AngioDynamics, Inc. v. C.R. Bard, Inc.

Opinion

UNITED STATES DISTRICT COURT NORTHERN DISTRICT OF NEW YORK

ANGIODYNAMICS, INC.,

Plaintiff, 1:17-cv-00598 (BKS/CFH)

v.

C.R. BARD, INC. and BARD ACCESS SYSTEMS, INC.,

Defendants.

Appearances: For Plaintiff: Philip J. Iovieno Nicholas A. Gravante, Jr. Mark A. Singer Cadwalader, Wickersham & Taft LLP 200 Liberty Street New York, NY 10281

Adam R. Shaw Anne M. Nardacci Boies Schiller Flexner LLP 30 South Pearl Street, 11th Floor Albany, NY 12207 For Defendants: Andrew J. Frackman Edward N. Moss O’Melveny & Myers LLP 7 Times Square New York, NY 10036 James P. Nonkes Philip G. Spellane Harris Beach PLLC 99 Garnsey Road Pittsford, NY 14534 Hon. Brenda K. Sannes, United States District Judge: MEMORANDUM-DECISION AND ORDER I. INTRODUCTION Plaintiff AngioDynamics, Inc. (“AngioDynamics”) brings this antitrust action against Defendants C.R. Bard, Inc. and Bard Access Systems, Inc. (collectively, “Bard”), asserting a claim of illegal tying in violation of section 1 of the Sherman Act (codified at 15 U.S.C. § 1)

under “per se” and “rule of reason” theories of liability. (Dkt. No. 1). AngioDynamics seeks treble damages, a permanent injunction, and declaratory relief. (Id. at 29). On May 5, 2021, this Court issued an order (the “May 5 Order”) denying both parties’ cross-motions for summary judgment, and granting Bard’s motion in limine to exclude the causation opinion and benchmarking analysis offered by AngioDynamics’ causation and damages expert, Dr. Alan Frankel. AngioDynamics, Inc. v. C.R. Bard, Inc., No. 17-cv-00598, 2021 WL 1792394, 2021 U.S. Dist. LEXIS 85578 (N.D.N.Y. May 5, 2021). The Court acknowledged the parties’ disagreement over whether other aspects of Dr. Frankel’s opinions could serve as a basis for a damages estimate at trial, noted that this issue was not before the Court, and stated that “the Court only addresses the issue that has been presented and briefed by the parties.” Id. at *42

n.26, 2021 U.S. Dist. LEXIS 85578, at *138 n.26. Following the May 5 Order, the parties disagreed, and sought clarification from this Court, on whether there remained any basis for Dr. Frankel to offer testimony on any topic at trial, or for AngioDynamics to seek any monetary damages beyond nominal ones. (Dkt. Nos. 191, 192). The Court directed AngioDynamics to identify what specific testimony from Dr. Frankel it sought to introduce at trial, and explain how such testimony could provide a basis for the jury to calculate a “just and reasonable” estimate of AngioDynamics’ damages. (Dkt. No. 193 (quoting Bigelow v. RKO Radio Pictures, 327 U.S. 251, 264 (1946)). AngioDynamics filed a letter brief in accordance with the Court’s instructions, (Dkt. No. 195), Bard responded, (Dkt. No. 196), and after procuring the Court’s leave, (Dkt. Nos. 198, 199), AngioDynamics filed a further response, (Dkt. No. 200). The Court heard oral argument regarding this issue on June 11, 2021. For the reasons below, the Court determines that Dr. Frankel may not offer a damages

analysis at trial based on Indicators 1 through 6 (as defined below), but may offer a damages analysis based on his alternative calculation for Indicator 7 (as defined below) that is based on AngioDynamics’ own sales data. II. MOTIONS TO SEAL As a preliminary matter, both parties have filed motions seeking to maintain under seal certain limited portions of their submissions that reflect confidential sales and market share information provided by the parties and non-party Teleflex Incorporated (“Teleflex”), as well as figures that Dr. Frankel derived using those estimates. (Dkt. Nos. 194, 197). In prior sealing orders in this matter, the Court has approved sealing of the same or similar information. (Dkt. Nos. 171, 177, 186, 190). Applying the standards set forth in Lugosch v. Pyramid Co. of Onondaga, 435 F.3d 110 (2d Cir. 2006) as described in this Court’s prior sealing decisions in

