The opinion
IN THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF DELAWARE
In re Seroquel XR (Extended Release
Quetiapine Fumarate) Antitrust
Litigation
Master Docket No. 20-1076-CFC
This Document Relates to:
All Actions
MEMORANDUM ORDER
Pending before me is Defendants’ motion pursuant to Federal Rule of
Evidence 702 and Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579 (1993) to
preclude Plaintiffs’ economics expert Dr. Keith Leffler from offering at trial the
modeling opinions set forth in J] 13(E)-{F), 92—108 of Dr. Leffler’s Opening
Expert Report and {f 2, 66—79 of his Rebuttal Expert Report. D.I. 633.
I held oral argument on the motion on February 6, 2025. See generally
D.I. 825. In a Memorandum Order issued on February 19, 2025 (D.I. 828), I
required Dr. Leffler to testify at a hearing on March 11, 2025 about two
“dispositive questions that govern this motion”: “(1) whether it was reasonable for
Dr. Leffler to use in his model the same number of generic entrants regardless of
whether Handa won on validity or non-infringement and (2) whether it was
reasonable for him to have found that the number of expected generic entrants does
not materially change the alternative settlement entry date.” D.I. 828 at 9 (internal
quotation marks omitted).
Although the parties agreed at the outset of the March 11 hearing that I had
correctly deemed both questions to be dispositive for purposes of deciding
Defendants’ motion, 3.11.25 Tr. 9:8-19 (docketed as D.I. 834), I realize now, after
further reflection, that the second question has no bearing on the admissibility of
Dr. Leffler’s opinions under Rule 702(d). Dr. Leffler’s finding that the number of
expected generic entrants does not materially change the alternative settlement
entry date is a conclusion or product of his economic model, not a principle or
method applied by his model. Accordingly, the sole question that governs the
pending motion is the first question I posed in the February 19 Memorandum
Order.
I have considered carefully Dr. Leffler’s March 11 testimony, counsel’s
arguments at the February 6 and March 11 hearings, and the parties’ extensive
briefing (D.I. 634, D.I. 714, D.I. 769, D.I. 838, □□□ 839). The relevant factual
background and legal standards are set forth in my February 19 Memorandum
Order (D.I. 828), which I incorporate by reference.
I.
I find that the AstraZeneca forecasts identified in footnote 146 of Dr.
Leffler’s Opening Report and the Handa forecasts in PTX 1178 (D.I. 835-9) and
PTX 1180 (D.I. 835-10) provided a reasonable basis for Dr. Leffler to have used in
his model the same number of generic entrants—i.e., at least five entrants—
regardless of whether the parties expected Handa to prevail on noninfringement,
invalidity, or both defenses in the underlying patent case.
Defendants do not dispute that the AstraZeneca forecasts cited in
footnote 146 of Dr. Leffler’s Opening Report support his assertion in that footnote
that those forecasts “impl[y] that [AstraZeneca] expected 5 or more generics to be
competing after Handa’s exclusivity.” D.I. 835-2 J 103 n.146. Rather, they argue
that Handa would have expected the number of generic entrants to vary depending
on Handa’s expectations about the likelihoods of success of its noninfringement
and invalidity defenses in the underlying patent litigation. Defendants’ logic
makes perfect sense in the abstract. As I noted in my February 19 Memorandum
Order, “[a] verdict in favor of Handa only on infringement would have allowed
Handa—but not other generic manufacturers—to enter the market” whereas “fa]
verdict in Handa’s favor on either of its invalidity defenses . . . would have opened
the door for other generics to enter the market once Handa’s 180-day FDA
exclusivity period had run.” D.I. 828 at 6.
But in the real world, Handa was free to reject, ignore, or simply fail to
appreciate that logic. And the forecasts in PTX 1178 and PTX 1180 appear on
their face to show that on September 5 and 7, 201 1—that is, only weeks before the
challenged September 29, 2011 settlement—Handa expected that regardless of
how it prevailed in the underlying patent case at least five generic manufacturers
would enter the market following Handa’s exclusivity period. Of the Handa
forecasts in the record that were made before the September 29, 2011 settlement,!
the forecasts in PTX 1178 and PTX 1180 are closest in time to that date.
Accordingly, it was reasonable for Dr. Leffler to assume for purposes of his
economic model that, like AstraZeneca, Handa expected the same number of
generic entrants—i.e., at least five—regardless of whether Handa prevailed on its
noninfringement or invalidity defenses in the underlying patent case.
Defendants concede that PTX 1178 and PTX 1180 are “simplified
assessments of both a settlement (Case 1) and litigation (Case 2) outcome.”
