albuterol solution and the 0.083% albuterol solution were launched in the early 1990’s, shortly after Warrick was formed in 1993. (Id. ¶¶ 14, 24.) At that time, other manufacturers’ versions of al-buterol were already in the market. (Id. ¶ 24.) Albuterol sulfate was a multi-source drug for the remainder of the class period, facing competition from over 25 different manufacturers. (See PX 4007 at 4; see also DX 2919; DX 2920.)

Warrick set the AWP for albuterol sulfate between 10% and 20% below the AWPs of the branded versions of the drugs. (Weintraub Decl. ¶ 52.) According to Dr. Hartman, “[i]t is generally true that once the generic manufacturers set their AWPs, most manufacturers maintain them at constant levels.” (Hartman Decl. ¶ 32(c).) Warrick, however, did change the AWP on three occasions. First, in 1993 Warrick lowered the AWP on one size of the 0.083% solution in order to make the AWPs the same for all forms of the product on a per unit basis. (Weintraub Decl. ¶ 58.) Then, in 1995, Warrick raised the AWP for its 20 mL albuterol solution twice, from $12.50 to $13.95, and later to $14.99. (See Sept. 2006 Weintraub Dep. 462:17-463:4; Hartman Decl. Attach. G.4.b.) At that time, Warrick was the only producer of the solution because a competitor was having manufacturing problems. (Weintraub Decl. ¶ 59.) The increases in AWP were matched by an identical per centage increase in the direct sales price to customers, (see PX 4079), such that the average sales price for the 20 raL solution also increased in 1995. (See Hartman Decl. Attach. G.4.a.)

After 1995, no changes were made to AWP. (Weintraub Decl. ¶ 59.) However, Warrick’s selling prices for all NDCs of albuterol sulfate declined substantially over time as Warrick sought to match the price of its generic competitors. (See id.; Weintraub Decl. ¶ 31.) The spreads for every NDC in every year were all over 100%, reaching over 800% in 2008. (See Hartman Decl. Attach. G.4.C.)

As explained above for Proventil, Medicare reimbursed for albuterol sulfate based on the median AWP of all generics. Plaintiffs allege that generic manufacturers engaged in tacit collusion to set a high AWP and then marketed the resulting spread. When Warrick announced its price increases in 1995, it did send out memos to its customers that featured the AWP and the direct price for albuterol sulfate next to each other for easy comparison. (See, e.g., PX 437; PX 445; PX 4080; PX 4082.) A 1993 advertisement for albuterol sulfate similarly quotes the AWP and direct price, but markets the product on the basis of “Quality,” “Service,” “Reliability,” and “Trust.” (PX 719 at 1836.) Plaintiffs offered no further evidence that Warrick marketed the spread on albuterol sulfate. As noted before, defendants contend that there was no economic incentive to market the spread on generic albuterol.

Warrick did, however, provide at trial detailed information regarding the AWP for each branded and generic form of albu-terol sulfate during each year of the class period. (See DX 2919; DX 2920.) According to these charts, Warrick’s AWPs were almost always below the median. There are a few exceptions. For the 0.5% solution, Warrick’s AWP was at or above the median from 1996 through 1999. (See DX 2920.) The significance, as explained later, is that if Warrick had reported a true AWP then the median would have shifted downward and a lower price would have been used for Medicare reimbursement.

II. CONCLUSIONS OF LAW

The parties raise five threshold crosscutting issues that must be addressed before the Court reaches the merits of the claims against each drug manufacturer. First, defendants argue that the claims are time-barred. Second, they argue that plaintiffs can only bring a claim under Mass. Gen. L. ch. 93A, § 11. Third, plaintiffs assert that there is per se liability under Chapter 93A because AWP is not a true average of wholesale prices. Fourth, defendants raise a Daubert challenge to the admissibility of Dr. Hartman’s testimony. Finally, the Court must address difficult issues of liability and causation for multi-source drugs.

A. Statute of Limitations

The defendants assert that the plaintiffs’ claims are barred by the four-year statute of limitations for consumer protection claims. See Mass. Gen. Laws ch. 260, § 5A. “Ordinarily, actions in tort accrue at the time the plaintiff is injured.” Taygeta Corp. v. Varian Assoc., Inc., 436 Mass. 217, 763 N.E.2d 1053, 1063 (2002) (citation omitted). In this case, because the plaintiffs filed their first complaint in December 2001, the statute of limitations would ordinarily bar all claims for damages prior to December 1997. However, the plaintiffs seek to toll the statute of limitations for injuries prior to December 1997 by invoking the discovery rule or fraudulent concealment doctrine. The burden is on the plaintiffs to show that the limitations period should be tolled. Saenger Org., Inc. v. Nationwide Ins. Licensing Assocs., Inc., 119 F.3d 55, 65 (1st Cir.1997).

Massachusetts has recognized that the general rule that accrues time from the date of injury is unfair “in actions where the wrong is ‘inherently unknowable.’ ” Taygeta, 763 N.E.2d at 1063. Under the discovery rule, “a cause of action ... does not accrue until the plaintiff knew, or in the exercise of reasonable diligence should have known of the factual basis for his cause of action.” Wolinetz v. Berkshire Life Ins. Co., 361 F.3d 44, 47-48 (1st Cir.2004). The appropriate test for determining whether plaintiffs should have known about facts giving rise to their claims is an objective one. McIntyre v. United States, 367 F.3d 38, 52 (1st Cir. 2004). The first question is “whether sufficient facts were available to provoke a reasonable person in the plaintiffs circumstances to inquire or investigate further.” Id. If so, then the plaintiff is charged with the knowledge of “what he or she would have uncovered through a reasonably diligent investigation.” Id. The court must then determine if that information is sufficient “to permit a reasonable person to believe that she had been injured” and that the defendants caused that injury. Id.

Plaintiffs argue that until quite recently class members were unaware of the real prices being paid for oncology and other Medicare Part B drugs in the marketplace. Even if class members knew of the existence of some discounting prior to 1997, in plaintiffs’ view, that information would still be insufficient to put plaintiffs on notice of the systematic super-sized inflation of AWP and of the marketing of the spread. Defendants retort that by 1996, TPPs knew or should have known, through the exercise of reasonable diligence, that AWP did not equal ASP and that there was no predictable relationship between AWP and acquisition costs. Defendants have produced a variety of articles and government publications that they claim should have put plaintiffs on notice that AWP was not related to acquisition costs. Therefore, defendants contend that all of the plaintiffs’ claims prior to December 1997 are barred because they were filed over four years later.

By the early 1990’s, the more sophisticated payors generally understood that AWP was a 20 to 25 percent markup over WAC, and that some discounting off of WAC was generally available. However, payors, even the most savvy, were not typically aware that mega-spreads were available to physicians and that drug manufacturers were marketing those spreads. Furthermore, plaintiffs were not typically aware of the publicly available reports and articles that began surfacing about the AWP abuse early in the class period. Thus, plaintiffs typically had no actual knowledge of the abuse of the AWP system for PADs prior to December 1997.

The difficult question, then, is when the amount of publicly available information in the marketplace was sufficient to provoke a reasonable TPP to investigate further. Defendants’ expert, Dr. Bell, testified about the articles and reports that were published between the beginning of the class period and 1998, which defendants assert disclosed the existence of significant spreads. The Court must determine whether, despite Bell, the statute of limitations tolls.

In 1992, the OIG studied 13 chemotherapy drugs, using a sample of patients and physicians in New York state, and reported to HCFA that “AWP is not a reliable indicator of the cost of a drug to physicians.” (DX 1053 at 5.) The OIG showed that Doxorubicin (Rubex) could be purchased at a discount of 59% off of AWP (a Hartman spread of 144%). (Id. at 6.) In 1993, a GAO survey examined the impact of the Medicaid rebates on prices offered to HMOs and hospital group purchasing organizations (“GPOs”), finding that the groups were able to purchase drugs at discounts of 34% to 38% off of list prices. (Bell T1 Aff. Attach. A, ¶ 16.) From 1990 to 1993, news outlets discussed pharmaceutical discounts in connection with the Congressional hearings on the federal Medicaid Best Prices legislation. The Los Angeles Times, New York Times, Seattle Times, and Drug Topics reported the levels of discounts, some as high as 70%, available to different classes of trade and the federal government. {Id. Attach. B, ¶ 6.) In 1993, the Los Angeles Times and the Chicago Sun-Times reported that retail drugstores stated that they paid up to 1,200% or 1,245% higher prices for drugs than did HMOs and mail order pharmacies. {Id. 118.) These were primarily self-administered drugs.

In 1996, the OIG focused on the pricing of nebulizer drugs, and albuterol sulfate (at issue in this ligation) in particular. A February 1996 report concluded that Medicare, reimbursing based on AWP, was paying higher prices than Medicaid for two of three nebulizer drugs, resulting in costs of over $11.7 million. {Id. Attach. A, ¶ 18.) A June OIG report concluded, “Medicare’s allowances for albuterol sulfate substantially exceed suppliers’ acquisition costs for the drug.” (DX 1065 at I.) In May of that year, the OIG reported that Medicare could have saved $122 million if it used the Medicaid reimbursement standard, rather than AWP, to calculate drug allowances. (DX 1062 at 7.) Later in 1996, the New York Times and the Chicago Tribune reported the large discounts available to HMOs, (Bell T1 Aff. Attach. B, ¶ 9), and the Washington Post reported that AWP is a “price that is used as a baseline to negotiate prices and reimbursement rates.” {Id.)

In June 1996, Barron’s published an article entitled Hooked on Drugs: Why Do Insurers Pay Such Outrageous Prices for Pharmaceuticals? (DX 2641.) The article reported the pricing for “the top 20 Medicare drugs (which account for about 75% of the program’s drug spending), as well as for various intravenous solutions.” {Id. at 15.) The analysis showed that the manufacturer’s prices were 10%-20% below AWP for single-source drugs and 60%-85% below AWP for generic drugs. {Id.) The article concluded that manufacturers are producing drugs “that cost far less than the published Average Wholesale Price that Medicare and other insurers pay on claims.” {Id. at 16.) The article also addressed current investigations by the DOJ and the possible filing of suits under the False Claims Act. {Id. at 18.)

In January of 1997, the Washington Post printed an article entitled Battling the High Prices Medicare Pays for Drugs, which reported that “doctors can buy drugs from a supplier at less than the AWP, then bill Medicare for the full AWP price.” (DX 1726 at 2.) The article also explained that HCFA was proposing a change to reimburse doctors only for the amount they actually pay for drugs. {Id.) In June of 1997, leading up to the passage of the BBA, the Committee on the Budget of the House of Representatives issued a report that stated:

The Inspector General for the Department of Health and Human Services has found evidence that over the past several years Medicare has paid significantly more for drugs and biologieals than physicians and pharmacists pay to acquire such pharmaceuticals. For example, the Office of Inspector General reports that Medicare reimbursement for the top 10 oncology drugs ranges from 20 percent to nearly 1000 percent per dosage more than acquisition costs.

(DX 1071 at 1354.)

Of great significance here, in August of 1997, Congress passed the BBA which changed reimbursement to 95% of AWP. See BBA of 1997, Pub.L. 105-33, 111 Stat. 251. In December 1997, the OIG issued another report noting that “published AWPs ... bear little or no resemblance to actual wholesale prices that are available to the physician and supplier communities that bill for these drugs.” (DX 1075 at ii.) The OIG stated its belief that the 5 percent discount off of AWP “is not a large enough decrease” given the existence of spreads from 11% to 900%. (Id. at ii-iii.)

Under the discovery rule, the question is when there was sufficient information such that a reasonable TPP in the plaintiffs’ position would have been on notice to investigate the possibility that AWP had become unhinged from acquisition costs causing plaintiffs to overpay for drugs. See Taygeta, 763 N.E.2d at 1063. “Where events receive ... widespread publicity, plaintiffs may be charged with knowledge of their occurrence.” McIntyre, 367 F.3d at 60 (quoting United Klans of Am. v. McGovern, 621 F.2d 152, 154 (5th Cir.1980)). The relevant factors include the geographical scope of the coverage vis-a-vis the plaintiffs, the content of the stories, and the degree of press and media saturation. Cascone v. United States, 370 F.3d 95, 99 (1st Cir.2004). This is a “fact-intensive inquiry into the pervasiveness and content of the publicity and the particular circumstances of the relevant plaintiff(s).” In re Mass. Diet Drug Litig., 338 F.Supp.2d 198, 208 (D.Mass. 2004). I begin by looking at the most sophisticated named plaintiff, BCBSMA.

The plaintiffs cite to several cases, which they claim stand for the proposition that tens of articles in major news outlets and coverage on the national news may not constitute widespread publicity such that plaintiffs should have discovered their claims. 46 They argue that this case involves only a few articles and government reports, well below the threshold for constructive notice. However, BCBSMA is fundamentally a different plaintiff than the individual consumers in the cases cited by plaintiffs. BCBSMA is a sophisticated non-profit entity in the business of providing health care. Reimbursing for drugs was a substantial part of this mission. From 1991 to 1997, the period discussed here, BCBSMA was also the Medicare carrier for Massachusetts. (11/08/06 Tr. 18:2-11 (Mulrey).)

A reasonable plaintiff in BCBSMA’s situation would be closely following any information that reported on drug reimbursement under Medicare. Although staff at BCBSMA might not have read the scattered national news articles or the handful of OIG reports, a reasonable TPP in the position of BCBSMA, as a major insurer, would have monitored major Congression al actions regarding Medicare reimbursement policies. In August of 1997 when the BBA was signed into law, reducing Medicare Part B reimbursement to 95% of AWP, BCBSMA should have been alerted to the fact that it could have been overpaying for drugs using AWP. At that time, a reasonable investigation would have uncovered the OIG reports finding that spreads on certain oncology drugs reached nearly 1,000% and the Barron’s article highlighting the spreads and the recent investigations into AWP fraud. At the least, BCBSMA could have conducted the reasonable investigation undertaken by the Barron’s staff to uncover the fact that physician costs were well below AWP.

Plaintiffs’ appeal to the “importance of being unimportant” is not persuasive here. While the relative insignificance of Medicare Part B drugs may be a reason for not changing their reimbursement system, it does not negate the fact that they were on notice of the problems with AWP and could have taken legal action.

I find that in August of 1997 (the date of passage of the BBA) sufficient facts were available for BCBSMA and any similarly situated large TPP to discern the basis for both the Class 2 and Class 3 claims. 47

It is a much more difficult question as to whether the other class representatives, Pipefitters and Sheet Metal Workers, should have been put on notice of their claims at this time. At trial, it was clear that these Taft-Hartley funds were much less sophisticated organizations than BCBSMA. However, as less sophisticated entities, both organizations hired third parties to handle their medical benefits, including drug reimbursement. Under standard agency principles, “[w]hen an agent acquires knowledge in the scope of [his] employment, the principal ... is held to have constructive knowledge of that information.” Sunrise Props., Inc. v. Bacon, Wilson, Ratner, Cohen, Salvage, Fialky & Fitzgerald, P.C. 425 Mass. 63, 679 N.E.2d 540, 543 (1997) (citing DeVaux v. Am. Home Assurance Co., 387 Mass. 814, 444 N.E.2d 355 (1983)). As the Taft-Hartley funds stated at trial, a key reason for hiring outside consultants and administrators was to obtain experience and expertise in the provision of health benefits.

Pipefitters, a Class 3 representative, contracted with BCBSMA and thus was put on notice at the same time as BCBSMA. Sheet Metal Workers, a Class 2 representative, hired Southern Benefits Administrators (“SBA”) to handle its health benefits, including the Medicare Part B payments. SBA, like BCBSMA, is actively engaged in the health care and insurance industries, and should reasonably have been on inquiry notice at the same time as BCBSMA. Because Sheet Metal Workers relied on SBA as its agent to provide expertise on health care matters, Sheet Metal Workers also should have been on inquiry notice at that time.