this case, (Dkt. Nos. 137, 171, 177, 186, 190), the Court finds that the parties’ proposed redactions are sufficiently limited, narrowly tailored and necessary to protect the parties’ and Teleflex’s confidential, competitively sensitive business information in which the Court has previously found that they have privacy interests that outweigh the strong presumption of public access, with the exception of the proposed redactions to Exhibit 1 of AngioDynamics’ Iovieno Declaration, (Dkt. No. 195-2), and the corresponding redactions to AngioDynamics’ letter brief itself, (Dkt. No. 195 at 5), for which such a showing has not been made. Therefore, Bard’s sealing motion is granted, and AngioDynamics’ sealing motion is granted in part. AngioDynamics is directed to file a version of its submission with revised redactions consistent with this order by no later than June 25, 2021. III. DR. FRANKEL’S TESTIMONY1 A. Dr. Frankel’s Remaining Damages Indicators Dr. Frankel calculated AngioDynamics’ purported damages by estimating its lost profits resulting from Bard’s TLS policy, using a formula that includes, among other variables, the

estimated percentage of Bard’s TLS‐paired PICC sales that would have been won by AngioDynamics had Bard not engaged in alleged tying (“PCTAngio”). (Dkt. No. 132-5, at 24).2 To determine this percentage, Dr. Frankel “considered a number of potential indicators of AngioDynamics’ prospective success at selling PICCs to users of Bard TLS devices.” (Id. at 28). These “indicators”3 were: • The 5 percent to 10 percent of Bard preloaded PICCs that a Bard employee estimated AngioDynamics could “pick off” to serve “special populations” if Bard were to sell its TLS stylet on a standalone basis (“Indicator 1”);

• AngioDynamics’ share of PICCs sold for use in the interventional radiology (“IR”) suite which does not require the use of TLS technology (“IR PICCs”) (“Indicator 2”), its share of PICCs sold for use by registered nurses at a patient’s bedside without the use of TLS technology (“Non-Navigation Nursing PICCs”) (“Indicator 6”), and AngioDynamics’ combined share of IR PICCs and Non-Navigation Nursing PICCs (“Indicator 4”) between 2013 and 2018;

• AngioDynamics’ share of total sales of PICCs that are paired with Teleflex’s TLS devices (“Indicator 7”);

• The share of Bard’s TLS-paired PICC sales that AngioDynamics would capture absent Bard’s TLS policy assuming that, without that policy, Bard would have the same

1 The Court assumes familiarity with the factual background of this matter in general, and Dr. Frankel’s opinions in particular, as set forth in the May 5 Order. AngioDynamics, 2021 WL 1792394, 2021 U.S. Dist. LEXIS 85578. As such, the Court does not repeat that background here, and instead only summarizes the portions of Dr. Frankel’s opinions that are directly relevant to the present dispute. 2 At this stage, Bard does not appear to challenge any aspects of Dr. Frankel’s formula for calculating lost profits other than his approach for determining the value of PCTAngio. 3 The Court refers to Dr. Frankel’s various indicators using the labels assigned to them in AngioDynamics’ letter brief and Dr. Frankel’s reports. (Dkt. No. 194; Dkt. No. 132-5, at 42). additional relative success at selling PICCs to users of its own TLS devices as Teleflex currently does, compared to its success at selling unguided nursing PICCs (“Indicator 3”); and

• A 2014 analyst report’s estimate of the potential PICC market share AngioDynamics could obtain if it could succeed in offering its own TLS device (“Indicator 5”).

(Id. at 28-29; see also id. at 42 (graph summarizing the foregoing indicators)). Dr. Frankel then “compute[d] AngioDynamics’ lost profits using two of these shares”—Indicators 7 and 3—“as benchmarks and explain[ed] (and illustrate[d]) how those computations can be scaled to reflect lost profits using other assumed values of [PCTAngio].” (Id. at 29-32). Dr. Frankel’s report stated that he “view[s] the indicators . . . and the benchmarks [he used] to compute lost profits and damages as providing a reasonable range of possible outcomes had the challenged conduct not occurred.” (Id. at 35). However, neither Dr. Frankel’s report nor his rebuttal report specifically explained whether he believed any of the five other damages indicators he considered could serve as a reasonable, reliable value for PCTAngio in his lost profits formula. (Id. at 28-32; Dkt. No. 132-6, at 12-16).4 When asked about his seven indicators in his deposition, Dr. Frankel testified that the indicators were a “starting point” for his analysis and collectively provide “the range of information we know that might bear on” the correct value of PCTAngio, given that “it’s impossible to know with precision what percentage of Bard sales in the but-for world would have been Angio sales.” (Dkt. No. 132-3, at 228-29). He further noted that, in selecting Indicators 7 and 3 to use as benchmarks, his approach was to “calculate damages based on what