D.I. 838 at 7. They fault the forecasts, however, for “not even includ[ing] a Handa
settlement scenario addressing the outcome of other generics’ ongoing litigations”
and not “contemplat[ing] a noninfringed-but-valid outcome for Handa, despite
[Dr.] Leffler’s assumption (based on Belvis) that this was the most likely
scenario.” D.I. 838 at 8 (emphasis in the original). But Handa’s actual
expectations about the likelihood and economic effects of the various potential
outcomes of the underlying patent case did not have to be reasonable. And, in any
event, alleged shortcomings in these actual (as opposed to hypothetical) forecasts
| See PTX 1165 (835-8), PTX 1178, and PTX 1180.
are appropriate material for cross examination, not grounds for excluding
Dr. Leffler’s opinions at trial. The forecasts’ lack of “contemplat[ion of] a
noninfringed-but-valid outcome for Handa” arguably confirms—not undermines—
the reasonableness of Dr. Leffler’s decision to use five or more generic entrants in
his model regardless of expected litigation outcomes.
II.
My finding that the AstraZeneca forecasts identified in Dr. Leffler’s
Opening Report and the Handa forecasts in PTX 1178 and PTX 1180 provided a
reasonable basis for Dr. Leffler to hold constant the number of generic entrants
(i.e., at five or more) in the application of his model resolves the pending motion.
But it does not resolve whether and, if so, the extent to which, I will allow
Dr. Leffler to refer to PTX 1178 and PTX 1180 at trial. Defendants argued in their
post-hearing briefing that PTX 1178 and PTX 1180 “and any corresponding
testimony” should be “stricken for non-disclosure.” D.I. 838 at 22. I have
effectively denied that request, as I considered PTX 1178 and PTX 1180 in ruling
today on Defendants’ Daubert motion. I decided it was appropriate to consider
PTX 1178 and PTX 1180 for purposes of the motion because (1) they were
identified in footnote 76 of Dr. Leffler’s Opening Report (albeit not for anything
related to the number of generic entrants in Dr. Leffler’s model), D.I. 835-2 54
n.76; (2) Defendants asked Handa employees questions about them in depositions
taken months before Dr. Leffler issued his Opening Report, D.I. 840-1, D.I. 840-2;
and (3) they were apparently referred to in a document linked to an Excel version
of Dr. Leffler’s model used by Defendants during Dr. Leffler’s deposition,
D.I. 840-4.
I am, however, troubled by the fact that Dr. Leffler made no reference to
PTX 1178 or PTX 1180 in paragraph 71 or footnote 139 of his Rebuttal Report.
Dr. Leffler purported to address in paragraph 71 Defendants’ criticism that he
“oversimplified [his] model by using a single value for the likelihood of Handa
winning the litigation,” D.I. 835-3 § 71—that is, the exact same criticism that is the
basis for Defendants’ Daubert motion. He responded as follows:
The alternative settlement model is quite detailed as
presented in my initial report. Using the single
probability of Handa winning offered by Mr. Belvis of
about 80% was reasonable. The only way that separately
modeling the impact of non-infringement versus
invalidity impacts my model is if a different number of
generics were expected depending upon whether Handa
won on validity or non-infringement. In my initial
report, I used the same number of generic entrants
regardless of whether Handa won on validity or non-
infringement.'38 This is consistent with the forecasts of
AstraZeneca and Handa,'*? which are the proper evidence
for modeling an alternative no-payment settlement
agreement reached in September 2011.
D.I. 835-3 J 71. In footnote 139, Dr. Leffler identified three documents as the
AstraZeneca and Handa forecasts that constituted “the proper evidence for
modeling an alternative no-payment settlement agreement.” He did not identify
PTX 1178 or PTX 1180. Moreover, in paragraph 72 of his Rebuttal Report,
Dr. Leffler wrote that his “[r]eliance on Handa and AstraZeneca’s forecasts of four
or more [(not five or more)] generics expected in the market after 180 days
regardless of a Handa win on invalidity or non-infringement [wa]s appropriate
based on the parties’ expectations at the time of settlement.” D.I. 835-3 | 72
(emphasis added).
To make matters worse, Plaintiffs made no mention of PTX 1178 or
PTX 1180 in their brief filed in opposition to Defendants’ Daubert motion or
during the lengthy oral argument I held on the motion. Instead, Plaintiffs argued
that Defendants were “wrong” to fault Dr. Leffler for not separately considering
the chances of Handa winning on noninfringement and invalidity “[b]ecause the
number of expected generic entrants affects both the expected values of litigation
and of settlement” and thus according to Plaintiffs, “any modification of Dr.
Leffler’s model to separately consider the chances of winning on invalidity and
noninfringement must be made to both the expected values of settlement and
continued litigation.” D.I. 714 at 6 (emphasis in the original). As I said in my
February 19 Memorandum Order, that explanation makes no sense to me. D.I. 828
at 7. There is no logical reason I can think of (and neither Dr. Leffler nor
Plaintiffs’ counsel has ever explained) why both AstraZeneca and Handa in a
hypothetical but-for world would necessarily have had the same expectations about
the number of generic entrants in a settlement scenario that they would have had in
a continuing litigation scenario. And, as noted above, in the abstract, the parties’
expectations in a continuing litigation scenario would have varied depending on
how the parties expected Handa to prevail on its noninfringement and invalidity
defenses.
