Benhaim v St-Germain: Adverse Inference, Causation, and the Limits of What Negligence Can Prove — The Supreme Court of Canada’s Leading Delayed Cancer Diagnosis Decision

The hardest causation problem in all of medical malpractice is not proving that a physician was negligent. It is proving that the negligence made a difference. And nowhere is this harder than in delayed cancer diagnosis, where the physician’s failure to investigate a finding destroys the very evidence that would show what stage the cancer was at when it should have been found.

Marc Émond was forty-four years old when a shadow appeared on a chest X-ray and his physicians failed to follow up. He was forty-seven when lung cancer killed him. His partner, Cathie St-Germain, brought a malpractice claim against both physicians. The trial judge found them negligent. She then found that causation had not been established and declined to draw an adverse inference of causation, even though she acknowledged their negligence had made causation harder to prove. She awarded $70,000 for the anguish of their negligent handling of the treatment, but not the death damages a causation finding would have required.

The Quebec Court of Appeal reversed her, finding she had been legally required to draw the adverse inference. The Supreme Court of Canada reversed the Court of Appeal, finding that no such legal requirement exists. The adverse inference is a tool. It is available. It may discharge the plaintiff’s burden. But no rule of law requires the trier of fact to use it.

Benhaim v St-Germain, 2016 SCC 48, is the leading Canadian authority on what the adverse inference of causation is, when it is available, what it does, and, critically, what it does not require. This article examines the case in full and explains what it means for every delayed diagnosis malpractice claim brought in Canada.

Understanding Adverse Inference of Causation: The Concept Before the Case

The Problem It Addresses

To understand Benhaim, it is necessary to understand the specific causation problem it addresses, one endemic to delayed diagnosis malpractice.

In a typical negligence case, the harm is there to be examined. A surgeon who nicks a vessel leaves an identifiable injury. An emergency physician who fails to diagnose a heart attack can be evaluated against a clinical record of what the patient’s presentation was. The negligence and its consequences can, at least in principle, be assessed using the evidence that exists.

In a delayed cancer diagnosis case, the physician’s negligence is precisely the failure to create the evidence that would be needed to establish causation. A radiologist who sees an opacity and recommends CT scanning, whose recommendation is not followed, has left behind a radiological finding but no pathology, no staging information, no biopsy result. The question of what stage the cancer was at when it should have been investigated cannot be answered with certainty, because the investigation that would have provided the answer was never performed. The negligence itself is what destroyed the evidentiary foundation of the causation case.

This creates a specific and persistent unfairness: the more thoroughly a physician fails to investigate a concerning finding, the harder it becomes for the patient to prove that the failure caused harm. The negligence and the evidential gap are the same event.

What an Adverse Inference Is

In the causation context of medical malpractice, the adverse inference is a conclusion the trier of fact may draw from the defendant’s negligence having made it impossible for the plaintiff to gather the evidence of causation that would otherwise be available. The court may reason: because the defendant’s negligence prevented the creation of staging evidence, and because some affirmative evidence points toward causation, the ambiguity arising from the missing evidence will be resolved against the party whose negligence created it.

In Canadian law, the adverse inference in causation draws directly from Justice Sopinka’s unanimous judgment in Snell v Farrell, 1990 CanLII 70 (SCC), [1990] 2 SCR 311, which established that a causal inference may be drawn even without positive or scientific proof of causation, in the absence of affirmative evidence of an alternative cause adduced by the defendant. Benhaim v St-Germain extends this into the specific territory where the defendant’s negligence undermined the plaintiff’s ability to gather the evidence that would prove causation, and asks whether the inference becomes not merely available but obligatory in those circumstances.

For a detailed explanation of the Snell v Farrell framework and the “but for” test, including how the robust and pragmatic approach to causation operates, and when the material contribution exception applies, see the companion article in this series covering Snell v Farrell and Clements v Clements.

The Facts: Marc Émond, a Shadow on an X-Ray, and a Year of Silence

The Patient

Marc Émond was, by every measure, a healthy man. In November 2005, at forty-four, he was a lifelong non-smoker who exercised regularly and took careful care of himself. He had annual physicals at the Clinique Physimed, where he was under the care of Dr. Albert Benhaim, a general practitioner. He had no symptoms of any chest condition.

