Opinion
MDL 2804 Case No. 1:17-md-2804
08-26-2019
IN RE: NATIONAL PRESCRIPTION OPIATE LITIGATION THIS DOCUMENT RELATES TO: Track One Cases
ORDER
Before the Court is Defendants' Motion to Exclude Expert Testimony of Katherine Keyes, Anna Lembke and Jonathan Gruber re the "Gateway Hypothesis" of Causation (Doc. #: 1858). For the reasons stated below, the motion is DENIED.
I. Introduction
Jonathan Gruber is one of the nation's leading health economists. He received his Ph.D. in Economics from Harvard University and is the Ford Professor of Economics at the Massachusetts Institute of Technology, where he has taught for over 25 years. Gruber directs the Health Care Program at the National Bureau of Economic Research and served as President of the American Society of Health Economists from 2016-2018. He has twice won the International Health Economics Association Kenneth J. Arrow Award for the outstanding health economics paper and has served as Co-Editor for both the Journal of Public Economics and the Journal of Health Economics. Along with David Cutler, another of Plaintiffs' experts, he won the ASHEcon award for best health economist in the nation age 40 and under in 2006. Gruber has previously testified in Int'l Tobacco Partners, Ltd. v. Kline, 475 F. Supp. 2d 1078, 1082 (D. Kan. 2007) ("Gruber is clearly qualified to testify as an expert in this case.").
Katherine Keyes is an Associate Professor of Epidemiology at Columbia University, specializing in substance use and substance use disorders epidemiology. She received her Ph.D. in Epidemiology from Columbia University. She has published 225 peer-reviewed articles and book chapters. Her work appears in leading journals such as Pediatrics, JAMA Psychiatry, Lancet Psychiatry, American Journal of Epidemiology, and International Journal of Epidemiology, and is widely cited. Keyes has published two textbooks on epidemiological methods, both with Oxford University Press. She is an elected member of the executive board of the Society for Epidemiological Research and serves as Associate Editor of the journal Drug and Alcohol Dependence. Keyes has received numerous professional awards honoring her research achievements, including early career achievement recognitions from the Research Society on Alcoholism, the American Psychopathological Association, the World Psychiatric Association-Epidemiology and Public Health Section, and the NIH Office of Disease Prevention Early-Stage Investigator award.
Anna Lembke is an Associate Professor, Chief of the Addiction Medicine Dual Diagnosis Clinic, Medical Director of Addiction Medicine, and Program Director of the Addiction Medicine Fellowship, in the Department of Psychiatry and Behavioral Sciences at Stanford University School of Medicine. She received her M.D. from Stanford University, where she also completed a residency in psychiatry and a Fellowship in Mood Disorders, Department of Psychiatry and Behavioral Sciences. She has taught at Stanford since 2001. She teaches topics related to psychiatry, addiction, and pain. She lectures on addiction medicine as part of the Practice of Medicine series for Stanford medical students. For the last 15 years in her clinical practice, a significant proportion of her patients have taken prescription opioids for pain relief and she has treated them for misuse, dependence, and addiction. She has served on the Board of the California Society of Addiction Medicine since 2015 and has been a member of the American Society of Addiction Medicine since 2011. She has chaired the Addiction Medicine Task Force, Stanford University School of Medicine, and served as the president of the Addiction Medicine Fellowship Directors Association. Lembke has authored a book on the prescription drug epidemic: "Drug Dealer, M.D.: How Doctors Were Duped, Patients Got Hooked, and Why It's So Hard to Stop," published by Johns Hopkins University Press, along with over 50 peer-reviewed articles, chapters, and commentaries that have appeared in leading journals such as the New England Journal of Medicine, Journal of the American Medical Association, and Pain Medicine. The Court refers collectively to Keyes, Lembke and Gruber as "the Experts."
The Experts offer numerous opinions, but Defendants' Motion focuses specifically on what Defendants term the "Gateway Hypothesis." The Experts prefer the term "Gateway Effect" and define it in various ways, but the gist of their Gateway opinion is that "there is compelling evidence of harm from the oversupply of prescription opioids, both for medical users, and to non-medical users because of diversion. These harms include opioid use disorders and overdose; these harms are greater than other pain relief drugs, and are causally related to additional harms from opioids including transition to heroin addiction." Keyes Report at 3 (Doc. #: 1999-9). Put more simply, the Gateway Effect refers to evidence that some individuals exposed to prescription opioids tend to become addicted, and once addicted tend to later feed their addition with illicit opioids, such as heroin.
