We reported on the Northern District of California’s order granting summary judgment in Messick v. Novartis Pharmaceutical Corp. when it came out in February 2013, and a good and proper order it was. The plaintiff was treated with bisphosphonate therapy for approximately two years, and she alleged that she developed osteonecrosis of the jaw (“ONJ”) more than a year after she stopped the therapy. The problem for the plaintiff was that she had no reliable scientific evidence linking her drug therapy to her alleged disease. Sure, she had an expert, but the expert admitted that pathology results upon which he relied were not scientifically reliable. He also admitted that his “differential diagnosis” failed to take into account numerous other risk factors, asserting that “it just doesn’t happen” that someone like the plaintiff would experience ONJ without contribution from bisphosphonates.
“It just doesn’t happen” is what we say when discussing whether the Chicago Cubs will win the World Series or whether we will ever fully appreciate the talents of Miley Cyrus. When it comes to giving causation opinions to a reasonable degree of medical certainty, you will forgive us for expecting more. The district court expected more too, and it found the opinion lacking in reliable methodology and also relevancy because the expert could not say that bisphosphonate exposure caused the plaintiff’s ONJ.
You can imagine then our disappointment with the Ninth Circuit’s recent opinion reversing the district court’s sound and well-reasoned order. Messick v. Novartis Pharmaceuticals Corp., No. 13-15433, 2014 U.S. App. LEXIS 6257 (9th Cir. Apr. 4, 2014). The Ninth Circuit began its analysis by stating that “[t]he relevancy bar is low,” and it faulted the district court for excluding the opinion because, even though the expert could not say that the plaintiff’s ONJ was caused by bisphosphonate therapy, he did say that the ONJ was related to bisphosphonate therapy. Id. at **6-7.
The Ninth Circuit was willing to brush aside the difference between association and causation, but we’re not. A disease may be associated with a certain drug, but it does not take an epidemiologist to understand that such an association does not demonstrate that the drug caused the disease. Perhaps the condition calling for the drug therapy is also a risk factor for the disease. Take for example a hypothetical obesity drug. You might see cardiovascular events in people who take that drug, but that does not mean the drug caused any patient’s heart attack because obesity is itself a major risk factor. We have no idea how bisphosphonates are “related” to ONJ (if at all), but we do know that that an expert who cannot say that the drug caused the disease is no help to the jury, which makes the expert’s opinion irrelevant.
The Ninth Circuit’s discussion of the opinion’s reliability is equally off the mark, mainly in its acceptance of the expert’s “differential diagnosis.” There is no concept in drug and device litigation that has been more misused than the “differential diagnosis,” which is a process of elimination that physicians use to makes diagnoses and prescribe treatment. As our readers already know, there is no need for a diagnostic tool when you’ve already got a diagnosis, which plaintiffs in litigation generally already have. However, somewhere along the way, plaintiffs and their lawyers advanced “differential diagnosis” as a method to determine the cause of a disease already diagnosed, and many courts have regrettably accepted this concept.
But even accepting that state of the law, we think the expert’s “differential diagnosis” in Messick was flawed. For one thing, it is not clear from the opinion that he reliably “ruled in” bisphosphonate therapy as a potential cause of the plaintiff’s ONJ. We have already belabored that he could not say it was a cause, and the district court observed that the expert “never explained the scientific basis for his opinion.” Id. at *11. The Ninth Circuit nonetheless gave him a pass because of the “inherent uncertainty” in the medicine. As the court explained,
[T]he [American Association of Oral and Mixillofacial Surgeons] stated in a 2009 position paper that “the current level of evidence does not fully support a cause-and-effect relationship between bisphosphonate exposure and necrosis of the jaws.” But the paper goes on to explain that while “causality might never be proven, emerging experimental and epidemiologic studies have established a firm foundation for a strong association between monthly IV bisphosphonate therapy and the development of BRONJ [bisphosphonate-related ONJ]. [¶] Because of that inherent uncertainty, we do not require that an expert be able to identify the sole cause of a medical condition in order for his or her testimony to be reliable.
Id. at *13. This is a remarkable passage for several reasons: First, we are not asking an expert to identify the “sole cause” of anything; we would merely have experts identify reliable scientific bases for their opinions, and so would the Supreme Court and the Ninth Circuit itself. Second, the Ninth Circuit acknowledged that there is no scientific evidence of causation and that there may never be scientific evidence of causation. But the plaintiff’s expert still gets to testify and the plaintiff gets to proceed with meritless claims because of an “association” based on “emerging” science, as described in a position paper which itself is not a scientific study. We are not impressed by an association (see above), and the reliance on “emerging” science calls to mind a favorite quote: “Law lags science; it does not lead it.” Rosen v. Ciba-Geigy Corporation, 78 F.3d 316, 319 (7th Cir. 1996).
Third, the plaintiff bears the burden of proving causation, and the proponent of expert opinions bears the burden of establishing their admissibility. If there is significant uncertainty—and it appears there was not, since the expert himself admitted that he could not say that bisphosphonate therapy caused the plaintiff’s ONJ—that uncertainty should result in the opinion being excluded.
Finally, separate and apart from the expert’s failure to reliably “rule in” bisphosphonate therapy as a cause, he also failed to “rule out” five other risk factors, which the Ninth Circuit did not even mention. This “differential diagnosis” is flawed coming and going, which the district court readily detected. The Ninth Circuit unfortunately did not, and it came to the wrong result.