City of Huntington, West Virginia v. AmerisourceBergen Drug Corporation

Opinion

IN THE UNITED STATES DISTRICT COURT FOR THE SOUTHERN DISTRICT OF WEST VIRGINIA THE CITY OF HUNTINGTON, Plaintiff, v. CIVIL ACTION NO. 3:17-01362 AMERISOURCEBERGEN DRUG CORPORATION, et al., Defendants. ________________________________ CABELL COUNTY COMMISSION, Plaintiff, v. CIVIL ACTION NO. 3:17-01665 AMERISOURCEBERGEN DRUG CORPORATION, et al., Defendants. ________________________________ MEMORANDUM OPINION AND ORDER Pending before the court is plaintiffs’ motion in limine for an order ruling IQVIA data admissible. See ECF No. 1066. That motion is fully briefed and was argued before the court on January 6, 2021. After the hearing, plaintiffs filed a motion for leave to file a sur-reply, see ECF No. 1204, that defendants opposed. See ECF No. 1208. The motion for leave to file a sur- reply is GRANTED but the motion in limine is DENIED without prejudice. I. In the MDL, for the first trial (Summit and Cuyahoga Counties) plaintiffs retained Lacey Keller, a "data mining expert," to support their claims that the Manufacturers: (1) implemented inadequate suspicious order monitoring systems ("SOMS"); (2) failed to comply with the Controlled Substances Act ("CSA"), 21 U.S.C. §§ 801 et seq. and its implementing regulations, 21 C.F.R. §§ 1301 et seq.; and (3) failed to investigate, report, and halt orders of prescription opioids into Summit and Cuyahoga counties which they knew, or should have known, were suspicious. See In re: National Prescription Opiate Litig., MDL 2804, 2019 WL 3934470, at *1 (N.D. Ohio Aug. 20, 2019). Plaintiffs in these cases have also retained Keller as an expert. In her expert report, Keller concludes: * Based on my analysis of IQVIA data, I conclude that Case Defendants [ABDC, Cardinal, and McKesson] had sufficient information to understand overall prescribing trends in Cabell County. Had they reviewed this data, they would have seen that dosage units of opioids prescribed in Cabell County doubled in just five years (1997 to 2002) and quadrupled in twelve years (1997 to 2009). Defendants would have also seen that MMEs – the strength of the drugs – doubled in just three years (1997 to 2000) and increased seven-fold by 2010. * The over 200 million of opioid dosage units and over three billion MMEs prescribed in Cabell County during the IQVIA period outpaced Cabell County's approximately 95,000 residents. In the peak years, Cabell County physicians were prescribing enough opioids to give every man, woman, and child in Cabell County over 150 2 dosage units per year. Had Case Defendants reviewed IQVIA data, they would have seen that Cabell County prescriptions were generally double the statewide average and triple the national average dosage units per capita. They would have also seen that Cabell County was, at times, amongst the top 20 of the United States' more than 3,000 counties in terms of dosage units per capita. They also would have seen that Cabell County exceeded the prescribing levels in West Virginia, the leading state in the U.S. in terms of opioid dosage unit prescriptions per capita, where physicians prescribed over two billion opioids for its fewer than two million residents. * [T]hat Case Defendants could have used IQVIA data and dispensing data to diligently monitor the prescribing activity of individual Cabell County prescribers. Had Case Defendants reviewed IQVIA data, they would have been able to identify by name prescribers exhibiting outlier prescribing activity based upon volume, dosage, and the composition of their prescriptions. The 1 percent of opioid prescribers – between five and nine prescribers – wrote upwards of 43% of all opioid dosage units and 65% of MMEs each year, totaling nearly 80 million dosage units and over 1.6 billion MMEs. In other words, a half dozen prescribers wrote enough opioid prescriptions to give every man, woman, and child in the county over 60 pills per year for several years between 2007 and 2011. ECF No. 1066-2 at 59. In formulating her opinions, Keller relied on data compiled by, and available for purchase from, healthcare company IQVIA ("IQVIA Data"). See id. at 5, 60. IQVIA Data is submitted from pharmacies, mail order services, and long term care facilities and provides information about physicians and the drugs they prescribed. IQVIA Data is used to measure market and product demand and to track product demand over time. 3 Plaintiffs’ motion in limine seeks to have the court rule two datasets owned and maintained by IQVIA are admissible in this trial. The two datasets--IQVIA's XPONENT and XPONENT PLANKTRAK--are part of the National Prescription Audit ("NPA") and are referred to as "the industry standard for measuring the retail outflow of prescriptions through the front door into the hands of consumers." ECF No. 1066-2 at 5. According to plaintiffs, these datasets are relevant to the issues of (1) defendants' notice or constructive knowledge of physicians' opioid prescribing patterns in Cabell County and Huntington, and (2) these opioid prescribing patterns themselves. Although defendants argue in opposition that the IQVIA datasets are inadmissible hearsay, plaintiffs contend they are not for two reasons. First, plaintiffs argue that the datasets are not hearsay to the extent plaintiffs may use them to show that defendants had notice or constructive knowledge of opioid prescribing patterns in Cabell County and Huntington. Second, even if the datasets otherwise are hearsay, plaintiffs maintain they are admissible under the hearsay exception for "[m]arket quotations, lists, directories, or other compilations that are