II. Discussion
Each of the Experts is well-qualified to testify, and Defendants do not challenge their qualifications. But Defendants do ask the Court to exclude testimony of the Experts concerning the Gateway Hypothesis, which Defendants characterize (or mischaracterize) as positing that "the entire problem of illicit heroin and fentanyl use today was somehow 'caused' by the last quarter century of trained healthcare providers prescribing legal opioid medicines." Motion at 2 (Doc. #: 1858) (emphasis added). Defendants also specifically ask the Court to bar the Experts from opining "that there is a correlation between non-medical prescription opioid use and heroin abuse." Motion at 16 (Doc. #:1858).
Defendants offer three reasons in support of their motion. First, they assert "there is no scientific evidence that even purports to look at - let along [sic] find a causal link between - properly prescribed pain patients and today's heroin and fentanyl problems." Motion at 6 (Doc. #: 1858). Second, the studies the Experts rely on, which Defendants insist examine only a population of "non-medical" opioid users, "do not claim that heroin or fentanyl use is 'caused' by such prior non-medical use or abuse of prescription opioids in this population." Id. Third, Defendants argue the Gateway Hypothesis lacks relevance to Plaintiffs' case, or, in the parlance of Daubert, lacks fit, and is thus apt to mislead the jury.
As explained more fully below, by "non-medical" opioid users, Defendants refer to individuals who used prescription opioids, but either: (1) did so despite not having a prescription written for them (e.g., used Grandma's prescription opioids), or (2) received a valid prescription, but then used their opioids not in conformity with the prescribed usage, dosage, or duration (e.g., consumed their opioids in greater amounts or more often or longer than directed).
Defendants' first two arguments differ subtly, and that distinction rests entirely on the classification of two populations studied by Lembke and Keyes, as well as by the epidemiologists whose work they cite: (1) "medical users" of prescription opioids and (2) "non-medical users" of prescription opioids. The term "medical users" of prescription opioids refers to individuals who used prescription opioids entirely in accord with a prescription they received, themselves. In contrast, by "non-medical users," Defendants refer to individuals who used prescription opioids, but either (a) did so despite not having a prescription written for them (e.g., used Grandma's prescription opioids), or (b) received a valid prescription, but then used their opioids not in conformity with prescribed usage, dosage, or duration (e.g., consumed their opioids in greater amounts or more often or longer than directed). Defendants' first argument essentially contends that none of the literature the Experts cite studied medical users of opioids, and thus it affords no basis to conclude that medical use of prescription opioids caused widespread heroin use. Defendants' second argument contends that the literature stops short of concluding that even non-medical use of prescription opioids caused the harms associated with heroin and illicit fentanyl.
A. The Gateway Experts Appropriately Relied on Epidemiological Literature.
Defendants argue the Experts lack any valid basis for their opinions on the Gateway hypothesis. Defendants assert that none of the Experts "has offered a single study that even tests - let alone supports - the hypothesis that medical use of prescription opioids causes the abuse of illicit street drugs, and their belief otherwise is speculation." Motion at 10 (Doc. #: 1858). Defendants also assert that "[n]one of the [16 epidemiological articles Keyes reviewed] reports on a controlled clinical trial designed to answer the hypothesis (or any hypothesis) the Experts put forth." Motion at 9 (Doc. #: 1858); see also Reply at 5 (Doc. #: 2446). Defendants ask, therefore, that the Court exclude "the Experts' 'gateway' opinion" as unreliable. Motion at 11 (Doc. #: 1858).