Therefore, Plaintiffs cannot bring claims for any damages arising before December 1997. 48

B. Liability Under Section 9 or 11 of Chapter 93A

Defendants argue that the class representatives, BCBSMA, Pipefitters, and Sheet Metal Workers, although technically nonprofit entities, were acting in a business context and therefore can only proceed under § 11 of Chapter 93A. 49 The significance of this challenge is that defendants contend plaintiffs cannot satisfy additional requirements imposed by § 11. Plaintiffs respond that their claims are properly brought under § 9 because the class representatives are not-for-profit entities, acting in furtherance of their core missions.

The Massachusetts Consumer Protection Act, Mass. Gen. Laws ch. 93A, § 2, protects against unfair or deceptive acts or practices in the conduct of any trade or commerce. Chapter 93A distinguishes between claims actionable under § 9 and “business” claims actionable under § 11. Frullo v. Landenberger, 61 Mass.App.Ct. 814, 814 N.E.2d 1105, 1111 (2004) (citing Lantner v. Carson, 374 Mass. 606, 373 N.E.2d 973, 976 (1978)). Section 11 provides a cause of action to “individuals acting in a business context,” Lantner, 373 N.E.2d at 976, while § 9 grants a cause of action to “[a]ny person, other than a person entitled to bring action under section eleven of this chapter.” Mass. Gen. Laws ch. 93A, § 9. The two sections are mutually exclusive and plaintiffs’ claims can proceed under only one section. See Frai-lo, 814 N.E.2d at 1112 (“[A] plaintiff who acts in a business context has a cause of action exclusively under § 11.”); see also Continental Ins. Co. v. Bahnan, 216 F.3d 150, 156 (1st Cir.2000) (“By their terms, however, [sections 9 and 11] of chapter 93A ... are mutually exclusive.”).

The dividing line between a claim under § 9 and a business claim under § 11 is as clear as mud. See Frullo, 814 N.E.2d at 1112. By its text, § 11 applies to any “person 50 who engages in the conduct of any trade or commerce.” Mass. Gen. Laws ch. 93A, § 11. Trade and commerce include “the sale, rent, lease or distribution of any services and any property.” Id. § 1. Given this capacious definition, Massachusetts courts look to the circumstances of each individual case, to see whether the case arose in a “business context.” Linkage Corp. v. Trs. of Boston Univ., 425 Mass. 1, 679 N.E.2d 191, 207 (1997). Relevant factors include the nature of the transaction, the character of the parties involved, and “whether the transaction is motivated by business or personal reasons.” Id. (citing Begelfer v. Najarian, 381 Mass. 177, 409 N.E.2d 167, 176 (1980)).

Plaintiffs rely on two cases to support their contention that their claims fall un der § 9 of Chapter 93A. While the cases were both decided by federal courts outside of Massachusetts, the cases involve the same plaintiff, BCBSMA, and factual circumstances that are quite similar to this case. In In re Lorazepam & Clorazepate Antitrust Litig., 295 F.Supp.2d 30, 33 (D.D.C.2003), BCBSMA brought Chapter 93A claims against Mylan Laboratories, Inc. for unlawfully raising prices on generic drugs that were reimbursed by BCBSMA. In addressing this threshold issue of whether plaintiffs could sue under § 9 or § 11 of Chapter 93A, the Court held that BCBSMA “is a charitable institution not engaged in trade or commerce when it undertakes activities in furtherance of its core mission.... [P]ayment[s] for members’ prescription drug claims ... are clearly at the core of BCBS Massachusetts’s charitable mission.” Id. at 45. The court gave considerable weight to the fact that BCBSMA “is a creation of statutory law,” specifically prohibited from operating for profit. Id. at 46.

Similarly, in In re Cardizem CD Antitrust Litig., No. 99-mdl278, slip. op. (E.D.Mich. May 27, 2003), BCBSMA brought a Chapter 93A claim against certain pharmaceutical companies alleging unlawful, anti-competitive acts that caused BCBSMA to pay millions of dollars in overcharges on drug reimbursement. The Court held that:

BCBS Massachusetts has pled facts showing that it is a nonprofit corporation created by statute and regulated by the Commonwealth of Massachusetts, and that the activity in question — its customary payment or reimbursement for its members’ prescription drug benefits^ — falls within its charitable mission as set forth by statute and case law.

Id. at *7-8.

Defendants contend that these cases were wrongly decided because the Massachusetts case law discussing whether nonprofits are engaged in trade or commerce for the purposes of Chapter 93A involves entities being sued as defendants, rather than entities suing as plaintiffs. It is true that the same entity can sue as a plaintiff under § 11 in one case, and be immune to suit under § 11 in a different case. See Boston Hous. Auth. v. Howard, 427 Mass. 537, 695 N.E.2d 192, 194 (1998) (refusing to impose § 11 liability on the Boston Housing Authority while recognizing that it had been allowed to sue as a plaintiff under § 11 in prior cases). However, the determination of whether the entity is engaged in a business context must focus on the transaction at issue in the particular case. See Begelfer, 409 N.E.2d at 176 (“[B]usiness context must be determined from the circumstances of each case.”).

Although not dispositive, a party’s status as a non-profit influences this analysis. Boston Hous. Auth., 695 N.E.2d at 193. “In most circumstances, a charitable institution will not be engaged in trade or commerce when it undertakes activities in furtherance of its core mission.” Linkage Corp., 679 N.E.2d at 209; see also Trs. of Boston Univ. v. ASM Communs. Inc., 33 F.Supp.2d 66, 77 (D.Mass.1998) (“A nonprofit or charitable corporation, however, is not engaged in trade or commerce ‘if, in the transaction in question, the non-profit is merely engaged in the customary business necessary to meet its charitable purpose.’ ”) (citation omitted); Shin v. Mass. Inst. of Tech., 2005 WL 1869101, at *14, 2005 Mass.Super. LEXIS 333, at *22 (Mass.Super. Ct. June 27, 2005) (“Federal courts interpreting Massachusetts law have held that colleges and universities, as charitable corporations, are not engaged in ‘trade or commerce’ for purposes of c. 93A ‘when [they] undertake[ ] activities in fur therance of [their] core mission.’ ”) (citation omitted). This Court has specifically applied that test to a non-profit plaintiff that, like BCBSMA here, was attempting to sue under § 11. See Trs. of Boston Univ., 33 F.Supp.2d at 77 (finding that Boston University, a nonprofit entity, was not engaged in trade or commerce when it purchased term papers from a corporation because investigating cheating was “central to a university’s educational mission” and therefore could not bring a claim under § 11 of Chapter 93A).

Based on this caselaw and the record, I conclude that BCBSMA is a nonprofit organization acting pursuant to its legislative mandate, 51 and that the reimbursement for prescription drugs is a key part of its core mission. There is no evidence that BCBSMA profited from its reimbursement for those over-priced drugs during the non-time-barred portion of the class period. 52 A fortiori, the Taft-Hartley funds may bring their claims under § 9 of Chapter 93A because they were not motivated by the desire to make money from the drugs and were acting within their core mission. Class 3 consumers who made co-payments have a claim under § 9. Plaintiffs are, therefore, not engaged in trade or commerce for the purposes of this case and their claims are properly brought under § 9 of Chapter 93A. 53

C. Per Se Unfair or Deceptive Conduct Under Chapter 93A

Plaintiffs advance the theory that the defendants’ acts and practices are per se unfair or deceptive in violation of Chapter 93A. Plaintiffs base this contention on three sources of law: the Federal Trade Commission Act, Massachusetts Attorney General Regulations, and the federal Medicare statute.

Chapter 93A protects against “unfair or deceptive acts or practices in the conduct of any trade or commerce.” Acting in accordance with authority granted in Chapter 93A, § 2(c), 54 the Massachusetts Attorney General enacted 940 Mass.Code Regs. 3.16, which states:

An act or practice is a violation of M.G.L. c. 93A, § 2 if:

(1) It is oppressive or otherwise unconscionable in any respect; or

(2) Any person or other legal entity subject to this act fails to disclose to a buyer or prospective buyer any fact, the disclosure of which may have influenced the buyer or prospective buyer not to enter into the transaction; or

(3) It fails to comply with existing statutes, rules, regulations or laws, meant for the protection of the public’s health, safety, or welfare promulgated by the Commonwealth or any political subdivision thereof intended to provide the consumers of this Commonwealth protection; or

(4) It violates the Federal Trade Commission Act, the Federal Consumer Credit Protection Act or other Federal consumer protection statutes within the purview of [Mass.] G.L. c. 93A, § 2.

Courts have been hesitant to find that a violation of any statute is a per se violation of Chapter 93A, but instead take into account all the facts and circumstances to determine whether the statutory violation involves unfair or deceptive conduct. Darviris v. Petros, 59 Mass.App.Ct. 323, 795 N.E.2d 1196, 1201 (2003). 55 However, where a consumer protection statute falls within the heartland of 940 Mass.Code Regs. 3.16, conduct that violates that statute may be per se unfair and deceptive. See, e.g., Barnes v. Fleet Nat’l Bank, N.A., 370 F.3d 164, 176 (1st Cir.2004) (holding that a violation of the Truth in Savings Act, a federal consumer protection statute, constitutes a per se violation of Chapter 93A); In re Tavares, 298 B.R. 195, 203 (Bankr.D.Mass.2003) (holding that a violation of the criminal usury statute constitutes a per se violation of Chapter 93A because the usury statute is a public policy statute covered by 940 Mass.Code Regs. 3.16).

Plaintiffs first argue that defendants violated 940 Mass.Code Regs. 3.16(4) by contravening the Federal Trade Commission Act, 15 U.S.C. § 45(a)(1) (“FTCA”). To support this claim, plaintiffs reference a section of the FTC’s 1967 “Guides Against Deceptive Pricing,” which states that “if the list price is significantly in excess of the highest price at which substantial sales in the trade area are made, there is a clear and serious danger of the consumer being misled by an advertised reduction from this price.” 16 C.F.R. § 233.3. Plaintiffs contend that because the defendants refer to AWP as a list price and no defendant made any sales at AWP, then these list prices were in violation of this regulatory guideline.

The FTC Guidelines, though supportive of plaintiffs’ allegations in this case, do not establish that defendants’ acts were per se deceptive. 56 As defendants point out, the factual circumstances of this case do not squarely fit within the context of these Guidelines. Section 233.3 is titled “Advertising retail prices which have been established or suggested by manufacturers (or other nonretail distributors).” 16 C.F.R. § 233.3. The Deceptive Pricing Guides are directed toward the advertising and promotion of misleading prices to the “consuming public.” 16 C.F.R. § 233.3(b). Here, manufacturers are not advertising prices to the consuming public, but to doctors and pharmacies, and the manufacturers are not involved in the offering of discounts off of those prices to consumers. In these circumstances, the Guidelines do not create per se liability under Chapter 93A.

Plaintiffs make a brief argument that defendants violate Chapter 93A pursuant to the Attorney General’s Regulation, 940 Mass.Code Regs. 3.16(4), by running afoul of two Massachusetts health and welfare regulations, 940 Mass.Code Regs. 3.04, 57 3.05(1). 58 These regulations seem to be intended to protect “buyers” of a product. Again, the purchasers are primarily the doctors or pharmacists; the plaintiffs are TPPs who do not buy a product, but rather reimburse for it. Arguably, however, the regulations could apply to consumers who make co-payments when they purchase PADs, but plaintiffs make no effort to carve out these claims as distinct from the claims of TPPs in Class 3.

Finally, Plaintiffs argue that the Medicare Act is a federal consumer protection statute within the meaning of the Attorney General’s regulations, 940 Mass. Regs. Code 3.16(4). The types of federal statutes that courts have found to be consumer protection statutes under section 3.16(4) include: the Truth in Savings Act (TISA), Fair Debt Collection Practices Act (FDCPA), and Truth in Lending Act (TILA). 59 Notably, these statutes all focus on the conduct of buyers and sellers in the marketplace, and specifically reference the protection of consumers in these transactions. For example, by the text of the statute, it is the purpose of TISA “to require [ ] clear and uniform disclosure ... so that consumers can make a meaningful comparison between the competing claims of depository institutions.” 12 U.S.C. § 4301. Similarly, the purpose of the FDCPA is “to promote consistent State action to protect consumers against debt collection abuses.” 60 15 U.S.C. § 1692. Finally, one of the enumerated purposes of TILA is “to protect the consumer against inaccurate and unfair credit billing and credit card practices.” 15 U.S.C. § 1601.

To be sure, the purpose of the Medicare statute is to provide quality health care and insurance to those in need, mainly the elderly and the disabled. See, e.g., Fischer v. United States, 529 U.S. 667, 680, 120 S.Ct. 1780, 146 L.Ed.2d 707 (2000) (“The structure and operation of the Medicare program reveal a comprehensive federal assistance enterprise aimed at ensuring the availability of quality health care for the broader community.”); Furlong v. Shalala, 156 F.3d 384, 392 (2d Cir.1998) (“The underlying purpose of the Medicare statute is to provide affordable medical insurance for the aged and disabled.... ”). While defendants frustrate the purpose of the Medicare Act when they make health care less affordable, it does not necessarily follow that the Medicare Act is essentially a consumer protection statute of the same genre as, for example, the Truth in Lending Act. Specifically, the provision in the Medicare statute setting government reimbursement at 95% of AWP (42 U.S.C. § 1395u(o)) does not appear to have consumer protection as its primary focus. As such, it is not an “other Federal consumer protection statute[ ] within the purview of G.L. c. 93A, § 2.” 940 Mass.Code Regs. 3.16(4).

Consequently, defendants’ acts are not per se unfair or deceptive. Nevertheless, these statutes and regulations can be examined among the several factors used to determine whether defendants’ acts or practices were unfair or deceptive. See Billingham v. Dornemann, 55 Mass.App. Ct. 166, 771 N.E.2d 166, 176 (2002).

D. The Daubert Challenge

Fed.R.Evid. 702 allows an expert witness to testify “if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case.” Before according weight to the expert testimony, the trial court must first perform a gatekeeping function to determine whether the expert is qualified and whether the expert’s testimony is sufficiently reliable and “relevant to the task at hand.” Daubert v. Merrell Dow Pharms., 509 U.S. 579, 597, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). The trial judge must make this requisite gatekeep-ing determination for all proffered expert testimony that reflects “specialized knowledge,” whether “scientific” or not. See Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147-48, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999) (citing Fed.R.Evid. 702). The critical inquiry is whether the expert “employs in the courtroom the same level of intellectual rigor that characterizes the practice of an expert in the relevant field.” Id. at 152, 119 S.Ct. 1167.

If the court finds the expert to be qualified, it must then turn to the proffered expert testimony to determine its relevance, i.e., “whether those principles and methods have been properly applied to the facts of the case.” United States v. Monteiro, 407 F.Supp.2d 351, 357-58 (D.Mass. 2006) (quoting Fed.R.Evid. 702 advisory committee’s note). Beyond the normal requirement of relevance for all evidence, “expert testimony must be relevant ... in the incremental sense that ... if admitted, [it] likely would assist the trier of fact to understand or determine a fact in issue.” Ruiz-Troche v. Pepsi Cola of P.R. Bottling Co., 161 F.3d 77, 81 (1st Cir.1998); see also Fed.R.Evid. 702. The party offering the expert testimony need not prove the testimony is correct, but rather that it rests upon “good grounds, based on what is known.” Id. at 85 (quoting Daubert, 509 U.S. at 590, 113 S.Ct. 2786) (internal citations omitted).

