Aredia/Zometa Motions in Limine

One of the remand courts in the Aredia/Zometa litigation recently issued a decision on various motions in limine.  See Brown v. Novartis Pharmaceuticals Corp., 2012 U.S. Dist. LEXIS 104985 (E.D.N.C. July 27, 2012).  Here’s the rundown of the “good,” the “not-so-good,” and the “it depends.”

The good:

·                     Post-injury label changes: The court excluded evidence of the defendant’s changes to the label of Zometa in 2007 and Aredia in 2008.  Id. at *38-40. Plaintiff’s injury was diagnosed in 2005, and the court held that these years-later label changes qualified as subsequent remedial measures. Of course they do.  This has long been the view of the majority of circuit courts, and getting this type of ruling has become easier with amended FRE 407, which now reads in part:  "When measures are taken that would have made an earlier injury or harm less likely to occur, evidence of the subsequent measures is not admissible to prove . . . . . a need for a warning or instruction."

·                     Emails by members of an Advisory Board to the defendant:  The court excluded these as hearsay.  Id. at *34-37.  Defendant appears to have had an advisory board of outside doctors who gave the defendant advice on its draft of a white paper.  Plaintiffs sought to admit emails from two of the advisors, arguing that the emails weren’t hearsay because the doctors were somehow agents of the defendant.  But agreeing to participate in an advisory board is far different from being an agent.  There was no evidence that the defendant controlled the two outside doctors or that the doctors acted on the defendant’s behalf.  They just gave advice.  Plaintiffs also argued that the emails were business records, but the court saw no evidence that serving on these advisory boards was regularly conducted business activity or that, even if it were, the emails were a regular part of that business activity. 

The not-so-good:

·                     Adverse Event Reports (AEs):  The court allowed plaintiffs to use AEs as evidence of causation and for notice.  Id. at *25-30.  We’ve blogged many times – because it’s right – that AEs should not be used as evidence of causation.  They’re hearsay, anecdotal, incomplete, and unreliable.  Unfortunately, this court will allow plaintiffs to do it here, so long as a plaintiffs’ expert testifies that he relied upon AEs to form a causation opinion.  We don’t think that finding an expert to offer such an opinion changes the deficiencies of AEs.  For those interested, we’ve compiled a cheat sheet of AE decisions that we believe got it right.  Additionally, by the way, the court allowed plaintiffs to use AEs for notice to the defendant, a much less controversial ruling. 

·                     Failure to warn non-prescribers.  The court recognized that under North Carolina law a defendant is not liable in a failure to warn case if the defendant gave an adequate warning to the plaintiff’s prescriber.  Id. at *33.  That’s good.  But plaintiffs also wanted to show evidence of a failure to warn non-prescribers.  That seems wholly irrelevant.  The prescribers were either warned or not.  Now, while the court did not issue a final ruling, it did hold open the possibility of allowing non-prescriber failure-to-warn evidence if plaintiffs showed that the actual prescribers were not warned.  Id. at *33-34.  This seems like an invitation to mischief.  Failure to warn does not occur in the air.  It must involve the relevant learned intermediaries – the prescribers. 

The "it depends":

·                     Choice-of-law on punitive damages.  It depends because all lawyers approach this issue by first determining which state’s law they would prefer applied.  Here, New Jersey law, the home of the manufacturer, allows no punitive damages because the FDA approved the drugs, while North Carolina law, where the plaintiff purchased, used and allegedly was injured by the drug, would allow punitive damages.  The court applied Restatement (Second) of Conflict of Law and chose New Jersey.  In short, the court held that, while North Carolina law properly applied to the underlying claim because it was aimed at protecting the plaintiff, New Jersey law should apply to the punitive claim because it was aimed at deterring any wrongful conduct of the defendant, a New Jersey company.  Id. at *10-24.  No matter how you look at this, though, it was a nice victory for the defendant. 

Through the Past, Darkly

We’ve been reading the Keith Richards autobiography, Life, and it made us think about the past. The Rolling Stones were formed 50 years ago, in April 1962. We were Beatles fans in the 1960's. (Still are). When the Rolling Stones showed up on The Ed Sullivan Show they seemed sort of terrifying. (Still are.) Time is no longer on their side. Or ours.