4 In his rebuttal report, Dr. Frankel stated that “[t]he best available evidence to shed light on [the success AngioDynamics would have absent Bard’s TLS policy] is provided by AngioDynamics’ success at selling nursing PICCs for applications that do not require use of Bard stylets,” and that he “provided information concerning AngioDynamics’ average share of IR PICCs (and its average combined share of IR and nursing PICCs)” “[f]or additional context,” but that “the other nursing PICC applications are the closest analogues to the PICCs that would have been used with Bard stylets” absent Bard’s TLS policy. (Dkt. No. 132-6, at 13). [he] think[s] are two reasonable approaches that give different answers” and, together, “reasonably bookend a reasonable range for lost profits and damages.” (Id. at 230). He clarified that when his report stated that he viewed all the indicators together as providing a “reasonable range of possible outcomes,” he meant that he “first gathered all the available indicators that bear

on this question that I could find, or think of, as to what share [of Bard’s PICC business] Angio is likely to get” absent Bard’s TLS policy, then “selected two benchmarks which I thought was a reasonable basis to calculate damages.” (Id. at 235). He testified that he believes that his other five indicators are “interesting” as “[a]dditional context,” but are “not . . . as appropriate measures of damages as [Indicators 7 and 3],” and that “[i]f [he] didn’t have [Indicators 7 and 3] and all [he] had are some of these other measurements, [he] might consider using them,” but “[he] thought [Indicators 7 and 3] were better.” (Id. at 233). He specifically confirmed that Indicator 1 is not in the range that “[he] view[s] as reasonable for calculating damages”; that Indicator 2 is also not in that range “given that [he has] the other indicators”; that he does not “consider [Indicator 4] to be the most reasonable indicator to use as a damages benchmark”; that

he did not intend to include Indicators 5 or 6 as “providing a reasonable range of possible outcomes”; and that he intended to “focus on [his benchmarking analysis, based on Indicators 7 and 3] as the most reasonable things to use for calculating damages.” (Id. at 237-39). In its May 5 Order, the Court granted Bard’s motion to exclude Dr. Frankel’s benchmarking analysis based on Indicators 7 and 3. AngioDynamics, 2021 WL 1792394, at *41- 48, 2021 U.S. Dist. LEXIS 85578, at *136-59. The Court found that the data on which both benchmarks were based—an email from Teleflex’s counsel providing a “rough estimate” of the percentage of its standalone stylets that are paired with AngioDynamics PICCs, which was based on undefined “sales personnel input,” which did not specify the time period the estimate related to, and which Teleflex “cannot confirm”—was so unreliable as to warrant the methodology’s exclusion. Id., at *43-46, 2021 U.S. Dist. LEXIS 85578, at *142-53. Because the question of whether any of Dr. Frankel’s remaining indicators passed muster under Daubert as a basis for estimating AngioDynamics’ damages was not before the Court, the Court did not address that

issue. Id. at *42 n.26, 2021 U.S. Dist. LEXIS 85578, at *138 n.26. In a supplemental declaration submitted with AngioDynamics’ letter brief now before the Court, Dr. Frankel confirms that Indicators 1, 2 and 5 do not rely on the Teleflex email, and therefore are not impacted by the May 5 Order; that Indicator 3 “cannot be meaningfully computed without relying on the Teleflex” email, and that Dr. Frankel therefore no longer has a basis to present a damages model based on that indicator; and that in light of the May 5 Order, he has now “conservatively adjusted” Indicators 6 and 4 “to rely only on actual sales data records” rather than the Teleflex email. (Dkt. No. 195-4, at 5-6). B. Alternative Estimate of AngioDynamics’ PICCs Paired with Teleflex’s TLS Dr. Frankel’s rebuttal report, which was disclosed prior to the May 5 Order and Dr. Frankel’s deposition, presented an alternative method of calculating Indicator 7—

