Photo of Bexis

Today’s guest post is by the Reed Smith team of Shana E. Russo, Jennifer A. Eppensteiner, and Kathy I. Oviedo.  It is about multi-district litigation (“MDL”) practice, and specifically, compares and contrasts various ways of selecting plaintiffs for possible bellwether trials.  It also useful in providing links to orders entered in a number of MDLs concerning this selection process.  As always, our guest posters are 100% responsible for what they wrote, deserving all of the credit and (any) blame.


Bellwether Trials in Multidistrict Litigations

As this blog has pointed out repeatedly, the bellwether method in mass tort litigation is fraught with problems.  Still, it is the system we have; and it is likely that for now, the bellwether method, in some form, is what we will have to deal with for the foreseeable future.  When it comes time to negotiate bellwether discovery and trial pool selection, knowing what has worked in other MDL courts may help.  In an attempt to save readers from the dockets, this post identifies recent case management orders (“CMOs”) dealing with bellwether selection, which can be used as a guide as you design your discovery and trial plans.

Judge Eldon E. Fallon from the United States District Court for the Eastern District of Louisiana, has categorized three bellwether selection methods:  (1) random selection; (2) judicial selection; or (3) party selection.  Hon. Eldon E. Fallon, et. al., Bellwether Trials in Multidistrict Litigation, 82 Tul. L. Rev. 2323, 2343 (June 2008).  Most CMOs reflect a hybrid of these methods; examples below.

Random Selection

This is the only statistically valid method of truly representative selection.  When using random selection, individual cases can be selected either from the entire universe of the MDL or from within smaller subsets identified as reflecting certain variables.  Judge Fallon suggests that random selection “can be problematic” because “there is no guarantee that the cases selected to fill the trial-selection pool will adequately represent the major variables”; however, that is only because the vast majority of plaintiffs in MDLs have non-triable, worthless cases.  As confirmed by empirical data, “a random selection process can produce a sample of cases that are more like many other cases in the docket.”  Loren H. Brown, et. al., Bellwether Trial Selection in Multi-District Litigation: Empirical Evidence in Favor of Random Selection, Akron L. Rev.: Vol. 47: Iss. 3, Article 2, at 665 (2014).  Garbage in; garbage out.  Random selection should not be dismissed as a viable selection method, and can be used effectively on its own or as part of a hybrid system.

The court in the Abilify MDL used random selection to fill its bellwether pool.  With the help of a litigation management services firm, the court selected one-hundred cases from the pool of completed Plaintiff Profile Forms to form its second discovery pool.  See Order Regarding Selection of the Second Group of Potential Trial Cases, In re: Abilify (Aripiprazole) Prod. Liab. Litig., Case No. 3:16-md-02734 (N.D. Fla. June 22, 2018) (Dkt. at 906).  The court and the litigation management services firm then narrowed down the pool by “identify[ing] all Plaintiffs who checked the box for (a) either Bipolar Disorder or Major Depressive Disorder/Depression as the diagnosis leading to their Abilify prescription . . ., and (b) ‘Gambling’ as an obsessive/compulsive/impulsive behavior they claim was caused by Abilify. . . .”  Id.  From that pool, the court randomly selected forty cases to proceed to pretrial discovery.  Id.  Each side was then given five strikes to remove specific cases from the pool, reducing the pool to thirty cases.  Id. at p. 2.  Those thirty cases then proceeded to case-specific discovery limited to medical, financial, and gambling records.  Id.  Each side was then allowed to strike an additional five cases, reducing the pool to twenty cases, which proceeded to full fact discovery.  Id. at p. 3.  With fact discovery completed, the pool was further reduced by each side striking an additional five cases, leaving a total of ten cases to comprise the trial pool and proceed to expert discovery.  Id.  Perhaps not surprisingly, there was global settlement of the matters before any case from the second discovery pool proceeded to trial.  See Joint Notice of Proposed Settlement Program, In re: Abilify, Case No. 3:16-md-02734 (N.D. Fla. Feb. 15, 2019) (Dkt. at 1125).  Random selection may also be as simple as picking names out of a hat.  See Order re: Bellwether Trial Selection at 2, In re Prempro Prods. Liab. Litig., MDL No. 1507 (E.D. Ark. June 20, 2005).

Other recent examples of random selection include:

  • In re: Johnson & Johnson Talcum Powder Products Marketing, Sales Practices and Prod. Liab. Litig., Case No. 3:16-md-02738 (D.N.J. May 15, 2020) (at 13317); see also May 26, 2020 Order (Dkt. at 13428).
  • Order No. 19, In re: Zimmer M/L Taper Hip Prosthesis or M/l Taper Hip Prosthesis with Kinectiv Technology and Versys Femoral Head Prod. Lib. Litig., Case No. 1:18-md-02859 (S.D.N.Y. April 1, 2019) (Dkt. at 102).
  • In re Benicar (Olmesartan) Prods. Liab. Litig., Case No. 15-md-2606, CMO No. 15 (D.N.J. Nov. 23, 2015) (Dkt. at 193).