He also had, through routine X-ray work at various points in his life, accumulated a medical file containing chest X-ray reports from 1994, 1998, and 1999. These prior records were retrievable. They would prove relevant in ways neither Marc Émond nor his physicians initially appreciated.

November 2005: The Opacity Appears

On November 9, 2005, Marc Émond attended his annual physical with Dr. Benhaim. Despite the absence of any chest symptoms, Dr. Benhaim recommended a chest X-ray. Dr. Michael O’Donovan, a radiologist, reviewed the film. His report noted “a 1.5 to 2 cm ill-defined opacity” in Mr. Émond’s right lung, “the etiology of which is uncertain.” No other associated abnormalities were visible.

Dr. O’Donovan’s report made specific recommendations: for purposes of comparison, Dr. Benhaim should consult Mr. Émond’s previous chest X-rays if they were available; otherwise, a follow-up X-ray should be done; and he suggested “probably a CT scan.”

Dr. Benhaim did not attempt to find Mr. Émond’s previous chest X-rays. Nor did he consult Mr. Émond’s medical file, which contained reports on chest X-rays taken in 1994, 1998, and 1999. He did not arrange the CT scan. He ordered a follow-up chest X-ray.

January 2006: The Follow-Up That Resolved Nothing

On January 17, 2006, the follow-up chest X-ray was performed. Dr. O’Donovan reviewed it and noted no change in the opacity. He suspected the opacity was a chronic lesion and suggested that Mr. Émond undergo a third X-ray in four months’ time. That third X-ray was never taken. The cancer, whatever stage it was at, continued to evolve without clinical attention for almost another full year.

December 2006: Discovery, Too Late

Mr. Émond had his next annual physical on December 4, 2006. A chest X-ray was taken. Dr. O’Donovan observed that the lesion had increased in size to approximately 2.5 centimetres. He suspected cancer and suggested further tests. Mr. Émond underwent a CT scan on December 19, 2006 and a PET scan in January 2007. Those tests confirmed stage IV lung cancer. He was diagnosed in January 2007. The cancer was incurable. He still had no symptoms.

Beginning in February 2007, Mr. Émond received palliative chemotherapy. The chemotherapy treatments were suspended between June and November 2007, as Mr. Émond was feeling well: he exercised regularly and biked over 5,000 kilometres that summer. He began to experience symptoms in the fall of 2007. Chemotherapy resumed in November 2007 but was stopped again in February 2008 because it was unsuccessful.

Marc Émond died on June 6, 2008. He was forty-seven years old. He was survived by his partner, Ms. St-Germain, and their eight-year-old son.

The Legal Proceedings: Three Courts, One Central Question

The Claim and the Central Defence

Cathie St-Germain brought a malpractice action against Drs. Benhaim and O’Donovan in her own name, in her capacity as tutor to her minor son, and as Marc Émond’s universal legatee. She alleged that the physicians’ failure to promptly investigate the 2005 opacity had caused Mr. Émond’s death by allowing the cancer to progress from a curable to an incurable stage.

The physicians’ defence was direct. They argued that a diagnosis in November 2005 would not have saved Marc Émond because his cancer was already at an advanced stage at that time. Dr. Ferraro, the defence expert, testified that Mr. Émond’s cancer was already at stage III or IV in November 2005, and his chances of survival at that time were low. The delay was negligent, but it was not the cause of death.

This framed the essential causation question: what stage was Marc Émond’s lung cancer in November 2005? If Stage I or II, prompt diagnosis and treatment might probably have cured him. If Stage III or IV, the prognosis was poor regardless of diagnosis timing. The problem was that no one knew what stage it was in November 2005; the investigation that would have established the staging was never performed, because of the negligence itself.

The Expert Dispute: Statistics Versus Clinical Reasoning

Three expert witnesses testified at trial. The judgment acknowledges that each formed an opinion on the basis of incomplete information and that each opinion involved some degree of speculation and estimation.