AngioDynamics’ share of total sales of PICCs that are paired with Teleflex’s TLS devices—that did not rely on the Teleflex email that the Court found to be unreliable. Specifically, according to Dr. Frankel’s rebuttal report: Using AngioDynamics sales data . . . I reported that in 2018 AngioDynamics sold non-IR PICCs to customers AngioDynamics was able to identify as using Teleflex TLS devices. Those sales alone represent percent of Teleflex separate stylet usage in 2018.

(Dkt. No. 132-6, at 14-15). While Dr. Frankel’s rebuttal report was the first time this calculation was presented, the underlying data—the list of customers using Teleflex TLS customers and the total number of non-IR PICCs sold to each customer in 2018—was disclosed in Dr. Frankel’s opening report. (Dkt. No. 132-5, at 30 n.62). In his deposition, Dr. Frankel was questioned on this alternative method, and he described the process in greater detail: My staff . . . asked AngioDynamics, through counsel, to give us their best available list of hospitals that they knew were using Teleflex TLSs and stylets to place Angio nursing PICCs. And they did their best to assemble that list for us. And then my staff went through the data. And that was a process because . . . the right hospital names in the data wasn’t always obvious, and so we did our best to match those up. I think they may have been confirming the communications with Angio to make sure we were on the right track. And that generated the percent of stylets that that are sold separately.

(Dkt. No. 132-3, at 273-74).5 While Dr. Frankel’s rebuttal report presented this alternative methodology as a way of confirming the reliability of his original analysis based on the Teleflex email (a contention which the Court rejected, AngioDynamics, 2021 WL 1792394, at *45-46, 2021 U.S. Dist. LEXIS 85578, at *151-53), AngioDynamics now contends that, with the Court having excluded that original analysis, this alternative method provides an independently reliable calculation of PCTAngio, which the jury may properly use in Dr. Frankel’s formula to calculate AngioDynamics’ lost profits. (Dkt. No. 195, at 9-10). IV. DISCUSSION6 A. Dr. Frankel’s Remaining Damages Indicators AngioDynamics argues that the remaining five indicators listed in Dr. Frankel’s report that are unaffected by the May 5 Order (including Indicators 4 and 6 as adjusted to rely only on actual sales data rather than the unreliable Teleflex email) each serve as a “reasonable and

5 At oral argument on Bard’s original motion to exclude Dr. Frankel’s benchmarking analysis, the parties disputed whether the percent figure Dr. Frankel independently calculated should be reduced in order to represent the percentage of total (as opposed to standalone) Teleflex TLS stylets that are sold for use in AngioDynamics’ PICCs. (Dkt. No. 185, at 44, 78-79). In a supplemental declaration, Dr. Frankel has since clarified that this dispute resulted from a typographical error in his rebuttal report, and that the percent figure does, in fact, represent the percentage of total Teleflex TLS stylets that are sold for use in AngioDynamics’ PICCs. (Dkt. No. 195-4, at 5 n.12). Bard does not appear to challenge Dr. Frankel’s explanation. 6 To the extent necessary, the Court applies the legal standards for analyzing expert testimony under Daubert and calculating antitrust damages as described in the May 5 Order, and does not repeat those standards here. AngioDynamics, 2021 WL 1792394, at *38-39, 41-48, 2021 U.S. Dist. LEXIS 85578, at *126-28, 136-59. reliable” estimate of the additional PICC sales AngioDynamics would have won from Bard but for Bard’s TLS policy, and that any of them can be plugged into Dr. Frankel’s formula as PCTAngio in order to calculate an estimate of AngioDynamics’ lost profits resulting from that policy. (Dkt. No. 195, at 1-9). Bard argues that Dr. Frankel may not use the remaining indicators