The Stones have been around long enough that their songs constitute the soundtrack of the Babyboomers. Maybe that is why Martin Scorsese repeatedly uses Stones tunes in his films. "Jumping Jack Flash" was recorded in 1968, an annus horribilis (Tet offensive, assassinations, Nixon election) if ever there was one. But the song that best captures the feel of 1968 -- "Gimme Shelter" -- came out the following year. In 1971, we pilfered Sticky Fingers (an album with the scandalous zipper -- try doing that with a cd or an iTune) from our sister's room. The 1980's was not a terrific decade for the Stones, though Steel Wheels at the end was a redemption.

We were listening to Voodoo Lounge while preparing for a Ninth Circuit oral argument. The issue was whether the district court erred in admitting evidence of prior bad acts. The panel was less than happy with our prosecutorial zeal back then. ‘Mr. AUSA, what percentage of cases does your office win? More than 95%, right? So why did you need to gild the lily and parade in front of the jury another heroin deal that took place ten years earlier?’ Well, gulp, because the defendant was guilty as Hell, always followed the same playbook, probably played a role in murdering a federal agent, and needed to go away with an LWOP (life without parole). Why the sympathy for the devil? Further, Federal Rule of Evidence 404(b) permits introduction of “crimes wrongs, or other acts” to prove “motive, opportunity, intent, preparation, plan, knowledge, identity, absence of mistake, or lack of accident.” One of the judges said something like, ‘The past is not always prologue.” Yeah, but sometimes it is. Anyway, evidence of prior acts is hot stuff. Appellate courts play close attention. It can clinch a conviction or it can be the biggest mistake in a trial. For a prosecutor to introduce prior acts is to play with fire.

Nowadays we usually argue against introduction of prior acts. Does that make us hypocritical? We (naturally) demur. Most of the time, when plaintiffs seek to drag in evidence of prior injuries, accidents, etc., there is little purpose beyond prejudice. The prior acts are not really similar. Or it is not even clear that they happened at all, or that they happened the way the plaintiffs want to argue they happened. Plus, a product liability plaintiff does not have to prove liability of a corporate defendant beyond a reasonable doubt. Tell you what: if plaintiff lawyers shoulder the reasonable doubt burden, then we defense hacks won’t file so many in limine motions seeking to preclude other acts evidence. Okay? No? We didn’t think so.

A court recently did a pretty careful job of going through the other acts analysis in the product liability case of Hershberger v. Ethicon Endo Surgery, Inc., 2012 U.S. Dist. Lexis 45220 (S.D. W.Va. March 30, 2012). Some of the analysis is sound. But we view the overall outcome with mixed emotions. The defendants moved to prevent introduction of investigative reports related to Ethicon stapler performance, as well as allegations of other incidents concerning stapler performance and litigation related to Ethicon staplers. The defendants’ motion was premised on arguments that that the evidence was: (1) irrelevant under Rules 401 and 402 of the Federal Rules of Evidence and the substantial similarity test; (2) far outweighed by undue prejudice under Rule 403; and (3) inadmissible hearsay.

Rule 404 was not discussed by the Hershberger court. Now discuss among yourselves why that is so.

The plaintiff argued that the prior incidents, which came from the defendants’ adverse event files, were relevant. The plaintiff proffered a summary of forty-five incidents and offered to produce to the Court the investigative files in full. Then, following the usual path, the defendants produced a chart of their own in reply to the plaintiff’s chart, setting forth the “inherent differences” between the allegations in this case and the forty-five incident files. Hershberger, 2012 U.S. Dist. LEXIS 45220 at *3. As is the case with adverse event reports, all the information was “necessarily derived second-or third-hand from medical personnel reporting an incident directly to Ethicon or through regional sales representatives.” Id. The main issue is whether the prior events were sufficiently similar so as to prove negligence, design defect, notice of defect, or causation.