In response, Plaintiffs initially observe that a "review of scientific literature is a sound methodology." Opp. at 5-6 (Doc. #: 2197); see In re Heparin Prods. Liab. Litig., 803 F. Supp. 2d 712, 738 (N.D. Ohio 2011) ("Courts have admitted expert testimony as reliable where experts extrapolate their opinions from their knowledge and experience combined with a review of the relevant scientific literature."). Plaintiffs next lay out in detail numerous published studies analyzing heroin users' prior nonmedical prescription opioid use, which the Experts relied upon in their reports. Opp. at 5-16 (Doc. #: 2197). Plaintiffs maintain that Defendants' insistence on a sharp division between medical and non-medical use is both untenable and beside the point. Opp. at 11, 24, n. 68 (Doc. #: 2197). Plaintiffs cite studies showing that "the majority of [non-medical use by addicts] involved a history of medical use." Plaintiffs assert that "the medical users of prescribed opioids become the nonmedical users of those same addictive drugs, who then become part of the population of heroin users described in the Compton article. This is the Gateway Effect, in no uncertain terms." Opp. at 11 (Doc. #: 2197). In other words: a patient who starts with a valid opioid prescription and abides by its limitations (a "medical user") may then become addicted and breach those limitations (becoming a "non-medical user") and may then turn to illicit opioids. In this way, the original, legitimate opioid prescription can serve as a gateway to heroin.
S. McCabe, et al., Trends in Medical and Nonmedical Use of Prescription Opioids Among U.S. Adolescents: 1976-2015. Pediatrics 139, Number 4 (2017) at 9.
Defendants rely on the Compton article for their argument that the epidemiological literature does not support the Experts' conclusions about causation. See infra.
Finally, responding to Defendants' critique of the underlying literature for failing to include controlled clinical studies, Plaintiffs respond that such studies are ethically impermissible given the harmful and illegal nature of heroin and illicit fentanyl. Opp. at 8 (Doc. #: 2197).
The Court agrees with Defendants that the literature the Experts rely on focuses predominantly on illicit opioid users who were initially non-medical users of prescription opioids. However, the Court does not agree that this fact provides a basis on which to exclude the Experts' reports and testimony. As a report of the National Academies of Sciences, Engineering, and Medicine ("NASEM") noted, "the prescription and illicit opiate epidemics are intertwined; indeed, a majority of heroin users report that their opioid misuse or OUD began with prescription opioids." NASEM Report at 6 (Doc. #: 2197-1). The categories "medical user" and "non-medical user" are similarly intertwined. In many cases, medical users transitioned into non-medical users because they became addicted - a result that Plaintiffs assert, and some evidence suggests, had something to do with assurances from opioid manufacturers downplaying the risks of addiction.
Defendants seek to frame the Experts' Gateway Hypothesis as relating only to medical users, quoting a bit of Lembke's deposition testimony in which she states, "the group I'm referring to in the Gateway Effect is . . . those individuals who started with a medical prescription and then became addicted through that medical prescription . . . ." Motion at 3-4 (Doc. #: 1858) (quoting Lembke Dep. at 83:22-84:1 (Doc. #: 1858-4). Despite this deposition excerpt, the Experts do not actually confine themselves in this manner, and the Court does not understand their opinions as relating only to initial medical use. In the scope of the opioid epidemic, there is no meaningful distinction between (i) patients who started using prescription opioids as prescribed, but then began to overconsume because of dependence or addiction, and (ii) non-medical users who somehow overconsumed for other reasons. Likewise, some evidence suggests those who used prescription opioids without a prescription had the opportunity to do so because of the overabundance of these drugs in the medicine chests of their relatives or friends. Defendants' first contention, that the Experts' testimony must be excluded because none of the literature the Experts cite studied "medical users" of opioids, is, therefore, not well-taken. Defendants may explain to the jury why the distinction between these two populations in the literature is important; the distinction goes to the weight of the Experts' testimony, not its admissibility. See In re Heparin Prods. Liab. Litig., 803 F. Supp. 2d 712, 738 (N.D. Ohio 2011) ("Courts have admitted expert testimony as reliable where experts extrapolate their opinions from their knowledge and experience combined with a review of the relevant scientific literature.").
