On June 26, 2012, the Ontario Superior Court released its much-anticipated decision in Andersen et al. v. St. Jude Medical, Inc. et al. http://canlii.ca/t/frtzt.
The proceeding, a product liability class action involving the safety of mechanical prosthetic heart valves and annuloplasty rings coated with Silzone, was first certified in 2003. Some nine years later, it has become the first class action in Canada involving a pharmaceutical drug or medical device to go the distance – a common issue trial resulting in a judgment with Reasons for Decision.
The essential allegation in the common issues trial was that the devices were unsafe and negligently designed. The devices were recalled from the Canadian market in 2000. The trial was lengthy and complex. It lasted 138 court days conducted over 18 months, involved almost 2300 documents, heard from 40 witnesses (23 of whom were experts from 14 different scientific and medical disciplines), and resulted in the submission of written argument briefs in the hundreds of pages. At the conclusion of the marathon trial, Justice Lax decided that the action should be dismissed.
Although there were nine common issues before the trial court, the Decision focused on breach of duty and causation. The Court examined the Defendants’ conduct in designing, testing and marketing the Silzone valve (Common Issue 1). It considered issues of general causation, including whether Silzone had an adverse effect on tissue healing (Common Issue 2). It assessed whether the risk of medical complications for patients with Silzone valves was greater than the risk associated with the use of other devices (Common Issue 3). The other common issues largely dealt with remedies which, given that the proceeding was dismissed, proved to be irrelevant.
Breach in Premarket Design, Manufacture or Testing
At common law in Canada, a manufacturer is required to perform a risk-utility assessment when designing and testing a product, and is required to use reasonable care in doing so. This assessment involves weighing both the gravity and likelihood of reasonably foreseeable risks relative to the overall utility of the product in question. In Andersen, Justice Lax found that the evidence demonstrated that the pre-market design and testing of the devices by the Defendants met the required standard of care. In doing so, she relied on both the expert evidence and the evidence that Silzone devices met industry and regulatory standards (as implicitly demonstrated by both the U.S. FDA and Health Canada approvals).
Breach in Post-Market Surveillance, Warning and Recall
Justice Lax held that because a device manufacturer is the expert on its own product, it has a continuing duty to inform physicians when dangerous side-effects concerning its products are discovered. In the case of the Silzone valve, warnings about the relevant side-effects were contained in the product labelling. The product was eventually recalled. The issue at trial therefore focused on whether the timing of the recall was reasonable in the circumstances.
Justice Lax found that the risk of complications was not materially increased by the Silzone valve when compared to the risk of complications in similar devices. For the sake of completeness, however, she also held that the evidence demonstrated that the Defendants effectively investigated complaints and that the information available up to the date of the recall supported a reasonable belief that the valve posed no additional risk.
Effect of the Valve Coating on Tissue Healing
The general causation issue concerned an analysis of the Silzone coating material, of which silver was an ingredient. The Court held that there was no reliable evidence in support of the Plaintiffs’ theory that silver was toxic or of the mechanism by which the coating interfered with healing of, or caused damage to, existing heart tissue.
The Court also examined whether Silzone devices created a statistically greater risk to patients based on an examination of epidemiological evidence. Justice Lax held that where epidemiological evidence demonstrated a “doubling” of risk of complications above those created by other devices, individual causation could be presumptively proven on a balance of probabilities, absent evidence to rebut that presumption. On the other hand, where the evidence demonstrated a statistically lesser increase in risk, individual causation was presumptively disproved, absent individual evidence to rebut that presumption. In other words, the “doubling of risk” threshold based on epidemiological evidence determined who had the burden of proving individual causation.
In reviewing the evidence, Justice Lax adopted the “recognized hierarchy” of reliability for medical studies: most reliable, randomized control trials; next, cohort studies (or non-randomized observational studies); and last, reported case series. Importantly, Justice Lax rejected the Plaintiffs’ assertion, based on cases such as Snell v. Farrell,  2 S.C.R. 311 (S.C.C.), that she should take a “robust and pragmatic” approach to the evidence and that she should make an inference of causation based on “the totality of the evidence” in the absence of reliable scientific evidence. In the result, Justice Lax ruled, with a few narrow exceptions, that Silzone did not materially increase the risk of complications.
Waiver of Tort
In product liability cases brought in Ontario after pronouncement of Serhan Estate v. Johnson & Johnson, 2004 CanLII 1533 (ON SC); 2006 CanLII 20322 (ON SCDC), plaintiffs have typically invoked the doctrine of “waiver of tort.” By doing so, they propose to give up the right to sue and recover damages in tort, if any, and instead seek recovery on the basis of restitution, claiming disgorgement of the revenues or profits gained by a wrongdoer from their wrongful conduct. A crucial but unresolved aspect of the doctrine of waiver of tort is whether its application requires proof of all elements or an underlying tort, or whether it is capable of being asserted as an independent cause of action. In previous decisions, Ontario courts declined to resolve the issue on the basis that adjudication of such an important point should only be made with the benefit of a “full factual record” such as that secured through trial.