The court reasoned that substantial similarity requires a showing that “(1) the products are similar; (2) the alleged defect is similar; (3) causation related to the defect in the other incidents; and (4) exclusion of all reasonable secondary explanations for the cause of the other incidents.” Id. at *5. Crucially, “nowhere in the incident files is there any indication that reasonable secondary explanations for the other incidents were eliminated. The most obvious reasonable secondary explanation, which to the Court’s knowledge was not ruled out in any of the forty-five other incidents, is surgical team error. Put simply, the forty-five other incidents are not clearly the result of a faulty Ethicon stapler, and thus, they are not substantially similar so as to permit their admission as direct evidence of negligence or a defect.” Id. at *7. So far so good. But “the threshold for admissibility is significantly lower if the other incidents are used to prove Ethicon’s notice or knowledge of a product issue.” Id. The court then came up with different rulings for different claims:

• Negligence – The similar incidents evidence was relevant to “show that Ethicon had notice or knowledge or other staplers that reportedly malfunctioned, and Ethicon was therefore better positioned to foresee the injury that is the subject of this litigation.” Id. at *9. Yet while the similar acts are relevant to notice, they cannot constitute “direct proof negligence in this case-that is, arguing that Ethicon was negligent in marketing defective staplers in the past and it therefore was negligent in this case-is strictly prohibited in light of the Court’s above ruling regarding the (dis)similarity of the incidents.” Id.
• Product defect – “[E]vidence of similar incidents is wholly inadmissible pursuant to the Court’s similarity finding. In a strict liability cause of action, Ethicon’s knowledge or notice of a potential defect in its product is irrelevant.” Id.
• Breach of warranty – “The Court perceives no relevant use of the other incident evidence to Plaintiff’s breach of warranty claim. Presenting other incidents evidence to demonstrate that Ethicon’s stapler was not fit for its intended foreseeable uses is impermissible under the Court’s similarity ruling.” Id. at *10.
• Punitive damages – “Such evidence is directly relevant to the jury’s determination of whether Ethicon acted in a wanton, willful, or reckless manner.” Id. [Ouch.]

Thus, the court deemed the evidence of prior adverse events relevant only for notice/negligence and punitive damages. But still remaining is whether, under Rule 403, the probative value of the prior acts is substantially outweighed by the danger of prejudice or jury confusion. Once again, the court applied that analysis differently to the different claims. As to the negligence claim, the court found “that introduction of other similar incidents is likely to waste time and confuse the jury. The applicability of other incidents to the negligence claim is somewhat obscure, and there is significant danger that the jury will consider the other incidents as direct evidence of negligence. Further, the probative value of the other events is minor-these records amount to uncorroborated and incomplete customer complaints constituting approximately 0.003% of Ethicon’s sales during the relevant time period.” Id. at *11.

No surprise, but that ruling gives us satisfaction. We’ve argued in the past that it is a very unfair and dangerous thing to permit plaintiffs to argue, and juries to deduce, that adverse events show negligence or product defects. For one thing, adverse events are not probative of causation. For another, in more and more litigation the adverse event process becomes tainted by litigation. Lawyer ads (that ain’t too proud to beg) can inflate adverse event reporting. Lawsuits themselves become adverse events. Some people complain about data-mining. Data-planting is just as bad. Accordingly, to the extent the Hershberger court rules that the adverse events are irrelevant or prejudicial, it's all right by us.

But you can't always get what you want. Are the good bits in Hershberger shattered by the ruling on punitive damages? The court held that “[a]s the punitive damages claim however, the risk of confusion is diminished-Ethicon’s knowledge and notice of other product complaints is directly relevant to punitive damages.” Id. at *12. (By the way, since the adverse events are admitted for knowledge and notice, the court holds that the hearsay objection fails.) That ruling leaves us scratching our heads. If the adverse events are unduly prejudicial with respect to the negligence claim, they are probably also prejudicial on punitive damages. We do not know whether or how the punitive damages issue would be staged at trial, but if the adverse events come in during the main liability phase, and if the plaintiff can argue that knowledge of such adverse events shows that the company was wanton or reckless, isn’t the prejudice nearly complete? Would a jury instruction really solve the problem? Would a defendant even ask for such a limiting instruction? Moreover, “to the extent Plaintiff may wish to cross-examine a defense witness regarding the general subject matter of the other incidents, i.e. the existence of reports of faulty staplers, such use may well be permissible.” Id. at *13. So not only will the adverse events come in, they might come in at a particularly dramatic time, and it might look like the defense was trying to hide them.