The plaintiff’s expert evidence relied significantly on statistical evidence. The key figure: approximately 78% of lung cancers discovered incidentally (as Marc Émond’s was, through an asymptomatic routine X-ray rather than investigation prompted by symptoms) are at Stage I at the time of discovery. Stage I lung cancer, treated promptly, has a cure rate of approximately 70%. The statistical argument was that Marc Émond was most probably in the 78%, meaning he had Stage I cancer in November 2005, and the thirteen-month delay allowed it to progress to Stage IV and kill him.

The defence expert, Dr. Ferraro, argued from clinical reasoning. He testified that Marc Émond’s cancer could not have progressed from Stage I to Stage IV in approximately thirteen months (November 2005 to December 2006) because this variety of cancer evolves slowly. Moreover, viewing the November 2005 X-ray retrospectively with knowledge of the December 2006 and January 2007 results, he could identify a prominence in the hilar region consistent with cancer already at Stage III or IV. If it was Stage III or IV in November 2005, the prognosis would have been poor regardless of diagnosis timing.

No expert could resolve the staging question with certainty. The negligence had prevented the creation of the staging evidence that would have answered it.

Stage One — The Quebec Superior Court: Negligence Found, Causation Not Established

Justice Marcotte of the Quebec Superior Court (2011 QCCS 4755) conducted a thorough review of the evidence. On negligence, her findings were clear: Dr. Benhaim failed to consult prior X-rays as recommended, failed to arrange a CT scan, and allowed a potentially significant finding to go unexplored. Dr. O’Donovan failed to adequately follow up. Both were negligent. This finding was not contested on appeal to the Supreme Court.

On causation, her findings went against the plaintiff. She found that in November 2005 and January 2006, Marc Émond’s cancer was likely already at Stage III and likely incurable. She was not convinced that but for the delay in diagnosis, he would probably have survived. She chose to believe Dr. Ferraro’s expert evidence over the plaintiff’s experts, for reasons she carefully articulated. The 78% statistical argument was insufficient to override the clinical evidence specific to Mr. Émond’s case, in her assessment.

Crucially, the trial judge explicitly recognized that she could draw an adverse inference of causation against the physicians because their negligence had made it impossible to prove causation with precision. She did not think this inference was warranted on the evidence before her. She awarded $70,000 in damages to Ms. St-Germain personally and in her capacity as universal legatee for the anguish caused by the defendants’ negligent handling of Mr. Émond’s treatment, not the death damages that a finding of causation would have required.

Stage Two — The Quebec Court of Appeal: The Inference Was Mandatory

The Quebec Court of Appeal (Kasirer, Fournier and Bélanger JJ.A.; 2014 QCCA 2207) reversed on causation.

The majority (Kasirer and Bélanger JJ.A.) found that the trial judge had erred in law by failing to draw an adverse inference of causation. Their view was that where two criteria are established, namely the defendant’s negligence undermined the plaintiff’s ability to prove causation, and the plaintiff has adduced at least some affirmative evidence of causation, the trier of fact is required to draw the adverse inference. The trial judge had recognized both criteria were present and declined to draw the inference anyway. The majority found that declining was a legal error, not a permissible exercise of judicial discretion.

Fournier J.A. concurred in allowing the appeal but on different grounds. He found that causation had been established on the evidence available, without the adverse inference being necessary, and that the trial judge had made a palpable and overriding error in her assessment of the facts. His central point: Mr. Émond survived more than 31 months after the November 2005 X-ray (14 months without treatment, then 17 months with treatment), but the life expectancy of untreated Stage III or IV lung cancer is 8 to 12 months. If Mr. Émond had Stage III or IV cancer in November 2005, he should have been dead well before January 2007. His survival to June 2008 strongly indicated his cancer had not been at an advanced stage in November 2005.

The result of the Court of Appeal’s decision was that both physicians were found liable for Marc Émond’s death, with substantial damages to flow from that finding. The physicians appealed to the Supreme Court of Canada, which granted leave on October 29, 2015 (2015 CanLII 69434).

The Supreme Court of Canada: A 4-3 Decision on a Question of Legal Obligation

The Majority: Wagner J. (McLachlin CJ, Karakatsanis and Gascon JJ.)

Justice Wagner wrote the majority judgment, allowing the physicians’ appeal and restoring the trial judge’s decision. The four-justice majority rested its analysis on a single precise legal proposition: no rule of law requires the trier of fact to draw an adverse inference of causation where the defendant’s negligence has undermined the plaintiff’s ability to prove causation, even where there is some evidence of causation.