“[A]n expert must vouchsafe the reliability of the data on which he relies and explain how the cumulation of that data was consistent with standards of the expert’s profession.” SMS Sys. Maint. Servs. v. Digital Equip. Corp., 188 F.3d 11, 25 (1st Cir.1999). However, the court may reject testimony for which the data relied upon is flawed or the methodology used is “internally inconsistent or unreliable.” Ed Peters Jewelry Co., Pic. v. C & J Jewelry Co., Inc., 124 F.3d 252, 260 (1st Cir.1997); see also Coastal Fuels of P.R., Inc. v. Caribbean Petroleum Corp., 175 F.3d 18, 34 (1st Cir.1999) (affirming district court’s rejection of damages expert’s testimony because of “considerable and unjustified variance” between testimony and Rule 26 report and because expert “unintentionally misled [court to believe] that he had performed certain crucial calculations” he had not actually done).

An economist’s failure to consider certain data is not fatal to admissibility if the expert sufficiently explains her choice of data for her analysis. See Cummings v. Std. Register Co., 265 F.3d 56, 65 (1st Cir.2001). Such shortcomings in an expert’s analysis go “to the weight, not the admissibility, of the testimony,” and the opposing party is free to argue at trial that the trier of fact should discredit it. Id. at 65; accord McMillan v. Mass. Soc. for the Prevention of Cruelty to Animals, 140 F.3d 288, 303 (1st Cir.1998).

1. The Hartman Speed Limit

A key dispute at trial was the validity and reliability of the expert opinion of Dr. Raymond S. Hartman, a health care economist who rendered opinions on liability and damages for Class 2 and Class 3. Dr. Hartman is an economist specializing in microeconomics and econometrics and holds a Ph.D. from the Massachusetts Institute of Technology. He has spent over thirty years teaching, consulting, and publishing in the field of applied economics. Over the last five to ten years, he has focused his work almost exclusively on health care economies. In the course of that work, he has testified as an expert in several other pharmaceutical pricing cases. I find that Dr. Hartman is qualified.

A key assertion of Dr. Hartman’s testimony is that drug prices exceeded the expectations of Class 3 TPPs as to the difference between the published AWP and the provider’s acquisition cost. Dr. Hartman starts with the premise that TPPs typically reimburse at AWP minus x% for physician-administered drugs based on expectations in the marketplace about the provider’s acquisition costs. He explains:

What have Class 3 TPPs come to understand x% is and should be, so that physicians can cover their costs and perhaps earn a “reasonable margin” rather than “egregious profit” on the drugs they administer? This would be the rule of thumb that they would use when bargaining with providers. If manufacturers then secretly increased spreads such that reimbursement rates negotiated by TPPs with the expectation of an average spread of x% led in reality to “egregious” overcharges and profits unbeknownst to TPPs, by a rule of reason approach, it would seem that those secret spreads constitute fraud injuring the Class members.

(Hartman Decl. ¶ 92.) Because of the “prohibitive costs of acquiring and acting upon information gathered by NDC, TPPs reasonably look to these rules of thumb to simplify reimbursement across all physician-administered drugs within standard computerizable algorithms based upon discounts off AWP.” (Id. ¶ 150(b).) To determine TPP expectations of the average spread between ASP and AWP in the physician-administered context, Hartman uses three approaches.

Hartman begins by examining the actual pricing history of certain single-source drugs that did not face competition to determine manufacturer expectations as to the margin which must be given to providers to ensure them reasonable profit and cover administrative fees. He explains: “Successful break-through innovator drugs serve as reasonable yardsticks for ‘but-for’ spreads or baseline spreads, precisely because they reflect the manufacturer’s understanding that AWP Inflation (or Spread Inflation or increased Return to Practice) was unnecessary to move market share for single-source branded drugs reimbursed by Class 3.” (Id. ¶ 138.) This baseline spread in the “but for” world, where there is no fraud, is called the “Yardstick Threshold Spread.” To determine the expected spreads, Dr. Hartman calculated the ASP by NDC for each single-source, physician-administered drug and compared it to the AWP. The Hartman “spread” is measured by the percentage markup over ASP, equaling (AWP-ASP)/ASP. 61 He concluded that a reasonable range of spreads expected in the market, negotiated into contracts between manufacturers and doctors, and untainted by the AWP scheme is 18%-22% using First DataBank; and 18%-27% using Red Book. (Id. ¶ 143(d).) “To be conservative”, he chose 30% as his Threshold Yardstick Spread for single-source drugs, and uses the same yardstick for six months after the first generic launch, after which he assumes competition in the multi-source world controls pricing. 62 (Id. ¶ 155.) Outside of Medicare, multi-source PADs are typically reimbursed by TPPs based on a MAC benchmark, which generally does not rely on AWP. That is why there are no viable class allegations involving multi-source drugs for Class 3.

If a manufacturer raises its AWP and/or lowers its ASP, such that the realized spread exceeds the 30% Threshold Yardstick Spread for a particular NDC for a given year, Hartman concludes that the manufacturer has increased the spread on that NDC in that year to move market share. (Hartman Deck ¶ 148.) At the trial, the parties referred to this 30% spread as the Hartman “speed limit.” Hartman concluded there was liability and causation whenever the spread between AWP and ASP exceeded 30%, and he calculated damages for Class 3 using that empirical yardstick. (Id. ¶¶ 148,154.)

As a first cross-check on the reliability of this theory, Hartman reviewed publicly available sources providing market-wide information concerning the relationship between AWP and ASP for branded and generic PADs, including the OIG reports. According to Hartman, this review revealed “reasonably anticipated spreads” of ll%-25%. (Id. ¶ 144; see also Hartman Rebuttal ¶¶ 46-47 (“The preponderance of spreads for single-source drugs reported through 2003 was 20%.”)-)

As a second check, he used the “revealed preferences method,” which is predicated on the theory that “economic agents reveal their preferences, and implicitly the information they relied on, by their actual market decisions and behavior.” (Hartman Decl. ¶ 137.) Defendants have not challenged the “revealed preferences method” as unreliable. Using this method, Hartman calculated the average spread expected by TPPs by examining the contracts between TPPs and providers to determine what the parties expected the spread between AWP and ASP to be. 63 Based on a review of contracts, Dr. Hartman found that TPPs have negotiated to reimburse PADs in the range of AWP minus 16% to AWP plus 15%. This contractual range is consistent with the literature in the field that discuss.es a range of AWP ± 15%. 64 (Id. ¶ 123.) According to Hartman, it is also consistent with the range of contractual reimbursement for self-administered drugs (SADs). Hartman states that TPPs reimburse SADs at AWP minus x%, where x% = (13% — 18%). (Id. ¶ 91.)

Dr. Hartman believes that the x% discount in the contracts between providers and TPPs is reflective of information in the marketplace about provider acquisition costs. According to Dr. Hartman, the “better informed” TPPs believed that the WAC represented the average acquisition cost of providers, and “would believe therefore that the average spread earned by providers reimbursed at AWP would be a 20 to 25 percent markup above acquisition cost.” (Id. ¶ 129(b)-(c).) Remember that the standard formulaic mark up between WAC and AWP (with the sole exception of J & J’s Remicade) is 20 to 25 percent, and this markup is widely published in the commercial compendia and (all agree) was well known in the industry. (See 11/15/06 Tr. 71:6-22; 94:3-20 (Rosen-thal).)

Dr. Hartman also believes that the range of TPP contractual reimbursement is consistent with Medicare’s reimbursement, stating: “Given the herd behavior revealed among TPPs, reliance upon Medicare reimbursement is common, which has reimbursed up to 15% off AWP (implying spreads reflected in negotiating positions of TPPs of 18%).” (Hartman Decl. ¶ 150(b).) Medicare reimbursement rates for single-source Part B drugs were not decreased from AWP-5% to AWP-15% until 2003. However, Medicare did try to reimburse at AWP-15% in 1991 and to go to a cost-based system in 2000. According to Dr. Hartman, the industry understood that Medicare did not believe that AWP reflected true acquisition costs.

Based on all this marketplace data (the contracts themselves, the literature, and the range of actual and proposed Medicare Reimbursement rates over the period from 1989 to the present), Dr. Hartman concludes that it is reasonable to infer that TPPs generally believed that spreads between AWP and provider drug acquisition cost were on the order of 0%-25% over the class period. (Id. ¶ 129(d).)

Hartman has calculated total damages for Class 2 based on the plaintiffs’ legal position that any spread violated Chapter 93A in the Medicare program. He determined damages for Class 3 from 1991 to 2003 using his 30 percent yardstick.

2. Defendants’Critique

Defendants challenge four hypotheses articulated by Dr. Hartman: (l) the pharmaceutical market is structured such that expectations by payors about acquisition costs affect reimbursement rates; (2) payors expected that spreads were 30% or less for every NDC in every year; (3) changes in payor expectations would lead to dollar-for-dollar changes in reimbursement; and (4) in the “but for” world, AWPs would drop to within 30% of ASPs. (11/27/07 Tr. 108:10-109:1 (McFadden).) Defendants rely on the expert testimony of Dr. Bell, Dr. Gaier, and Dr. Daniel L. McFadden. 65

a. Payors’ expectations

Defendants challenge the hypothesis that payors’ expectations about provider acquisition costs affect reimbursement rates. They highlight the fact that Dr. Hartman did not conduct a survey of pay-ors to determine what they believe but rather relied on three surrogates to determine payor expectations: (1) comparator drugs; (2) publicly available information on spreads; and (3) contracts. Dr. McFadden believes that TPPs adjust rates to attract and retain providers and that reimbursement rates do not depend on expectations about providers’ acquisition costs. He contends that three indicators that might measure the value of providers’ costs to TPPs are inconsistent with the Hartman hypothesis and consistent with his alternative: TPPs did not attempt to acquire information about acquisition costs; TPPs did not negotiate prices individually with every provider; and most payors, including BCBSMA, have not changed their reimbursement structures in response to new information about acquisition costs. (McFadden Dir. ¶ 16(a).) Based on the record in this trial, I disagree that these factors undercut Dr. Hartman’s theory of liability.

First, as the plaintiffs have proven, TPPs do not seek cost data because, with respect to most of the drugs at issue in the litigation, there was no readily available market data providing physicians’ costs that could be gleaned from commercial services. 66 Entities like PBMs or consultants did not collect such data. The pric ing of specialty drugs was complex, opaque, and confusing. Accordingly, to procure accurate pricing data, a TPP would have to individually collate invoice data doctor-by-doctor, NDC-by-NDC, on a quarterly basis. Even Medicare did not do this until it had a statutory mandate. Indeed, providers were under a contractual obligation not to disclose discounts and rebates. There is nothing in the record to indicate that providers divulged information about acquisition costs to TPPs, and they would have been contractually precluded from doing so.

Second, the fact that TPPs don’t negotiate prices drug-by-drug with every provider does not mean that TPPs did not care about cost data. With the thousands of drugs and thousands of transactions, a TPP would reasonably choose to contract with providers based on a benchmark like AWP with the general expectation that the price would reflect acquisition costs as well as the margin necessary to cover a provider’s administration costs while providing a reasonable profit. Thus, as both Dr. Gaier and Dr. Bell testified, payors and providers could not, and did not, consider reimbursement on a drug-by-drug basis; rather, they focused on reimbursement levels overall. (Bell T1 Aff. ¶ 71; 11/29/06 Tr. 41:18-42:5 (Gaier).)

Dr. McFadden argues that TPPs failed to react to data about providers’ acquisition costs once they became available because reimbursement rates are driven by the recruitment of providers by drug profitability and the establishment of competitive prices to attract customers, not on expectations about acquisition costs. (McFadden Dir. ¶¶ 45-48.) As an illustration, McFadden uses the market for meals in restaurants and the market for new cars. (See id. ¶ 27.) But these comparisons are not apt. The record established that there was no competitive market with normal supply and demand forces setting the drug reimbursement rates because AWP was embedded in the Medicare statute. Moreover, the AWP for branded drugs was a fictitious price effectively controlled by the drug manufacturers.

Dr. Bell argues that TPPs did not consider knowledge of actual acquisition costs to providers to be important. For example, he highlights the following quote from Robert Farias, Director of Planning and Administration for Harvard Pilgrim Health Care:

Q: And indeed, if Harvard Pilgrim were to learn more information about what providers paid to acquire drugs, that would not change the amount that Harvard Pilgrim is reimbursing for drugs. Is that a fair statement?

A: That’s a fair statement.

(Farias Dep. 43:10-16 (objection omitted).) Moreover, Dr. Bell points out that TPPs had no expectation as to what the margin was. For example, he quotes Kelly Ell-ston, assistant vice president for claims and care management for Union Labor Life:

Q: It would be impossible to say that Union Labor Life expects that they’ll make a percentage profit of 5 percent, 10 percent, 20 percent, 30 percent, 40 percent?

A: That’s not in our calculations.

Q: That’s something that is entirely irrelevant to Union Labor Life’s calculations of the amounts that it’s going to reimburse. Is that correct?

A: Correct.

(Ellston Dep. 89:18-90:5.)

Plaintiffs point to testimony which states the opposite. For example, Joanne Ro- raasko of Blue Cross Blue Shield Montana testified at her deposition:

Q: Is it important to Blue Cross Blue Shield of Montana that AWP prices be as accurate as possible?

A: Absolutely.

Q. And why is it important that AWP be accurate and 135 reliable?

A. So we’re compensating the physicians appropriately for the drug they’re administering.

Q. Does Blue Cross Blue Shield of Montana consider AWP prices to be an accurate, reliable pricing benchmark?

A. Yes, that’s what we use today.

(Romasko Dep. 94-95.)

The TPPs’ failure to react when they received true data about actual acquisition costs is better explained by the insurmountable barrier to devising an alternative system. Even after it became clear by the mid-to-late 1990’s that there were mega-spreads for oncology drugs and other Medicare Part B drugs, it was not feasible for each TPP to devise its own ASP system either by doing its own survey of drug pricing, or by developing a pragmatic pricing methodology to handle the millions of annual drug transactions when there were different prices per NDC and per dose. Significantly, it took three years for Medicare to devise an alternative pricing structure, because of the complexity of simultaneously increasing the prices paid for physician services.

TPPs were also concerned that a unilateral decrease in reimbursement rates would drive doctors away or would induce providers to move their patients into the more expensive hospital setting. As Dr. Rosenthal testified, shifting the pricing paradigm from AWP to another approach is like turning the RMS Queen Elizabeth. 67 Thus, the evidence that TPPs did not nimbly react to new pricing information does not support Dr. McFadden’s alternative hypothesis that reimbursement rates for providers are determined primarily by competition for providers rather than expectations as to provider costs. I find that TPP knowledge about physician acquisition costs was material to the establishment of reimbursement rates.

b. Spreads of thirty percent

McFadden also challenges Hartman’s second hypothesis that buyers expected that spreads were 30% or less for every NDC for physician-administered and pharmacy-dispensed drugs with and without therapeutic competition. He argues that TPPs would expect that prices would drop and spreads would rise in response to increased competition between manufacturers for drugs with therapeutic equivalency and that drug prices would differ by distribution channel and customer. {See McFadden Dir. ¶ 32.)

As a threshold matter, it is undisputed that the market understood and expected a 20 to 25 percent formulaic markup from WAC to AWP. Thus, the dispute hinges on Dr. Hartman’s premise that the market understood and expected rebates or discounts no greater than 3.8% below WAC, resulting in a spread at or below 30%. (See 11/27/06 Tr. 135:7-136:3 (Hartman).) However, Dr. McFadden overstates Dr. Hartman’s hypothesis. TPPs surely should have expected that price competition between therapeutic equivalents would prompt manufacturers to give entities that can move market share rebates and discounts, thereby increasing the spread. Still, TPPs did not know the degree of the spread because there was little available pricing information. There is no evidence that the TPPs had any knowledge about the existence of the huge spreads between AWP and ASP for the drugs on trial until the late 1990’s.

c. Changes in reimbursement

Dr. McFadden attacks Hartman’s third hypothesis that changes in expectations would lead to dollar-for-dollar changes in reimbursement. Disagreeing, McFadden states that reimbursement rates are a function of competitive conditions, business objectives, and provider demands. {See McFadden Dir. ¶¶ 34-39.) This seems to be an attack on the plaintiffs’ strategy for calculating damages for all spreads above the speed limit. The BCBSMA scenario demonstrates that Dr. McFadden is correct that market factors other than cost, like the relative power of the TPPs and physicians, will also affect pricing, even with a full understanding of cost. Nonetheless, I find that expectations are a substantial factor in setting reimbursement rates and reject the argument that payors, like Medicare, knowingly permitted providers to retain mega-spreads to achieve other objectives. 68

d. Living in the “but for” world

Finally, defendants challenge the fourth hypothesis that in the “but-for” world, total reimbursements would have been lower. Dr. McFadden argues that drug prices would be higher than in the “as-is” world because manufacturers would not achieve greater market share by discounting, and providers would have an incentive to choose the product with the highest AWP (i.e., 95% of $100 is greater than 95% of $90). However, the proof is in the pudding. Dr. Rosenthal’s testimony demonstrates that the MMA resulted in lower drug costs, even with the increase in administration fees. Moreover, if there were transparent and accurate pricing, AWPs would have likely been much lower because they would have been related to true market prices.

