Docket: T-1048-07
Citation:
2017 FC 88
Ottawa, Ontario, April 4, 2017
PRESENT: The
Honourable Mr. Justice O'Reilly
|
BETWEEN:
|
|
ELI LILLY
CANADA INC., ELI LILLY AND COMPANY, ELI LILLY AND COMPANY LIMITED, AND ELI
LILLY SA
|
|
Plaintiffs
(Defendants by
Counterclaim)
|
|
and
|
|
TEVA CANADA
LIMITED
|
|
Defendant
(Plaintiff by
Counterclaim)
|
AMENDED PUBLIC JUDGMENT AND REASONS
(Redacted from Confidential Judgment
and Reasons issued January 30, 2017)
I.
Overview
[1]
Teva Canada Ltd (formerly Novopharm Ltd) seeks
damages from Eli Lilly Canada, Inc as compensation for having been prevented
from coming to market in 2006-2007 with a generic version of a medicine called
olanzapine. Olanzapine is useful primarily in the treatment of schizophrenia. Teva
alleges that it suffered losses for having been kept out of the olanzapine
market as a result of Lilly’s application for an order prohibiting Teva from
obtaining a Notice of Compliance (NOC) from the Minister of Health, and from the
corresponding regulatory stay, pursuant to the Patented Medicines (Notice of
Compliance) Regulations, SOR/93-133. Teva was kept out of the olanzapine
market until the proceedings under the Regulations were completed on June 5,
2007, when Justice Roger Hughes ruled against Lilly (Eli Lilly Canada Inc v
Novopharm Ltd, 2007 FC 596).
[2]
Subsequently, Lilly commenced an action against
Teva for infringement of its patent for olanzapine. I found that Lilly’s patent
for olanzapine was invalid (Eli Lilly Canada Inc v Novopharm Ltd, 2009
FC 1018). I also concluded that Teva was entitled to damages under s 8 of the Regulations
in an amount which would be determined in a separate proceeding (see Annex II for
all provisions cited). An appeal from my decision was allowed in part (Eli
Lilly Canada Inc v Novopharm Ltd, 2010 FCA 197). In a later decision, I
again concluded that Lilly’s patent was invalid (Eli Lilly Canada Inc v
Novopharm Ltd, 2011 FC 1288). The latter decision was upheld by the Federal
Court of Appeal; the Supreme Court of Canada denied leave to appeal.
[3]
In effect, the dispute has come to me for a third
time to decide a new question: What is the amount of the damages, if any, to
which Teva is entitled under the Regulations for the time it was kept off the
market? To answer that question, one must create a hypothetical world in which Lilly
would not have brought an application whose effect was to deny Teva access to the
olanzapine market. In that hypothetical world, Teva might have come to market on
the strength of an NOC for generic olanzapine as early as March 3, 2006, when the
Minister would have granted an NOC to Teva but for the proceedings initiated by
Lilly. Lilly disputes that start date on the basis that Teva had earlier abandoned
its claim for damages, and that Teva was not actually in a position to bring
its product to market until the spring of 2007, at the earliest.
[4]
Lilly also raises a number of grounds on which
Teva’s damages should be discounted, including: Teva has included losses that
are not attributable to the operation of the Regulations; Teva has failed to
take account of the likely presence of another generic manufacturer, Apotex
Inc, in the market at the same time; Teva has overstated the actual profits it
would have realized in the various provinces; and Teva’s claim fails to include
the full amounts that Teva would have given to pharmacies to promote its
product (so-called “trade-spend”). Each of these
factors requires separate analysis.
[5]
Accordingly, the main issue, the amount of
damages owed to Teva, raises five distinct questions:
1.
What is the period of liability?
2.
What was the size of the olanzapine market?
3.
What was the generics’ share of the olanzapine
market?
4.
What was Teva’s share of the generic olanzapine
market?
5.
What is the real amount of Teva’s losses?
[6]
The parties do not ask me to make any
calculations. They ask me simply to make the factual findings necessary for
those calculations to be made.
II.
The Legal Framework
[7]
Under the Regulations, a drug company holding a
patent on a particular drug (the “first person”)
can commence proceedings to prohibit another company wishing to market a
generic version of that drug (the “second person”)
from obtaining an NOC until the latter has addressed the former’s patent or
until the patent has expired. The second person can address the patent by
alleging that its product will not infringe the patent or that the patent is
invalid. Until the Court has ruled on those allegations, the second person
cannot enter the market. The Regulations impose an automatic stay for 24 months
or until the first person’s prohibition application has been dismissed.
[8]
If the first person fails to persuade the Court
that the second person’s allegations are unjustified, the first person will not
obtain its prohibition order and the second person will be free to obtain its
NOC. However, the Regulations recognize that the automatic stay will keep the
second person off the market for the duration of the proceedings even where the
second person’s allegations are ultimately found to be justified. Accordingly,
the Regulations state that where the Court dismisses a first person’s
application for a prohibition order, the first person is liable to the second
person for any losses suffered during the relevant period. The relevant period
begins on the date certified by the Minister as being the date the second
person would have obtained its NOC but for the proceedings initiated by the
first person, unless the Court finds that another date is more appropriate. The
relevant period ends on the date the first person’s application was dismissed.
[9]
Teva shoulders the legal burden of establishing
all of the elements of its claim for damages: this includes showing that its alleged
losses were a product of the operation of the Regulations (Pfizer Canada Inc
v Teva Canada Limited, 2016 FCA 161 at para 64). To be recoverable,
there must be a causal connection between the damages Teva seeks and the proceedings
initiated by Lilly. The essential question is: What would have happened if
Lilly had not applied for an order of prohibition against Teva?
[10]
Lilly has an evidentiary burden to respond to
Teva’s evidence, and bears the legal burden in respect of its defences. For
example, in its defence, Lilly maintains that Teva had earlier abandoned its
claim for damages and that the start date for the period of liability is much
later than the date certified by the Minister. Lilly bears the burden of proof
on those issues (Pfizer Canada Inc v Teva Canada Limited, above, at para
65).
III.
Evidentiary Issues
[11]
The parties disputed two major evidentiary
issues. The first was whether fact witnesses could testify about what they
thought would or would not have happened in the but-for world. For the most
part, this involved Lilly objecting to testimony proffered by Teva’s fact
witnesses on the basis that asking those witnesses hypothetical questions
invited them to provide opinion evidence, an area that should be left to
qualified experts. Teva argued repeatedly that it had no choice but to ask fact
witnesses about the but-for world given its burden of proving what would have
happened if Lilly had not initiated proceedings under the Regulations. Teva
pointed me to previous cases where this kind of evidence had been allowed. During
the trial, I allowed some hypothetical questions to be put to fact witnesses,
reserving on Lilly’s objection until I had an opportunity to consider the
objection further and review the authorities.
[12]
The second evidentiary issue related to alleged
hearsay evidence. Lilly objected to the admission of certain documents for the
truth of their contents. Again, for the most part, I reserved on Lilly’s
objections until I had an opportunity to consider the admissibility of the
documents in question, considering their necessity and reliability in the
context of the case as a whole.
[13]
With respect to the first issue, I agree with
Lilly that the opinions of fact witnesses are not admissible. During the trial,
I suggested to counsel that the best way to provide the relevant evidence to
the Court would be to explore with fact witnesses what they did in the real
world. The witnesses could then be asked whether they knew of any reason why
they would have acted differently in the but-for world. This would confine fact
witnesses to their own knowledge and experience, as opposed to asking them, in
an open-ended fashion, what they would have done or what they thought would
have happened in the but-for world. The latter involved inadmissible opinion
evidence; the former related to admissible facts within the witness’ knowledge.