as a basis for a damages estimate after failing to do so in his reports and (in Bard’s view) conceding their unreliability in his deposition, and also brings substantive arguments challenging the reliability of each remaining indicator. (Dkt. No. 196, at 1-9). In his reports, Dr. Frankel chose only two of the seven indicators he presented to use as his benchmarks, and provided damages calculations based on those benchmarks alone. To the extent Dr. Frankel believed that any or all of the other five indicators could serve as a basis for a “just and reasonable” estimate of AngioDynamics’ damages as the antitrust laws require, Bigelow, 327 U.S. at 264, he could easily have explained his reasoning and presented alternative damages calculations based on those indicators. He did not do so. In his deposition, he clarified the reasoning behind this approach: he viewed the seven indicators merely as a “starting point”

that represented “the range of information we know that might bear on” the question of what percentage of Bard’s PICC business AngioDynamics could capture had Bard not engaged in alleged tying; his benchmarking analysis used the two indicators that he thought, viewed together, could “reasonably bookend a reasonable range for lost profits and damages”; and he included the remaining five indicators in his report as “interesting” “context,” but did not rely on them in calculating AngioDynamics’ damages. (Dkt. No. 132-3, at 228-39). It is only now, after the Court has excluded the benchmarking analysis upon which Dr. Frankel’s damages analysis rested, that Dr. Frankel and AngioDynamics do an about-face and argue that the indicators Dr. Frankel previously agreed were “not as reasonable” as those he used in his benchmarking analysis, (id. at 231), are not actually mere “context” or a “starting point,” (id. at 228, 233), but in fact constitute independent, alternative and reliable bases for damages calculations, which, viewed together, present a reasonable range of damages estimates for the jury to consider. Whatever the reasons for Dr. Frankel’s decision not to use his other five

indicators—Indicators 1, 2, 4, 5, and 6—as a basis to calculate damages in his reports, and to disclaim reliance on them in his deposition, he cannot now recast these indicators as damages theories. Dr. Frankel’s reports provide absolutely no analysis establishing that any of the remaining indicators are reliable approximations of the additional PICC sales AngioDynamics would have won from Bard but for Bard’s TLS policy, and his deposition testimony explicitly suggests that he did not consider them sufficiently reliable approximations to use as a basis for his damages model. With expert discovery closed, it is now too late for him to attempt to rehabilitate these indicators by providing such an analysis. Indeed, a brief examination of each indicator highlights the need for Dr. Frankel to have done more work at the outset to establish their reliability before attempting to use them as a basis

for a damages theory. Indicator 1 is a statement from a Bard District Manager (which Bard’s Senior Vice President of Sales and Marketing then copy-pasted into an internal email) averring his belief that AngioDynamics’ “strategy is to pressure the market to pressure us into launching this wire, so it will open the door for them to go after special patient populations in our 3CG accounts and pick off 5-10%, which is better than the 0% they are getting at 3CG accounts now.” (Dkt. No. 195, at 5-6). While AngioDynamics correctly argues that economics experts are entitled to rely on a defendant’s contemporaneous internal estimates in creating a damages model, (id.), on its face, a single email from a middle-management employee at one point in time, without more foundation—including additional information about what precisely the employee meant by “special patient populations,” the basis for the employee’s estimate, and whether the estimate accurately reflected the views of Bard as an organization or merely a single employee’s off-the-cuff speculation—hardly constitutes the sort of reliable evidence on which an economics expert may rely, for similar reasons that the “rough estimate” in the Teleflex email

failed to pass muster. AngioDynamics, 2021 WL 1792394, at *43-46, 2021 U.S. Dist. LEXIS 85578, at *142-53.7 Indicator 5, a 2014 third-party analyst report’s estimate of the potential PICC market share AngioDynamics could capture if it could succeed in offering its own TLS device, fares little better; again, Dr. Frankel merely cites the report’s projection without any investigation into the underlying methodology, inputs, or assumptions that would be necessary to establish the estimate’s reliability. (Dkt. No. 132-3, at 295 (Dr. Frankel’s testimony that he “recall[ed] reading the analyst report at some point, but [he] didn’t make any further investigation,” and “didn’t place much weight on it” as “it was just . . . an additional market observer’s point estimate.”)).8