Defendants also launched numerous challenges to the accuracy of Dr. Hartman’s data, but these criticisms affect the weight of his testimony but not its admissibility. Sometimes those disagreements were valid, and I took them into account in calculations. 69 I conclude, however, that the 30% yardstick methodology used by Dr. Hartman was reliable and admissible under Fed.R.Evid. 702. The yardstick is consistent with the undisputed evidence in the market establishing an industry-wide markup between WAC and AWP of 20% to 25%.

Given my adoption of Dr. Hartman’s basic methodology for determining liability and damages, I further find that his calculation of aggregate damages for the class is sufficiently reliable and reject defendants’ argument that aggregate damages are inappropriate on this record. See 3 Alba Conte & Herbert Newberg, Newberg on Class Actions § 10.2 (4th ed. 2002) (“The evidentiary standard for proof of monetary relief on a classwide basis is simple — the proof submitted must be sufficiently reliable to permit a just determination of the defendant’s liability within recognized standards of admissible and probative evidence.... Individual damage issues should not, except in extraordinary situations, have any adverse effect on the propriety of aggregate class judgments as a proper means for determining the defendant’s liability to the class.”).

E. The Merits: Chapter 93A Unfair or Deceptive Acts

1. The Standard

Massachusetts Gen. Laws Ch. 93A, § 2 prohibits “unfair or deceptive acts or practices in the conduct of any trade or commerce.” While a practice may be both unfair and deceptive, a finding need only be made that the practice was unfair to constitute a violation of Chapter 93A. See, e.g., Serv. Publ’ns, Inc. v. Goverman, 396 Mass. 567, 487 N.E.2d 520, 527 (1986). One difference between unfair conduct and deceptive conduct may turn on whether the plaintiff had knowledge of the conduct. See Commonwealth v. DeCotis, 366 Mass. 234, 316 N.E.2d 748, 753-55 (1974). In DeCotis, the Attorney General brought a Chapter 93A claim against the proprietors of a mobile home park for charging a “resale fee” for any tenant who moved out of the park and sold their home to another buyer. The court held that there were two groups of injured home owners: those that knew of the resale charge when they committed to living in the park, and those that had no such knowledge. Id. 316 N.E.2d at 753-54. As to the first group, the court noted:

Although deception may not have been involved where the disclosure by the defendants to the prospective tenant was timely and complete, we believe that the practice of charging a fee for no service whatsoever was an unfair act or practice within the intent of § 2(a) of G.L. c. 93A and that it was therefore unlawful.

Id. at 754. The court explained that the prospective tenants were in a vulnerable situation because they were still better off selling the home than trying to relocate it. Id. at 755. “The willingness of tenants to pay resale fees, and even to contract knowingly to pay those fees, does not make the collection of such a fee fair. It merely demonstrates the extent to which the defendants had their tenants at their mercy.” Id.

Chapter 93A gives no definition of “unfairness,” and Massachusetts courts have refrained from establishing such a definition. Instead, “whether an act is unfair or deceptive is best discerned from the circumstances of each case.” Buster v. George W. Moore, Inc., 438 Mass. 635, 783 N.E.2d 399, 413-14 (2003) (internal quotations and citation omitted). The Massachusetts courts have, however, enumerated several factors to be considered when determining whether a practice is unfair: “(1) whether the practice ... is within at least the penumbra of some common-law, statutory, or other established concept of unfairness; (2) whether it is immoral, unethical, oppressive, or unscrupulous; [and] (3) whether it causes substantial injury to consumers (or competitors or other businessmen).” Mass. Eye & Ear Infirmary v. QLT Phototherapeutics, Inc., 412 F.3d 215, 243 (1st Cir.2005) (quoting PMP As-socs., Inc., v. Globe Newspaper Co., 366 Mass. 593, 321 N.E.2d 915, 917 (1975)). Additional consideration may be given to the “equities between the parties,” “what a defendant knew or should have known,” and “a plaintiffs conduct, his knowledge, and what he reasonably should have known.” Swanson v. Bankers Life Co., 389 Mass. 345, 450 N.E.2d 577, 580 (1983); see also Mass. Sch. of Law v. Am. Bar Ass’n, 142 F.3d 26, 41 (1st Cir.1998) (to state a 93A claim, “the defendant’s conduct must be not only wrong, but also egregiously wrong”).

The Court should focus “on the nature of challenged conduct and on the purpose and effect of that conduct as the crucial factors in making a [Chapter 93A] fairness determination.” Mass. Employers Ins. Exch. v. Propac-Mass, Inc., 420 Mass. 39, 648 N.E.2d 435, 438 (1995) (characterizing the much worn phrase “level of rascality” as “uninstructive”); see RGJ Assocs. v. Stainsafe, Inc., 338 F.Supp.2d 215, 234-35 (D.Mass.2004) (quoting Mass. Employers Ins. Exch.). Adherence to industry standards or customs is one factor that supports a finding of no unfairness under Chapter 93A. See, e.g., James L. Miniter Ins. Agency Inc. v. Ohio Indem. Co., 112 F.3d 1240, 1251 (1st Cir.1997) (considering defendant’s adherence to industry standard in finding no unfairness); USM Corp. v. Arthur D. Little Sys., Inc., 28 Mass.App. Ct. 108, 546 N.E.2d 888, 898 (1989) (using “conformity with accepted methods within the business community” as one factor in concluding that there was no Chapter 93A violation). Nonetheless, the existence of an industry-wide practice does not constitute a complete defense to unlawful conduct in a Chapter 93A action. DeCotis, 316 N.E.2d at 753.

2. The Inflation of AWP

The key question in this litigation is whether causing the publication of an AWP that greatly exceeds the average sales price charged to a doctor or pharmacist for certain drugs covered by Medicare Part B is an unfair or deceptive trade practice under Chapter 93A. Under the plain meaning canon of statutory construction, I have construed the statutory term AWP in 42 U.S.C. § 1395u(o) to mean the average price at which wholesalers sell drugs to their customers, including physicians and pharmacies. See In re Pharm. Indus. Average Wholesale Price Litig., 460 F.Supp.2d 277, 278 (D.Mass.2006).

The overwhelming evidence at trial established that AWPs are fictitious and are rarely, if ever, prices paid by doctors for PADs or by pharmacies for SADs. Nonetheless, defendants argue that they had no intent to deceive the patients or payors who ultimately paid for their products when they caused their AWPs to be published in the compendia. The manufacturers have emphasized that both the government and TPPs understood that AWP was a fictitious number and were not deceived by the published AWP.

It is true that by the late 1990’s most sophisticated TPPs and the government understood that AWP did not represent a true average of wholesale prices, but that there was a spread of 20 or 25 percent between the AWP and wholesale list (or acquisition) price. However, this knowledge does not exonerate defendants.

I find that the defendants unfairly and deceptively caused to be published false AWPs (or their formulaic counterparts: false WACs or WLPs) knowing that TPPs and the government did not understand the extent of the mega-spreads between published prices and true average provider acquisition costs. Moreover, defendants knew that neither the government nor the TPPs could do much to change the AWP reimbursement benchmark because they were locked into the nationwide reimbursement scheme established by statute or contract.

Unscrupulously taking advantage of the flawed AWP system for Medicare reimbursement by establishing secret mega-spreads far beyond the standard industry markup was unethical and oppressive. It caused real injuries to the insurers and the patients who were paying grossly inflated prices for critically important, often life-sustaining, drugs. Defendants caused these injuries by not reporting a true average wholesale price, that approximated provider actual acquisition costs or was within well established industry expectations (i.e., the Hartman 30 percent “speed limit”). Instead, the spreads were as high as 1,000%. This is exactly the sort of false and misleading information for which Chapter 93A is intended to provide relief. See OIG Compliance Program Guidance for Pharmaceutical Manufacturers, 68 Fed.Reg. 23,731 at 23,733 (May 5, 2003) (specifying, at the end of the class period, that manufacturers are under a legal duty not to submit “false, fraudulent, or misleading information” where “reimbursement by Medicare and Medicaidf ] for the manufacturer’s product depends, in whole or in part, on information generated or reported by the manufacturer, directly or indirectly, and the manufacturer has knowingly ... failed to generate or report such information completely and accurately”).

While I find that the mega-spreads prior to 2001 were deceptive as well as unfair, I also find that once the cat was out of the bag, and the mega-spreads became widely known, the conduct was still egregious under the unfairness prong of Chapter 93A because neither the TPPs nor the government could move quickly or effectively to fix the problem. In retrospect, at least, it has become clear that the Medicare statute itself created a perverse incentive by pegging the nationwide reimbursement for billions of drug transactions a year to a price reported by the pharmaceutical industry, thus putting the proverbial pharmaceutical fox in charge of the reimbursement chicken coop. The different pharmaceutical companies unfairly took advantage of the system by setting sky high prices with no relation to the marketplace.

While establishing mega-spreads itself constitutes egregious misconduct, marketing those spreads so that doctors would choose a drug based on profit rather than therapeutic value is particularly outrageous and unethical. Even the industry understood that spread-marketing violated industry standards. Both BMS and J & J instructed their sales teams that the spread should not be a promotional or marketing tool, although these instructions were often ignored. Moreover, in 2003, the OIG belatedly issued guidelines condemning this practice. Id. at 23,737 (“If a pharmaceutical manufacturer purposefully manipulates the AWP to increase its customers’ profits by increasing the amount the federal health care programs reimburse its customers, the anti-kickback statute is implicated.”). Although these guidelines were issued at the end of the class period, they defeat any notion that the federal government’s failure to change the AWP pricing benchmark signaled acquiescence in spread-marketing or the reporting of mega-spreads.

Throughout the class period, the pharmaceutical industry understood that if the size of the spreads and the marketing of the spreads became public, a public relations nightmare would ensue. As such, the manufacturers insisted on confidentiality in physician contracts and lobbied to undermine government surveys. See In re Lupron Mktg. & Sales Practices Litig., 295 F.Supp.2d 148, 168 n. 19 (D.Mass.2003) (pointing out that “[i]f everything [about Lupron] was known to everybody, why did defendants emphasize secrecy?”).

Significantly, the defendants well understood the devastating impact the mega-spreads had on old and sick patients required to make co-payments they could ill afford, and set up programs to help some needy patients by subsidizing their costs. The spiraling drug costs incurred by third-party payors and the government, however, were never a concern.

3. Causation

In order to warrant an award of damages under Chapter 93A, “there must be a causal connection between the seller’s deception and the buyer’s loss.” Hershenow v. Enter. Rent-A-Car Co. of Boston, 445 Mass. 790, 840 N.E.2d 526, 532 (2006) (internal citation and quotation marks omitted). To establish causation, Dr. Rosenthal testified that class members paid more for drugs based on a false AWP than they would have if defendants had reported a true AWP. Cf. Hershenow, 840 N.E.2d at 535 (finding no causation because “[t]he [illegal provision of the car rental contract] made neither rental customer worse off during the rental period than he or she would have been had the [provision] complied in full with the requirements of [Massachusetts law]”). She confirmed this finding by examining current reimbursements under the MMA, demonstrating cost savings on many of the Medicare Part B drugs at issue. As noted above, the fact that the TPPs have been slow to change their reimbursement systems does not negate causation. 70 Even Dr. Bell admitted that TPPs faced several significant impediments to quickly changing reimbursement practices.

Furthermore, several pharmaceutical witnesses confirmed causation by testifying that they knew that TPPs and consumers were paying more for a drug every time the AWP was raised. Plaintiffs’ damages were not only foreseeable, defendants were well aware of them throughout the class period.

For Class 2, defendants argue that the method BCBSMA uses to set its premiums for its Medigap policies demonstrates that there is no loss to BCBSMA. 71 Defendants argue that the contribution to reserves is an actual profit to BCBSMA, over and above the costs of prescription drugs, and that the costs are in effect passed on to the insureds such that BCBSMA suffers no injury.

The evidence does not support this conclusion. First, there is as much as a two year lag period between the time when BCBSMA incurs a cost and the time when those costs may be incorporated into the rate setting process used to determine premiums. Second, Mr. Arruda, the BCBSMA executive, testified that insurance is a risky business and the contribution to reserves is used to cover unforeseen risks. Defendants have failed to prove that the purpose or effect of the contribution to reserves is to recover money paid out for current claims. See In re Terazosin Hydrochloride Antitrust Litig., 220 F.R.D. 672, 690 (S.D.Fla.2004) (“[T]o the extent that any third-party payer did charge its insureds a higher premium be cause of a drug company’s monopolistic activities, the charging of a higher premium in the future cannot be accurately described as a ‘pass on’ of those charges. The record is clear that the purpose of a future projection is, as the name implies, to estimate anticipated future costs.”).

Defendants caused injury to both Class 2 and Class 3 plaintiffs.

4. Class 2 Liability and Damages

With respect to Class 2, Dr. Hartman’s damage and liability calculation was based on the plaintiffs’ legal conclusion that any spread between AWP and ASP was per se unlawful under the Medicare statute because the statutory term AWP in 42 U.S.C. § 1395u(o) means the average price at which wholesalers sell drugs to their customers, including physicians and pharmacies.

What Congress understood and intended AWP to mean is not the same as what the industry understood. It is undisputed that AWP was calculated by a 20 to 25 percent markup from WAC or WLP, and that this formula was widely known and published in the class period (although not as well understood by less sophisticated market participants). The unfair and deceptive standard of misconduct required by Chapter 93A is different from a strict liability statutory violation. Because information about the 20 to 25 percent spread was widespread in the industry, a violation of the Medicare statute by publishing an “AWP” that was not a true average of wholesale prices does not trigger per se liability under Chapter 93A. 72 Therefore, I reject plaintiffs’ zero tolerance approach to liability and damages in Class 2.

5. Multi-source drugs

Certain BMS, 73 Schering-Plough, 74 and Warrick products were multi-source for at least part of the class period. The interchangeability of these drugs, coupled with the “J-Code” reimbursement system, makes it practically impossible for the plaintiffs to know which drug company’s product was dispensed to any party at any point. Plaintiffs must demonstrate that the unfair conduct caused them harm. The method for reimbursing multi-source drugs and the difficulty in product identification create extremely difficult legal issues for the branded and generic multi-source drugs in Class 2. 75

I begin with describing the system of reimbursement of multi-source drugs. During the time-period at issue, 1998 through 2003, Medicare reimbursed multi-source drugs at 95% of the lesser of the median of the generic AWPs or the lowest brand AWP. See 42 C.F.R. § 405.517 (2003) (DX 1852). For these Part B drugs, the branded drugs nearly always had higher AWPs such that in actuality Medicare reimbursed based on the median generic AWP.

a. Causation

For the branded multi-source drugs, defendants argue that reimbursement was based upon the median of the generics rather than the AWP of the manufacturer’s branded drug, so that the manufacturer did not cause plaintiffs’ injuries. It is true that, given the statutory reimbursement scheme and the fact that generic AWPs were below the brand AWPs, payments for these drugs were never based upon the brand name drug’s AWP. However, the flip side is that if BMS had reported a true AWP for its branded multi-source drugs, Medicare would have reimbursed based on that branded drug’s AWP, rather than the inflated median, and plaintiffs would have paid less. 76 Thus, when BMS reported an inflated AWP for a branded multi-source drug, it caused a higher reimbursement rate to be used. This resulted in injury to every plaintiff who purchased any version of its multi-source drug, regardless of the manufacturer.