[14]
I note that the Federal Court of Appeal has
ruled that fact witnesses can, in a limited way, give evidence about a
company’s “general intentions in the hypothetical
world” and “the general steps it took to prepare
itself for entry into the market” (Pfizer Canada Inc v Teva Canada
Limited, above, at para 106). However, the evidence Teva sought from its
fact witnesses went beyond general intentions or preparatory steps: the
witnesses were sometimes asked what they actually would have done or what would
have actually happened in the but-for world. The witnesses were being
improperly invited to provide opinion evidence.
[15]
However, the approach I suggested was not
possible for some witnesses. For example, Teva’s witness Dr Brian Des Islet, Executive
Director of Scientific Affairs at Teva, was invited to imagine a scenario in which
Teva had obtained its NOC on March 3, 2006, and was asked whether Teva would
have launched with material produced by a particular process (Process 1) or by
another (Process 2). Counsel for Lilly objected to the question on the basis
that it invited Dr Des Islet, a fact witness, to provide an opinion. Since only
Process 2 material was used in the real world, asking Dr Des Islet whether Teva
would have relied on Process 1 material in the but-for world would have invited
him to offer a speculative opinion. The question of whether Teva could have and
would have launched with Process 1 material can be answered only after
considering all the relevant evidence; it is my responsibility to answer it
based on the evidence before me. The fact that the burden falls on Teva to
prove that it could have and would have launched in March 2006 with Process 1
API does not mean that it is entitled to ask a fact witness to answer that
question.
[16]
Accordingly, I have not considered the testimony
of fact witnesses in which they offered opinions about what would or would not
have happened in the but-for world. I rely solely on the opinions of experts
and my own inferences drawn from the evidence.
[17]
With respect to Lilly’s hearsay objections, I
again agree with Lilly that hearsay evidence cannot be admitted unless it falls
within a recognized exception (eg, business records) or it meets the
criteria of necessity and reliability. This basic proposition has been firmly
underscored by the Federal Court of Appeal (Pfizer Canada Inc v Teva Canada Limited,
above, at paras 95-103). Lilly’s principal objection relates to a report
prepared by Deloitte that, among other things, contains information about
Teva’s trade-spend rate on its venlafaxine product. The author of the report
was not called as a witness.
[18]
In my view, the Deloitte report does not meet
the test for the business records exception. That exception requires that the
author of the record have a duty to create it, and did so contemporaneously and
based on personal knowledge (Ares v Venner, [1970] S.C.R. 608 at p 626; see
also Canada Evidence Act, RSC 1985, c C-5, s 30). Since the author of
the Deloitte report is unknown and the details surrounding the report’s
preparation were not in evidence, the report cannot meet these criteria.
[19]
Lilly also challenges on hearsay grounds
documents purporting to contain information about the rebates and other
incentives (“trade-spend”) that Teva would have
paid to pharmacies and other retailers of another one of its products,
venlafaxine. The documents in question were not prepared contemporaneously with
the financial transactions they allegedly record. Those who authored the
documents were not called as witnesses. These, too, amount to inadmissible
hearsay not falling within the business records exception.
[20]
Even though there are some indications that
these impugned documents are reliable, they are not admissible under the principled
exception to the hearsay rule because the criterion of necessity is not met.
Evidence relating to trade-spend was provided by way of direct evidence from Teva’s
fact witnesses – Ms Oksana Tressel, Mr Doug Sommerville, and Mr Barry Fishman.
It is not necessary to look to the Deloitte report or the other impugned
documents to determine what Teva’s trade-spend rate was for venlafaxine.
[21]
Therefore, I have confined myself to the
evidence that is properly before me. I have not considered opinions offered by
fact witnesses or inadmissible hearsay.
A.
Issue One – What is the period of liability?
[22]
Teva maintains that the period of Lilly’s liability
commences on the date certified by the Minister as being the day on which Teva
would have obtained its NOC if Lilly had not initiated proceedings under the
Regulations: March 3, 2006. The parties agree that the period ends on the date
on which Justice Hughes rendered his decision dismissing Lilly’s application:
June 5, 2007.
[23]
Lilly’s main position is that, because Teva
abandoned its claim to damages, there is no period of liability. Alternatively,
Lilly disputes Teva’s reliance on the Minister’s certified date of March 3,
2006, and suggests that the appropriate start date is March 22, 2007 because
Teva could not actually have put its product on the market any sooner. In
effect, Lilly would reduce the period of liability from the approximately 15
months asserted by Teva, to either zero or a maximum of 3.5 months.
[24]
I am not persuaded by Lilly’s arguments relating
to abandonment or an alternate start date. Read in context, the evidence does
not show that Teva abandoned its claim to damages in this proceeding. Further, the
evidence demonstrates that but for the prohibition proceedings, Teva would have
been able to put its product on the market upon receiving its NOC on March 3,
2006. Therefore, the period of liability is from March 3, 2006 to June 5, 2007.
(1)
Abandonment
[25]
Lilly relies on the sequence of events relating
to Teva’s first Notice of Allegation (NOA) in which Teva alleged that Lilly’s
olanzapine patent, Canadian Patent No 2,041,113 (the ‘113 patent), was invalid.
Teva served that NOA on August 5, 2004. Lilly responded to the NOA by invoking
the Regulations and applying for a prohibition order against Teva (T-1734-04).
Lilly filed its evidence, but Teva later withdrew its NOA and served a new one.
Lilly sought its costs in the first proceeding; in response to Lilly, Teva
submitted that it had been prejudiced by the delay resulting from the
withdrawal, in part, because it had “abandoned its
claim to s 8 damages”.
[26]
Lilly points to Teva’s submissions on costs
filed in that earlier proceeding and contends that Teva made an unequivocal
undertaking, both to the Court and to Lilly, that it had unequivocally abandoned
any claim to s 8 damages. According to Lilly, Teva’s claim is blocked by the
doctrines of abandonment and estoppel. Further, Lilly says, the Court should
take account of Teva’s submission in the earlier proceeding when assessing
damages under s 8(5) of the Regulations here. Lilly also notes that the
proceeding relating to Teva’s second NOA had already been initiated at the time
Teva made its submissions on costs. Therefore, says Lilly, Teva’s earlier
position on abandonment should carry over from the earlier proceeding and apply
here, and should prevent Teva from advancing any damages claim against Lilly.
[27]
I disagree with Lilly on this point. Reading
Teva’s submission on abandonment in context, I find that it related solely to
the first proceeding. Teva’s submission responded to Lilly’s request for costs.
As I read it, Teva was simply pointing out that one of the consequences of the
withdrawal of its first NOA was a relinquishment of a claim to damages within
that proceeding. Teva did not agree that it would not to seek s 8 damages in
the second proceeding or, indeed, in this action.
[28]
Further, in my first decision in this action, I
concluded that Teva’s claim to s 8 damages would be decided in a separate
proceeding. While that decision was appealed successfully in part, the Federal
Court of Appeal made clear that its ruling did not affect the issue of s 8
damages (2010 FCA 219 at para 13). It left the issue to be decided in a
subsequent proceeding.