B. The Experts' Opinions on Causation
Defendants next argue that even those studies that address non-medical use of prescription opioids do not provide a sufficiently reliable basis upon which the Experts may found an opinion that prescription opioid use causes illicit opioid abuse. Motion at 11 (Doc. #: 1858); Reply at 4-5 (Doc. #: 2446). Defendants challenge any suggestion of causation by quoting a 2016 article examining prescription opioid use and heroin use by Dr. Wilson Compton, which states that "conclusions about cause and effect are uncertain." Motion at 11 (Doc. #: 1858). Defendants also rely on their own expert, Dr. Robert Lyerla, whose reports states that "[i]n epidemiology, and in particular for behavioral health conditions, causal relationships, such as the one that the gateway hypothesis suggests, are difficult to substantiate." Motion at 15 (Doc. #: 1858) (quoting Lyerla Report at 5 (Doc. #: 1858-10)). Defendants also point out that only a very small percentage of people who use prescription opioids later move on to illicit opioids. Motion at 15 (Doc. #: 1858); Reply at 6 (Doc. #: 2446).
Wilson M. Compton et al., Relationship between Nonmedical Prescription-Opioid Use and Heroin Use, N. Engl. J. Med., 374, 156-57 9 (2016).
Defendants also point to Keyes' own somewhat equivocal deposition testimony regarding causation ("the literature is 'consistent with' a causal association") to attack the reliability of her opinion. Motion at 13 (Doc. #: 1858) (quoting Keyes Dep. at 305 (Doc. #. 1858-11)). Defendants cite a Sixth Circuit case for the proposition that "to say that something is 'consistent with causing' an effect is to 'testify to a possibility rather than a probability.'" Turpin v. Merrell Dow Pharms., Inc., 959 F.2d 1349, 1360 (6th Cir. 1992); Motion at 13 (Doc. #: 1858). Defendants further state that Gruber "admits that the literature on which he relies 'are not causal studies' and that when he says these studies 'establish[] the link' between opioid use and heroin abuse, he doesn't 'mean that to say that these studies are causal evidence of that link.'" Motion at 14 (Doc. #: 1858) (quoting Gruber Dep. at 332-333, 271 (Doc. #: 1858-5). Defendants likewise attack Lembke's statement that there is a "clear link" connecting prescription opioid use and illicit opioid abuse, asserting Lembke is purposefully avoiding language that sets forth a more certain conclusion concerning causality. Motion at 14 (Doc. #: 1858) (quoting Lembke Report at 9 (Doc. #: 1997-10)).
Plaintiffs respond to each of these attacks, noting Defendants chose to examine certain phrases used by the Experts but ignored others. For example, whereas Defendants contend the Experts' own words concede an inability to reliably opine on causation, Plaintiffs point out that Gruber "based his opinions concerning causation on standard methods in his field of health economics, including regression analysis, that he clearly and repeatedly testified do meet standards of proving causation." Opp. at 23 (Doc. #: 2197) (emphasis added). Plaintiffs assert Defendants selectively quoted Gruber, who testified in deposition that his research team established a "causal relationship, not just a correlation," between increased opioid shipments and prescription overdose mortality for the period up to 2010, and that thereafter, "the evidence is clear from the time period we present to make the causal case that in those counties with high shipments, that's where the illicit deaths went up the most," and "we then, as is standard empirical practice, tried to rule out other factors that could explain that." Opp. at 23-24 (Doc. #: 2197) (quoting Gruber Dep. at 202:7-203:5, 207:16-208:4; 234:4-235:21 (Doc. #: 2197-32)).
In response to Defendants' argument that the studies the Experts rely upon do not show causation, Plaintiffs cite studies by Muhuri, Lankenau, and Cicero finding that up to 86%, 85%, and 79.5% of heroin users used prescription opioids before moving on to heroin. Keyes Report at 26 (Doc. #: 1999-9); Lembke Report at 85 (Doc. #: 1999-10). Plaintiffs explain the Experts rely on these data as a basis for their opinions regarding causation, pointing out various statistical attributes of the data, such as the strength of association, dose-response relationship, and temporal relationship - all attributes that support a causal relationship between prescription opioid use and illicit opioid use. Opp. at 12-13, n. 12 (Doc. #: 2197).
Lembke Report at 85 (Doc. #: 1999-10), citing Lankenau, et al., Initiation into Prescription Opioid Misuses Amongst Young Injection Drug Users, Int. J. Drug Policy 2012; 23(1):37-44, at p.41. (Eight-six percent of urban people who used injected heroin in New York and Los Angeles in 2008 and 2009 had used prescription opioids nonmedically before using heroin.).
