Written By Cheryl Woodin and Alexander Payne

The Court of Appeal for Ontario in Levac v James, 2023 ONCA 73 [Levac] has unanimously upheld a trial judgment in a common issues trial regarding an infectious disease outbreak in respect of which the allegedly negligent cause was linked to a very high statistical increase in risk, and for which no alternative and non-negligent cause was presented.

The Court of Appeal’s decision clarifies the circumstances in which statistical evidence, particularly epidemiological evidence, may be used to infer causation. Epidemiology is the study and analysis of the distribution of patterns and determinants of health and disease conditions in a defined population—it examines associations between health risks and outcomes in a population. The Court of Appeal concluded that where there is (1) a proven breach of the standard of care in negligence; and (2) a proven injury, then “powerful” epidemiological evidence can be used to infer causation, even if it cannot be proven with scientific certainty.

Background and Trial Decision

The treating physician and appellant, Dr. James, was an anesthesiologist who administered epidural injections into the area around his patients' spines as a pain relief treatment. Some of his patients developed meningitis, an acute inflammation of the protective tissue surrounding the spine typically caused by a bacterial or viral infection, or other serious infections, after receiving injections.

A class proceeding was commenced on behalf of a class of patients who developed signs or symptoms clinically compatible with bacterial meningitis, epidural abscess or cellulitis of a bacterial origin and/or bacteremia (collectively, the Injuries) after receiving an epidural injection administered by Dr. James.

The class was divided into subclasses, including (1) patients who were infected by a rare strain of CC59 Staphylococcus aureus bacteria that genetically matched the bacteria that colonized Dr. James himself (the Genetically Linked Patients); and (2) the remaining patients, who suffered Injuries but could not scientifically prove that they were infected by the same rare CC59 strain that colonized Dr. James (the Remaining Patients).

Following a five-week common issues trial, the trial judge, Justice E.M. Morgan, found against the anesthesiologist on all the common issues: negligence (duty of care, standard of care and breach, and causation), fiduciary duty, limitation period and entitlement to punitive damages.

In the context of the negligence analysis, the trial judge made detailed factual findings about Dr. James’ failure to take reasonable precautions to prevent the transmission of healthcare-associated infections, including failures to use aseptic techniques. The trial judge also found that Dr. James failed to report suspected infections linked to his practice. The trial judge accordingly found that there was a breach of the standard of care.

The causation aspect of the negligence analysis was hotly disputed at trial, with Dr. James submitting that “causation is a crucial hurdle on which the Class’s claim falters”.¹ For the Genetically Linked Patients, the trial judge had “no hesitation” in concluding that the Genetically Linked Patient’s Injuries were caused by Dr. James, given the genetic match between the Staphylococcus aureus CC59 strain infecting the Genetically Linked Patients and colonizing Dr. James.²

The analysis regarding the Remaining Patients required an inference of causation, however, using what the trial judge referred to as the “risk ratio” approach discussed by Justice Lax in Andersen v St. Jude Medical, Inc., 2012 ONSC 3660 [Andersen] and in other cases, including Stanway v Wyeth Canada Inc., 2012 BCCA 260. In Andersen, Justice Lax found that a breach of the standard of care that more than doubled the risk of harm (i.e., a risk ratio of 2.0) presumptively established causation for the class, subject to proof to the contrary.³

In Levac, the epidemiological evidence at trial revealed that patients of Dr. James had either a 49 times or 69 times greater risk of developing a serious infection than pain clinic patients not exposed to Dr. James’ substandard infection prevention and control (risk ratios of 49.0 and 69.0, respectively). The trial judge concluded that this evidence is “so overwhelming that it cannot be ignored”,⁴ and that a rebuttable presumption of causation was established.

The Appeal Decision

The causation criterion was the central issue on appeal, with the appellant challenging the trial judge’s use of statistical evidence to infer causation in negligence in respect of the Remaining Patients.

The Court of Appeal for Ontario upheld the trial judge’s causation analysis on the basis that (1) there was “powerful” circumstantial evidence on which to conclude that a statistical association represented a causal link on a balance of probabilities; and (2) Dr. James had not put forward a viable, non-negligent explanation for the outbreak as a whole.

The Court of Appeal found no error in the judge’s reliance on statistical evidence in drawing a class-wide rebuttable inference that Dr. James' substandard infection prevention and control caused the Remaining Patients’ Injuries. It will be open to Dr. James to rebut the presumption in the subsequent individual issues phase of the proceeding.

Takeaways and Analysis

The Court of Appeal’s decision in Levac clarifies the circumstances in which epidemiological evidence may give rise to a rebuttable presumption of causation, and the role of epidemiological risk ratios in the analysis.

Epidemiological, statistical data has increasingly featured prominently in medical device and pharmaceutical class proceedings,⁵ notwithstanding that Canadian courts, including the Supreme Court of Canada, have generally encouraged restraint in the use of statistical evidence to establish causation.⁶

There is currently no specifically-delineated risk ratio that gives rise to a rebuttable presumption of general causation. While several cases (including Levac) used the 2.0 risk ratio as a benchmark, in Wise v Abbott Laboratories, Ltd., 2016 ONSC 7275, Justice Perell cited a 2015 decision of the British Columbia Court of Appeal in cautioning against equating legal degrees of proof with mathematical probabilities, noting that “there are no hard and fast rules for inferring causation in any case.”⁷

In Levac, the epidemiological evidence was largely undisputed, and in the trial judge’s words “overwhelming.” Regardless of whether the 49.0 or 69.0 risk ratio applied, it was multiple times higher than the Andersen 2.0 risk ratio benchmark.

In this context, which included a finding of clear breach of the standard of care, manifested injury and the absence of any other plausible non-negligent causal alternatives, the Court of Appeal upheld the use of “powerful” statistical epidemiological evidence to rebuttably infer causation.

The role of epidemiological and other statistical evidence in drawing legal inferences of causation is a critical but sensitive area of the law which requires judicial diligence both in fact finding and legal analysis. The Court of Appeal's decision illustrates the importance of clear factual and legal safeguards around the use of such evidence and leaves open the question whether the relatively modest 2.0 risk ratio benchmark applied in Andersen could be appropriate in another case. 

Download PDF