To be sure, we’re not trying to take a mostly good ruling and paint it black. But we wonder how this ruling will play out at trial. Predicting the future is harder than replaying a past that will not fade away.

More Reasons to Discount Adverse Event Reports

            A topic of repeated interest for us is how courts treat evidence of adverse event reports.  And we haven’t been shy about how we feel -- AER's should not be used in a civil trial.  AER's are voluntary, biased (sometimes litigation-generated), and should not be used to calculate incidences of drug risk. AER's are not the stuff of good science.  We've even put together an Adverse Event Report Cheat Sheet  filled with cases laying out compelling reasons for excluding AER's.  

            So, why are we re-hashing this already well trod ground?  Well because in railing against the use of AERs in the courtroom, we’ve often distinguished the FDA’s reliance on AERs on the basis that the FDA can act in the public interest on a lower quantum of evidence than is necessary to establish causation in a products liability suit.  A recent report by the FDA, however, has left us wondering just how helpful AERs are to even the FDA.  The purpose of the report, entitled Understanding Barriers to Medical Device Quality was “to assess and understand gaps in medical device quality.”  FDA Report at 3.  While the FDA didn’t rely exclusively on AERs, they were a driving factor in the FDA’s apparent conclusion that medical device quality has significantly lagged behind medical device technology.  While we will leave a more thorough discussion of the four corners of the report to our friends at the FDA Law Blog, we thought it worthwhile to point out how even the FDA’s use of voluntary AERs can lead to false conclusions.

 

For starters, the report states that AERs for medical devices are on the rise.  The FDA reports that “serious adverse event reports related to medical device use have outpaced industry growth by 8% per annum since 2001.”  FDA Report at 3.  The report then suggests that the increase in volume of AERs must mean a decrease in the quality of medical devices.  Huh?  Even the FDA had to admit that looking at the quantity of AERs can be misleading:

  

Several factors may contribute to the growth in volume of adverse event reports.  These include greater outreach by FDA emphasizing reporting requirements, along with greater manufacturer sensitivity to reporting requirements following notable recalls.  Some of this growth may also be due to growth in the number of medical devices in use.

FDA Report at 12.  That last one seems really obvious – if more people are using more devices, there are going to be more AERs.  But does that tell us anything about safety?  No.  So, it is not surprising that the FDA itself recommends “adjusting absolute numbers of adverse events and recalls for device usage, or the number of devices on the market.”  FDA Report at 38.  If the FDA has yet to make that necessary adjustment, then it hasn’t answered the real question – whether the number of AERs as a percentage of total device usage has increased?  And, if the FDA is reaching misguided conclusions based on misleading AER evidence, how can anybody suggest that juries won’t be similarly misled?  

            The other growth factors cited by the FDA – greater FDA outreach and heightened sensitivity by manufacturers – demonstrate another problem we have with using AERs to prove anything about causation in drug/device product liability litigation.  AERs are voluntary.  So, as the FDA Report suggests, manufacturers err on the side of over-reporting to avoid FDA enforcement actions.  So too do users over-report (by phone calls to manufacturer consumer service lines or by filing lawsuits) in the face of chiefly adverse publicity concerning this or that device/drug.   Where a manufacturer has been stung by a Warning Letter or the public has been inundated with ads about a product recall or a public health alert – AERs go up.  The voluntary reporting bias is huge and it is another variable that should be controlled for if AERs are going to be used in safety analyses.  Also, it contradicts plaintiffs’ experts’ theory that AERs are under-reported and therefore, if anything skewed in favor of defendants. 

            The FDA Report also states that AERs are not evenly distributed across all medical devices.  Rather, “cardiovascular, in vitro diagnostic (IVD), and general hospital/surgical devices account for nearly 60% of adverse event reports.”  FDA Report at 3.  Really?  Is anyone surprised that cardiovascular devices “have increased as a share of total serious adverse event reports even faster than for adverse events overall.”  FDA Report at 15.  Could it because the cardiovascular patients are generally older and less healthy than the population as a whole and therefore more likely to experience an adverse event?  Remember, AERs are generated without regard for cause.  So, can we draw any meaningful conclusion about the safety of cardiovascular devices without accounting for the unique patient population?  No.  Again, the raw numbers are misleading.