generally relied on by the public or by persons in particular occupations." Fed. R. Evid. 803(17).

4 II. A. Notice Evidence that is offered to show notice rather than to seek to prove the truth of the matter asserted is admissible as non-hearsay. See United States v. Cone, 714 F.3d 197, 219 (4th Cir. 2013) (allowing emails to be admitted “for the non-hearsay purpose of showing that [the defendants] were on notice as to the counterfeit nature of the goods they sold”). “‘Hearsay’ is a statement, other than one made by the declarant while testifying at the trial or hearing, offered in evidence to prove the truth of the matter asserted.” Fed. R. Evid. 801(c). Thus, an essential element of whether a statement is hearsay at all is determining for what purpose(s) the statement is offered. If not offered for the truth of the matter it asserts, it is not hearsay. Significantly, “‘evidence is not hearsay when it is used only to prove that a prior statement was made and not to prove the truth of the statement.’” United States v. Ayala, 601 F.3d 256, 272 (4th Cir. 2010) (quoting Anderson v. United States, 417 U.S. 211, 220 n.8 (1974)). Thus, a statement is not hearsay when the statement is offered to show that the County was on notice of the statement. See Green v. Adm'rs of the Tulane Educ. Fund, 284 F.3d 642, 660 (5th Cir. 2002) (holding that testimony as to three other complaints of sexual harassment was not hearsay because it was offered to prove that the employer was on notice of the complaint rather than for the truth of the matter asserted), abrogation on other grounds recognized by McCoy v. City of Shreveport, 492 F.3d 551, 559 (5th Cir. 2007); see also Southerland v. Sycamore Cmty Sch. Dist. Bd. of Educ., 125 F. App'x 14, 22 (6th Cir. 2004) (holding that rumor testimony and notes were admissible as non-hearsay because they were not offered to prove the truth of the matters they asserted, but instead were used to show that government officials had knowledge of the problem); Dixon v. Int'l Fed'n of Accountants, No. 09cv2839, 2010 WL 1424007, at *2 n.4 (S.D.N.Y. Apr. 9, 2010) (holding that evidence of complaints by other employees against the plaintiff 5 was non-hearsay because it was not offered for the truth of the matter asserted—that the plaintiff actually did the things complained of—but offered to show that plaintiff's employer was on notice of the complaints). A statement is also not hearsay if it is offered to prove knowledge or show the effect on the listener or listeners' state of mind. United States v. Safari, 849 F.2d 891, 894 (4th Cir. 1988). Thus, when used to show notice, knowledge, or effect on the listener, Mohler's statement is not hearsay. . . . Bland v. Fairfax Cnty., Va., No. 1:10cv1030 (JCC/JFA), 2011 WL 2490995, at *4 (E.D. Va. June 20, 2011). The notice argument is a nonstarter. First, it is clear that plaintiffs want to offer the evidence for the truth of the matter asserted, i.e., physicians’ opioid prescribing patterns in Cabell County and Huntington. See ECF No. 1152 at 4-5. Furthermore, although there is evidence that defendants had purchased some products from IQVIA, there is no evidence that defendants had purchased the two datasets at issue here. According to plaintiffs, they do not need to show defendants had actual knowledge of the data. They argue that the IQVIA datasets are admissible because defendants should have known of them, whether or not they actually did. Plaintiffs maintain that Keller's testimony shows that defendants should be charged with constructive knowledge of physicians' opioid prescribing patterns in Cabell County and Huntington. In so arguing, plaintiffs rely on Keller's deposition testimony that "IQVIA was widely available for purchase by the pharmaceutical industry. . . . And what we're trying to demonstrate here is the information that could 6 have been gleaned from distributors, had they accessed the IQVIA Xponent data or some other similar dataset." ECF No. 1124-2 at 12; see also id. ("We're saying based off what we have seen in the IQVIA data, there was -- you could understand the prescribing trends in Cabell County."). “[B]efore [a] plaintiff can argue non-hearsay notice she must show that the defendant was at least inferentially put on notice by the report.” George v. Celotex Corp., 914 F.2d 26, 30 (2d Cir. 1990). Keller's statements are insufficient to establish that defendants should have been on notice. In determining whether to admit newspaper articles to prove a criminal defendant's notice, one court stated that it must consider "the specific content of the articles and the connection, if any, between the articles and the defendant." United States v. Buck, No. 13 Cr. 282, 2017 WL 5201447, at *3 (S.D.N.Y. Oct. 30, 2017). The court noted that a decision to admit such evidence would depend “upon the relevance of the specific articles and the likelihood the defendant would have read them.” Id. (emphasis added). Thus, showing a connection between the IQVIA data and defendants is important. See Austin