The Law on Adverse Inference: May, Not Must

Wagner J. traced the adverse inference concept in the causation context to Snell v Farrell, where Sopinka J. had established that a causal inference may be drawn in the absence of positive or scientific proof, without shifting the legal burden of proof to the defendant. Building on Snell, Wagner J. confirmed the following framework:

Where a defendant’s negligence has undermined the plaintiff’s ability to prove causation (where the very investigation that would have established the causal picture was the thing the defendant should have performed but did not) and where the plaintiff has nonetheless adduced some affirmative evidence of causation, the adverse inference is available. The trier of fact may draw it. In appropriate circumstances, it may discharge the plaintiff’s burden of proving causation on the balance of probabilities.

But the inference is not triggered automatically by the existence of these conditions. Whether the inference is warranted, and how it is to be weighed against the contrary evidence, are matters for the trier of fact, not a command to the trier of fact. In Wagner J.’s words at para 42: “This Court held in Snell that, in such circumstances, an adverse inference of causation may discharge the plaintiff’s burden of proving causation. Those circumstances do not trigger such an inference. Whether an inference of causation is warranted, and how it is to be weighed against the evidence, are matters for the trier of fact.”

The adverse inference is permissive precisely because it is a component of the fact-finding process: not a rule of law but an ordinary step in evaluating evidence. As Wagner J. quoted from Snell: “It is not strictly accurate to speak of the burden shifting to the defendant when what is meant is that evidence adduced by the plaintiff may result in an inference being drawn adverse to the defendant. Whether an inference is or is not drawn is a matter of weighing evidence.” That discretion cannot be removed by converting the inference into an automatic presumption triggered by defined conditions.

The Civil Code Framework: Article 2849

Because Benhaim arose in Quebec, Wagner J. also situated the adverse inference within the Quebec civil law framework. The Civil Code of Québec (arts. 2846 to 2849) provides a comprehensive set of evidentiary rules, and the rules on presumptions govern the inferences drawn by triers of fact in Quebec proceedings.

Wagner J. held that, in Quebec civil law, the adverse inference discussed in Snell constitutes nothing more than a presumption of fact as defined in art. 2849 of the Civil Code. Under that provision, presumptions that are not established by law “are left to the discretion of the court which shall take only serious, precise and concordant presumptions into consideration.” The decision to draw an adverse inference of causation is therefore a factual decision governed by art. 2849, not a special rule for medical malpractice that operates differently from the ordinary law of presumptions.

This framing is important. The majority rejected any interpretation of Snell that would lower the threshold for drawing presumptions in medical malpractice cases below the art. 2849 threshold of serious, precise, and concordant presumptions. The inference is available, but only where the evidence meets this standard, assessed according to a legal, not a scientific, standard.

The Trial Judge’s Decision Was Fact, Not Law

Having established that the adverse inference was discretionary rather than mandatory, the majority turned to the standard of review. The trial judge’s decision not to draw the inference was a finding of fact: an assessment of the evidence before her and a determination of the weight the inference should carry against the contrary clinical evidence. Findings of fact attract appellate deference; they may be overturned only upon a palpable and overriding error, a clear, obvious error going to the very core of the factual determination.

The majority found the trial judge had made no such error. She had reviewed the expert evidence carefully. She had assessed the 78% statistical argument and found it too general to compel a finding of causation in this specific case: the statistic applied to cancers discovered incidentally through CT screening, a more precise diagnostic tool than the X-ray that identified Mr. Émond’s opacity, making the statistic a poor proxy for his situation. She had accepted the defence expert’s clinical reasoning about the slow-growing nature of this particular cancer and the retrospective reading of the 2005 X-ray. These were permissible findings of fact.

The majority emphasized that the trial judge had recognized the adverse inference was available to her. She had not misunderstood the law. She had exercised her discretion, on a careful review of all the evidence, not to draw it. That exercise was entitled to appellate deference.

On Statistical Evidence Specifically

Wagner J. addressed the role of population-level statistics in causation analysis. Statistics are derived from observations across large numbers of cases. The relationships they reveal between conditions and outcomes are functions of repetition on a large scale. Legal fact-finding, by contrast, must be concerned with what actually happened between the parties before the court, not with what happens in most cases, most of the time.

