Putting in the Work

Onglyza Products Cases, No. A165387 (D1d4 Apr. 19, 2023)

Plaintiffs claim they were injured by a diabetes drug. In connection with the drug’s NDA, the FDA made the manufacturer do an outcomes study on high-cardiovascular risk patients with type 2 diabetes. The study—a randomized, double-blind, placebo-controlled study of 16,492 patients—looked at three primary endpoints: cardiac arrest, heart attack, and stroke. It also looked at 10 secondary endpoints. The only statistically significant endpoint was hospitalization for heart failure. The study’s authors noted that the difference was small, unexpected, and merited further investigation. FDA required the drug’s label to disclose a warning for potential increased risk of heart failure.

Nobody replicated the study. Later observational studies of large groups could not find an association, and studies of pooled data from other clinical trials of the drug did not reveal a correlation. 

To establish general causation—that the drug in question is capable of causing their injury—Plaintiffs called two experts, a cardiologist and a biostatistician. The cardiologist testified that the drug could generally cause heart failure. His analysis was heavily reliant on the single study. He purported to go through a series of criteria called the Bradford Hill criteria that are used in epidemiology to ascertain whether a causal effect has been shown. But the trial court found that the expert had employed the factors in a manner that over-weighted the single study to the exclusion of all other evidence. Finding that the cardiologist employed “a shifting results-based methodology that fails to logically and consistently weigh all relevant evidence,” the court excluded the expert under Sargon.

And because the biostatistician admitted that (a) his opinions were somewhat dependent on the cardiologist’s; and (b) he was not, on his own, able to opine about general causation, the trial court found that Plaintiffs didn’t meet their burden and granted summary judgment on general causation.

The Court of Appeal affirms. As the Court explains, a “trial court does not abuse its discretion in excluding expert testimony on general causation when the expert’s opinion is based on a single study that provides no reasonable basis for the opinion offered.” The study’s own authors rejected the notion that the study alone established a causal relationship. And, as the court explains, going factor by factor, the cardiologist employed the Bradford Hill criteria in a slanted manner that appeared driven to reach a desired outcome. This was not helped by the doctor’s testimony at a Sargon hearing under Evidence Code § 402, which was often inconsistent with the methods in his report. 

 

Affirmed. 

 

As I have complained about in the past, correctly applying the Sargon standard can be hard work, because the Court needs to engage with the underlying methodology to assess its reliability. The Court here does a particularly good job on that. It takes a close look at the methods applied by the expert and asks (1) are these the types of methods the expert should employ?; and (2) are they being applied consistently with the ways they are is applied in the field, outside of litigation. Notably, the Court does not just abdicate any gatekeeping responsibility because there is some ill-defined dispute over the science. Examining the reliability of a scientific expert’s methods is not the same thing as resolving a controversy between two or more scientists applying reliable methods. The first is an issue for the judge, under Sargon, the second is the ultimate question for the jury.

 

It’s also worth noting how much work the standard of review does in these cases. As we’ve seen, getting a reversal based on a Sargon error in permitting an expert to testify is hard uphill climb. But getting an affirmance of an expert’s exclusion on an SJ is much less so.