v. Hill, Civil Action No. 11-2847, 2014 WL 3797284, at *2 (E.D. Pa. Aug. 1, 2014) (admitting document “for the limited purpose of demonstrating what effect, if any, it had on the defendants provided that plaintiff establishes defendants were aware of this 7 document and that it was sufficient to put them on notice. . . .”) (emphasis added); R.K. v. Kanaskie, CASE NO. 02-61534-CIV- MORENO/SIMONTON, 2006 WL 8450679, at *11 (S.D. Fl. Apr. 28, 2006) (“[T]his document is [n]ot hearsay, not offered for the truth of the matter asserted. . . . Therefore, it is not hearsay, and may be used to establish notice, assuming that there is evidence that the document was received or reviewed by a defendant.”) (emphasis added) (internal citation and quotation omitted). And the cases cited by plaintiffs do not really help their position. In those cases, the link between the evidence to be admitted and the party opposing admission is either far less tenuous or the evidence was not admitted because of an inability to show that it was reasonable to put the party on notice. See, e.g., Gardner v. Q.H.S., Inc., 448 F.2d 238, 244-45 (4th Cir. 2017) ("[The witness] should have been permitted to be cross-examined as to the article in Consumers Report provided that the article had been published and was read by him before the time of the fire in this case. If the president had not read the article, the jury should be cautioned that the question of counsel on cross- examination did not constitute evidence, and