Keyes Report at 26 (Doc. #: 1999-9), citing Cicero, et al., The Changing Face of Heroin Use in the United States: A Retrospective Analysis of the Past 50 Years, JAMA Psychiatry 2014; 71 (7): 821-826. Describing Cicero's findings, Keyes writes: "From the 1960s through 1970s, less than one third used prescription opioids before heroin. From the 1990s on, as the supply of opioids increased, so too did the proportion of individuals who use heroin who began opioid use with prescription opioids (among those initiating in the 1990s, 50%; 200s, 85%, 2010s, 78%).
Lembke Report at 85 (Doc. #: 1999-10), citing Muhuri, et al., Associations of Nonmedical Pain Reliever Use and Initiation of Heroin Use in the United States, CBHSQ Data Rev. 2013; (August):1-16, at p. 1. (79.5% of persons who recently began using heroin had used prescription opioids nonmedically before initiating heroin use.).
Lembke and Keyes both also rely on their own prelitigation publications, while Lembke further relies on her own first-hand experience from 15 years of clinical treatment of patients who suffered dependence and addiction after taking prescription opioids. Opp. at 16-19 (Doc. #: 2197).
See Lembke, Drug Dealer, MD: How Doctors Were Duped, Patients Got Hooked, and Why It's So Hard to Stop (2016) at 22-23 ("For increasing numbers of people, especially young people, prescription drugs are the first exposure to addictive substances and the first stepping stone to future addictive use.") (Doc. #: 2197-24); Keyes, et al., Understanding the Rural-Urban Differences in Nonmedical Prescription Opioid Use and Abuse in the United States, A. J. Public Health 104:e52-e59 at e53 (2014) ("Increased medical use of prescription opioids has resulted in increased access to opioids for nonmedical use, either through the nonmedical use of legitimately acquired prescriptions or through formal or informal distribution networks.").
The Court finds not well-taken Defendants' assertion that the Experts have no methodologically reliable basis to offer their opinions concerning causation. It is true, as Dr. Lyerla's excerpt suggests, and as corroborated by the Federal Judicial Center's Reference Manual on Scientific Evidence, that epidemiological studies by their nature commonly do not conclusively establish causation. Ref. Manual at 598 (Doc. #: 2197-12). But there is still reliable evidence from which one may reasonably infer that some heroin addiction results from opioid use. The strongest evidence is based on studies centered on non-medical users, but as the NASEM report stated, "[a] preponderance of evidence suggests that the major increase in prescription opioid use beginning in the late 1990s has served as a gateway to increased heroin use." NASEM Report at 215 (Doc. #: 2197-1). The NASEM report further noted "the interrelated nature of the prescription and illicit opioid epidemics means that one cannot be addressed separately from the other." Id. at 248. The Muhuri, Lankenau, and Cicero studies provide a sturdy basis for the Experts' opinions; the Experts' reliance on observational studies simply reflects that these studies are the best evidence the discipline can point to where, as in the case of addiction to illicit and deadly substances, controlled clinical studies are not feasible. And Defendants' reference to the low proportion of prescription opioid users who go on to use heroin is a red herring, for the Experts do not opine that most people who use prescription opioids became addicted to heroin; rather, they opine that most people who are addicted to heroin first used prescription opioids.
When assessing the methodological reliability of epidemiological studies suggesting causation, courts in the Sixth Circuit have relied on the so-called Bradford Hill criteria, a set of factors whose presence or absence indicate whether a correlation between variables indicates causality. In re Welding Fume Prods. Liab. Litig., 2005 WL 1868406, n. 75 (N.D. Ohio August 8, 2005) (the Bradford Hill factors "guide epidemiologists in making judgments about whether a cause-effect relationship may be inferred from an association."); In re Gadolinium, 2010 WL 1796334 at *22 (N.D. Ohio May 4, 2010) (approving experts' use of Bradford Hill criteria). As explained in the Reference Manual on Scientific Evidence, "epidemiology cannot prove causation; rather, causation is a judgment for epidemiologists and others interpreting the epidemiologic data." Ref. Man at 398 (Doc. #: 2197-12). The nine Bradford Hill criteria are: temporal strength (exposure to a causal agent precedes development of the supposed effect), strength of the association, dose-response relationship (the greater the exposure, the greater the likelihood of harm), replication of the findings, biological plausibility (coherence with existing knowledge), consideration of alternative explanations, specificity of the association, and consistency with other knowledge. Id. at 600. Findings of correlation do not have to meet any particular number of these criteria in order to support an inference of causation. Id.