7 AngioDynamics points to the Court’s statements in its May 5 Order that “Bard’s internal emails and documents, and this email in particular, reflecting Bard’s own employees’ ‘beliefs about the impact of Bard’s allegedly anticompetitive activity are clearly relevant to the question of whether that activity caused injury to AngioDynamics,’” and that “[a]rguments about the proper inferences to be drawn from those statements or the weight they should be given merely raise factual disputes that are best reserved for trial.” (Dkt. No. 195, at 5 (quoting AngioDynamics, 2021 WL 1792394, at *35, 2021 U.S. Dist. LEXIS 85578, at *117)). The Court made these statements, however, in the context of determining whether there was record evidence from which AngioDynamics could demonstrate that it suffered any injury as a result of Bard’s conduct, not in the very different context of quantifying that injury, for which, notwithstanding a somewhat “relaxed standard of proof,” Hygrade Milk & Cream Co. v. Tropicana Products, Inc., No. 88-cv-2861, 1996 WL 257581 at *16-17, 1996 U.S. Dist. LEXIS 6598, at *52-54 (S.D.N.Y. May 16, 1996) (citing MCI Comms. Corp. v. Am. Tel. & Tele. Co., 708 F.2d 1081, 1161 (7th Cir. 1982)), the party presenting expert testimony bears the burden of establishing the reliability of that testimony “by a preponderance of the evidence,” Lara v. Delta Int’l Mach. Corp., 174 F. Supp. 3d 719, 734 (E.D.N.Y. 2016). The Court’s finding that the Bard employee’s statement passed the minimal test of relevance with respect to the question of whether AngioDynamics suffered any injury does not imply a finding that the specific, 5-10% estimate the employee gave is sufficiently reliable to serve as a basis for quantifying that injury. 8 By contrast, in a case AngioDynamics relies on, In re SemCrude L.P., 648 F. App’x 205 (3d Cir. 2016), the court found an expert’s reliance on a Goldman Sachs valuation permissible where, among other things, the expert “had previously worked for Goldman Sachs” such that it could not “be said that [the expert] did not ‘know the methodology used to create the [Goldman Sachs Report] or the assumptions on which the [Goldman Sachs Report’s] price and volume estimates were based,’” and the expert “did not simply adopt the Goldman Sachs’ evaluation as his own” but As for Indicators 2, 4 and 6, all of which measure AngioDynamics’ success in segments of the PICC market in which TLSs are less of a factor, or not a factor, there is at least an arguable basis to believe that these indicators could serve as reasonable, if imperfect, estimates of how AngioDynamics would fare in the nursing PICC segment if Bard’s TLS policy was not at

play in that segment. However, in light of Dr. Frankel’s failure to conduct any analysis establishing that either the IR or non-navigation nursing market segments are sufficiently comparable to the navigation nursing market segment to serve as valid comparators, combined with his explicit decision not to use these indicators as a basis for any damages calculation in his reports or deposition, the Court cannot find that he has met even his relatively low burden of establishing their reliability as a damages basis.9 For the foregoing reasons, the Court finds that the five indicators unaffected by the Court’s May 5 Order do not constitute evidence from which a “just and reasonable” estimate of AngioDynamics’ damages may be made. Bigelow, 327 U.S. at 264. Therefore, Dr. Frankel will not be permitted to offer a damages model based on these indicators at trial.

B. Alternative Estimate of AngioDynamics’ PICCs Paired with Teleflex’s TLS AngioDynamics also contends that Dr. Frankel’s alternative estimate for Indicator 7— AngioDynamics’ share of total sales of PICCs that are paired with Teleflex’s TLS devices— which he derived from AngioDynamics’ own sales information rather than relying on the Teleflex email, constitutes an admissible estimate of PCTAngio for purposes of calculating lost profits. (Dkt. No. 194, at 9-10). Bard objects to this analysis as a “new and undisclosed” expert

“used his own analysis and judgment to adjust” the valuation. Id. at 214 (quoting ZF Meritor, LLC v. Eaton Corp., 696 F.3d 254, 293 (3d Cir. 2012)). 9 Because the Court will not permit Dr. Frankel to testify as to a damages model based on his remaining indicators, the Court need not decide the extent to which the adjustments Dr. Frankel has made to Indicators 4 and 6 to rely on actual sales data rather than the unreliable Teleflex email, as well as the resulting damages calculations, constitute belated, undisclosed expert opinions, as Bard argues. (Dkt. No. 196, at 6-7, 9). opinion, argues that it is “nothing more than a rough estimate based on an alleged poll conducted for purposes of this litigation by AngioDynamics,” contends (as it did in its motion in limine) that “Dr. Frankel did zero work to analyze whether Teleflex TLS customers are comparable to Bard TLS customers,” and raises several other arguments challenging Dr. Frankel’s application