The 78% statistic, that 78% of incidentally discovered cancers are at Stage I, is relevant evidence and may inform causation analysis. But it is not self-sufficient proof of causation in an individual case. Statistics are silent about whether the particular patient would have conformed to the trend or been an exception. Without an evidentiary bridge to the specific circumstances of the plaintiff, statistical evidence is of limited assistance. What inferences follow from a general statistic in a particular case is a matter for the trier of fact, assessed against all the evidence, and entitled to considerable deference on appeal.

Moreover, the majority noted the trial judge had a reasonable basis for giving the 78% statistic limited weight: it originated from data about cancers discovered through CT screening, and the statistic’s methodology made it a poor fit for a cancer identified through an X-ray at a routine physical where the patient remained asymptomatic throughout his progression to Stage IV.

The Broader Legal Context

The majority situated its analysis within the framework of Snell v Farrell and Clements v Clements. Causation is a practical question of fact to be answered by common sense. It does not require scientific certainty. An inference may be drawn from the available evidence without positive proof. But the trier of fact must assess all the evidence, including the evidence that weighs against the inference, and reach a conclusion that the whole of the evidence, viewed practically with common sense, supports. The Snell framework gives trial judges tools to establish causation without scientific proof; the adverse inference is one of those tools. But tools must be deployed by judgment, not triggered mechanically by preconditions.

The Dissent: Côté J. (Abella and Brown JJ.)

Justice Côté wrote the dissenting reasons for three justices, with Abella and Brown JJ. concurring. On the central legal question, Côté J. agreed with the majority: the adverse inference was permissive, not mandatory as a matter of law. The Court of Appeal had erred in characterizing the trial judge’s failure to draw the inference as an error of law. On this core legal principle, all seven justices were unanimous.

Where Côté J. parted from the majority was on the factual question of whether the trial judge’s failure to draw the inference constituted a palpable and overriding error, and she found that it did, for three specific reasons.

Error One: Misapprehension of Dr. Ferraro’s Expert Evidence

Côté J. identified a critical problem with how the trial judge had characterized Dr. Ferraro’s testimony about the rate of cancer progression. The trial judge found that lung cancer could not have progressed from Stage I to Stage IV in approximately twelve months. But Dr. Ferraro never said this was “impossible”; he said it was “possible but unlikely” or “very, very unlikely,” and he acknowledged explicitly that there was no medical literature confirming impossibility because observing untreated Stage I cancer for comparison purposes would be unethical.

More significantly, the trial judge found that Dr. Ferraro had identified shadows in both the hilar and mediastinal regions on the November 2005 X-ray. But Dr. Ferraro’s expert report, examination in chief, and cross-examination all referred only to a shadow in the hilar region, not the mediastinal region. The mediastinal reference appeared only once, during re-examination, and Côté J. concluded it was a mistake, not a change of position. This distinction was clinically significant: a shadow in the hilar region alone is consistent with Stage I or II cancer, while a shadow in the mediastinal region indicates Stage III or higher (when cancer has spread to the lymph nodes of the mediastinum). The trial judge’s reliance on an erroneous finding of mediastinal involvement was a palpable and overriding error.

Error Two: Omission of Mr. Émond’s Survival Period

The dissent found that the trial judge had disregarded key objective evidence: Mr. Émond survived more than 31 months after the November 2005 X-ray (14 months without treatment, then an additional 17 months with treatment) even though the life expectancy of patients diagnosed with Stage III to Stage IV lung cancer is 8 to 12 months without treatment, with palliative chemotherapy extending life by a few additional months.

If the defence theory was correct, namely that Mr. Émond had Stage III or IV cancer in November 2005, he should have been dead well before January 2007, when the diagnosis was finally confirmed. Instead he was alive, asymptomatic, exercising regularly, and had biked over 5,000 kilometres in the summer of 2007. His survival trajectory was entirely inconsistent with the defence staging hypothesis. This evidence, supported by uncontested statistical data showing that 78% of fortuitously discovered cancers are at Stage I, should have compelled a finding that his cancer was not at an advanced stage in November 2005. Failing to give this evidence adequate weight was a palpable and overriding error.

