On May 2, 2024, Paul presented at the Medico-Legal Society of Toronto’s annual Spring Bouquet event, held at the Ontario Bar Association’s offices on Toronto Street. The presentation analyzed the Court of Appeal for Ontario’s decision in Levac v James, 2023 ONCA 73, and what it means for causation in medical malpractice claims where the alleged negligence is “systemic” in character.
Levac arose from a 2012 outbreak of serious infections at a Toronto pain management clinic. It is the kind of fact pattern that does not arise often in medical malpractice litigation, and the Court of Appeal’s decision is one of the more important Ontario causation cases of recent years. The framework it endorses for the use of epidemiological evidence has implications that extend well beyond the specific clinic involved.
The causation framework
The starting point in any causation analysis is the “but for” test. In Donleavy v Ultramar Ltd., 2019 ONCA 687, the Court of Appeal restated the test in terms that are useful to keep in mind when reading Levac. Causation is established under the “but for” test if the defendant’s negligence caused the whole of the plaintiff’s injury, or contributed in some not insubstantial or immaterial way to the injury sustained. The test requires a substantial connection between the injury and the defendant’s conduct.
That is the doctrine. The harder question, in cases like Levac, is how the doctrine is to be applied where the precise mechanism of harm cannot be scientifically demonstrated, but the statistical evidence of a connection between the breach and the injury is overwhelming. That is the question Levac answers.
The facts
Dr. Stephen James was an anesthesiologist who administered epidural steroid injections for pain management at the Rothbart Centre for Pain Care in Toronto. Beginning in 2012, some of his patients developed meningitis, epidural abscesses, or other serious infections after their injections. Toronto Public Health discovered the outbreak in November 2012 and launched an investigation.
The investigation concluded that the outbreak was caused by inadequate Infection Prevention and Control (IPAC) practices for the injections. IPAC refers to the practices and procedures intended to prevent or reduce the risk of transmission of microorganisms in health care settings. No specific deficiency could be scientifically demonstrated to have caused the spread of infection. What could be demonstrated was that the rate of infection associated with Dr. James’ injection practice was many times greater than the reported risk of infection for such procedures generally, in a way that was inconsistent with the use of appropriate IPAC.
A second piece of the evidentiary record proved important. Dr. James was found to be colonized with a rare strain of Staphylococcus aureus designated CC59. That strain matched the bacteria infecting six of his patients. The genetic match supplied a direct evidentiary link in the cases of those six patients. The question on appeal was how the trial judge had handled causation for the remaining patients, where no such direct genetic link existed.
The class proceeding
In September 2014, Anne Levac, one of Dr. James’ patients who had suffered an infection, launched a class action under the Class Proceedings Act, 1992, SO 1992, c 6, on behalf of approximately 20 former patients of the clinic who had received injections from Dr. James. The claim alleged that Dr. James, along with the now-defunct Rothbart Centre, its medical director, and its nursing staff, had caused the outbreak by implementing substandard IPAC, by failing to report or investigate the early infections, and by failing to remediate IPAC after learning of those early infections.
A five-week common issues trial was held in 2021 (the decision below is reported as Levac v James, 2021 ONSC 5971). The trial judge found against Dr. James on all of the common issues: duty of care, standard of care and breach, causation, fiduciary duty, the limitation period defence, and entitlement to punitive damages.
For the six patients with the matching CC59 strain (the “Genetically Linked Patients” in the trial judge’s terminology), causation was proven on the direct microbiological evidence. For the remaining patients, the trial judge drew a rebuttable inference of causation, grounded in the statistical evidence that Dr. James had exposed his patients to a risk of infection that was either 49 or 69 times higher than the risk for comparable patients not exposed to his substandard IPAC.
The appeal
Causation was the central issue on appeal. The appellant challenged the trial judge’s use of statistical evidence to draw a class-wide rebuttable inference of causation for the Remaining Patients.