of the data he obtained from AngioDynamics. (Dkt. No. 196, at 7-8 & n.27). As Bard correctly points out, Dr. Frankel’s alternative analysis was first presented in his rebuttal report as a way of confirming the reliability of the percent figure that he derived using the unreliable Teleflex email, and at that time, Dr. Frankel did not present an independent damages calculation based on the percent figure he calculated through his alternative analysis. (Dkt. No. 132-6, at 14-15). However, unlike the five indicators previously discussed— for which Dr. Frankel presented no analysis or justification in either his original or rebuttal report—Dr. Frankel’s reports laid out his justification for believing that Indicator 7 could serve as the best possible approximation of the additional sales AngioDynamics would capture from Bard in a world without Bard’s TLS policy, as well as the methodology by which he used

Indicator 7 to estimate AngioDynamics’ lost profits. (Dkt. No. 132-5, at 29-30; Dkt. No. 132-6, at 13). Dr. Frankel’s opening report also disclosed a “partial list of several dozen customers that purchase AngioDynamics’ PICCs and use Teleflex TLS devices” that AngioDynamics had identified, and that “in 2018 that partial list of customers purchased non-I.R. PICCs from AngioDynamics.” (Dkt. No. 132-5, at 30 n.62). Dr. Frankel’s rebuttal report described (albeit briefly) how he used that “partial list of customers” provided by AngioDynamics as an alternative means of calculating Indicator 7 without relying on the Teleflex email, (id. at 14-15), and Bard had the opportunity to extensively question him on that alternative method and the data underlying it during his deposition, (Dkt. No. 132-3, at 268-83). Under these circumstances, where Dr. Frankel’s methodology for deriving lost profits from Indicator 7 and his several alternative ways of calculating that indicator were explained and justified in his two reports, and Bard had a full opportunity to cross-examine him on these issues during his deposition, the Court does not find that Dr. Frankel’s failure to engage in the mechanical exercise of plugging the

percent into his lost profits formula to derive a damages estimate, in itself, bars him from presenting this analysis at trial. Moreover, the Court finds that the reliability issues with respect to this methodology, while potentially significant, are not so extreme as to require the methodology’s exclusion, in contrast to Dr. Frankel’s benchmarking analysis based on the “rough estimate” in the Teleflex email that the Court previously excluded. Because the Teleflex email provided no visibility into how that estimate was derived, Dr. Frankel knew absolutely no information about it or the data underlying it other than the final number, which Teleflex’s counsel made clear was a “rough estimate” that could not be confirmed because Teleflex does not track the requested data. AngioDynamics, 2021 WL 1792394, at *43-46, 2021 U.S. Dist. LEXIS 85578, at *142-53. By

contrast, here, Dr. Frankel himself has derived an estimate of Indicator 7 by relying on: (1) a specific (and possibly underinclusive) list of customers provided by AngioDynamics that it has confirmed use Teleflex’s TLS with AngioDynamics’ PICCs, and (2) data regarding the total number of PICCs sold to each of those customer in a particular year, specifically excluding those sold for use in the IR suite, which do not require the use of any TLS. In its letter brief and at oral argument, Bard has raised compelling and well-taken arguments challenging the reliability of this data, including that the list of customers AngioDynamics provided was compiled through polling of its sales representatives rather than through data maintained in the ordinary course of business, that Dr. Frankel does not know the specific basis for these sales representatives’ knowledge about their customers’ pairing of AngioDynamics’ PICCs with Teleflex’s TLSs, and that Dr. Frankel has not independently confirmed the reliability of the data AngioDynamics provided. However, given the relatively lenient threshold for admissibility under Daubert and the wide latitude given to experts in selecting the data on which they rely, the Court finds that these

are issues of weight best addressed at trial through cross-examination. AngioDynamics, 2021 WL 1792394, at *43-44, 2021 U.S. Dist. LEXIS 85578, at *143-48 (reviewing case law and explaining that “experts have wide latitude with respect to the data on which they rely, and challenges concerning the reliability of an expert’s data often present questions of weight, rather than admissibility,” unless the data falls into an “extreme category of cases in which the data relied on by an expert is so patently unreliable as to render it inadmissible,” as in ZF Meritor). Finally, the Court has considered Bard’s arguments (both in its submissions now before the Court, and in its original motion in limine) that Dr. Frankel has failed to conduct any analysis establishing that Teleflex’s customers are sufficiently similar to Bard’s as to constitute an appropriate benchmark. As the Court explained when discussing these arguments in its May 5