[29]
I disagree with Teva’s position, though, that my
earlier decision in this case and the subsequent response of the Federal Court
of Appeal renders the issue of abandonment res judicata. I specifically stated
that the start and end date of the period of liability could be decided in the
separate proceeding on damages, apart from the trial on liability and
infringement (see my Order of December 16, 2009). As I read the bifurcation
order, I was free, but not bound, to address all the issues relating to liability
for damages in the first phase. The bifurcation order relieved the parties from
adducing evidence on the quantum of damages during that phase, but it did not
specify what would be decided in the second phase. I preferred to leave the
issue of liability for damages to the second phase as I believed it was closely
connected to the question of quantum. While I disagree with Lilly on the issue
of abandonment, I do not fault it for raising the issue here. The issue of abandonment
is not res judicata.
(2)
The Start Date
[30]
Lilly contends that the start date of the period
of liability should be March 22, 2007. Lilly’s position is based on the
following:
1.
Teva had to seek and obtain a notifiable change
for the process it planned to use for its product before it entered the market.
Its notifiable change was approved by Health Canada on March 22, 2007.
2.
The bulk Active Pharmaceutical Ingredient (API)
Teva received during the relevant period was provided by its supplier, Dr
Reddy’s Laboratory, |||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.
3.
Dr Reddy’s supply of API was unreliable until
late 2006.
4.
Teva had not completed its validation tests
until March 22, 2007.
[31]
I disagree with Lilly. The evidence does not
support its claim that Teva would have had difficulty entering the market once
it obtained its NOC on March 3, 2006.
[32]
Regarding the notifiable change, the evidence
shows that Teva used one process (Process 1) for manufacturing API for
regulatory purposes. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.
Dr Reddy’s then created Process 2 and Teva sought a notifiable change for the
new process. Teva submitted its request on September 14, 2006. Health Canada agreed
to review the request on November 9, 2006, and completed its review on March
22, 2007. From that date, Teva was able to sell its product with Process 2 API.
[33]
Lilly’s expert, Dr Gordon Munro, concludes from
this evidence that Teva could not have implemented the change from Process 1 to
Process 2 until March 22, 2007, and that no Process 2 product could have been
marketed until after that date (see Annex I for a summary of experts’
qualifications).
[34]
Based on this evidence, Lilly submits that
Teva’s product could not have been marketed commercially until March 22, 2007.
[35]
I disagree with Lilly’s position. While the
evidence shows that Teva could not have marketed its product with Process 2 API
prior to March 22, 2007, it equally demonstrates that Teva could have sold
olanzapine tablets containing Process 1 API as of March 3, 2006. Dr Reddy’s was
in a position to supply it. Mr Rajesh Sadanandan, an employee of Dr Reddy’s who
was responsible for European sales of Dr Reddy’s API products at the relevant
time, explained that Process 1 was developed ||||||||||||||. During the 2005 to 2007
period, Dr Reddy’s was capable of producing about 1800 kg a year. In 2007, Dr
Reddy’s started selling Process 2 material, in addition to Process 1 API.
[36]
Dr Brian Des Islet, Executive Director of
Scientific Affairs at Teva, testified that Dr Reddy’s had supplied Teva with
Process 1 material for the batches that were submitted to Health Canada for
regulatory purposes and, in fact, has continuously supplied Teva with API for
its olanzapine product. He explained that Teva asked Dr Reddy’s to develop a
different process for producing the API, which resulted in Teva’s seeking a
notifiable change to switch from Process 1 to Process 2. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||.
[37]
Dr Des Islet was invited to imagine a scenario
in which Teva had obtained its NOC on March 3, 2006, and was asked whether it
would have launched with Process 1 or Process 2 API. I have discussed the
inadmissibility of this evidence above. However, Dr Des Islet was able to state
that the only material with which Teva had regulatory approval to launch on March
3, 2006 was Process 1 API. Further, he was aware that Dr Reddy’s was able to
supply the US market with Process 1 material, suggesting that Dr Reddy’s could
likely have met Teva’s needs, as well.
[38]
Lilly argues that Teva could not have come to
market with Process 1 API |||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.
However, there is little or no evidence that ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.
Lilly’s expert, Dr Munro, stated in a supplemental report that he “would consider” |||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||.
However, on cross-examination, Dr Munro stated that he was merely making an
observation, one which did not affect his overall opinion about when Teva might
have been in a position to launch its product. He conceded that it was not a
matter within his main area of expertise and that he could not render an
opinion on infringement. Therefore, I cannot conclude that Teva would have been
legally prevented from entering the market with tablets containing Process 1
API on March 3, 2006. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.
[39]
Regarding the US injunction, Lilly points to
packing slips and invoices showing that Teva received bulk API from Dr Reddy’s
Laboratories in the US prior to June 2007. Any such shipments, says Lilly,
would have violated a US court order. However, it is unclear whether the
evidence on which Lilly relies shows the actual provenance of shipments of API
to Teva in Canada, or whether they simply show where the documents originated
or where the billing occurred. On this evidence, I cannot conclude that the
shipments of API were illegal.
[40]
Regarding supply of API, Lilly relies on Dr
Munro’s opinion that shipments from Dr Reddy’s from India were slow and
unpredictable, at least until June 2007. Dr Munro noted that early orders from
Teva to Dr Reddy’s took approximately six months to arrive. Prior to June 5,
2007, Teva had received only 61 kg of API. Delays thereafter were reduced to
about six weeks. Still, Dr Munro concludes that Teva’s access to API did not
stabilize until June 2007. These problems related to Process 2 material. Dr
Munro was asked whether Teva, given that it had successfully launched to a sole
source market in June 2007, could have done the same in March 2006. Dr Munro
characterized the question as hypothetical and refused to answer it. He based
his opinion solely on the documents he had reviewed, which related to the
events in the real world in 2006-2007: he was unwilling to consider what might
have happened in the but-for world.
[41]
Dr Munro also points out that Teva’s validation
batches, essential for regulatory approval, were not completed until March
2007. Based on that information, he concluded that Teva would not have been in
a position to launch its product in March 2006. He was asked whether Teva would
have carried out its validation process earlier if it knew it could come to
market in March 2006, but Dr Munro dismissed that as a hypothetical scenario
for which he had no evidence on which to base an expert opinion.
[42]
The data for these batches showed that the
yields varied from 86% to 97%, a range that Dr Munro considered unusually low. Dr
Munro also noted that the lot numbers appeared to be out of sequence,
suggesting that Teva might not have tested its batches consecutively, as
required. Again, Dr Munro expressed concern that the documentation revealed
variability in the manufacture of Teva’s product that would have impaired its
commercial launch. However, he could not say that this variability would have actually
affected the reliability of Teva’s validations or interfered with its entry
into the market.
(3)
Conclusion on the Liability Period
[43]
The liability period begins on March 3, 2006 and
ends on June 5, 2007, the date on which Justice Hughes rendered his decision
dismissing Lilly’s application for an order prohibiting the Minister from
issuing an NOC to Teva.
B.
Issue Two – What was the size of the olanzapine
market?
[44]
The parties agree that the size of the
olanzapine market in the but-for world would have been the same as it was in
the real world. Entry of a generic manufacturer into the market would not have
affected the overall olanzapine market. Since Teva launched its product only in
certain dosage forms (2.5, 5, 7.5, 10 and 15 mg), it is the overall market for
those particular products that should be considered.
C.
Issue Three – What was the generics’ share of
the olanzapine market?
