This week, NEJM released a meta-analysis that, after omitting six studies where there were no cardiac events, found a barely statistically significant relative risk of 1.43 of myocardial infarction from long-term use of diabetes drug Avandia (rosiglitazone). (A slightly higher relative risk of death from cardiac causes was not statistically significant, though also just barely.) Reuters reports that plaintiffs’ lawyers are deciding whether they can create a mass tort litigation against GlaxoSmithKline, though some already confidently predict that they will (which suggests how important the supposed due diligence required by the rules and being conducted now is). Note that a relative risk of 1.43 means that 70% of people who suffered heart attacks while taking Avandia would have had heart attacks anyway. Avandia’s warning label already warned of the risk of cardiovascular adverse events. (Cross-posted from Point of Law)
Update: Interesting analysis at one of the medical blogs.
9 Comments
If the verdict is decided by the “preponderance of evidence”, then a 70% chance the heart attack was not caused by Avandia would seem to guarantee a defense victory. In practice, though, that is definitely not the case. I understand how juries may rely on sympathy or factors other than the statistics to make their decision, but I don’t understand how civil law applies. What legal grounds are used to justify a case where even the plaintiffs favored evidence does not show a preponderance of evidence?
Note: There are some assumptions here–maybe you can identify heart attacks caused by Avandia or the 70% figure is not right. Even if they’re not true, the broader question remains.
So the standard for all medicines is now to have no potentially fatal side effects, no matter how rare or how positive the primary effects?
Brilliant.
Hot diggity, Jackpot. I knew there had to be a winner somewhere in my pillbox.
The extra “risk” in Vioxx and now Avandia are barely significant statistically.
The rational inference is that these drugs do not have a medically significant risk. Such a risk only would be indicated by elevated risk factors of at least 3.
Life expectancy has been rising rapidly. Big Parma is very inept at killing us. Impoverishing us is another matter.
I should add, my question above is not rhetorical. Does anyone out there well versed in the law who knows how you would show a preponderance of evidence here? I would like to know.
DD,
Lawyers need EVIDENCE? New one on me…
Evidence makes cases a bit easier, but all a lawyer needs to win is the emotions of the jury.
Drew: You raise a point that I and several others have often raised, which is that failure to demonstrate a relative risk greater than 2.0 is grounds for summary dismissal. Plaintiffs often get around this by pointing to a single study where there was a higher relative risk (ignoring the ones where the risk was lower or nonexistent), and having an expert make a conclusory statement that a drug caused this particular adverse event. I’ve discussed this in more detail in the context of Vioxx litigation, where the relative risk for low-dose Vioxx is also below 2.0.
Thanks for the helpful answer Ted! The more I learn about the mechanics of civil law the stranger it seems.
I don’t think these numbers must, of necessity, foreclose liability under “preponderance of the evidence” standard in any single case because that standard requires a showing by preponderance, of causation with regard to the particular plaintiff, rather than a group at large (although the study of the group at large is certainly an impediment to a Daubert challenge). However, I cannot imagine what evidence in any particular case would overcome the lack of preponderance in the general sense in my own mind if I were a juror. And of course posting that last sentence on the web should ensure I will not become a juror in any of the eventual cases.