Error Three: The Inference-Drawing Process Itself

Taking the two errors together, Côté J. found that the trial judge’s failure to draw an adverse inference of causation was itself a palpable and overriding error. The factual evidence, namely Mr. Émond’s survival period and the 78% statistic specific to fortuitously discovered cancers, supported a serious, precise, and concordant presumption under art. 2849 of the Civil Code that his cancer was not at Stage III or higher in November 2005. It was clearly wrong not to draw that presumption.

Côté J. accordingly would have dismissed the physicians’ appeal, affirming the Court of Appeal’s conclusion that causation had been established, though arriving there through the palpable and overriding error standard rather than through any mandatory inference rule.

The Significance of the 7-0 Agreement on the Law

The 4-3 split on the outcome should not obscure the unanimous agreement on the legal framework. All seven justices agreed that the adverse inference of causation is permissive, not mandatory: that no rule of law requires a trier of fact to draw it when its preconditions are met, and that a trial judge who declines to draw it has made a finding of fact entitled to appellate deference. The case before the Supreme Court was in every legal respect agreed upon. The disagreement was entirely about whether the specific facts of this case constituted a palpable and overriding error in the trial judge’s application of that agreed framework.

What Benhaim v St-Germain Establishes: Key Principles for Practice

Adverse Inference of Causation: Available, Not Automatic

Where a defendant’s negligence has undermined the plaintiff’s ability to prove causation, and where the plaintiff has nonetheless adduced some affirmative evidence of causation, an adverse inference of causation may be drawn by the trier of fact. This inference may be sufficient to discharge the plaintiff’s burden of proving causation on the balance of probabilities. But the trier of fact is not required to draw it. The conditions for the inference’s availability do not trigger the inference automatically.

Whether the inference is warranted, and how much weight it should carry against contrary evidence, are matters of judicial discretion to be resolved by assessing all the evidence. This is the same discretion the trier of fact exercises in every other aspect of fact-finding. The adverse inference is not a special rule; it is the ordinary operation of inference-drawing under medical malpractice circumstances where the defendant’s own fault created the evidential gap.

The Legal Burden Remains With the Plaintiff

The availability of the adverse inference does not shift the legal burden of proof to the defendant. The plaintiff must establish causation on the balance of probabilities. The adverse inference is a tool for discharging that burden but does not relocate it. A defendant who adduces affirmative contrary evidence provides the trier of fact with material to weigh against the inference, and that weighing may result in the inference being insufficient to establish causation.

In Quebec: the adverse inference corresponds to a presumption of fact under art. 2849 of the Civil Code, which requires the presumption to be serious, precise, and concordant before it can be taken into consideration. This standard is consistent with the Snell approach in common law provinces: the inference requires more than a theoretical causal possibility or bare statistical probability detached from the individual circumstances of the plaintiff.

The Decision Not to Draw the Inference Is a Finding of Fact

A trial judge who recognizes the adverse inference is available and decides, on a review of all the evidence, not to draw it, has made a finding of fact. That finding attracts appellate deference. An appellate court may only overturn it if the trial judge made a palpable and overriding error, an obvious, serious error that goes to the core of the factual conclusion. The 4-3 split in Benhaim itself illustrates how high that threshold is: four justices found no such error; three found it clearly present. Appellate review of a trial judge’s decision not to draw the adverse inference is, in practice, difficult to pursue and rarely available.

Population Statistics Are Evidence, Not Compulsion

Statistical evidence about the probability of a particular staging at a given time, like the 78% figure for incidentally discovered lung cancers being at Stage I, is relevant and admissible evidence of causation. It may support the adverse inference. But it does not compel a finding of causation in an individual case. Population-level statistics must be weighed against the individual clinical evidence about the specific patient’s disease and circumstances. They are evidence for the trier of fact to assess, not a mathematical formula that generates causation at a defined probability threshold. And the trier of fact’s assessment of that statistical evidence is itself entitled to substantial appellate deference.