[45]
The parties agree generally on the methodology
for determining the generic portion of the olanzapine market. However, they
dispute the speed with which a generic company could have entered the
olanzapine market in each province. This depends on the date the generic could
have obtained approval and listing on the provincial formularies. In each
jurisdiction, one must consider what happened in the real world – the delay
between Teva’s obtaining its NOC and its entry on the market in each province –
and then assess whether something different would have occurred in the but-for
world.
[46]
The parties’ principal disputes relate to the
provinces of British Columbia, Alberta, Saskatchewan, and Manitoba. However,
the circumstances in those provinces can be understood only after considering the
situation in Ontario in 2007. Therefore, even though the parties largely agree
on the appropriate listing date in Ontario, I will review the Ontario scenario
first.
(1)
Ontario
[47]
Assuming that Teva obtained its NOC on March 3,
2006, the parties agree that Teva’s generic olanzapine would have been listed
on the Ontario formulary on May 19, 2006. The more challenging question is the
price at which it would have been listed, and what effect certain events in
Ontario would have on the prices in other provinces.
[48]
Generally speaking, Teva would have sought to
list its product in all provinces, including Ontario, at 70% of Lilly’s brand
price. For Ontario, however, in the real world, the situation was complicated
by the introduction of new legislation, known as Bill 102, in October 2006.
Teva’s expert on this subject, Mr Ian Hilley, explained that under Bill 102,
generic prices were generally set at 50% of brand prices. However, some
exceptions were permitted. The Executive Officer of the Ontario formulary commonly
allowed exceptions for sole-source generic products in which the manufacturers
had invested heavily. For example, Teva achieved an exception for its
sole-source venlafaxine product, which was listed at 70% of brand price instead
of 50%.
[49]
However, Teva sought, but was denied, an
exception for its olanzapine product. Teva was unable to justify a listing at ||||||||||||||||||||||||||||||||||
of the brand price. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Teva obtained a published listing at 75% of brand price, ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.
Accordingly, the ||||||||||||||||
under the agreement was |||||||| of Lilly’s price.
[50]
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Once Bill 102 came into effect, Teva would have sought, but would likely not have
received, an exception to the rule. Therefore, as of January 1, 2007, Teva would
have had to ||||||||||||||||||||||||||||||
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.
In the but-for world, as in the real world, Teva would have had to set its effective
price in Ontario at no more than |||||||| of Lilly’s price.
[51]
On the strength of its published 75% listing price
in Ontario, Teva looked to increase its pricing in other provinces to an
equivalent level. As will be seen below, this was a confounding factor in
Teva’s efforts to obtain listings in those provinces. Since the Ontario ||||||||||
agreement would not have been in place in the but-for world, one must consider
what would have transpired in its absence. In addition, the landscape was
changing in 2007, largely as a result of developments in Ontario. One must be
mindful, therefore, in reconstructing the but-for world of 2006-2007, that
certain real world events had not yet transpired.
[52]
In my view, some real-world events in 2007 and
thereafter – the impact of Bill 102, Lilly’s competitive offer to Ontario,
Teva’s corresponding |||||||||| agreement, and the ongoing
litigation between the parties – do not reflect the situation in the 2006-2007 but-for
world.
(2)
British Columbia
[53]
In the real world, Teva’s product was not listed
in British Columbia until May 2011, almost four years after Teva obtained its
NOC. Lilly argues that Teva would have experienced the same difficulties
getting its BC listing even if it had obtained its NOC in 2006 instead of 2007.
Accordingly, says Lilly, Teva did not experience lost sales in BC during the
relevant time frame.
[54]
I disagree. There were particular circumstances
in 2007 that delayed Teva’s listing on the BC formulary. I am satisfied that
those circumstances would not have prevailed in 2006 and, therefore, that Teva
would have obtained its listing on the BC formulary fairly promptly had it
obtained its NOC on March 3, 2006.
[55]
Teva’s expert, Ms Jane Costaris, estimated that Teva
would likely have obtained its BC listing for generic olanzapine on March 17,
2006. By contrast, Lilly’s expert, Mr Ghislain Gauthier, concluded that Teva
would not have obtained its listing in BC until sometime after the liability
period. On many points relating to formulary listings, I prefer the evidence of
Ms Costaris who has substantial real-world experience making submissions to
provincial formularies.
[56]
In Ms Costaris’ opinion, in 2006, BC officials
would have been keen to list Teva’s generic olanzapine quickly because doing so
would have saved the province over $8 million each year given that Teva’s generic
product would have been listed at 70% of the cost of Lilly’s product. In
reality, as pointed out above, it actually took nearly four years to list
Teva’s product. However, as Ms Costaris explained, that delay was the result of
a series of special circumstances in 2007 that would not have existed in 2006.
[57]
In August 2007, after Teva had obtained its NOC
and while Lilly’s appeal of Justice Hughes’ decision was pending, Lilly entered
into an agreement with BC to provide olanzapine at a price lower than Teva’s in
order to preserve its listing. After Lilly’s appeal was dismissed, BC issued a
request for proposal (RFP) in respect of olanzapine. It was the first time BC
had invoked this type of procedure. Lilly succeeded on the RFP and entered into
further agreements with BC. Teva did not participate in the RFP; rather, it
chose to challenge the process. It did not succeed. However, on May 5, 2011,
Teva’s product was ultimately listed in BC, just 16 days after it resubmitted
its application.
[58]
In the but-for world, Lilly would not have
initiated its application to prevent Teva from obtaining its NOC. It follows
that Lilly would have had no right to appeal Justice Hughes’ decision in
respect of that application. Therefore, to the extent that BC officials’
concerns about the ongoing litigation between the parties under the Regulations
impaired Teva’s ability to get its product listed, I must disregard them. In
fact, numerous documents in evidence cited Lilly’s appeal as a significant factor
standing in the way of Teva’s listing. While he had not mentioned that factor
in his report, Mr Gauthier conceded its significance on cross-examination.
[59]
Lilly suggests that even in the absence of
proceedings under the Regulations, its infringement action, commenced
immediately after Justice Hughes’ decision, would also have stood in the way of
Teva’s obtaining its listing in BC. However, I see no evidence of concern on
the part of BC officials about patent infringement; rather, their concerns, as
expressed in correspondence with Teva, related solely to the proceedings under
the Regulations – that is, whether Teva’s NOC was valid.
[60]
In relation to the August 2007 agreement between
BC and Lilly, Mr Gauthier stated that a similar agreement would likely have
been entered into had Teva obtained its NOC in 2006. However, Mr Gauthier’s
opinion was based on the precedent of a 2005 product-listing agreement (PLA)
between BC and Lilly in respect of a drug called “Actos”.
On cross-examination, Mr Gauthier conceded that the 2005 PLA was not an
exclusive listing agreement and was therefore unlike the 2007 agreement between
BC and Lilly. He also failed to consider a 2006 BC report that stated that the
province had historically “not actively negotiated drug
prices with manufacturers”, but would begin to do so in 2007.
[61]
With respect to the RFP process, the evidence
shows that this was likely a result of the introduction of Bill 102 in Ontario,
which came into force on January 1, 2007. Dr Aidan Hollis, Teva’s economics
expert, described the BC tender process as an unprecedented event that responded
to contemporaneous developments in Ontario. BC did not invoke a RFP process in
2006 in respect of risperidone, a drug with similar uses to olanzapine, and did
not rely on that type of process again until 2015. In 2007, the Competition
Bureau published a report recommending alternative drug plan approaches that
would save costs, and BC’s Deputy Minister of Health established a
Pharmaceutical Task Force to explore various drug procurement and delivery
options. These actions, according to Dr Hollis, were responsive to
circumstances particular to the 2007 environment.