            The FDA purports to “possess a wealth of data pertaining to medical device quality. . . [that] may be used to quantify the magnitude of medical device quality problems and to better understand the root causes of these problems.”  FDA Report at 38.  But what they really have is a stack of voluntary, selective, biased reports that a device may have caused or contributed to an adverse event.  So, perhaps the most important takeaway from this report for those of us in the drug/device litigation arena is the FDA’s almost-admission that more is needed for AERs to be meaningful:

Bolstering the data that FDA currently collects with a few key additional pieces could greatly increase its utility and provide a fuller view of medical device quality.  Gathering additional risk data from companies will likely paint a clearer picture of the associated level of risks for devices.

FDA Report at 38.  So, until the additional data is collected, the current picture remains murky at best and murky isn’t good enough in the courtroom.

Securities Suits Can Be Brought On Non-Statistically Significant Complaints

So says the unanimous Supreme Court today in Matrixx Initiatives, Inc. v, Siracuso, 09-1156, slip op. (U.S. March, 22, 2011).  We're not securities lawyers, so we're primarily concerned with how Matrixx could affect product liability litigation.

We don't think it will have all that much, since the court makes pretty sure that "materiality" for securities litigation does not mean proof of medical causation that would stand up in a court of law.  Instead it basically shies away from a "bright line" test.  Slip op. at 11-12.

We've discussed earlier how the government (especially FDA) can, and does, act on data that could not establish causation in a court of law.   The Supreme Court's decision also makes that clear:

Not only does the FDA rely on a wide range of evidence of causation, it sometimes acts on the basis of evidence that suggests, but does not prove, causation.  For example, the FDA requires manufacturers of over-the-counter drugs to revise their labeling “to include a warning as soon as there is reasonable evidence of an association of a serious hazard with a drug; a causal relationship need not have been proved.”  21 CFR §201.80(e).  More generally, the FDA may make regulatory decisions against drugs based on postmarketing evidence that gives rise to only a suspicion of causation.

Slip op. at 13-14 (citation omitted).  Indeed, specifically with respect to the civil "more likely than not" civil liability standard, the court quotes the Manual on Scientific Evidence - “[R]isk assessors may pay heed to any evidence that points to a need for caution, rather than assess the likelihood that a causal relationship in a specific case is more likely than not.”  Slip op. at 14 n.9.

Frankly, we'd rather the Court not have cited expert witness cases allowing "testimony on causation based on evidence other than statistical significance," slip op. at 12, especially the pre-Daubert abomination, Wells v. Ortho Pharmaceutical Corp., 788 F.2d 741 (11th Cir. 1986).  But the Court did go out of its way to state that "We need not consider whether the expert testimony was properly admitted in those cases, and we do not attempt to define here whatconstitutes reliable evidence of causation."  Slip op. at 12. But since we've never been able to have Daubert construed as limiting experts solely to statistically significant epidemiological proof (as opposed to the infamous "differential diagnosis"), we'd actually have been jealous if the securities defense bar had been able to pull off that result.

Basically, Matrixx comes down on the side of a case-by-case assessment of whether statistically significant adverse events are considered "material" in any particular case.  The bottom line:

Application of [the prevailing securities law] standard does not mean that pharmaceutical manufacturers must disclose all reports of adverse events.  Adverse event reports are daily events in the pharmaceutical industry. . . . The fact that a user of a drug has suffered an adverse event, standing alone, does not mean that the drug caused that event. The question remains whether a reasonable investor would have viewed the nondisclosed information as having significantly altered the "total mix" of information made available.  For the reasons just stated, the mere existence of reports of adverse events - which says nothing in and of itself about whether the drug is causing the adverse events - will not satisfy this standard.  Something more is needed, but that something more is not limited to statistical significance and can come from “the source, content,and context of the reports.

Slip op. at 15-16 (various citations omitted) (our bold emphasis added, other emphasis original).