whether he admitted reading it or not, the jury should be cautioned that they might not rely upon the article for the truth of the legal or factual conclusions it contained."); Franchina v. City of Providence, 881 F.3d 32, 50 (1st Cir. 2018) ("The 8 transcript, then, was admissible as non-hearsay if, as Franchina argues, it was offered for the purpose of establishing that the City was, or should have been, on notice of Franchina's alleged workplace harassment."); George v. Celotex Corp., 914 F.2d 26, 30 (2d Cir. 1990) ("Philip Carey could not have been put on notice by the Hemeon Report because there was no proof at trial . . . that it reasonably should have seen it as part of the published literature in the industry."). B. Rule 803(17) hearsay exception Federal Rule of Evidence 803(17) provides a hearsay exception for “[m]arket quotations, lists, directories or other compilations that are generally relied on by the public or persons in particular occupations.” The United States Court of Appeals for the Fourth Circuit has made clear that the rule is intended to cover “established factual information.” In re C.R. Bard, Inc., 810 F.3d 913, 923 (4th Cir. 2016) (holding that document did not fall under 803(17) because it bore “no resemblance to the factual list-type documents enumerated in Rule 803(17)”). Therefore, Rule 803(17) applies to “objective compilations of easily ascertainable facts,” not reports

containing “conclusions reached by analysis by a specialized marketing company.” JPC Management, Inc. v. Incredible Pizza Co., No. CV 08-04310 MMM (PLAx), 2009 WL 8591607, at *24 (C.D. Cal. July 14, 2009). 9 Rule 803(17) is a narrow exception to the hearsay rule, which applies by its terms to “[m]arket quotations, lists, directories, or other compilations that are generally relied on by the public or by persons in particular occupations.” That enumeration suggests that the exception is designed to include compilations of information such as reports of stock market prices, telephone directories, and sales information for products. The advisory committee's note supports that interpretation, explaining that the exception applies to matters such as “newspaper market reports, telephone directories, and city directories.” The note adds that the “basis of trustworthiness is general reliance by the public or by a particular segment of it, and the motivation of the complier to foster reliance by being accurate.” The courts have generally taken a similarly narrow view of the scope of Rule 803(17), applying it to compilations of data, not to narrative and potentially subjective assessments in either general or specialized publications. See United States v. Woods, 32 F.3d 361, 363–64 (3d Cir. 2003) (database showing location of manufacture of automobiles within exception); United States v. Masferrer, 514 F.3d 1158, 1162 (11th Cir. 2008) (Bloomberg market price quotes for various markets admissible); Conoco, Inc. v. Dep't of Energy, 99 F.3d 387, 393 (Fed.Cir. 1996) (characterizing Rule 803(17) as reaching "market reports, telephone directories, weather reports, mortality tables, or like documents"); United States v. Cassiere, 4 F.3d 1006, 1018–19 (1st Cir. 1993) (real estate listing of properties sold, sale prices, and dates sales closed admissible); United States v. Goudy, 792 F.2d 664, 675 (7th Cir. 1986) (bank directory showing bank routing numbers admissible); United States v. Grossman, 614 F.2d 295, 297 (1st Cir. 1980) (catalogue of a particular company's products admissible). Bianco v. Globus Med., Inc., Case No. 2:12-CV-00147-WCB, 2014 WL 119285, at *1 (E.D. Tex. Jan. 12, 2014). According to Keller: The IQVIA Xponent® data is one of many datasets owned and maintained by IQVIA, a healthcare information company formerly known as IMS Health and Quintiles. 10 The data is part of the National Prescription Audit ("NPA") and "is the industry standard for measuring the retail outflow of prescriptions through the ‘front door’ into the hands of consumers." The data is said to "integrate real-world data" to provide a "timely, accurate picture of [. . .] competitive performance and demand," as well as a "deep understanding of [. . .] key drug dispensing channels." IQVIA data is considered by some pharmaceutical companies as the "gold standard in terms of understanding prescription trends." IQVIA describes the key prescription information offerings of Xponent data as: "A suite of sub-national reporting providing granular prescription performance perfectly aligned to help manage customer operations, sales targeting, and representative incentive compensation." IQVIA is not reported through a government