The causation question for generic mul-ti-source drugs, in particular Warrick’s albuterol sulfate, is considerably more difficult. Warrick argues that because no generic manufacturer can unilaterally affect the median AWP, a manufacturer of a generic drug could not have legally caused plaintiffs’ injuries.

Dr. Hartman responds that “[although the median itself is not readily subject to strategic manipulation by any single generic manufacturer, the distribution of AWPs for generic sources of the drug is subject to the strategic manipulation of all generic manufacturers, and, thereby the median AWP.” (Hartman Deck ¶ 32 (emphasis in original).) He continues that “all manufacturers of a multi-source drug have the incentive to maintain the median AWP as high as possible, to increase the spreads of all these manufacturers relative to potential therapeutic competitors.” (Id.) He refers to this result as a “tacit informal Nash equilibrium in the dispersion of generic AWPs.” (Id.) “A set of strategies is called a Nash equilibrium if, holding the strategies of all other firms constant, no firm can obtain a higher payoff (profit) by choosing a different strategy. Thus, in a Nash equilibrium, no firm wants to change its strategy.” (Id. ¶ 32(d) n. 49 (quoting Dennis Carlton and Jeffrey Perloff, Modem Industrial Organization 157 (3d ed.2000)).) In plaintiffs’ view, this tacit collusion in the AWP-setting is sufficient to find that the generic manufacturers of albuterol jointly caused the harm to the class members. Dr. Hartman rounds out his theory by positing that once the generic manufacturers have jointly inflated the median AWP, they “compete amongst themselves on spread through the reduction of their ASPs.” (Hartman Decl. ¶ 32.)

Warrick takes issue with Dr. Hartman’s claim that all manufacturers have an incentive to maintain a high median in order to compete with therapeutic alternatives. As Dr. Addanki points out, that does not make sense for albuterol because it is primarily pharmacy-dispensed. While the pharmacist may well be able to choose which generic it dispenses, there is no evidence that he might unilaterally substitute a therapeutic alternative not prescribed by the doctor. (See Addanki Am. Deck ¶ 42.) In Dr. Addanki’s opinion, the spread, therefore, cannot influence what drug is prescribed, and manufacturers have no reason to collude to inflate the median.

Warrick states that when it first set the AWP for albuterol, it was merely following the standard industry practice of listing the generic AWP 10%-20% below the branded drug’s AWP. In other words, it contends that there was no market-based motive for inflating the AWP in the generic context. Regardless of motive, it is still true that all twenty-nine or so generic manufacturers of albuterol did independently post inflated AWPs which caused the median itself to be inflated, which in turn caused substantial overpayments by TPPs and patients. Given that there are no claims or evidence of conspiracy or joint enterprise, the pertinent legal question is whether Warrick can be said to have individually caused the plaintiffs’ injuries.

Plaintiffs must prove that Warrick was a “but for” cause of their injuries when purchasing albuterol sulfate. In other words, plaintiffs must show that they would not have suffered the same injury if Warrick had reported a true AWP. Given the procedure for calculating a median, this could only possibly be true when Warrick’s AWP was at or above the median. In that case, reporting a true AWP (which would be well below the median) would cause the median to shift down to the AWP of the next manufacturer in the ordered series. If that manufacturer had reported a lower AWP (rather than the same AWP), then the median would drop. In this situation, Warrick would affect the reimbursement for every version of albuterol sulfate sold.

However, when Warrick’s AWPs were below the median, moving them farther down to a true AWP would have had no effect on the median. In that case, reporting a true AWP could not change the median used for reimbursement and plaintiffs would sustain the same injury as when Warrick published an inflated AWP. Warrick would not be a “but for” cause of plaintiffs’ injury. In sum, Warrick was a legal cause of plaintiffs’ injury only when reporting a true AWP would have actually shifted the median.

Looking at the manufacturer data provided by Dr. Addanki, there are only two years between 1998 and 2003 in which Warrick’s AWP was at or above the median, and the effect of reporting a true price would be to lower the median. 77 For the 0.5% solution, Warrick’s AWP in 1998 was the median and in 1999 Warrick’s AWP was above the median. In both those years, reporting a true AWP would have resulted in the median shifting slightly downward. 78 Therefore, there is liability for albuterol in both 1998 and 1999. For all other years, legal causation has not been proven.

BMS and Schering argue that for both the brand and the generic multi-source drugs, plaintiffs’ claims must still fail because plaintiffs cannot identify the manufacturer of any particular drug for which they reimbursed. However, this is of no consequence because when the manufacturer causes the median to be inflated, it affects reimbursement for every manufacturer’s version of the drug. It does not matter who manufactured any particular drug. Because defendants failed to report a true AWP, plaintiffs paid a higher reimbursement amount every time they reimbursed for every manufacturer’s version of that multi-source drug.

b. Apportionment

Plaintiffs urge the Court to find that defendants are jointly and severally liable for the whole harm suffered by plaintiffs with respect to each multi-source drugs — a hard pill to swallow. The theory of joint and several liability has been applied by Massachusetts courts in the context of Chapter 93A actions. See, e.g., Kattar v. Demoulas, 433 Mass. 1, 739 N.E.2d 246, 258 (2000); Int’l Fidelity Ins. Co. v. Wilson, 387 Mass. 841, 443 N.E.2d 1308, 1318 (1983); Piccuirro v. Gaitenby, 20 Mass. App.Ct. 286, 480 N.E.2d 30, 35 (1985); see also Pepsi-Cola Metro. Bottling Co. v. Checkers, Inc., 754 F.2d 10, 19-20 (1st Cir.1985) (affirming the district court’s use of joint and several liability for a Chapter 93A claim). Joint and several liability is appropriate when “the independent tor-tious conduct of two or more persons is a legal cause of an indivisible injury.” 79 Restatement (Third) of Torts: Apportionment of Liability, § A18 (1999). Under joint and several liability, “a plaintiff may sue and recover all damages from any defendant found liable.” Id. § A18 cmt. a.

Here, many of the manufacturers of a multi-source drug independently caused the injury to all payors that reimbursed for that multi-source drug. Had any one of the manufacturers of branded multi-source drugs reported a true AWP, the reimbursement amount would have been lower. With respect to generics, if any manufacturer with an AWP at or above the median had reported a true AWP, the reimbursement amount would have been lower (in most cases). Therefore, joint and several liability is appropriate if there is no way to divide the injury to TPPs and consumers paying for drugs based on a J-code. However, given that plaintiffs purchased a discrete quantity of drugs from each manufacturer, this may be a case where the injury is divisible, rather than indivisible.

The earlier Restatement (Second) of Torts provides explanation and guidance regarding what constitutes a divisible injury:

d. Divisible harm. There are other kinds of harm which, while not so clearly marked out as severable into distinct parts, are still capable of division upon a reasonable and rational basis, and of fair apportionment among the causes responsible. Thus where the cattle of two or more owners trespass upon the plaintiffs land and destroy his crop, the aggregate harm is a lost crop, but it may nevertheless be apportioned among the owners of the cattle, on the basis of the number owned by each, and the reasonable assumption that the respective harm done is proportionate to that number. Where such apportionment can be made without injustice to any of the parties, the court may require it to be made.

Restatement (Second) of Torts § 433A cmt. d (1965); see also Bass v. Gen. Motors Corp., 150 F.3d 842, 846 (8th Cir.1998) (explaining that a divisible injury is one that “can be clearly separated and attributed either to the manufacturer or the original tortfeasor”). The caselaw provides little guidance in this area. Examples of divisible injuries include an injury that one party causes and another subsequently aggravates in some measurable way, Vanguard Sav. & Loan Ass’n v. Banks, 1995 WL 30589, *2-3, 1995 U.S. Dist. LEXIS 737, *7-8 (E.D.Pa. Jan. 24, 1995), and the flooding of tribal lands where it was possible to ascertain the percentage of water attributable to each contributing irrigation district, United States v. Imperial Irrigation Dist., 799 F.Supp. 1052, 1069-70 (S.D.Cal.1992).

“An injury is indivisible if, according to the applicable rules of causation, ... each relevant person caused the entire injury.” Restatement (Third) of Torts: Apportionment of Liability, § 7 cmt. e; see also Watts v. Laurent, 774 F.2d 168, 180 (7th Cir.1985) (explaining that an indivisible injury is one that “cannot be apportioned in any sensible way among the several defendants”). Examples of indivisible injuries seem even farther afield; they include brain damage, broken bones, and paraplegia, Richardson v. Volkswagenwerk, A.G., 552 F.Supp. 73, 84 (W.D.Mo.1982); death, Huddell v. Levin, 395 F.Supp. 64, 77 (D.N.J.1975); a fire started by multiple parties, Wausau Bus. Ins. Co. v. Turner Constr. Co., 143 F.Supp.2d 336, 345 (S.D.N.Y.2001); and lost profits from a breach of contract, Peacock v. Landquest, Ltd., 1993 U.S. Dist. LEXIS 4371, at *15 (W.D.Mich. Feb. 5, 1993). The parties cited no cases, and the Court could find none, that deal with a situation similar to this case.

Here, it is likely that the class-wide harm can be divided and apportioned based on the reasonable assumption that the harm is proportionate to the number of pills sold at the inflated AWP. Defendants bear the burden of demonstrating that the injury is divisible and proving the magnitude of the damages that they caused, through their relevant market share. See Restatement (Third) of Torts: Apportionment of Liability, § 26 cmt. h (“A party alleging that damages are divisible has the burden to prove that they are divisible.... The burden to prove the magnitude of each part is on the party who seeks division.”). Plaintiffs “have advised the Court that they will accept entry of judgments revised to reflect BMS’s, Schering’s, and Warrick’s individual market shares for each of their multi-source drugs for Massachusetts, measured on an annual basis for each year of the Class Period.” (Pis.’ Post-Trial Omnibus Trial Br. 59.) Defendants should therefore provide information on their relevant market shares for the purpose of calculating Class 2 damages for multi-source drugs. Otherwise, joint and several liability will be imposed. 80

6. Drug-by-Drug

In order to examine each defendant and each drug, I have identified three salient factors relevant to a finding of unfair conduct under Chapter 93A for both Class 2 and Class 3.

First, the most important inquiry asks: were there egregious spreads above the 30% yardstick expected in the industry? In particular, I focus on the extent and duration of the spreads to evaluate egregiousness.

Second, I will look at the company’s history of creating the spread. Did the manufacturer actually increase the AWP and/or list price, as opposed to just increasing the spread through discounts and rebates? Creating the spread by increasing the AWP comes at no cost to the pharmaceutical company and places the full financial burden of the spread on the payor and patient. This approach to expanding the spread is strong evidence of unethical conduct. Also relevant to this analysis is the legitimacy of the list price from which the markup is derived: Is it a real list price at which substantial sales were made or an unfair and deceptive price used to jack up the AWP? Finally, evidence that an AWP increase was intended to thwart Congress’s change in reimbursement rates will constitute evidence of unethical behavior.

Third, did the defendant engage in a proactive scheme to market the spread to doctors by encouraging them to purchase drugs because of their profitability rather than their therapeutic qualities? See OIG Compliance Program Guidance, 68 Fed. Reg. at 23,737 (“Active marketing of the spread includes, for example, sales representatives promoting the spread as a reason to purchase the product or guaranteeing a certain profit or spread in exchange for the purchase of a product.”).

The weight given to each of these factors depends on the particular circumstances of each manufacturer and each drug for each year; no single factor is necessarily determinative, but the size and duration of a mega-spread is the most significant factor. With these criteria in mind, I turn to each defendant.

a. AstraZeneca

Under these three criteria, I find that AstraZeneca engaged in unfair and deceptive conduct. First, from 1996 until 2002, spreads on Zoladex ranged from 40% to over 169%, exceeding the 30% yardstick in every year for both NDCs. Thus, the extent and duration of the spreads were significant. Second, from 1996 through 1999, AstraZeneca continued to increase WAC and the corresponding AWP such that beneficiaries and TPPs were forced to pay higher amounts despite the falling sales price of Zoladex. It is particularly troubling that AstraZeneca raised AWP in 1998 in order to torpedo Medicare’s attempt to reign in costs by reducing reimbursement to 95% of AWP in the BBA. Finally, AstraZeneca actively marketed the spread to physicians by repeatedly emphasizing the “Return to Practice” that could be obtained by prescribing Zoladex. Plaintiffs presented letters, emails, spreadsheets, and call notes from several years to document this campaign to sell Zoladex based upon profitability.

AstraZeneca raises several arguments to counter plaintiffs’ Chapter 93A claim. First, it claims that it did not keep the spreads secret. AstraZeneca notes that throughout the class period it reported an accurate average manufacturer’s price (“AMP”), a close proxy for ASP, to CMS for purposes of Medicaid. However, AMP data is confidential information that is unavailable to TPPs or consumers. Similarly, AstraZeneca points out that data about the actual price paid by physicians for Zoladex was available in the reports of IMS Health, a private pharmaceutical data provider. As noted earlier, that IMS data did not provide a clear representation of the spreads on Zoladex. AstraZeneca next points out that in 1996 it made efforts to start a “MAP” program under which a TPP would buy PADs through a specialty pharmacy. 81 AstraZeneca says it discussed the spreads with TPPs to persuade TPPs to use this system.

While it is true that some data regarding the acquisition costs of Zoladex was leaking into the public domain, this did not mitigate the unfairness of using a grossly inflated AWP (or WAC). 82 As explained earlier, TPPs faced significant structural impediments to changing the reimbursement system for a single drug. Furthermore, Medicare reimbursement was statutorily based on AWP, so TPPs were stuck paying for Zoladex based on the inflated AWP provided by AstraZeneca.

Second, AstraZeneca denies being in an “arms race” with Lupron that hurt the TPPs and consumers. After all, if it stopped offering physician discounts, it claims that physicians would have purchased the more expensive Lupron. While that may have been true, at least in the short run, one fraud does not excuse another. While AstraZeneca may initially have tried to do the right thing, it soon entered the fray by manipulating and marketing the spread with gusto.

Finally, AstraZeneca argues that for Class 2, it was CMS, and not AstraZeneca, that caused plaintiffs’ injury because CMS determined the allowed amount for Zola-dex. Although CMS was responsible for the twenty percent calculation owed by the patient, the allowed amount was clearly set by a statute that was known and understood by AstraZeneca. Every time a plaintiff reimbursed for Zoladex based upon the AWP, AstraZeneca caused a loss by reporting a false and grossly inflated AWP.

AstraZeneca’s conduct supports all of the factors I enumerated, and I therefore easily find that its actions were unfair to consumers and TPPs under Chapter 93A. Accordingly, I find liability for Zoladex during the years 1998-2002. 83 Using Dr. Hartman’s calculations, I find damages to Class 3 plaintiffs of $751,338 in 1998, $799,284 in 1999, $858,145 in 2000, $1,008,700 in 2001 and $1,033,962 in 2002. (See PX 4028, Attach. J.l.a.) Dr. Hartman will have to calculate the Class 2 damages consistent with this opinion.

b. Johnson & Johnson

1. Procrit

Two factors militate in favor of Chapter 93A liability for J & J on Procrit. First, J & J actively marketed the spread on Procrit despite having a policy prohibiting such conduct. J & J was touting “revenue” as a reason to give Procrit. The sales force was educated on the importance of the economics of prescribing Pro-crit and instructed on how to explain profit to a customer. Second, from 1991 through 1996, J & J did not raise its list price and AWP, but the first increases it reported in 1997 and 1998 summed to approximately 5%, the exact reduction that the BBA was implementing for Medicare at that time. Further price increases were then taken in the subsequent years 2000-2002.