[62]
In contrast, Dr Iain Cockburn, for Lilly,
believed that the only reason Teva’s product was not listed in BC in 2007 was
because of the listing agreement between Lilly and the BC government entered
into in August 2007. Given that a similar agreement would likely have been
achieved in the but-for world, Dr Cockburn concluded that Teva’s product would
not have been listed in BC during the liability period. However, Dr Cockburn conceded
that BC’s concern about the outstanding litigation between the parties was
another possible, perhaps even more dominant, reason why Teva’s product was not
listed promptly in that province.
[63]
Looking at this evidence as a whole, I am
satisfied that the real-world events that unfolded in BC in late 2007 cannot be
relied on to determine what would have happened in the liability period of
March 2006 to June 2007. During the relevant time-frame, BC would probably not
have entered into a PLA with Lilly, had a stay placed on Teva’s listing pending
the outcome of litigation, or established an RFP process for supplying
olanzapine. In my view, in March 2006, BC would likely have listed Teva’s
product promptly. I agree with Ms Costaris that BC would likely have listed Teva’s
product on or about March 17, 2006. Mr Gauthier conceded that Ms Costaris’
estimated date was reasonable, assuming that BC and Lilly had not entered into
a PLA.
(3)
Alberta
[64]
In the real world, Teva’s generic olanzapine
product was listed in Alberta on September 1, 2007, nearly three months after
Teva obtained its NOC.
[65]
Mr Gauthier concluded that Teva would likely
have received its listing on July 1, 2006, had it obtained its NOC on March 3,
2006, a delay of about four months.
[66]
According to Ms Costaris, if Teva had obtained
its NOC on March 3, 2006, it would have obtained its Alberta listing about two
months later, on May 1, 2006. In her opinion, generic olanzapine would have
received a fast-track listing given the quantity of annual savings that the province
would realize, about $5.5 million. Ms Costaris stated that generic olanzapine’s
listing in the real world was probably delayed somewhat because of issues about
price. However, she notes, those issues arose out of the regulatory environment
in Ontario in 2007 in which Teva negotiated |||||||||||||| agreement with Ontario.
In turn, Teva requested a price in Alberta of 75% of brand instead of 70%. That
would not have happened, according to Ms Costaris, in the 2006-2007 but-for
world.
[67]
I agree with Ms Costaris’ opinion. In his
analysis, Mr Gauthier relied on listing dates for products that were not
analogous to olanzapine, used a methodology based on average dates for listing that
tended to overstate the delay for individual products, did not account for the
delay caused by a requested change in price, and could not plausibly explain
why his estimated listing dates were longer than those that occurred in the
real world.
[68]
The listing date for Teva’s olanzapine product
would likely have been May 1, 2006, at 70% of brand price.
(4)
Saskatchewan
[69]
In the real world, olanzapine was listed in
Saskatchewan on November 1, 2009, nearly thirty months after Teva obtained its
NOC.
[70]
Ms Costaris and Mr Gauthier agree that the delay
in Saskatchewan was at least partly the result of Teva’s request for a price
increase after it had made its submission for listing. As in Alberta, Teva’s
revised pricing was a product of the situation in Ontario in 2007. Accordingly,
Ms Costaris estimates that Teva’s generic olanzapine would have been listed no
later than October 1, 2006 if Teva had obtained its NOC on March 3, 2006. Ms
Costaris compared olanzapine with another drug (Apo-Feno-Super) which was the
subject of a NOC on April 1, 2006. The (negative) decision regarding listing of
Apo-Feno-Super was issued six months later, on October 1, 2006. It is likely,
therefore, according to Ms Costaris, that the decision to list Teva’s
olanzapine product would have been made at the same time.
[71]
Mr Gauthier points to the discussions and the
eventual agreement between Lilly and Saskatchewan according to which Lilly
offered rebates in return for a commitment to continue to list its olanzapine
product. The agreement was signed on December 12, 2008. Mr Gauthier believes
that Lilly would have sought the same kind of arrangement with Saskatchewan if
Teva had obtained its NOC earlier. Since the rebate Lilly was offering would
have resulted in a lower price for Saskatchewan than Teva was seeking, Mr
Gauthier concludes that there would been a significant delay in listing Teva’s
product in the but-for world to the point that it would not have been listed in
Saskatchewan within the relevant time frame.
[72]
Again, I find Ms Costaris’ opinion more
persuasive. First, the issue about pricing would not have arisen until after
Bill 102 came into effect in Ontario in 2007: it would not have affected Teva’s
position in Saskatchewan in March 2006 when Teva would have obtained its NOC.
Therefore, in the but-for world, Teva would probably not have experienced the
delays that it encountered in the real world.
[73]
Second, Lilly’s agreement with Saskatchewan was
not signed until eighteen months after Teva acquired its NOC. Assuming that
Lilly would have sought a similar arrangement in the but-for world, Lilly would
likely not have achieved it until after the expiry of the liability period; it
would not have been a factor affecting Teva’s listing in Saskatchewan.
[74]
Therefore, I find that the appropriate listing
date for Saskatchewan is October 1, 2006, and that Teva’s product would have
been listed at 70% of brand price.
(5)
Manitoba
[75]
In the real world, Teva’s product was listed in
Manitoba on October 14, 2010, more than three years after Teva obtained its
NOC.
[76]
In Ms Costaris’ opinion, Teva’s product would
have been listed in Manitoba on September 14, 2006 in the but-for world, about
six months after Teva obtained its NOC. In arriving at that date, she assumed
that Teva would not have sought a price increase as it did in the real world
and, therefore, that Teva’s listing would not have been delayed. In addition,
she assumed that Teva would have applied for its listing prior to Manitoba’s
introduction in 2007 of a new process called a utilization management agreement
(UMA).
[77]
Mr Gauthier stated that, in his view, Teva would
not have obtained a listing in Manitoba in the but-for world any earlier than
it did in the real world. He assumed that Teva would have sought a higher price
in Manitoba compared to other provinces, just as it did in the real world. He
also concludes that the discussions that led to the UMA that was ultimately
reached between Manitoba and Lilly in 2009 would also have occurred in the
but-for world and delayed Teva’s listing. Mr Gauthier did agree, however, that
Manitoba was not actively seeking UMAs in 2006.
[78]
I agree with Ms Costaris that the UMA process
would not have existed in 2006. Therefore, I accept her projected listing date
of September 14, 2006 at a price of 70% of brand.
(6)
The Other Provinces
(a)
Quebec
[79]
The experts agree that Teva’s product would have
been listed in Quebec on October 11, 2006 at a price of 70% of brand. While Mr
Gauthier believes that Teva would have been obliged to disclose to Quebec its
subsequent |||||||||| arrangement in Ontario and to give
Quebec the equivalent net price of |||||||| of brand, the evidence shows
that in the real world that did not happen. Prices in Quebec did not ||||||||||||||||||||||||
until February 2008, well outside the liability period.
(b)
New Brunswick
[80]
The experts agree that Teva’s olanzapine product
would have been listed in New Brunswick on June 9, 2006. The product would have
been listed at 70% of brand price.