Sure, we'd rather the Court have bought into a bright line statistical significance standard - because we'd have tried to import it into Daubert - but after the oral argument, we didn't see that as in the cards, and it definitely wasn't.

One final observation:  anybody who claims that the current court is "pro-business" is just spouting plaintiff-side propaganda.

A Grab Bag Of Daubert Rulings In A Remanded Aredia/Zometa Case

We have reported periodically on rulings in the Aredia/Zometa litigation, including Daubert and summary judgment rulings by the MDL judge and choice of law rulings by Judge Arthur Spatt of the E.D.N.Y. in Deutsch v. Novartis Pharmaceuticals Corp., a case remanded for trial from the MDL. Judge Spatt has now decided Daubert motions in Deutsch and another remanded case, and his very long opinion is a bit like yesterday’s list of silly law songs : there is something in there for everyone. Deutsch v. Novartis Pharmaceuticals Corp., 2011 U.S. Dist. LEXIS 22755 (E.D.N.Y. March 8, 2011).

MDL rulings are law of the case. The MDL judge ruled on Daubert challenges to several plaintiffs’ experts. Novartis challenged those experts based both on arguments made to the MDL judge and on new arguments. Judge Spatt generally refused to consider the arguments previously decided by the MDL judge based on the law of the case doctrine, finding the doctrine particularly applicable when cases are transferred from an MDL. 2011 U.S. Dist. LEXIS 22755 at *18-21. “Reversing or otherwise undermining the decisions by the MDL court could lead to the type of inconsistent pretrial rulings that Congress sought to avoid [in the Multi-District Litigation Act], and therefore frustrate the very purpose of consolidation.” Id. at *21.

As a general matter, we agree that issues decided by the MDL judge should not be relitigated on remand. In Deutsch, however, the MDL rulings became the third rail, and anyone raising an argument that came within a few yards of an issue decided by the MDL court got zapped.

Reliability of opinions based on non-controlled epidemiologic studies. Novartis challenged many expert opinions because they were based only on epidemiologic studies rather than on randomized controlled clinical trials or studies. But the MDL court had rejected this argument, and Judge Spatt found that binding under the law of the case doctrine. Id. at *22. The Deutsch court, however, then went on to address this argument and agreed with the MDL judge, although it is hard to tell how much this decision was tilted by deference to the MDL court. Id. at *23-35. Score one for their side.

Challenges to opinions of experts ruled qualified in the MDL. After the MDL judge rejected Novartis’s Daubert challenges to certain plaintiffs’ experts, Novartis tried challenging those experts’ opinions that were not considered by the MDL judge. Again, Novartis hit the third rail. Because the MDL court found them qualified in certain areas, the Deutsch court allowed the experts to testify about additional scientific topics that were related to their areas of expertise. See, e.g., id. at *50-51. Score one for their side.

Opinions about Novartis’s corporate conduct. Plaintiffs wanted their medical and scientific experts to testify about Novartis’s corporate conduct. Judge Spatt ruled that the experts could explain Novartis internal documents, such as whether they indicated certain risks, but could not explicitly or implicitly testify about the intent, motivations, or state of mind of Novartis or its employees or whether Novartis acted in bad faith. Id. at *60-64, 77-78. Score one for our side.

Opinions based on a doctor’s personal experience. The court barred one of the plaintiffs’ experts from opining that, based only on his experience with his own patients, a reduced dosing regimen led to a lower incidence rate of ONJ. He had no data to support the reliability of his opinion, nor had he followed a particular methodology, and his personal observations were insufficient to support the conclusion he reached. Id. at *74. Score one for our side.

Adverse event reports. The MDL judge allowed an expert to testify about general causation, despite Novartis’s objection that his opinion was based on adverse event reports. Unfortunately, that was considered law of the case, which made adverse event reports another third rail. Novartis tried to challenge experts’ reliance on adverse events reports, but the Deutsch court zapped those arguments. See, e.g., id. at *85. We have previously explained that adverse event reports are not reliable evidence, and the court’s ruling is simply wrong. The court tried to soften its ruling by noting that the expert did not “cite un-reviewed [adverse event] reports for their truth, but rather for the accepted theory that a vast number of reports contributes to the plausibility of causation.” Id. at *85. Hogwash; even a vast number of unreliable reports do not amount to anything reliable. And we all know that this distinction will disappear when these opinions are trotted out to the jury. Score one for their side.