Nevertheless, the spread for Procrit did not exceed 30% in any year for any of the 15 Procrit NDCs. In fact, most spreads were below 25%. As Dr. Rosenthal noted, Procrit is one of the drugs for which AWP seems to work well because the AWP closely tracks the ASP. Given this reasonable relationship throughout the class period, I find that J & J’s conduct regarding Procrit, while troubling, was not outrageous or egregious under Chapter 93A.

2. Remicade

The story for Remicade is somewhat similar. There is some evidence, though much less than for Procrit, that J & J was marketing Remicade based on its profitability. 84 J & J also increased the WAC and corresponding AWP for Remicade each year from 1999-2001. However, the spreads only exceeded 30% by 2.1% in 1999 and 1.9% in 2001. And according to the calculations of J & J’s expert, Jayson Dukes, those two spreads drop below 30% when a weighted average AWP is used for spread calculation rather than the June 30 AWP that Dr. Hartman uses. Using the factors, I find that there is no liability for Remicade.

Yet plaintiffs argue that a different expectations threshold should be used for Remicade: 25% rather than 30%. Remi-cade is unique because unlike substantially all other physician administered drugs and particularly all the drugs in this trial, J & J set the AWP for Remicade at 30% above the WAC. This is 5%-10% more than the expected markup that nearly all experts testified was common in the marketplace. Furthermore, J & J’s John Hoffman explained that part of the reason for setting the AWP at this level was that it was a price that the payors could bear. Plaintiffs therefore suggest that it is appropriate to use a 25% expectations threshold to determine liability.

This is a close call. Although Remi-cade’s AWP markup was higher than the generally understood industry standard markup, J & J’s minimal discounting resulted in a spread that was reasonably within the range of payor expectations. The Remicade spread hovered near 30% in every year, such that the AWP was predictably related to the actual acquisition costs. Furthermore, the 30% AWP markup was published by the industry compen-dia, contributing to the expectation that the spread would be approximately 30%. As such, there were no secret or deceptive spreads. Given that AWP closely tracked ASP throughout the period, and the spreads were all at or about 30%, I conclude that there is no liability for Remi-cade.

c. Bristol-Myers Squibb

At trial, BMS repeatedly argued that its WLP was a legitimate list price, and thus neither unfair nor deceptive. 85 BMS justified its pricing strategy, whereby it never decreased the list price despite heavy discounting, because it could always sell to a significant proportion of its customers at that list price. Dr. Bell provided detailed calculations of the percentage of sales that were made at or about WLP, in order to demonstrate the legitimacy of the list prices.

Plaintiffs argue that regardless of the number of sales at WLP, BMS had a duty to disclose to payors that sales were being made at substantial discounts off WLP. Plaintiffs emphasize that as the ASP dropped, and BMS held WLP constant, insurers and patients were paying deceptively inflated prices. Plaintiffs contend that they were misled by the WLP, did not have information about the confidential discounts, and did not know of the mega-spreads.

BMS relies on the FTC’s Guides Against Deceptive Pricing, which provide that a list price “will not be deemed fictitious if it is the price at which substantial (that is, not isolated or insignificant) sales are made.” 16 C.F.R. § 233.3(d). The FTC does not define “substantial” and there are no cases interpreting the Guides Against Deceptive Practices since its adoption in 1967. Both parties cite to cases that interpret an earlier form of the guidelines that required list prices to reflect the “usual and customary prices at which products are sold.” See Regina Corp. v. Fed. Trade Comm’n, 322 F.2d 765, 767 (3d Cir.1963); compare Helbros Watch Co. v. Fed. Trade Comm’n, 310 F.2d 868, 870 (D.C.Cir.1962) (construing 60% of sales at the list price as not sufficiently substantial prior to the adoption of the Guides Against Deceptive Pricing), with Federated Nationwide Wholesalers Serv. v. Fed. Trade Comm’n, 398 F.2d 253, 261 (2d Cir.1968) (construing 40% of sales as “substantial and significant” prior to the adoption of the Guides Against Deceptive Pricing). The sparse caselaw applying this language is inconclusive. I find and hold that if more than 50 percent of all sales were made at or about the list price, the list price will not be deemed fictitious.

For list prices, like WLP, it is expected that there may be some discounting, but that most customers are paying at or about the list price. Since the BMS AWPs were simply a formulaic 20 to 25 percent markup over WLP, the standard industry practice, I do not find Chapter 93A liability when a substantial number of sales were made at the WLP. However, when discounting became so prevalent that the list price no longer reflected the price that most people paid, it became unfair and deceptive to continue publishing such a list price upon which the AWP is based. See 16 C.F.R. § 233.3(a) (“To the extent that list or suggested retail prices do not in fact correspond to prices at which a substantial number of sales of the article in question are made, the advertisement of a reduction may mislead the consumer.”).

BMS also argues that oncologists are not involved in the negotiation of drug-prices because most belong to regional buying groups or GPOs that negotiate with the manufacturers. The amount of sales made to GPOs as opposed to physicians directly was never resolved at trial. Nevertheless, as Dr. Rosenthal explained, GPOs negotiate with the manufacturers for volume discounts on drugs which are then passed on as lower acquisition prices to doctors so that the effect is essentially the same. (11/15/06 Tr. 46:4-25 (Rosen-thal).)

Finally, BMS challenges Dr. Hartman’s damage calculations, arguing that he did not exclude capitated contracts that are not based upon AWP. Defendants’ expert Dr. Gaier calculated that as much as 43% of the reimbursements for BCBSMA were not based upon AWP. (See Gaier Aff. Attach. 37; Gaier Aff. ¶¶ 56-60.) Dr. Hartman disputes the reliability of Gaier’s calculations and finds that it contradicts his own claims analysis. Hartman reviewed BCBSMA Medigap payments as well as actual Medicare claims to obtain a sample of claims. Using this data, he found that most claims are paid based on AWP. (Hartman Rebuttal ¶¶ 68-70.) Furthermore, in his damages analysis, Hartman attempted to exclude all capitat-ed contracts that were not AWP-based. (11/21/06 Tr. 68:18-20 (Hartman).) I cannot say that his method was unreliable.

1. Etopophos

BMS’s Etopophos, a single-source drug throughout the class period, merits little discussion. Dr. Hartman calculated a spread above 30% in only one year, 1996. In that year, 99.9% of sales were made within 5% of list price. I find no liability for Etopophos.

2. Paraplatin

Throughout the class period, Paraplatin was a single-source drug that was often used in combination with Taxol. BMS marketed the spread on both drugs, emphasizing the profitability of the combined regimen in meetings with physicians primarily from 1998 through 2002. BMS made annual increases in the WLP, but was able to maintain a substantial amount of sales at that price. From 1993 to 2002 between 83% and 99% of sales were made within 5% of WLP each year. Discounting did result in spreads for certain NDCs, which reached as high as 67% for one NDC in 1999. However, the spreads were not consistently above 30% for any NDC. Given the legitimate list prices and relatively low and sporadic spreads, I find no liability for Paraplatin.

3. Taxol

Taxol lost exclusivity in 2000 when a competing brand entered the market, followed by generic competition in 2001. BMS was actively marketing the spread on Taxol from 1998 through 2002, as demonstrated by substantial evidence presented at trial. Up until 2001 when Taxol was subject to generic competition, only two Taxol NDCs had annual spreads that exceeded 30%, and both by less than 1%. However, in 2001 the spreads began to rise and less than 42% of sales were made near list price. In 2002, less than 1% of sales were made at list price and spreads reached as high as 500%. I therefore find that BMS’s conduct in marketing and manipulating the spread for Taxol violated Chapter 93A for the years 2001 and 2002. Using Dr. Hartman’s calculations, I find damages to Class 3 plaintiffs of $183,454 in 2001. (See Hartman Deck, Attach. J.2.a.) These damages arise from the six month period in 2001 after Taxol became subject to competition, during which time Dr. Hartman assumes that AWP pricing is still in effect. The Court does not assess damages for Class 3 after this time because pricing was no longer typically based on AWP in provider reimbursement benchmarks. The Court will await Dr. Hartman’s revised calculations with respect to Class 2. 86

4. Vepesid

Vepesid was dispensed in two forms, capsule and injectable. Although there was no evidence that BMS was aggressively marketing the spread on Vepesid, BMS did provide an online “Cost Differential” report, which could calculate the “AWP cost differential” for any BMS drug. Of significance, there were huge spreads on the injectable form throughout the class period. Beginning in 1996, the vast majority of sales were made at prices less than 50% of WLP (except in the year 2000). Spreads above 30% existed on at least half of the injectable NDCs for every year from 1994 until 2002. In 1998 and 1999 spreads reached over 1000%. This raises the question of whether the existence of mega-spreads alone, without any proactive spread marketing or increase in the published AWP, is sufficient to create liability under Chapter 93A.

After hearing all the evidence in this trial, I find that these mega-spreads are shocking and on their own prove a sufficient degree of unfairness and deception to impose Chapter 93A liability because they are so oppressive and injurious to the insurers and patients who must pay such inflated prices. I therefore find that the existence of the mega-spread, by itself, is a violation of Chapter 93A. For Vepesid, there is one year, 2000, for which over 55% of sales were made within 5% of list price. I exclude that year, and find liability for Vepesid injectable from 1998-1999 and 2001-2002. In contrast, Vepesid capsules only exceeded the 30% yardstick once, by 0.1%, and I therefore find no liability for the capsule form. Because Vepesid became multi-source prior to 1998, there are no Class 3 damages. The Court will await Dr. Hartman’s revised calculations for Class 2. 87

5. Cytoxan

Cytoxan also was produced in two forms, an injectable solution and tablets. There was no evidence of spread marketing on either form of Cytoxan. However, the injectable form had spreads of over 100% on certain NDCs in every single year from 1993 until 2002, peaking at 676% in 1999. Moreover, by 1999, virtually no sales were made at WLP, and as early as 1996, the vast majority of sales were made at prices less than 50% of WLP. (DX 2524.) I find that these mega-spreads alone are sufficient to find liability for the injectable form of Cytoxan from 1998-2002.

Spreads for the tablet form of Cytoxan were much smaller, and much less consistent. In 1999 two of the five total Cytoxan tablet NDCs had spreads of 31% each and later in 2002 the same NDCs had spreads of 34% and 39%. I find that these small, sporadic spreads are insufficient to assess liability for Cytoxan tablets.

Because Cytoxan became multi-source prior to 1998, there are no Class 3 damages. The Court will await Dr. Hartman’s revised calculations for Class 2.

6. Blenoxane

When Blenoxane became subject to generic competition in 1996 the spreads began to rise, and sales at WLP began to plummet. In every year from 1998 to 2002, at least two of the four Blenoxane NDCs had spreads exceeding 30%. The highest spreads ranged from 72% in 1998 to 199% in 2002. Although the spreads aren’t as shocking as some of the mega-spreads in this trial, the spreads were large and consistent throughout that time period.

Furthermore, WLP was no longer a true list price, with less than 16% of sales made within 5% of WLP in each year. Indeed, by the year 2000, virtually no sales were made at WLP, and by 2001, the vast majority of sales were made at prices less than 50% of WLP. I therefore find that the manipulation of these spreads was unfair under Chapter 93A, that WLP was not a true price, and that there is liability for Blenoxane from 1998 to 2002. However, there are no damages for Blenoxane. For Class 3, Blenoxane became subject to mul-ti-source competition in 1996 such that reimbursements were not based on AWP after December 1997. For Class 2, Dr. Hartman’s survey data indicated that for Blenoxane “there were no incidents of visits to doctors’ offices that were reimbursed under Medicare.” (11/20/06 Tr. 54:10-22 (Hartman).)

7. Rubex

Rubex was a multi-source drug for the entire period. Like many of the other BMS drugs, there is no evidence that BMS was marketing the spread on Rubex. However, from 1994 until 2002, spreads were consistently above 30% for at least two of the six NDCs, with the highest spreads ranging from 55% to 438%. Although for most of these years, the vast majority of the sales were made at less than 50% of list price, in 2001 62% of sales were made within 5% of WLP. I therefore find that the WLP in 2001 was a true list price and thus there is no liability. For the remaining years, however, the spreads are large and consistent. I therefore find liability for Rubex in 1998-2000, and 2002. Because Rubex has been multi-source throughout the class period, there are no damages to Class 3. The Court will await Dr. Hartman’s revised calculations regarding Class 2 damages.

d. Schering-Plough

1.Intron-A and Temodar

Of the four Schering-Plough drugs at issue, two can be addressed very briefly. First, regarding Temodar, all spreads are below 30% and I therefore find no liability. For the Intron-A NDCs that Dr. Hartman considered to be physician-administered, the spread exceeds 30% in only 3 years, 1996, 2001 and 2002. The highest spread is 32.6% in 2001. The only evidence of spread marketing on Intron-A is an internal memorandum trumpeting the profitability of Intron-A to sales representatives. However, the spread that can be calculated from that 1998 memorandum is only 14% and there were no spreads over 30% from 1997 to 2000. Given the isolated, minor spreads and little evidence of spread marketing, I find no liability for Temodar or Intron-A.

2.Proventil

Proventil, Schering-Plough’s branded al-buterol solution, is a somewhat strange case. Although there are spreads consistently in the 30%-60% range for 1992-1997, from 1998 until 2003 there is only a single occurrence of a spread exceeding 30%. In 2002, one of the four Proventil NDCs had a spread of 163%. During that five year period, Schering-Plough took several direct price and corresponding AWP increases on Proventil, but Proven-til’s ASPs rose as well except for the year 2002. 88 There is no evidence of any spread marketing on Proventil. It is therefore a close question as to whether there should be liability for Proventil in 2002. Although the spread in 2002 is of significant magnitude, it is an isolated, anomalous occurrence on one of the four Proventil NDCs. As such, I do not find that it rises to the level of unfairness prohibited by Chapter 93A.

3.Warrick’s albuterol sulfate

Finally, Warrick produces a generic version of albuterol sulfate. There is some evidence suggesting spread marketing, mainly advertisements listing the price and AWP for albuterol. As a generic, the spreads were large and consistent from 1993 through 2003. From 1998 until 2003, three of the seven NDCs had spreads of over 200% in every single year with some reaching over 600%. The spread for one NDC of Warrick’s albuterol reached as high as 867%.

I return to the conclusion that I reached for the multi-source BMS drugs: the persistent existence of mega-spreads is by itself unfair to insurers and patients who are paying based on a median AWP that has no relation to real acquisition costs. Warrick continued this practice despite knowing that patients were overpaying for the drug. Given the limited years for which plaintiffs have shown that Warrick’s AWP caused an inflated median price, I find liability for albuterol sulfate in Class 2 only for the years 1998 to 1999. Furthermore, the basis for damage calculations is limited to the difference between the actual median AWP and the “but for” median AWP had Warrick reported a true AWP. As a multi-source drug throughout the class period, there are no damages in Class 3. 89

F. Class 2 Damages

For Class 2, the information provided at trial is insufficient to calculate exact damages. {See Hartman Decl. Attach. 4 (providing only an aggregate of Class 2 damages, 1991-2003, using a 30% liability threshold).) The Court needs a breakdown of the damages for each drug, using the 30% threshold, for each of the years from 1998 until 2003 for which I have found liability. Defendants may provide their market shares in Massachusetts so that the Court can apportion the damage amount on that basis. If necessary, the Court will hold a damages phase of the bench trial.

