Judge Virginia Hopkins of the U.S. District Court for the Northern District of Alabama granted Novartis’s motion for summary judgment and entered final judgment in Jones v. Novartis Pharmaceuticals Corp., No. 2:13-cv-00624-VEH, 2017 WL 553134 (N.D. Ala. Feb. 10, 2017).  The district court held that plaintiffs had offered no admissible expert testimony that Novartis’s prescription osteoporosis medication, Reclast®, caused her alleged injuries.

The decision relied on the district court’s January 26, 2017 opinion granting in full or in part Novartis’s four Daubert motions to exclude plaintiff’s expert testimony.  Jones v. Novartis Pharm. Corp., 235 F. Supp. 3d 1244 (N.D. Ala. 2017) (“Opinion”).

Plaintiff Ernesteen Jones alleges that she experienced an atypical femur fracture (“AFF”) as a result of her treatment with Novartis’s prescription osteoporosis medication, Reclast®.  The district court excluded in their entirety the opinions of plaintiff’s bone expert Dr. William Hinshaw and statistician Wayne Taylor, including their opinions regarding both general causation and specific causation.  The district court also excluded the causation opinions of two of plaintiff’s treating physicians, finding those doctors unqualified to offer causation opinions and that in any case, they had not employed a reliable methodology.  Id. at 103-118.  Finally, the court granted in part Novartis’s motion to strike the testimony of plaintiff’s regulatory expert, Dr. Suzanne Parisian, finding both that Dr. Parisian was not qualified to offer a number of her opinions and that her methodology was unreliable.  Among other things, the court held that Dr. Parisian could not testify about:  (1) causation and a “causal association”; (2) Novartis’s state of mind, intent, motive, and knowledge; (3) whether and how a different warning might have impacted the prescribing decision of plaintiff’s doctors; and (4) whether and when Novartis may have received “notice” of a risk of AFFs.  Opinion at 40.

In excluding Dr. Hinshaw’s general causation opinion, the district court found that the Bradford Hill methodology, on which Dr. Hinshaw relied, requires as a first step the existence of an established association between a substance and a disease, but that Dr. Hinshaw admitted he could not point to a study in the peer-reviewed literature finding a statistically significant association between Reclast® and AFFs.  The court rejected Dr. Hinshaw’s attempt to extrapolate from an alleged class-wide association between bisphosphonates (“BPs”) and AFFs based on Dr. Hinshaw’s admission that there are significant differences, including regarding the impact on material properties of bone, between the other class BPs, which are administered orally on a weekly or monthly basis, and Reclast®, which is a once-a-year I.V. infusion.  Opinion at 46-58.  The Court next excluded Dr. Hinshaw’s specific causation opinion, (i) as unreliable in the absence of an admissible general causation opinion, and (ii) because, the court found, he inappropriately relied on Bradford Hill for specific causation when that methodology is directed to assessing general causation.  Opinion at 61-64, 69-70.  The Court further held that to the extent Dr. Hinshaw claimed to have conducted a differential diagnosis, it was admittedly only “limited,” and in any case, Dr. Hinshaw had failed to disclose any differential diagnosis in his expert report in violation of Federal Rule of Civil Procedure 26.  Opinion at 64, 66-69.

Plaintiff’s expert statistician, Dr. Taylor, opined that Reclast® clinical trial data provides “evidence of causality” based on a reanalysis of such data that he conducted.  The district court found that Dr. Taylor was not qualified to perform his reanalysis because he lacked a medical degree and expertise on BPs or AFFs.  The court also rejected the reliability of Dr. Taylor’s methodology in a number of respects, including (i) his selection of a supposed background rate for AFFs, (ii) his attempt, without the necessary qualifications, to reclassify events reported in a study as AFFs, (iii) his inclusion of an event that plainly failed to meet the AFF definition, (iv) his comparison of a point estimate to a lower bound confidence interval, and (v) his use of a one-tailed test rather than a two-tailed test to calculate a confidence interval.  Opinion at 75-99.