The Court of Appeal upheld the analysis. The reasons rest on two propositions.
First, the trial judge had identified powerful circumstantial evidence on which to conclude that a statistical association represented a causal link on a balance of probabilities. The risk ratios of 49.0 and 69.0 were not modest statistical signals; they were, in the trial judge’s words, “so overwhelming that it cannot be ignored.” Combined with the proven breach of the standard of care and the existence of an infectious disease outbreak temporally and geographically tied to Dr. James’ practice, the statistical evidence supported the inference the trial judge drew.
Second, Dr. James had not put forward a viable, non-negligent explanation for the outbreak as a whole. The court treated the absence of an alternative explanation as part of the analysis. Where the breach of the standard of care is established, the outbreak is established, the statistical signal is overwhelming, and no alternative explanation has been advanced, the inference of causation can properly be drawn.
The Court of Appeal accordingly upheld the class-wide rebuttable inference. Dr. James retained the right to rebut the presumption in the subsequent individual issues phase of the proceeding, but the common issues finding stood.
What Levac clarifies
The decision sits at an interesting point in the development of Ontario causation law. The Supreme Court of Canada has generally encouraged restraint in the use of statistical evidence to prove causation, and that restraint reflects a real concern: statistical association is not the same as causal connection, and treating it as such can lead to liability findings that the underlying science does not support. Epidemiological evidence has, however, become increasingly prominent in medical device and pharmaceutical class proceedings, where the question of general causation often depends on it.
Levac clarifies when epidemiological evidence can support a rebuttable presumption of causation in the negligence context. The decision does not announce a specific risk ratio threshold. The judgment in Wise v Abbott Laboratories, Limited, 2016 ONSC 7275, in which Justice Perell granted summary judgment dismissing a product liability claim involving the topical testosterone product AndroGel, had already cautioned against equating legal degrees of proof with mathematical probabilities. As Justice Perell put it, “there are no hard and fast rules for inferring causation in any case.” Wise used the 2.0 risk ratio as a benchmark for thinking about whether the exposure more likely than not caused the disease in an individual plaintiff, but it did not adopt 2.0 as a legal threshold.
Levac sits in the same general framework. The risk ratios in Levac were 49 and 69, which is not an order of magnitude above the Wise benchmark; it is an order of magnitude or more above what any other reported Ontario case has used to ground a rebuttable inference. The Court of Appeal did not say that 2.0 is enough, and it did not say that 49 is required. It said that, in this case, the combination of a clear breach of the standard of care, an outbreak of infectious disease, overwhelming statistical evidence, and the absence of any plausible non-negligent causal alternative was sufficient to support the inference the trial judge drew.
Practical observations
For plaintiff counsel, three points follow.
The evidentiary architecture matters. Levac did not turn on the statistics alone. The statistical evidence was one of four pillars: breach of the standard of care, an outbreak, overwhelming risk ratios, and the absence of an alternative explanation. Cases that present only one or two of those pillars will not readily extend the Levac framework. The combination is what the Court of Appeal endorsed.
Expert evidence in epidemiology becomes important earlier than it might in a single-plaintiff case. Where a claim is framed as systemic, the experts who can speak to risk ratios, study design, and confounders are not optional. They are central to the case the plaintiff must put on at the common issues stage.
And the class proceedings vehicle remains the natural home for this kind of analysis. Levac was decided as a class action, not as 20 individual actions. The common issues framework allowed the trial judge to make the causation finding on a class-wide basis, with the rebuttable inference doing real work for the plaintiffs across the class. Without that vehicle, each plaintiff would have had to prove causation individually, and the Wise concerns about statistical evidence in single-plaintiff cases would have applied with full force.
The role of epidemiological and other statistical evidence in drawing legal inferences of causation will probably always be contentious, because it is by necessity inferential. In some cases, however, it can be extremely convincing. Levac is the leading Ontario authority on when that line is crossed.