III. ORDER

1. The Court orders dismissal of the J & J defendants.

2. The Court orders dismissal of Scher-ing-Plough (not including Warrick).

3. The Court finds liability for:

a. AstraZeneca: Zoladex (1998-2002)

b. BMS: Taxol (2001-2002); Vepesid (1998-1999, 2001-2002); Cytoxan (1998-2002); Blenoxane (1998-2002); Rubex (1998-2000, 2002)

c. Warrick: albuterol sulfate (1998-1999)

4. By August 1, 2007, the Court orders plaintiffs to provide calculations of the Class 2 damages consistent with these findings.

5. By August 1, 2007, in order to apportion damages for Class 2, the Court allows BMS to provide market share data in Massachusetts for Taxol, Vepesid, Cy-toxan, and Rubex for the years 1998-2002.

6. By August 1, 2007, in order to apportion damages for Class 2, the Court allows Warrick to provide market share data for Warrick’s generic albuterol sulfate for the years 1998-1999.

SO ORDERED.

APPENDIX A

Glossary of Terms

AMP.Average Manufacturer’s Price

ASP .Average Sales Price

AWP.Average Wholesale Price

BBA.Balanced Budget Act of 1997

BCBSMA. Blue Cross/Blue Shield of Massachusetts

CMS.Centers for Medicaid and Medicare Services

DHHS .Department of Health and Human Services

DME .durable medical equipment

DOJ .Department of Justice

EAC.Estimated Acquisition Cost

GAO.Government Accountability Office

GPO.group purchasing organization

HCFA . Health Care Financing Administration

IPA.independent practice association

LCA.Least Costly Alternative

MAC.Maximum Allowable Cost

MMA.Medicare Prescription Drug, Improvement and Modernization Act

NDC.National Drug Code

OIG. Office of Inspector General

PAD.physician-administered drug

PBM.pharmacy benefit manager

SAD. self-administered drug

TPP .Third Party Payor

WAC.Wholesale Acquisition Cost

WLP.Wholesale List Price

APPENDIX B

“Class 2: Third-Party Payor MediGap Supplemental Insurance Class” is defined as:

All Third-Party Payors who made reimbursements for drugs purchased in Massachusetts, or who made reimbursements for drugs and have their principal place of business in Massachusetts, based on AWP for a Medicare Part B covered Subject Drug that was manufactured by AstraZeneca (AstraZeneca, PLC, Zeneca, Inc., AstraZeneca Pharmaceuticals L.P., and AstraZeneca U.S.), the BMS Group (Bristol-Myers Squibb Co., Oncology Therapeutics Network Corp., and Apothecon, Inc.), SmithKline Beecham Corporation d/b/a GlaxoSmithKline, the Johnson & Johnson Group (Johnson & Johnson, Cento-cor, Inc., Ortho Biotech, McNeil-PPC, Inc., and Janssen Pharmaceutica Products, L.P.), or the Schering Plough Group (Schering-Plough Corporation and Warrick Pharmaceuticals Corporation).

In re Pharm. Indus. Average Wholesale Price Litig., 233 F.R.D. 229, 231 (D.Mass. 2006).

APPENDIX C

“Class 3: Consumer and Third-Party Payor Class for Medicare Part B Drugs Outside of the Medicare Context,” is defined as:

All natural persons who made or who incurred an obligation enforceable at the time of judgment to make a payment for purchases in Massachusetts, all Third-Party Payors who made reimbursements based on contracts expressly using AWP as a pricing standard for purchases in Massachusetts, and all Third-Party Pay-ors who made reimbursements based on contracts expressly using AWP as a pricing standard and have their principal place of business in Massachusetts, for a physician-administered Subject Drug that was manufactured by AstraZeneca (AstraZeneca, PLC, Zeneca, Inc., As-traZeneca Pharmaceuticals L.P., and AstraZeneca U.S.), the BMS Group (Bristol-Myers Squibb Co., Oncology Therapeutics Network Corp., and Apothecon, Inc.), SmithKline Beecham Corporation d/b/a GlaxoSmithKline, the Johnson & Johnson Group (Johnson & Johnson, Centocor, Inc., Ortho Biotech, McNeil-PPC, Inc., and Janssen Pharmaceutical Products, L.P.), or the Schering Plough Group (Schering-Plough Corporation and Warrick Pharmaceuticals Corporation). Included within this Class are natural persons who paid coinsurance (i.e., co-payments proportional to the reimbursed amount) for a Subject Drug purchased in Massachusetts, where such coinsurance was based upon use of AWP as a pricing standard. Excluded from this Class are any payments or reimbursements for generic drugs that are based on MAC and not AWP. In re Pharm. Indus. Average Wholesale Price Litig., 233 F.R.D. 229, 231 (D.Mass. 2006).

1 . See Appendix A for a glossary of terms used in this order.

2 . GlaxoSmithKline settled all claims prior to trial. AstraZeneca settled the claims involving Medicare beneficiaries prior to the jury trial scheduled for June 4, 2007. However, the claims involving these classes were not settled.

3 .These costs vary by disease, dosage, time period, frequency of treatment, weight of the patient and other factors. (See Rosenthal Dir. ¶ 14.)

4 . Except where otherwise noted, I use the term "spread” and "markup” interchangeably.

5 . See PX 4030 at ¶ 38 n. 55 for the specific National Drug Codes ("NDCs”) for each drug in the chart.

6 . See Appendix B for Class 2 definition.

7 . See Appendix C for Class 3 definition.

8 .The class does not include persons who make flat co-pays for every drug no matter what the price (like $10) because they are not affected by an inflated AWP. Also, very few people pay for the drugs at issue entirely out of pocket because they are so expensive. (Ro-senthal Dir. ¶ 24.) About 10 percent of the population with employer-sponsored coverage pays coinsurance for a physician's office visit, and the typical coinsurance rate is 20 to 25 percent.

9 . See PX 183 ("Effective immediately Bristol-Myers Oncology Division products factor used in determining the AWP should be changed from 20.5% to 25%.”). Red Book made the requested change.

10 . The drugs involved in this litigation represent a tiny percentage of the thousands of pharmaceutical products available in the United States market. In 2002, approximately 3 percent of Medicare’s spending was on physician-administered drugs. (DX 1275, Berndt Rep. ¶ 187.) Dr. Berndt estimates that "in 2002, expenditures on physician-administered drugs were likely less than 1.5 percent of national health expenditures, and considerably smaller in earlier years.” (Id.)

11 . HCFA stated:

Estimated acquisition costs would be based on individual carrier estimates of the costs that physicians, or other providers as appropriate, actually pay for the drugs. Carriers could survey a sample of the physicians who furnish the drugs to obtain cost information. As an alternative, carriers could request that physicians periodically provide cost information when they submit claims for payment for the drugs.

56 Fed.Reg. at 59,502.

12 . Multi-source drugs are drugs that no longer have patent protection so that several manufacturers can produce generic versions of the drug.

13 . Most states use AWP (or its formulaic counterpart WAC) as the pricing benchmark for Medicaid as well.

14 . A single-source drug is a drug that is under patent protection and produced by one manufacturer, so that there are no competitors producing generic versions of the drug. However, some single-source drugs face therapeutic competition from drugs that have different chemical compositions, but can be used for the same indication.

15 . Indeed, throughout this litigation these contracts were marked “highly confidential” pursuant to a protective order.

16 . See, e.g., Black Deck ¶ 6 (explaining As-traZeneca's Patient Assistance Program which provided products to economically disadvantaged patients free of charge); Kane Decl. ¶¶ 23-24 (explaining Schering's "Commitment to Care” and "SP Cares” programs which provide free drugs to qualified low income patients); 11/16/06 Tr. 25:15-26:16 (Dooley) (explaining J & J’s patient assistance program that helped low income patients pay for their portion of drug costs); 11/14/06 Tr. 115:19-117:19 (Hoffman) (explaining that Centocor provides free Remicade to patients with incomes of 300% or less of the federal poverty level); Pasqualone Aff. ¶ 22 (explaining BMS's "ProCert” program which provided reimbursement assistance to health care providers).

17 . Although AWP can influence MACs, most MACs are proprietary and the specific formulas that are used to develop those MACs are unknown. (11/15/06 Tr. 119:7-120:4 (Rosen-thal).) Dr. Hartman assumes that the MACs are not based on AWP. (Hartman Decl. ¶ 155(b).)

18 . For single-source drugs that lose patent protection, Dr. Hartman continues to calculate damages for six months after the first generic launch. At that point he assumes that multiple generic launches occur and MAC pricing is used. (Hartman Decl. Í 155(a).)

19 .The important role of PBMs in the area of self-administered drugs is described at length in the memorandum on class certification. See In re Pharm. Indus. Average Wholesale Price Litig., 230 F.R.D. 61, 71-73 (D.Mass. 2005).

20 . Dr. Hartman refers to this expectation that AWP exceeded ASP by 30 percent as his “expectations yardstick.” Dr. Hartman uses the yardstick to find liability whenever a drug exceeds that threshold. This expectations yardstick is explained in detail in the Conclusions of Law, infra.

21 . Leucovorin is a cancer drug, not at issue in this trial, that was highly profitable to physicians under the AWP system. (See DX 2641 at 15.)

22 . See Bell TI Aff. App. B (listing articles in The New York Times, Wall Street Journal, Boston Globe, and other popular press).

23 . Every drug has a unique identifying 11-digit number called an NDC. The first 5 digits identify the firm marketing the drug, the next 4 digits identify the specific strength, dosage form, and formulation of the product, and the final 2 digits identify the package size and package type. (DX 1275, Berndt Report ¶ 192.)

24 . The statute defines "average sales price" as “the manufacturer's sales to all purchasers" divided by "the total number of such units of such drug or biological sold by the manufacturer." 42 U.S.C. § 1395w-3a(c)(l). The average sales price "shall include volume discounts, prompt pay discounts, cash discounts, free goods that are contingent on any purchase requirement, chargebacks, and rebates. ..." Id. § 1395w-3a(c)(3).

25 . Dr. Gaier argues that the results under the MMA are inconclusive and introduced a graph demonstrating that the annual percent age growth in Medicare Part B expenditures per beneficiary have increased in 2004 and 2005. (11/29/06 Tr. 21:2-22:10 (Gaier); DX 1496.) While he believes that graph is "evidence that the overall Medicare Part B expenditures are going up," Dr. Gaier concedes that he doesn't have the data to analyze why that is happening and acknowledges that there are a variety of possible reasons. (Id. 21:15-23:8.)

26 . CMS established the Health Care Common Procedure Coding System ("HCPCS”) which provides 5 digit J-codes to be used for billing injectable drugs. (See DX 1275, Berndt Report ¶ 193.) For multi-source drugs, multiple NDCs for drugs sold by various manufacturers are reimbursed under the same J-code. (See id.)

27 . Some private insurers pay physicians for drugs on a capitated basis, i.e., the physician and the plan negotiate over a drug budget for each patient, with the physician bearing the risk that payments received may not be adequate to cover his other costs of services provided. (Bell T1 Aff. ¶ 67.) This lawsuit does not cover these capitated agreements. Dr. Hartman's damage assessment excludes any reimbursements unrelated to AWP.

28 . If the staff model HMO clinics included Medicare patients, Medical West, Inc. generated revenue from this spread. (See 11/8/06 Tr. 103:10-103:22 (Coneys) (expressing uncertainty about whether Medicare patients were treated by the staff model HMO).)

29 . A Taft-Hartley fund provides health and welfare benefits for union members. The fund, pursuant to federal law, is "administered jointly by employer-designated trustees and union-designated trustees.” Levy v. Local Union No. 810, 20 F.3d 516, 517-18 (2d Cir. 1994) (citing 29 U.S.C. § 186(c)(5)(B) (West 1978)).

30 . These drugs are Zoladex (AZ), Cytoxan in-jectable (BMS), Paraplatin (BMS), Rubex (BMS), Taxol (BMS), Procrit (J & J), and albuterol/Proventil (SPW). (PX4012.)

31 . Defendant “AstraZeneca'' collectively in- eludes AstraZeneca Pharmaceuticals L.P., Zeneca, Inc., and AstraZeneca U.S. AstraZ-eneca Pharmaceuticals L.P. and Zeneca, Inc. are U.S. subsidiaries of AstraZeneca, PLC, a limited liability company domiciled in the United Kingdom. AstraZeneca U.S. maintains its headquarters in Wilmington, Delaware.

32 . WAC is sometimes referred to as the list price or catalogue price.

33 . Dr. Hartman continues to assess damages after 1999, when the LCA went into effect.

34 . TAP later pled guilty to conspiring to violate the Prescription Drug Marketing Act, based on conduct during this time period including allegations of encouraging urologists to bill for free samples, inflating AWP to market the spread, and providing kickbacks to doctors who prescribed Lupron rather than Zoladex. (Hartman Decl. ¶ 24, citing Med-PAC report.)

35 . Both Zoladex and Lupron were LhRh agonists.

36 . Plaintiffs’ expert, Dr. Hartman, calculates the ASP for each drug NDC using data from the manufacturers. Dr. Hartman includes price reductions and discounts in his calculation of ASP. Hartman's definition of ASP is generally consistent with the definition in the MMA. (Compare Hartman Decl. ¶ 3, with 42 U.S.C. § 1395w-3a (defining ASP as the manufacturer’s total sales divided by the total number of units sold, and including "volume discounts, prompt pay discounts, cash discounts, free goods that are contingent on any purchase requirement, chargebacks, and rebates”).)

37 . Dr. Hartman recalculated the ASPs for Zo-ladex, omitting the inclusion of "free goods.” (See Hartman Rebuttal ¶ 6.) The updated ASPs and resulting spreads are found at PX 4028.

38 . Apothecon was a BMS subsidiary that manufactured and sold primarily generic drugs. BMS sold Apothecon's assets in 2000. There are no Apothecon drugs at issue in this case.

39 . On May 11, 2005, after the end of the class period, OTN became an independent, privately-held company. (Akscin Decl. ¶ 2.)

40 . WLP is essentially the BMS term for WAC. (Sea Hartman Decl. ¶ 44.)

41 . Paraplatin became subject to generic competition in November 2004 after the class period ended. (Bell BMS Aff. § 13.)

42 . Frank Pasqualone, Senior Vice President of the Oncology division at BMS, testified: “I do not recall specifically why that happened; however, in my experience there are times when generic supply becomes constrained and we are, therefore, able to make non-contract (spot) sales at higher transaction prices.” (Pasqualone Aff. ¶ 41.)

43 . Dr. Hartman’s original spreads for Temo-dar included several over 30%. He later revised his calculations of ASP to exclude the donation of free goods to a patient assistance program, AmeriCares. (See Hartman Rebuttal ¶¶ 7-8, Attach. A; Addanki Am. Decl. ¶¶ 69-70.)

44 . As he did with Temodar, Dr. Hartman recalculated the ASPs for Intron-A to exclude free good donations. These percentages reflect those revised spread calculations. (See Hartman Rebuttal ¶¶ 7-8, Attach. A; Ad-danki Am. Decl. ¶¶ 69-70.)

45 . This method of calculating liability and damages also applies to the BMS multi-source drugs.

46 . See In re Mass. Diet Drug Litig., 338 F.Supp.2d at 205-10 (deferring until trial the issue of whether wide-spread publicity over several years regarding the potential danger of diet drugs was sufficient to put plaintiff consumers on notice of their claims); Thompson v. Metropolitan Life Ins. Co., 149 F.Supp.2d 38, 50-53 (S.D.N.Y.2001) (denying summary judgment because twenty-four articles over a sixty year period and a segment telecast on 60 Minutes was insufficient to find that a class action of African-American insureds had constructive notice of discriminatory overcharges by defendant insurer); Sawyer v. Indevus Pharms., Inc., 2004 WL 1739405, *12-16, 2004 Mass.Super. LEXIS 274, *36-49 (Mass.Super.Ct. July 26, 2004) (denying summary judgment because widespread coverage of problems with diet drugs was not sufficient as a matter of law to put plaintiffs on notice of their injuries).