agency but is proprietary and purchased by financial and pharmaceutical companies. . . . Given that IQVIA reflects the prescribing history of physicians, the dataset could allow anyone who purchases it to determine how frequently a physician prescribed particular drugs, as well as what formulations and in what dosages, and how prescribers ranked among other prescribers. The data also allows for the identification of opioid prescribing patterns of individual physicians compared to their cohorts based on specialty, geography (e.g., city, county, zip code, state), and time period. ECF No. 1066-2 at 5. Keller's description of the IQVIA data as “the industry standard” is taken from IQVIA's own information. Defendants argue that the IQVIA datasets do not recite established factual information because IQVIA itself asserted at the time it produced these datasets in the MDL that the data therein represents "an estimate" and a "reasonable approximation" and is the product of "independent judgment, expertise and opinion of IQVIA representatives." ECF 1124-2 at 31 (quoting 11 Defendants' Ltr. from IQVIA dated July 25, 2018). According to IQVIA: Although there is an inclination to view numerical data as fact, the IQVIA Information represents an estimate of measured activity and should be treated accordingly. As more fully described in our Published Specifications for Information Services (“Published Specifications”), IQVIA Information reflects projections, estimates, forecasts that are the result of a combination of confidential and proprietary technologies, statistical methodologies and a significant number of sources. These estimates reflect the independent judgment, expertise, and opinion of IQVIA representatives to arrive at a reasonable approximation of market activity. The IQVIA Information is intended to support sales, marketing and research applications, and it is highly reliable for those purposes. The IQVIA Information, although appropriate for its intended purpose of supporting business and marketing analyses in industries such as the pharmaceutical industry, contains data that is susceptible to error or variance, and is not intended to be used as direct evidence or to establish any fact. Accordingly, IQVIA offers no assurances that the IQVIA Information will be suitable for use as evidence in any litigation. Id. Given IQVIA’s disclaimer, it is impossible for the court to conclude that the datasets are merely a compilation of objective facts as contemplated by Rule 803(17). See In re Dual--Deck Video Cassette Recorder Antitrust Litig., No. CIV 87-987-PHX RCB, 1990 WL 126500, at *4 (D. Ariz. July 25, 1990) (“The type of publications contemplated by Rule 803(17) are those which deal with compilations of objective facts not requiring for their statement, a subjective analysis of other facts.”); White Indus., Inc. v. Cessna Aircraft Co., 611 F. Supp. 1049, 1069 (W.D. Mo. 1985) (“[T]he kinds of publications contemplated by the rule are those which deal with compilations of relatively straightforward 12 objective facts not requiring, for their statement, a subjective analysis of other facts.”) (emphasis in original). Furthermore, on the basis of the record as it now stands, it is difficult to conclude that the datasets are “generally relied on by . . . persons in particular occupations.” See, e.g. Allen v. Hylands Inc., CV 12-1150-DMG (MANx), 2015 WL 12720304, at *13 (C.D. Cal. Aug 20, 2015) (“Plaintiffs state that the compilation is intended to inform ‘the public, consumers in the market for homeopathic OTC drugs, and persons in the medical and drug manufacturing occupations,’ but asserting a compilation’s intended audience is different from demonstrated that it is generally relied on by that audience.”). At this juncture, the court cannot conclude that the IQVIA datasets fall within 803(17). III. The motion in limine is DENIED without prejudice. Plaintiffs are free to seek admission of the datasets at trial if Keller and/or another witness can lay an appropriate foundation

for admission under Fed. R. Evid. 803(17), see BMG Rights Mgmt. v. Cox Commc’ns, Inc., 881 F.3d 293, 314 (4th Cir. 2018) (no abuse of discretion in admission of two studies under Rule 803(17) where expert testified that studies “were widely cited in the industry” and were “the most substantial published publicly available studies” on the issue), or some other rule. 13 The Clerk is directed to send copies of this Memorandum Opinion and Order to counsel of record who have registered to receive an electronic NEF. IT IS SO ORDERED this 12th day of April, 2021. ENTER:

David A. Faber Senior United States District Judge