Id. at 600.
Here, the Bradford Hill criteria weigh in favor of admitting the Experts opinions on causation. Plaintiffs cite a 2014 peer-reviewed study showing a high dose-response relationship between prescription opioid exposure and the development of opioid use disorder (OUD); the Lankenau and Cicero studies found that high percentages of heroin users ingested prescription opioids before initiating heroin use, satisfying the temporality requirement; the intensity of the correlation is high, as those same studies show; Gruber employed a regression analysis to rule out alternative causes, Opp. at 23 (Doc. #: 2197); and the substantial pharmacological similarity of prescription opioids and illicit opioids is evidence of biological plausibility, Keyes Report at 25 (Doc. #: 1999-9). Accordingly, the Experts' opinions on causation are rooted in a reliable methodology sufficient to withstand Daubert scrutiny. Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 596 (1993) ("Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible evidence.").
Opp. at 11-13, n. 36 (Doc. #: 2197) (citing Edlund, et al., The Role of Opioid Prescription in Incident Opioid Abuse and Dependence Among Individuals with Chronic Noncancer Pain, Clin. J. Pain 30(7): 557-564 (2014)).
Lankenau, supra, n. 6; Cicero, supra, n. 7. --------
Plaintiffs' essential theory of their case is as follows: as a result of Defendants' aggressive marketing, medicine cabinets around the country and in Cuyahoga and Summit Counties contained excessive prescription opioids. People became addicted to them, either pursuant to medically prescribed usage or otherwise - it matters not which. When those cabinets ran bare, the addicted individuals went looking for other substances to feed their addiction. When they were unable to obtain these other substances from medical professionals, they found cheaper, more readily available, illicit substitutes.
This trajectory - from addiction to prescription opioids to addiction to illicit opioids - finds support in the NASEM report, the Lankenau and Cicero studies, and Keyes' and Lembke's first-hand clinical experience. Accordingly, Defendants Daubert challenge to the Gateway Hypothesis based on unreliability fails. Defendants may explain their critiques of the statistical methodologies the Experts rely on to the jury.
C. Relevance and the Danger of Misleading the Jury
Defendants' final argument repackages the preceding two arguments and presents them in terms of the risk of misleading the jury: "When Plaintiffs are seeking to recover damages that arise from persons who allegedly became addicted to opioids prescribed by their doctors, it is fundamentally misleading to rely on evidence about persons who are not known ever to have received a doctor's prescription for opioids and who instead abused these medications." Motion at 17 (Doc. #: 1858).
The Court concludes it would exceed the bounds of its gatekeeping role under Daubert to exclude the Experts' testimony, which is based upon both epidemiological studies and the Experts' own published literature and clinical experience. Under Daubert, a district judge should allow the parties to rely upon "the capabilities of the jury and of the adversary system generally," rather than "wholesale exclusion" of fairly supported, relevant testimony by the Court. In re Welding Fume Product Liab. Litig., 2005 WL 1868046, at *5 (quoting Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 596 (1993). A jury is capable of understanding and evaluating the salience of the distinction between medical and nonmedical users, and of assessing all the competing experts' opinions regarding correlation and causation. A court should not exclude otherwise admissible testimony "for fear that they will not grasp its complexities or satisfactorily weigh its inadequacies." Id. (quoting Ruiz-Troche v. Pepsi Cola Bottling Co., 161 F.3d 75, 85 (1st Cir. 1998). The Experts' opinions are not without vulnerabilities, but they are neither methodologically illegitimate nor fundamentally misleading. The Defendants can present countervailing evidence and otherwise undermine these Experts through the adversarial give and take of a trial.
III. Conclusion
For the foregoing reasons, the Motion is DENIED.
IT IS SO ORDERED.
/s/ Dan Aaron Polster August 26 , 2019
DAN AARON POLSTER
UNITED STATES DISTRICT JUDGE