(c)
Nova Scotia
[81]
Ms Costaris testified that Teva’s listing date
in Nova Scotia would have been June 1, 2006. She based her opinion on the
treatment of other products submitted for listing around the same time. Mr
Gauthier sets the date a month later, July 1, 2006, based on average listing
delays in late 2005 to late 2006. As mentioned, I find that Mr Gauthier’s
approach tends to overstate delays to listing. Accordingly, I think Ms
Costaris’ date is more accurate. The listing price would have been 70% of brand
price.
(d)
Prince Edward Island
[82]
Ms Costaris testified that Teva’s product would
have been listed on October 16, 2006 at the latest. Mr Gauthier agreed with her
that an earlier date of July 1, 2006 was quite possible. Based on these
opinions, my view is that the safer date is October 16, 2006, and that the
price would have been 70% of brand.
(e)
Newfoundland and Labrador
[83]
Ms Costaris opined that Teva could have obtained
a listing in Newfoundland and Labrador by January 1, 2007 or earlier. Mr
Gauthier chose the earlier date of September 1, 2006. I am satisfied that Mr
Gauthier’s date is sound. Again, the listing price would have been 70% of
brand.
D.
Issue Four – What was Teva’s share of the
generic olanzapine market?
[84]
Lilly maintains that Teva would have had to
share the generic olanzapine market with another generic company, Apotex Inc.
Lilly says that Apotex would likely have entered the market on June 23, 2006,
the date on which Apotex would have received its NOC for olanzapine if Lilly
had discontinued its prohibition proceedings against Apotex. In the real world,
Apotex entered the olanzapine market in October 2009 after Lilly’s patent had
been found to be invalid. Lilly was successful in its prohibition application
against Apotex, which prevented Apotex from marketing its product any sooner.
[85]
I have no basis for finding that in the but-for
world in which Teva obtained its NOC on March 3, 2006, Lilly would have
discontinued its proceeding against Apotex. Lilly commenced prohibition
proceedings against numerous other generic companies in respect of olanzapine,
and most of them were initiated after Lilly’s application against Teva was
dismissed on June 5, 2007. In respect of Apotex specifically, Lilly entered into
discussions with that company, but the discussions did not result in any
agreement to allow Apotex onto the market. In fact, the evidence indicates that
Lilly rarely enters into agreements to authorize generic companies to
participate in a market alongside Lilly. The last time Lilly signed an
authorized generic agreement was in 1995.
[86]
Lilly’s actions in the real world do not suggest
to me that it would have acted differently in the but-for world. In the but-for
world, the Regulations would still have existed (Apotex Inc v Sanofi-Aventis,
2014 FCA 68 at paras 171, 186). Therefore, Apotex would have had to address the
Regulations before it could have entered the market. Lilly would have had the remedies
available to it under those Regulations in relation to olanzapine and, in my
view, would have invoked those remedies against other generic companies wishing
to market olanzapine, including Apotex.
[87]
Further, there is no evidence that Apotex was
even in a position to come to market in 2006. Mr Gordon Fahner, Apotex’s Vice
President of Global Finance, was called as a witness by Lilly, but he was
unable to provide evidence about his company’s access to bulk olanzapine API
during the relevant period.
[88]
Therefore, the evidence shows that Teva would
have been the sole generic company on the market during the liability period.
E.
Issue Five – What is the amount of Teva’s
damages?
[89]
I have already dealt with the pricing of Teva’s
product in the various provinces. There remain two additional issues relating
to the quantification of Teva’s losses. First, should the calculation of Teva’s
losses include compensation for so-called “pipefill”?
Second, how much should Teva’s losses be reduced to account for the monies it
would have paid to pharmacies and other retailers of its product (ie,
trade-spend)?
(1)
Pipefill
[90]
Teva claims that its losses should include an
amount for pipefill – that is, the quantity of sales Teva would have made to
distributors in the but-for world, an amount that would not be captured by
retail sales figures (from IMS Health Canada, referred to as IMS data). Teva contends
that losses attributable to pipefill sales have been compensated in other s 8
cases.
[91]
The concept of pipefill is that a manufacturer
must first get its product into the channels of trade before any sales to
customers can take place. There is a delay, therefore, between the shipment of a
product from the factory and retail sales. Teva maintains that its losses
should include a pipefill adjustment to include the shipments it would have
made during the liability period.
[92]
I disagree. As I understand it, pipefill does
not represent lost sales during the liability period. Rather, pipefill
represents the differential between retail sales and the quantity of product
leaving the factory. That differential represents sales that would have been
made outside the liability period. It is true that pipefill may represent some lost
sales in the sense that in the but-for world Teva would have moved a certain
amount of inventory into the distribution stream which, in due course, would be
sold to customers. In the but-for world, those sales would have been made, but
they would have been made outside the liability period. Accordingly, for
present purposes, they should not be included in Teva’s losses.
[93]
I concur with Dr Iain Cockburn’s opinion on
pipefill. He explained that an adjustment to retail data would be appropriate
to deal with the issue of timing – the delay between manufacture and retail
sales. He stated:
The drug is
manufactured. It may sit in inventory at the factory for a while. Then it is
shipped by the manufacturer to the wholesalers. That is what we are calling
ex-factory sales. It may sit at the wholesaler for some time until it is
ordered by the retailer. Having arrived [at] that retailer, it may then sit on
the pharmacy shelf for some time until it is finally dispensed to an
individual. I think the challenge that was being addressed here is, if you look
at the IMS prescriptions data, you need to try to understand what would have
been sold or moved through the distribution chain at an earlier point in time.
You need to make an adjustment for timing.
[94]
The economic loss associated with that delay,
according to Dr Cockburn, is confined to the “time
value of money” or the “opportunity cost”
based on the fact that those sales would have been made earlier in the but-for
world. The appropriate adjustment, in effect, would be to award Teva an amount
representing interest on the money involved.
[95]
Dr Cockburn distinguished this issue of timing
from a pipefill adjustment in the sense described above. The latter, proposed
by Dr Hollis, represents accumulation of inventory in the hands of wholesalers
and not captured by the retail data. In Dr Cockburn’s analysis, if Teva were
able to recover for the amount of product that was held in inventory, it would
be credited with market entry at a steady state of sales; that is, each tablet
leaving the factory would be counted as a sale. In Dr Cockburn’s view, with
which I concur, such an approach would over-compensate Teva because it would
not correspond with the amount of Teva’s lost sales in the but-for world. As he
explained, the accumulation of product in wholesalers’ warehouses will
eventually be sold. Since it will be sold, it cannot constitute lost sales.
[96]
As mentioned, Teva maintains that compensation
for pipefill has been awarded in other s 8 cases. In my view, unlike here, the
issue of pipefill in those cases was not strenuously contested. Teva points to Apotex
Inc v Sanofi-Aventis, 2012 FC 553 at paras 221-226, aff’d 2014 FCA 68; Teva
Canada Limited v Pfizer Canada Inc, 2014 FC 248 at paras 186-190; and Apotex
Inc v Takeda Canada Inc, 2013 FC 1237 at paras 119-120. I will address each
of these cases, in turn.
[97]
In Apotex v Sanofi, the experts before Justice
Judith Snider agreed that the calculation of lost sales should include some
measurement for the time-lag between sales to distributors and sales to
individual customers. After considering the three experts’ methodologies, she
concluded that a simple adjustment of two months of sales should be added to
the calculation of Apotex’s losses. That figure represented the difference
between Apotex’s sales in the real world as compared to the data captured in IMS
records. It is not clear whether the adjustment Justice Snider made was
intended to include pipefill or simply to address the issue of timing. On the appeal
of Justice Snider’s decision, the Federal Court of Appeal did not address the
issue of pipefill.