Judicial Selection

 Judge Fallon explains that although “[b]eing an unbiased neutral, the transferee court’s selections are likely to be more focused on cases that are truly representative of the litigation,” this option should be avoided as the “transferee court simply does not have the resources available, or the familiarity with each individual case, to conduct this task adequately.”  Fallon, at 2349.  However, there are plenty of resources available, including court-appointed experts, if necessary, to those MDL judges willing to use them to select bellwether candidates.  Recent litigation provides an example of how the court can guide bellwether selection while making the process less burdensome.  Take, for example, In re: Taxotere (Docetaxel) Products Liability Litigation, where the parties submitted a list of ten cases to the court, where the court had original jurisdiction and venue (that is, no Lexecon issues).  See Case Management Order No. 3, In re: Taxotere (Docetaxel) Prod. Liab. Litig., Case No. 2:16-md-02740 (E.D. La. July 21, 2017) (Dkt. at 669).  These ten cases then proceeded to the first phase of discovery.  Id.  The parties then “nominate[d] to the Court and rank[ed] in order of preference the four cases that [would] proceed to the second phase of discovery,” effectively giving each side one strike.  Id.  The Court then selected a “Primary Plaintiff” for the first trial date, and the court ranked the remaining three Plaintiffs second, third and fourth as alternates in the event that the “Primary Plaintiff” was dismissed.  Id.  Judicial selection may also be just a small part of the plan.  For example, in In re: Davol, Inc./C.R. Bard, Inc., Polypropylene Hernia Mesh Prod. Liab. Litig., the parties picked the discovery and then trial pool cases.  The parties the proposed the trial pool cases, and the court determined the order of the trials, alternating between defendant’s proposed cases and plaintiff’s.  See Case Management Order No. 10, In re: Davol, Inc./C.R. Bard, Inc., Polypropylene Hernia Mesh Prod. Liab. Litig., Case No. 2:18-md-02846 (S.D. Ohio November 20, 2018) (Dkt. at 62).

Party Selection

Judge Fallon finds party selection preferable because “[t]he attorneys are in the best position to know, or ascertain, the true census of the litigation,” and “[i]n addition, they have the most staff resources available.”  Fallon, at 2349.  We disagree.  Judge Fallon may well be right that party selection allows the parties to feel more invested in the selection process, which might make them more likely to extrapolate results to a wider set, particularly if they chose the case that either wins or loses.  But, on the plaintiffs’ side, that buy-in happens only because they get to hide their vast majority of weak to nonexistent cases from the adversary process.  Party selection comes at the expense of representativeness, as both sides will seek advantage by selecting cases favorable to them.  Thus, when the case tried was chosen by the opposing side, it is easier to dismiss the case as an outlier and be hesitant to rely on it for settlement purposes.  Experience dictates that the plaintiffs habitually abuse the party selection process, since they have the power to dismiss unfavorable cases, often times late in the game, without sanction.

If allowing a party-driven bellwether selection plan, an MDL court needs to deal with potential issues, particularly biased selection, proactively.  Even when relying on party selection, the court retains the power to direct the selection process and can influence how certain injury groups are represented.  In the Invokana (Canagliflozin) Products Liability Litigation, for Group A cases, each party identified six plaintiffs who actually used the drug (many MDL plaintiffs cannot even prove exposure), three of which who then developed each of the two major claimed injuries (another pro-plaintiff thumb on the scale) to undergo “Bellwether Core Discovery.”  See Case Management Order No. 20, Invokana (Canagliflozin) Prod. Liab. Litig., Case No. 3:16-md-02750 (D.N.J. July 27, 2017) (Dkt. at 218).  Thereafter, plaintiff and defense designees exchanged their lists.  Id. at p. 2.  The CMO “strongly encouraged” the parties “to select cases that they have a good faith belief are representative cases that should be robustly discovered and then taken to trial.”  Id.  Following discovery, the court directed the parties to narrow their selections to four cases, two from each injury category, subject to further discovery and bellwether trials.  Id. at p. 5.  The court then selected three bellwether cases to serve as the first bellwether trials.  Id.  Notably, in order for the court to alleviate any Lexecon waiver issues, it directed the parties to “waive applicable venue and forum non conveniens challenges and stipulate that the trial of any of the final . . . bellwether cases . . . [could] be conducted in the District of New Jersey without remanding any case to the transferor forum under Lexecon v. Milberg Weiss. . .”  Id. at p. 7.

Other recent examples of party selection include:

 Identifying the Appropriate Selection Method

The MDL bellwether system is notorious for creating six-week show trials designed to “ring the bell.”  Even a case that starts out “representative” loses that distinction when litigated with way more time and effort than any possible intrinsic value it might have.  However, it is what it is.  Understanding the flaws in the system will allow counsel to more thoughtfully approach future CMO negotiations.  As recognized by Judge Fallon, “the sheer number and type of feasible trial-selection processes are limited only by the ingenuity of each transferee court and the coordinating attorneys.”  Fallon, at 2343.  Courts may assert their preferred methodology on the parties without giving the parties an opportunity to negotiate a deal.  However, if that opportunity arises be sure to protect against potential abuses in the system.  For example, consider that when deciding the process to select cases for trial, striking cases rather than selecting them may be the best approach, as it removes the other side’s best cases, and plaintiffs in MDLs tend to have far more bad cases than good ones.