I am equally troubled that Dr. Leffler appears to have offered at the
March 11 hearing opinions and explanations he purports to have based on
PTX 1178 and PTX 1180 that were not previously disclosed to Defendants. For
example, when asked at the hearing why he used the same number of generic
entrants in his model regardless of whether Handa expected to win on invalidity or
noninfringement, Dr. Leffler testified: “Because, in reviewing all these forecasts,
some of which clearly included the litigation scenario, they think they -- they are
telling me they think there will be five or more.” 3.11 Tr. 122:3-6 (emphasis
added). In his Rebuttal Report and deposition, Dr. Leffler offered a different
answer to the question—namely that he used the “same number” of generic
entrants—not necessarily five or more generic entrants—regardless of litigation
outcome. See D.I. 835-3 ¥ 71 (Rebuttal Report); D.I. 835-4 at 205:6-207:6
(deposition).? Dr. Leffler also disclosed at the hearing—apparently for the first
2 To be clear, J am not exactly sure what Dr. Leffler was saying in the cited
deposition lines, but Plaintiffs stated in their briefing, and I will accept, that
Dr. Leffler intended to communicate in those lines of testimony that “[b]ecause the
time—that he relied on and accorded weight to PTX 1178 and PTX 1180 because
those forecasts were “relatively contemporaneous” with the date of the challenged
settlement. See, e.g.,3.11 Tr. 106:2-4; 113:9-18.
Plaintiffs insist that “Dr. Leffler disclosed all his opinions in his reports and
backup.” D.I. 839 at 18. In support of this assertion, they point first to
paragraph 103 of Dr. Leffler’s Opening Report for the proposition that Dr. Leffler
“used 5+ generic entrants to calculate the expected value of settlement and
continued litigation regardless of the patent defense.” D.I. 839 at 18.
Paragraph 103, however, addresses AstraZeneca’s expectations, not Handa’s
expectations. D.I. 835-2 | 103. There is no dispute about the disclosure of
AstraZeneca’s forecasts as a basis for Dr. Leffler’s decision to hold constant the
number of generic entrants. The only issue is whether Plaintiffs sufficiently
disclosed to Defendants Dr. Leffler’s reliance on Handa’s forecasts in PTX 1178
and PTX 1180 to set the number of generic entrants in his model. Plaintiffs also
argue that the number of expected generic entrants and Defendants’ assumptions
about that number do not materially change the alternative settlement entry date.
D.I. 839 at 19. But whether the number of expected generic entrants does or does
number of expected generic entrants affects both the expected values of litigation
and of settlement, any modification of Dr. Leffler’s model to separately consider
the chances of winning on invalidity and noninfringement must be made to both
the expected values of settlement and continued litigation.” D.I. 714 at 6
(emphasis in the original) (citing lines 205:6—207:6 of Dr. Leffler’s testimony).
not materially change the alternative settlement entry date has no bearing on
whether Dr. Leffler or Plaintiffs disclosed Dr. Leffler’s reliance on PTX 1178 and
PTX 1180 in deciding the number of generic entrants to use in his model. Finally,
Plaintiffs point to a reference to PTX 1178 and a reference to PTX 1180 ina
document linked to the backup Excel files Dr. Leffler produced with his Opening
Report. D.I. 839 at 10-11, 19. Whether those references constitute a sufficient
disclosure of PTX 1178 and PTX 1180, however, is far from obvious. Neither my
clerks nor I were able to access the linked document from the Excel files in the
USB drive filed with the Court by Plaintiffs.
I need not decide today the permissible scope of Dr. Leffler’s testimony at
trial. But given the questions raised by the absence of any reference to PTX 1178
and PTX 1180 in paragraph 71 and footnote 139 of Dr. Leffler’s Rebuttal Report
and deposition and in Plaintiffs’ brief filed in opposition to the pending motion, I
will allow Defendants to file a stand-alone motion in limine separate and apart
from (and not subject to the normal page limitations for their inclusion in) the
Pretrial Order to address Defendants’ request to strike PTX 1178, PTX 1180, and
“any corresponding testimony.” If Defendants elect to file that motion, they must
identify with specificity the “corresponding testimony” they seek to preclude, and
they must provide sufficient briefing and case law citations to enable me to address
10
whether, and if so how, I should apply Federal Rule of Civil Procedure 37(c)
and/or the Pennypack factors in resolving the motion.
3B OK 2
NOW THEREFORE, at Wilmington on this Thirty-first day of March in
2025, itis HEREBY ORDERED that Defendants’ Motion to Exclude the Opinions
of Dr. Keith Leffler (Daubert Motion No. 1) (D.I. 633) is DENIED. It is
FURTHER ORDERED that if Defendants intend to file a motion in imine to
preclude and/or limit Plaintiffs from adducing at trial evidence of PTX 1178 and
PTX 1180, they shall do so no later than April 7, 2025. Any response to such
motion shall be filed no later than April 15, 2025, and any reply shall be filed no
later than April 22, 2025. Finally, it is FURTHER ORDERED that if such a
motion is filed, Plaintiffs shall make Dr. Leffler available at the Pretrial Conference
for questioning about the motion.
ih