[98]
Dr Cockburn reviewed Justice Snider’s decision
and could not tell whether the adjustment she awarded was for timing or for
pipefill.
[99]
In Teva v Pfizer, Justice Russell Zinn
pointed out the difference between IMS data showing the number of sales made to
individual customers, and the measurement of sales based on the amount of
product leaving the manufacturer’s factory, so-called ex-factory data. Justice
Zinn concluded that IMS data underreports actual sales and, therefore, that ex-factory
data should be used to calculate sales. In his view, using ex-factory data eliminated
the need to calculate pipefill separately. A calculation of an amount
specifically reflecting pipefill was not made.
[100] In Apotex v Takeda, the parties agreed that an adjustment had
to be made for pipefill because IMS data would not capture it. As Justice
Michael Phelan described it, the question involved a “reporting
system delay” and required a decision on when sales would reach a steady
state in each jurisdiction (at para 110). The two experts before him suggested
adjustments of 0.9 months and 1.5 months respectively. It is not clear to me
that this adjustment represented pipefill in the sense in which the parties have
used that term in the case before me.
[101] I find the preceding authorities to be somewhat ambiguous on the
issue of pipefill. In none of them was the issue seriously contested or a
quantum specifically calculated.
[102] Within the liability period, where there is a differential between
data on retail sales and figures on the amount of product leaving the
manufacturer’s factories, that difference represents future sales of the
product, sales that will take place outside the liability period. Do those
constitute lost sales for purposes of s 8 of the Regulations?
[103] It is clear that it is only losses suffered during the liability
period that are compensable under the Regulations. It follows, in my view, that
where a generic company lost sales that would have been made after the end of
the liability period, those losses are not compensable. Accordingly, with the
greatest respect for the learned judges of this Court who may have acceded to representations
to the contrary, I find that a figure representing pipefill should not be added
to Teva’s losses.
(2)
Trade-spend
[104] Trade-spend represents the after-sales amount paid by generic drug
manufacturers to their purchasers, mainly pharmacies and other retailers.
Trade-spend can take various forms: rebates, trade allowances, educational
subsidies, purchasing incentives, etc. Generic companies use trade-spend to encourage
trade with, and foster loyalty from, those who actually put pharmaceutical products
on their shelves. In a multi-source market, generics often compete with one
another through the amount of trade-spend they are willing to provide. The
higher the trade-spend, the greater the incentive for a retailer to sell the
generic product from a particular manufacturer.
[105] Lilly argues that in the but-for world Teva’s trade-spend for its
generic olanzapine product would have been in the range of ||||||||||||||||||||||||||||,
and that Teva’s losses should be discounted accordingly. Lilly’s position is
that Teva’s trade-spend should be set at the average rate across all products
in the real world (||||||||||||),
or even higher (||||||||)
due to the circumstances that would have prevailed in the but-for world.
[106] I disagree with Lilly’s submission. I am satisfied that in the 2006-2007
time frame, Teva’s trade-spend would have been lower for generic olanzapine
than it would have been than for its other products. In particular, given that
Teva would have been the sole generic on the market during the relevant period,
I am satisfied that its trade-spend rate would have been relatively low. Teva
says it should be no higher than 30%. I agree.
[107] Lilly’s expert, Ms Ann Woods, testified that a single-source
trade-spend rate is a fiction. In reality, she says, retailers expected an
overall rebate from manufacturers regardless of the molecule they were being
asked to sell. In her experience, manufacturers did not set trade-spend rates
molecule by molecule.
[108] Ms Woods opinion, while genuine, does not accord with the evidence
before me. Further, it does not correspond with findings in other s 8 damages
cases where this Court has concluded that single-source trade-spend rates are
very low, much lower even than percentage put forward by Teva in this case. For
example, Justice Phelan found that the trade-spend rate on a single-source
molecule in circumstances where there was a risk of an infringement action was
8.9% (Apotex v Takeda, above, at paras 161-162). Justice Zinn found that
the trade-spend rate on venlafaxine in a single-source market would have been
15% (Pfizer Canada Inc v Teva Canada Limited, above, at para 217).
[109] Teva’s witnesses, both factual and expert, suggest that the
trade-spend for olanzapine in the real world in late 2007 was higher than it
would have been in the but-for world because Teva was trying to use its
sole-source market position with olanzapine to leverage another Teva product,
venlafaxine, which was facing competition from other generic entrants at the
time. Teva claims it boosted its trade-spend on olanzapine, even though it was
a sole-source product, in order to secure long-term commitments from retailers
in respect of both olanzapine and venlafaxine. That was a circumstance peculiar
to the late 2007 time-frame; it would not have existed, says Teva, during the
liability period of March 2006 to June 2007.
[110] Rather, says Teva, it would have been the sole generic on the market
with an olanzapine product during the liability period. To determine its likely
trade-spend on olanzapine, Teva maintains one should look to the trade-spend
rate for venlafaxine during the period when Teva was the sole-source supplier
of that molecule – a rate close to ||||||||.
[111] Lilly disputes the relevance of evidence relating to the trade-spend
on venlafaxine. It points out that Teva sold venlafaxine under a license from
Wyeth, which means that it was not at risk of being sued for infringement (unlike
the situation with olanzapine) and could be reasonably certain of its exclusive
position on the market. Further, Teva’s agreement with Wyeth limited Teva only
to reasonable expenses; in other words, Teva was not free to set trade-spend as
high as it might have wished. Finally, Lilly notes that Teva was splitting its
profits with Wyeth, which might have inclined Teva to maximize its own profits
by reducing trade-spend. Accordingly, says Lilly, these circumstances would
have led Teva to set a low trade-spend rate for venlafaxine.
[112] These circumstances, according to Lilly, would have caused Teva to
set a lower trade-spend rate for venlafaxine than for olanzapine. In
particular, since Teva would have been at risk of being found liable for
infringing Lilly’s patent, Teva would likely have reduced its exposure to
damages (calculated according to profits) by increasing its trade-spend. In
addition, since Teva would likely have faced imminent competition from other
generics, Lilly maintains that Teva would have set elevated trade-spend rates
for olanzapine, just as it did when competition in the venlafaxine market was
on the horizon.
[113] In a different vein, Lilly also contends that the evidence relating
to trade-spend for venlafaxine is inadmissible or, at best, unreliable. The
evidence takes the form of a report prepared by Deloitte, which I have
discussed above and found to be inadmissible hearsay. However, there is
additional documentary evidence in the form of a summary of financial data
relating to venlafaxine prepared by Ms Tressel, a former financial officer at
Teva (2002 to 2011), who also testified about her knowledge of venlafaxine
trade-spend rates as reflected in her summary. Other Teva witnesses, Mr Doug Sommerville
and Mr Barry Fishman, also gave direct testimony about trade-spend rates for
venlafaxine and olanzapine.
[114] Ms Tressel confirmed that the trade-spend rate in her summary (||||||||||||||||||||||)
reflected the underlying data captured in Teva’s financial records for
venlafaxine and reported to Deloitte pursuant to Teva’s arrangement with Wyeth.