Order: As a starting premise, the general rule is that arguments regarding the comparability or appropriateness of a particular benchmark go to the weight, not the admissibility, of a benchmarking analysis . . . Moreover, a benchmark need not be perfectly comparable, so long as it allows the jury to calculate a “reasonable estimate of damages” as required under the antitrust laws . . . This is particularly true where a Defendant's anticompetitive conduct has rendered selection of a perfectly comparable benchmark difficult or impossible . . . Nonetheless, there are some circumstances under which, even applying the foregoing liberal standard, a benchmarking analysis may be so utterly deficient as to warrant its exclusion. One such circumstance is where the expert did not merely select arguably inappropriate benchmarks, but utterly “failed to perform any substantive analysis of those factors most relevant to comparability” . . . Similarly, courts have excluded benchmarking analyses that attribute all of plaintiff's claimed losses to a defendant’s misconduct, without making any effort to isolate the losses actually attributable to that conduct from the impact of other significant differences between the plaintiff and the chosen benchmark . . . Here, Dr. Frankel used PICC sales by Teleflex—which is Bard’s only competitor in the navigation-enabled TLS space, and which does not engage in anticompetitive activity— as a benchmark for the PICC sales Bard would have made had it not engaged in that activity. This approach at least attempts to approximate a but-for world that “[has] not been affected by [Bard’s alleged] antitrust violations,” and to use a benchmark company that is comparable to Bard in some important ways, including the industry they compete in, the types of products they sell, and the customer base they serve, i.e. “hospitals that [ ] place PICCs.” . . . Bard’s arguments, however, raise valid questions about the reliability of Dr. Frankel’s analysis, especially about whether he sufficiently considered important potential differences between the purchasing preferences of the two companies’ customers that could impact the appropriateness of his benchmarks.

AngioDynamics, 2021 WL 1792394, at *47-48, 2021 U.S. Dist. LEXIS 85578, at *155-59 (citations omitted). In light of the foregoing standards, the Court finds that Dr. Frankel’s use of Teleflex—which, notwithstanding potential important differences between the two companies, is likely the best, and possibly the only, comparator to Bard in the navigation-enabled TLS space— as a benchmark for Bard is not so utterly unreasonable or unjustified as to warrant an exception to the usual rule that comparability questions are issues of weight, rather than admissibility. As such, Bard’s arguments regarding comparability, however valid, are best addressed through cross-examination at trial. The Court has considered Bard’s remaining arguments attacking Dr. Frankel’s methodology, (Dkt. No. 196, at 8 n.27), and, as with Bard’s arguments regarding comparability, finds them to be questions of weight rather than admissibility, which are most appropriately addressed through cross-examination. Therefore, the Court will permit Dr. Frankel to present a damages analysis based on his alternative calculation for Indicator 7 that is derived from AngioDynamics’ own sales data. V. CONCLUSION For these reasons, it is hereby ORDERED that AngioDynamics’ motion to maintain certain portions of its submission under seal (Dkt. No. 194) is GRANTED in part to the extent set forth in this opinion, and AngioDynamics is directed to file a copy of its letter brief and exhibits (Dkt. No. 195) with revised redactions consistent with this opinion no later than June 25, 2021; and it is further ORDERED that Bard’s motion to maintain certain portions of its submission under seal (Dkt. No. 197) is GRANTED; and it is further ORDERED that Dr. Frankel may not offer a damages analysis at trial based on Indicators 1 through 6 (as defined in this opinion), but may offer a damages analysis based on his alternative calculation for Indicator 7 (as defined in this opinion) that is derived from AngioDynamics’ own sales data. IT IS SO ORDERED. Dated: June 11, 2021 ; Brrr ob cx Korres Syracuse, New York □□□□□□□□□□□□□□□□□□□□□□ U.S. District Judge