47 . A much debated issue at trial was whether the knowledge of the staff model HMOs, Medical West and Medical East, should be attributed to BCBSMA. Due to the fact that BCBSMA was on inquiry notice of its claims, triggering the statute of limitations bar in 1997, the dispute is inconsequential for purposes of determining when the statute of limitations was triggered. Furthermore, because BCBSMA sold the staff model HMO in 1997, any possible agency relationship or knowledge about drug prices learned through inter-company communications would not affect the knowledge of BCBSMA regarding drug prices available to physicians (as opposed to HMOs) after that point. To the extent that defendants have argued that BCBSMA is an atypical plaintiff because it had to face a unique statute of limitations defense, that argument is rejected.

48 .Massachusetts also tolls the statute of limitations to "the discovery of [the] cause of action” if "a person liable in a personal action fraudulently conceals the cause of such action from the knowledge of the person entitled to bring it.” Mass. Gen. Laws ch. 260, § 12. Plaintiffs argue that defendants engaged in fraudulent concealment primarily by keeping pricing information confidential and using confidentiality clauses in all of their contracts with physicians. Defendants argue that confidentiality clauses are a normal business practice for manufacturers in a competitive market. It is not necessary to resolve whether confidentiality clauses amount to fraudulent concealment because the secret was out by August of 1997 when the BBA was signed into law.

49 . Class 3 consumers who made co-payments clearly have a claim under § 9.

50 . "Person” is defined to include "natural persons, corporations, trusts, partnerships, incorporated or unincorporated associations, and any other legal entity.” Mass. Gen. Laws ch. 93A, § 1.

51 . See Arruda Aff. ¶ 14 (testifying to “BCBSMA's status as a not-for-profit organization organized under M.G.L. c. 176A and B”).

52 . There is a reasonable inference, though, that the staff model HMO may have profited from the inflated reimbursement for Medicare drugs prior to 1997.

53 . After reviewing the relevant law, plaintiffs also satisfy the requirements necessary to bring an action under § 11. Given the finding that § 9 is appropriate, I decline to fully address those issues.

54 . “The attorney general may make rules and regulations interpreting the provisions of subsection 2(a) of this chapter.” Mass. Gen. Laws ch. 93A, § 2(c).

55 . See, e.g., Swenson v. Yellow Transp., Inc., 317 F.Supp.2d 51, 55 (D.Mass.2004) (finding that the violation of federal and state motor carrier regulations, which were not directed to the protection of consumers, could not constitute a per se unfair or deceptive act); United States ex rel. Metric Elec., Inc. v. Enviroserve, Inc., 301 F.Supp.2d 56, 71 (D.Mass. 2003) (holding that conduct violating a state statute prohibiting unfair acts by insurers is not a per se violation of Chapter 93A as an unfair or deceptive act); Darviris, 795 N.E.2d at 1201 (finding that a violation of the patient’s bill of rights is not a per se violation of Chapter 93A); Reiter Oldsmobile, Inc. v. General Motors Corp., 378 Mass. 707, 393 N.E.2d 376, 378 (1979) ("Not every act made unlawful by statute is unfair or deceptive within the meaning of [Chapter] 93A.”).

56 . Moreover, guidelines have unclear precedential weight. See Federal Trade Comm’n v. Mary Carter Paint Co., 382 U.S. 46, 47-48, 86 S.Ct. 219, 15 L.Ed.2d 128 (1965) ("These, of course, were guides, not fixed rules as such, and were designed to inform businessmen of the factors which would guide Commission decision.”); B. Sanfield, Inc. v. Finlay Fine Jewelry Corp., 168 F.3d 967, 973 n. 4 (7th Cir.1999) (“We recognize that the federal guideline is merely that, and as such it does not have the same force as the Illinois regulation.”); In re John Surrey, Ltd., et al., 67 F.T.C. 299, 1965 WL 92786, 1965 FTC LEXIS 42, at *69-70 (Mar. 16, 1965) ("The Guides are not designed to be an encyclopedic restatement of the law regarding deceptive pricing, as it has been developed in Commission and court decisions under Section 5 of the Federal Trade Commission Act.... They are to be considered as guides, and not as fixed rules of ‘do's’ and 'don’ts,' or detailed statements of the Commission’s enforcement policies.”) (emphasis in original).

57 . Section 3.04 provides:

No claim or representation shall be made by any means which has the capacity or tendency or effect of deceiving buyers or prospective buyers as to the value or the past, present, common or usual price of a product, or as to any reduction in price of a product, or any saving relating to a product.

940 Mass.Code Regs. 3.04 (emphasis added).

58 . Section 3.05(1) provides:

No claim or representation shall be made by any means concerning a product which directly, or by implication, or by failure to adequately disclose additional relevant information, has the capacity or tendency or effect of deceiving buyers or prospective buyers in any material respect.

940 Mass.Code Regs. 3.05(1).

59 . See, e.g., Barnes, 370 F.3d at 176 (TISA); Dean v. Compass Receivables Mgmt. Corp., 148 F.Supp.2d 116, 119 (D.Mass.2001) (FDCPA); Martin v. Sands, 62 F.Supp.2d 196, 201 (D.Mass.1999) (FDCPA); Fidler v. Cent. Coop. Bank, 210 B.R. 411, 430 (Bankr. D.Mass.1997) (TILA), rev’d on other grounds, 226 B.R. 734 (Bankr.D.Mass.1998).

60 . The FDCPA is also a part of the Federal Consumer Credit Protection Act, which is specifically listed in section 3.16(4). See Martin, 62 F.Supp.2d at 201.

61 . Therefore, to use a hypothetical example, if the AWP of a drug is $100 and the ASP is $75, there is a spread of 33% = ($100 — $75)/$75. Some publications, including the OIG, calculate the spread as a discount off of AWP (rather than a markup above ASP) equal to (AWP-ASP)/AWP. In the hypothetical I created, Hartman's spread of 33% is thus equal to a discount of 25% off of AWP [ ($100-$75)/ $100], This can create some confusion, since the parties variously use percentages to refer to either a discount off of AWP or a spread. To use a real example, a November 1992 OIG report found that Doxorubicin (Rubex) could be purchased at discounts of 56% to 59% off of AWP (DX 1053 at App. Ill) which is equivalent to a markup or spread of 127% to 144% above ASP. See Hartman Decl. ¶ 77(c) n. 123 (explaining the calculation used to convert from a discount to a spread).

62 . Hartman notes that although he is not aware of any survey information that has “documented systematic spreads on generic physician-administered pharmaceuticals,” he finds "no compelling reason that pricing expectations ... would be more educated (i.e., different) than the observed relationship for single-source physician-administered drugs.” (Hartman Deck ¶ 147 n. 184.) He therefore uses the same 30% yardstick for all PADs.

63 . Recall that ASP is the actual average acquisition cost of providers, taking into account rebates, discounts, chargebacks, free samples, and the like. Hartman's definition is generally consistent with the definition in the MMA. (Compare Hartman Decl. ¶ 3, with 42 U.S.C. § 1395w-3a (defining ASP as the manufacturer’s total sales divided by the total number of units sold, and including "volume discounts, prompt pay discounts, cash discounts, free goods that are contingent on any purchase requirement, chargebacks, and rebates”).)

64 . The Dyckman survey, which reported on surveys of 33 large private health plans in the MedPAC report, found typical spreads of AWP plus or minus 15%; other independent studies in the MedPAC report found that private payors reimbursed from a low of AWP minus 20% to a high of AWP plus 10%. (See Hartman Decl. ¶ 123(b)-(c).) According to Hartman, some less informed TPPs reimbursed providers at prices greater than the AWPs because they believed that AWP was an actual average, as stated by publications like First DataBank. (See id. ¶¶ 105-06.)

65 . Defendants repeatedly point out that Dr. McFadden is a Nobel prize-winner. Plaintiffs repeatedly point out that he has no background in healthcare economics. Unfortunately, because of a family emergency, his oral testimony was truncated, and defendants have not relied much on his opinion in their briefs.

66 . AstraZeneca points out that the prices paid by physicians for Zoladex were available in the reports of IMS Health, which collects private pharmaceutical data that is available for purchase. Professor Gould, an economist, provided a chart comparing the IMS data on Zoladex sales prices to the ASPs calculated by Dr. Hartman. (See Gould Deck ¶ 20, Fig. 6.) According to the chart, the IMS data follows the WAC more closely than the ASP up until 1999. It is, therefore, likely misleading concerning the true prices that doctors were actually paying for Zoladex. From 1999 to 2003, the IMS data and Dr. Hartman’s calculations are nearly identical. This chart high lights the difficulty in gaining accurate information for PADs.

67 . Defendants made a very perfunctory argument that plaintiffs had a duty to mitigate their damages. Since this issue was not well-briefed by defendants, I will not address it any further.

68 . Defendants do challenge Dr. Hartman's failure to give significance to publicly available reports documenting spreads exceeding 30% on some multi-source PADs. {See Defs.' Mem. in Support of their Renewed Motion to Strike the Expert Test, of Dr. Hartman 4-5.) Dr. Hartman concluded that this information was "sufficiently idiosyncratic or limited so as to be insufficient for market participants to draw any conclusions regarding Defendants’ systematic abuse of the AWP system through spreads for Defendants’ multi-source drugs.” (Hartman Decl. ¶ 6 (emphasis in original).) Hartman says that the awareness of large spreads began to reach public awareness in 1996 primarily with respect to the generic drug albuterol. {Id. ¶ 77(d).) While this is a fair dispute as to what weight to give public reports, it does not undermine the reliability of Dr. Hartman’s methodology, but rather reflects a disagreement over the weight to be given to certain data.

69 . For example, Dr. Hartman included free goods in his calculation of ASP, picked certain dates for calculating a spread, or predicted certain data about ASPs in 2003 from past "trends.” The Court found defendants’ challenges persuasive in undermining the weight of Dr. Hartman’s testimony on these and other points.

70 . While there is a duty to mitigate, defendants mention this defense only in passing. In any event, mitigation was only possible once Medicare had developed an ASP methodology. Furthermore, I reject defendants argument that BCBSMA's decision to continue using AWP defeats causation. If defendants had reported true AWPs, plaintiffs would have paid less.

71 . Defendants did not present any evidence regarding the rate setting practice for BCBSMA's commercial plans.

72 . Interestingly, the 20 to 25 percent markup of AWP was well known in the industry at the time the BBA was enacted in 1997. As such, this practice is arguably relevant in construing the meaning of the statutory term AWP. Defendants never made this argument. Rather, defendants’ proposed statutory interpretation that Congress intended AWP to be a blank check to the industry to impose whatever markup it wanted is supported nowhere in the legislative or trial record.

73 . Blenoxane became multi-source in 1996, Vepesid injectable in 1994, Vepesid capsules in 2001, and Cytoxan tablets in 2000. Taxol faced branded competition in 2000 and generic competition in 2001. Cytoxan injectable and Rubex were multi-source throughout the class period.

74 . Proventil became multi-source in 1992.

75 . Remember, multi-source drugs are not included in the damage calculations in Class 3 because many such drugs were not reimbursed based on AWP, but on other benchmarks developed by TPPs like “maximum allowable cost.” (See Hartman Decl. ¶ 156.)

76 . Furthermore, the industry practice was that generic drugs pegged their AWPs at 10%-20% off of the brand AWP. Thus, by publishing an inflated brand AWP, the manufacturer contributed to the establishment of a median fictitious AWP.

77 . According to Dr. Addanki, he used information from Medispan to calculate the median of the generics and compare it to War-rick’s AWPs. (See DX 2920.) While it is not clear whether this was the actual median AWP used for reimbursement by Medicare, Warrick presented it as such and it was not disputed.

78 . In 1998, reporting a true AWP would have shifted the median from Warrick’s AWP of $0.7495 to Aligen’s AWP of $0.7325. (See DX 2920.) In 1999, reporting a true AWP would have shifted the median from $0.7410 to $0.7325. (See id.)

79 . Use of a joint and several liability theory does not require that all tortfeasors are joined in the action. See, e.g., Shantigar Found, v. Bear Mt. Builders, 441 Mass. 131, 804 N.E.2d 324, 332 (2004) ("Under our current system of joint and several liability, a plaintiff injured by more than one tortfeasor may sue any or all of them for her full damages.”) (citations omitted).

80 . Plaintiffs have also pressed a theory of market share liability. However, that theory of liability typically deals with products liability cases in which it is impossible to identify the actual product that caused a plaintiff's injury. See McCormack v. Abbott Labs., 617 F.Supp. 1521, 1525 (D.Mass.1985) (“The emergence of a market-share theory of liability in American jurisprudence stems from the recognition by several courts that the field of products liability has been changed drastically with the advent of mass production of fungible goods and complex marketing methods.”). Here, plaintiffs have proved that defendants have caused injury to each of the class members, and the only difficulty is determining the amount of damages.

81 . AstraZeneca dropped the MAP program in 1999 or 2000 because of concerns from purchasing physicians.

82 . While AstraZeneca claims that it continued to make some sales at WAC, it does not provide data that there were a substantial number of such sales.

83 . Dr. Hartman has calculated damages for Zoladex in 2003 using "trends” from 1998-2002. (See Hartman Decl. Attach. J. 6.) He does not provide any figure for the actual ASP or corresponding spread in 2003, so I decline to find liability in that year. Dr. Hartman also uses this "trend” approach to find liability and calculate 2003 damages for many other drugs at trial. I similarly decline to find liability in 2003 without actual evidence of the ASP and spread.

84 . After trial, plaintiffs did newly discover and seek to introduce evidence that J & J sales representatives were marketing the spread on Remicade using a slide presentation that contained an audible "Ka-Ching” sound on the slide showing the profit potential of Remicade. (See Docket No. 3687.) I allow the motion. While this is certainly strong evidence of spread marketing, it does not affect my conclusion that because of the small spreads and predictable relationship between AWP and ASP there is no liability.

85 . Recall that BMS only reported a WLP to the publications, and not an AWP. As explained earlier, WLP and AWP are formulai-cally related by a set markup. BMS approved and republished the AWPs listed by the publications. BMS’s argument that it cannot be responsible for the actions of the publishers is unavailing.

86 . Class 2 damage calculations for all BMS multi-source drugs, including Taxol, will be based upon the difference between the ASP for BMS’s brand drug and the corresponding median generic AWP. (See Hartman Decl. ¶ 72.)

87 . It appears from the record that reimbursements for the Vepesid and Cytoxan capsules can be differentiated from those of the injecta-bles because the two forms are associated with different J-codes. (See Hartman Decl. Attach. 3.) If so, Dr. Hartman shall calculate damages only for injectables. If not, the Court will permit BMS to demonstrate how to fairly apportion the damages.

88 . Dr. Addanki calculated that 83% of sales for Proventil were made within 5% of WAC between 1991 and 2004. Because it is an average across 14 years, that figure is not as helpful as a year-by-year calculation. The more relevant data point to this analysis, which was not presented at trial, is the percentage of Proventil sales near WAC in 2002.

89 . At trial, Warrick argued that Sheet Metal Workers' 0.083% albuterol reimbursements were based upon a billed charge, and not at all based upon AWP. However, defendants later conceded that the discrepancy in the charges was due to the fact that when albuter-ol was administered in combination with ip-ratropium bromide, the lower median AWP for the concentrated albuterol solution (rather than the unit dose) was sometimes used for reimbursement. (See Docket No. 3480.) Warrick now argues that Dr. Hartman’s original damages for albuterol were overstated because the 0.083% albuterol solution was sometimes reimbursed based upon a lower median than Dr. Hartman assumed. When asked about this possibility, Dr. Hartman noted that he believes his damages for albuterol may have been overstated given his use of manufacturer data. {See 12/18/06 Tr. 141:25-143:2 (Hartman).) This is a highly complex factual issue that was not discovered until the final days of trial and was thus poorly explained and vetted. If feasible, Dr. Hartman shall take this information into account in his recalculation of damages pursuant to this opinion.