It was Ms Tressel’s responsibility to assemble the relevant data. She
personally prepared the summary based on prior reports on sales and trade-spend
prepared by members of her team, and personally verified the accuracy of the
information contained in it. It was prepared in the ordinary course of business
for purposes of tracking trade-spend on venlafaxine. I am satisfied that the
summary meets the criteria for the business records exception to the hearsay
rule and is therefore admissible as proof of its contents.
[115] Ms Tressel stated that the trade-spend policy for venlafaxine was
set by Mr Sommerville and Mr Fishman, both of whom also testified.
[116] Mr Sommerville, Senior Vice-President and General Manager at Teva, was
responsible for setting trade-spend rates in 2006-2007. He testified that while
trade-spend rates varied with the product and the customer, rates for single-source
products were typically much lower (||||||||||||||) than for multi-source
products (||||||||||||||),
especially those produced at a low cost. On some occasions, trade-spend rates
for particular customers might be set relatively high in order to convince
those customers to list other products or to gain their long-term loyalty. Contrary
to Lilly’s argument, Mr Sommerville stated that when Teva was at risk of being
sued for infringement, it kept its trade-spend rates low.
[117] Mr Fishman, former President of Teva Canada, concurred. He testified
that the trade-spend on a single-source product would normally be less than ||||||||.
Mr Gordon Fahner, from Apotex, agreed that trade-spend on single-source
products was comparatively low. In effect, he said, the more competition the
higher the trade-spend.
[118] I am satisfied that Teva’s trade-spend on olanzapine in the but-for
world would have been at the low end of the scale – I am prepared to accept
Teva’s proposed rate of 29.4%.
[119] Teva would have been the sole source for a generic version of
olanzapine during the liability period. There were no competitors in a position
to enter the market.
[120] As mentioned, trade-spend on sole-source products is usually low. I
am persuaded that venlafaxine, a sole-source product in the same therapeutic
class that was marketed during the relevant time frame, is a good model for
calculating the trade-spend on olanzapine in the but-for world. I accept the
opinion of Mr Errol Soriano, who calculated the venlafaxine trade-spend rate
based on Ms Tressel’s summary, that the trade-spend for venlafaxine was 29.4%.
In his analysis, he adjusted his calculation to take account of a large sum
paid to Loblaws in respect of venlafaxine, and the rising rate of trade-spend
in the three months just before Teva faced competition in the venlafaxine
market.
[121] In my view, Mr Soriano’s adjustments were appropriate. Regarding the
former, Mr Sommerville explained that the payment to Loblaws was attributed to
venlafaxine for accounting purposes, but was actually meant to secure a
long-term commitment from Loblaws across Teva’s product line. Prior to that
payment, Teva’s trade-spend for venlafaxine at Loblaws was ||||||||.
With respect to the latter, Mr Sommerville explained that Teva learned in the
summer of 2007 that Ratiopharm was close to launching its own venlafaxine
product. Accordingly, Teva raised it trade-spend in August 2007. I accept that
the trade-spend on olanzapine would have remained low throughout the liability
period because competition was not expected during that time frame.
Accordingly, I agree with Mr Soriano’s calculation.
[122] By contrast, actual trade-spend rates for olanzapine in the real
world are not indicative of rates in the but-for world. As Mr Somerville
explained, in the real world, olanzapine was used to leverage commitments to
Teva’s venlafaxine product, a circumstance that would not have arisen in the
but-for world.
[123] Further, I do not accept that the possibility of a patent
infringement action by Lilly would have raised Teva’s trade-spend rate. I am
not persuaded that a generic company with a sole-source product would have
deliberately increased its trade-spend rate in order to lower its profits in an
effort to limit its potential liability for damages in the event that it might
be sued for infringement, and might lose.
F.
Other Costs
[124] There are a few comparatively minor factors that should also be
recognized in the calculation of Teva’s losses:
•
The provision of free goods;
•
Distribution allowances, prompt payment
discounts, and distribution expenses; and
•
Costs of sales.
[125] Ms Woods calculated a discount for free goods based on actual data
relating to olanzapine (||||||||||||||). Mr Soriano used an average figure across
products (||||||||||||);
I accept that Ms Wood’s calculation is more precise and better represents the
quantity of free goods that would have been provided in the but-for world.
[126] The parties’ experts agree on the issues of distribution allowances,
discounts, and expenses. I need not make any finding on those questions.
[127] With respect to other costs of sales, the only amounts really in
issue relate to employee bonuses, yield losses, and processing costs.
[128] Both Mr Sommerville and Ms Tressel testified that no bonuses would
have been paid to sales personnel or other employees if Teva’s olanzapine
product had been launched in 2006. The budget had already been fixed and the
unanticipated launch of a new product would not have given rise to new bonuses.
However, Mr Sommerville conceded that in the absence of litigation, a budget
would have been put in place to provide for employee bonuses. I accept the
evidence of Lilly’s expert, Mr Greg McEvoy, that the proper amount is ||||||||||||
of incremental gross sales.
[129] Lilly relies on the yield loss data from Teva’s validation batches,
while Teva contends that the figures from its commercial production are more
accurate. I agree with Teva on this point and accept Mr Soriano’s calculation
based on yield loss data from 2007-2008.
[130] Similarly, I accept Mr. Soriano’s estimate of the processing costs
of goods sold. He relied on figures from 2008 based on actual production in
2007.
G.
Interest
(1)
Pre-judgment Interest
[131] Teva seeks pre-judgment interest beginning on March 3, 2006,
calculated according to the Ontario Courts of Justice Act, RSO 1990, c
C43, s 127, which would provide a rate of 4.5%. Lilly agrees that the
presumptive rate is 4.5%, but urges the Court to set a lower percentage given
the decline in interest rates in recent years, relying on the discretion set
out in s 130(1) of the Ontario statute.
[132] I agree with Lilly that the rate of 4.5% sought by Teva is too high.
I find that a variable rate interest rate over the relevant period would be
more appropriate. This approach was proposed by Lilly’s expert, Mr McEvoy, and
accepted by this Court in Apotex Inc v Sanofi-Aventis¸ 2012 FC 553 at para
298.
(2)
Post-judgment Interest
[133] Post-judgment interest should be calculated from the date this
judgment is issued.
IV.
Conclusion and Disposition
[134] I find the following:
1. The period of liability is from March 3, 2006 to June 5, 2007.
2. The size of the olanzapine market in that time frame would have been
the same as it was in the real world.
3. The generic portion of the olanzapine market would have been the
same as it was in the real world. Teva’s olanzapine product would have been listed
at 70% of brand price on the various formularies, as follows:
•
Ontario – May 19, 2006
•
British Columbia – March 17, 2006
•
Alberta – May 1, 2006
•
Saskatchewan – October 1, 2006
•
Manitoba – September 14, 2006
•
Quebec – October 11, 2006
•
New Brunswick – June 9, 2006
•
Nova Scotia – June 1, 2006
•
Prince Edward Island – October 16, 2006
•
Newfoundland and Labrador – September 1, 2006
4. Teva was in a position to supply the entire generic market during
the liability period, and no other generics would have been present. No
allowance for pipefill should be included in lost sales.
5. Teva’s trade-spend rate for olanzapine would have been 29.4%.
6. Other expenses include free goods, distribution allowances, prompt
payment discounts, distribution expenses, and costs of sales, to be calculated
in accordance with paras 124-130 above.
7. Teva is entitled to pre-judgment interest calculated at a variable
rate from March 3, 2006, and post-judgment interest as of the date this
judgment is issued.
Email this Content