Meta-analysis. The court did a thorough analysis of one expert’s testimony about his meta-analysis of epidemiologic studies and allowed some opinions and barred others. Id. at 87-113. In doing this analysis, the court was mindful of the “strong risk of prejudice if a Court permits testimony based on an unreliable meta-analysis because of the propensity for juries to latch on to the single [risk] number” resulting from the meta-analysis. Id. at *106. Tie.

Opinions expressed not included in expert report but stated at expert’s deposition. One well-known danger of deposing an expert is that he or she will offer opinions beyond those stated in the expert report and then will be allowed to testify about them at trial. That happened here – Novartis moved to exclude the opinions not stated in the report; and the court denied the motion, saying that “Novartis cannot claim surprise by any of Prof. Ray’s opinions that it elicited and questioned him on at length.” Id. at *113. Score one for their side.

Alleged “cover up” activity after plaintiffs were treated. Plaintiffs offered an expert to opine that Novartis meddled with the conclusions in a scientific article published in 2008. The court rejected this testimony as irrelevant because the plaintiffs were treated long before 2008 and the article could have not influenced their care. Id. at *118-19. Score one for our side.

Dr. Suzanne Parisian. There must be a page in the plaintiffs’ manual directing counsel to try in every single case to use Dr. Suzanne Parisian, called by this court a “seasoned veteran of product liability litigation.” Id. at *125. The court split the baby here, allowing her to testify about FDA regulatory requirements but barring her testimony about corporate state of mind, pharmaceutical industry standards, ghostwriting, causation, and diagnosis. Id. at *125-41.

Dr. Parisian was also offered to testify about the reasonableness of Novartis’s conduct in complying with FDA regulations and interacting with the FDA. We think allowing that sort of testimony is a big, big mistake. But the court was indulgent. The court found that her report discussed her experience and FDA regulations and subsequently offered conclusions without analysis connecting it all together. Although recognizing that “other courts have excluded Dr. Parisian’s testimony based on the disconnect between her analysis and opinions,” id. at *132 (citing cases), the court decided to have a Daubert hearing to determine whether she employed a reliable methodology. Id. at *136. We think a judge should not convene a hearing to bail out an expert who could not prepare a proper report. What’s worse, the hearing also will consider whether “Dr. Parisian can testify as to her opinions within the confines of certain limitations so as to avoid prejudice to Novartis.” Id. at *136-37. In other words, a hearing to decide whether, as we put it before, “the notoriously hard-to-control Dr. Parisian” will behave herself on the witness stand. That sounds crazy to us. We predict the witness will say whatever the court wants to hear at the Daubert hearing and whatever plaintiffs want the jury to hear at trial. More or less a tie, but anything Parisian is allowed to testify to is junk.

Causation testimony by treating physicians. The court excluded two of plaintiffs’ treating physicians’ testimony on specific causation because they were not experts in the disease or its causes and did not perform a proper differential diagnosis, as they did not analyze or rule out other causes. Id. at *152-58. Another, more qualified treater was allowed to opine about specific causation as long as he limited his opinion to his experience as the patient’s treating physician and his opinions based on that treatment. Id. at *163-64. Score one for us.

Fair enough. But the court really stumbled in its very last ruling on plaintiffs’ treaters, a Dr. Gelfman. This doctor testified that he performed a differential diagnosis but could not “remember what factors he considered at the time . . . or what specific literature he relied upon outside of various case reports.” Id. at *165. He also acknowledged being unaware of certain risk factors. Id. at *166. That should have been the end of Dr. Gelfman, as a plaintiff has the burden of establishing the reliability of her experts’ testimony, and plaintiff here could not establish that the forgetful Dr. Gelfman did a proper differential diagnosis. The court, however, let him testify, ruling that these defects could be explored on cross-examination. No, no, no, no. A proper foundation for admissibility is required before anyone ever gets to cross-examination, and all the cross in the world can’t make up for that. Score one for their side.

Final score, 4-5-2. It’s pretty certain this poor jury will be sifting through a lot of garbage.