While a full record was before Justice Lax on waiver of tort, she declined to resolve the issue given her conclusion that there was no liability. She did, however, make several statements on the significance of the issue as well as opining on the nature of the factual record required for resolution of the waiver of tort issue. This included a statement that analysis of waiver of tort involved “philosophical and policy considerations” that did not require a trial for resolution, seemingly in contradiction to obiter dicta in earlier appellate decisions. Justice Lax observed: “the fundamental question for a Court to answer is whether the recognition (or not) of the waiver of tort doctrine is within the capacity of a Court to resolve, or whether it has such far reaching and complex effects that is best left to consideration by the legislature.” Hopefully, this statement will spur Ontario courts to resolve the waiver of tort issue without the necessity of a full trial, which will give much needed clarity.
Important Implications from the Decision
The Andersen v. St. Jude Medical Inc. case demonstrates that the requirement to establish all elements of negligence against a manufacturer is alive and well in common law Canada. When dealing with causation, Justice Lax cited authority for the proposition that plaintiffs must establish “if only by the slimmest balance of probability, that a named cause is likely. To demonstrate a possibility is not enough; probability must be established.”
The Decision also asserts that when a manufacturer conducts testing that is reasonable and that is in accordance with international standards, and in the absence of any scientific literature that casts doubt on the safety of a new product, the court should be reluctant to impose liability on the manufacturer. The court will look to materials or standards published by regulatory authorities such as Health Canada, the FDA, ISO, etc., in assessing the adequacy of testing conducted by a manufacturer.
Furthermore, even if a specific product design or attribute materially increases the risk of a medical complication, the plaintiffs have the burden of establishing that such increased risk is attributable to some act or omission by a manufacturer that falls below the applicable standard of care. It will be insufficient for plaintiffs merely to make allegations to satisfy their burden. Plaintiffs will have to discharge their onus by leading evidence to demonstrate specifically what different and/or more extensive pre-clinical and clinical studies would have been required before the product was marketed.
While a manufacturer may be required to undertake a risk-benefit analysis in relation to a new product, and is thereby required to weigh both the gravity and likelihood of a reasonably foreseeable risk relevant to the potential extent of a product’s utility, liability can be avoided. In conducting its risk-benefit analysis, a manufacturer is not required to assess whether the benefits of the putative “improvement” outweigh the benefits of the existing therapy, but rather, whether the potential benefits associated with the putative improvement outweigh the potential risks of the improvement.
Moreover, where a regulator approves a device (or drug) based on specific data, the court will not second-guess the regulator or the manufacturer provided its decision is reasonable: for example, where there is no suitable pre-marketing evaluation process available to assess the concern.
As to general causation, if available data involves small numbers, and especially where a 'control' group is absent, no conclusions as to cause-and-effect can be drawn. This finding has important consequences as to the nature of scientific evidence that must be adduced in a drug or device case to enable a court to conclude general causation in the plaintiffs’ favour. Other indicia of causation in a medical context must be present (for example, biological plausibility).
Thus, causation cannot be determined merely by examining a series of clinical case reports and concluding, for example, that the complication alleged with the St. Jude valve was caused by the Silzone coating. This finding applies even where there are a higher number of reports of complication in the patients who received Silzone-coated artificial valves than in those who received uncoated, but otherwise identical, artificial valves. To establish cause-and-effect or to establish that the putative cause "materially" increases the risk of the alleged adverse effect, the plaintiffs will be obliged to adduce credible epidemiologic evidence - the kind of epidemiologic studies and investigations at the higher end of the recognized hierarchy of epidemiological studies in the scientific literature.
Importantly, however, Justice Lax found that even where the quality of epidemiologic evidence is sound, it does not mean that all patients who suffered the alleged adverse effect would not have suffered it 'but for' the putative cause. This means that epidemiologic evidence alone cannot and does not establish 'specific causation' (i.e., that the alleged side effect was caused by the putative cause in the case of any individual patient). This finding will have material impact on arguing against class certification in product liability cases involving allegations of personal injury.
Lastly, the Court affirmed the generally recognized increase of 2.0 or greater in risk ratio required by the (predominantly U.S.) jurisprudence to establish cause-and-effect based on epidemiological evidence. The Court observed, however, that because defendants retain the right to rebut individual causation (whether the device caused plaintiff's specific injury) even where the risk ratio is greater than 2.0, there is an implicit acknowledgement that individualized causation evidence must be adduced at the individual stage of proceedings following the common issues trial.
The trial decision in Andersen v St. Jude Medical Inc., released on June 26, 2012, coupled with the decision of C. Horkins J. in Martin v. Astrazeneca Pharmaceuticals Plc, 2012 ONSC 2744 released on May 7, 2012 declining class certification, could well mark a significant change in fortune for defendants in pharmaceutical and medical device class actions in Ontario, and inferentially in the common law provinces of Canada. We will be watching closely for any appellate review of the decisions. It seems, however, that manufacturers can take comfort finally that the tide has turned for their class action risk in Canada.