The Framework Applies Across Common Law and Civil Law

Benhaim confirms that the adverse inference described in Snell v Farrell applies across all Canadian jurisdictions, including Quebec. In Quebec, it operates as a presumption of fact under art. 2849 of the Civil Code. In common law provinces, it operates as the ordinary causal inference from circumstantial evidence. The underlying principles are substantively equivalent: both require the inference to be grounded in affirmative evidence and weighed against all the evidence, both leave the decision to the trier of fact, and both are subject to appellate review only for clear and fundamental factual error.

How Benhaim Fits Within the Causation Framework

Benhaim and Snell v Farrell

The relationship between Benhaim and Snell is one of extension and precision. Snell established that a causal inference may be drawn without positive or scientific proof of causation, in the absence of affirmative contrary evidence from the defendant, holding that causation is a practical question of fact answered by common sense. Benhaim takes the specific scenario where the defendant’s negligence created the evidentiary gap and asks whether, in that scenario, the inference becomes automatic. The answer is no. The practical wisdom of Snell, that the trier of fact must assess all the evidence, remains intact. Benhaim confirms that Snell created a tool, not a rule.

Benhaim and Clements v Clements

Clements v Clements, 2012 SCC 32, confirmed that the “but for” test must be applied robustly and pragmatically: that scientific certainty is not required, that inference is appropriate, and that the material contribution to risk exception is narrow and multi-defendant specific. Benhaim operates within the “but for” test as confirmed by Clements: the adverse inference is one way of satisfying “but for” causation where the evidence is incomplete, not an exception to it. A plaintiff who successfully invokes the adverse inference has satisfied the “but for” test, not avoided it.

The Layered Framework: Snell, Clements, Benhaim

Together, these three decisions describe a layered, coherent framework for causation in Canadian negligence law, including medical malpractice. Snell established the foundational principle: causation by inference, without scientific certainty, where the evidence and common sense permit. Clements clarified the structure: “but for” is the standard, applied pragmatically; material contribution to risk is a narrow exception for multi-defendant impossibility. Benhaim added the specific rule for evidence-destroying negligence: the adverse inference is a powerful tool available to the trier of fact when negligence has prevented the creation of causation evidence, but it is not a mandatory presumption and the decision to draw or decline it belongs to the trier of fact.

Key Cases Cited by Wagner J. in the Majority

• Snell v Farrell, 1990 CanLII 70 (SCC), [1990] 2 SCR 311: foundational adverse inference and causation framework; burden remains on plaintiff
• St-Jean v Mercier, 2002 SCC 15, [2002] 1 SCR 491: Quebec civil law application of Snell; causation under the Civil Code
• Clements v Clements, 2012 SCC 32, [2012] 2 SCR 181: “but for” as the primary test; robust pragmatic approach; material contribution as narrow exception
• Ediger v Johnston, 2013 SCC 18, [2013] 2 SCR 98: causal inference in obstetrical malpractice; recently confirmed that the trier of fact may draw an inference against a defendant who does not introduce sufficient contrary evidence
• Laferrière v Lawson, 1991 CanLII 87 (SCC), [1991] 1 SCR 541: legal and scientific causation are distinct; loss of chance not recognized; factual causation for legal purposes is for the trier of fact
• Housen v Nikolaisen, 2002 SCC 33, [2002] 2 SCR 235: palpable and overriding error as the standard for appellate review of findings of fact and inferences of fact

Practical Implications for Delayed Cancer Diagnosis Claims

What the Adverse Inference Does and Does Not Do

The adverse inference of causation, as defined by Benhaim, is a tool that helps plaintiffs in delayed diagnosis cases where the defendant’s negligence prevented gathering of staging evidence. It permits the trier of fact to draw a causal conclusion from incomplete evidence, in circumstances where the incompleteness arises from the defendant’s own negligence. It allows an inference to substitute for the scientific staging evidence that does not exist, when the evidence that does exist supports causation and no compelling contrary evidence has been adduced.

It does not guarantee a finding of causation whenever negligence prevented staging evidence from being gathered. It does not shift the burden of proof to the defendant. It does not compel a trier of fact to resolve ambiguity in the plaintiff’s favour when the defendant has adduced affirmative contrary evidence. And it does not override a trial judge’s careful assessment of the evidence on appeal without a palpable and overriding error.

Building a Case That Supports the Inference

For plaintiffs in delayed cancer diagnosis cases, Benhaim identifies the conditions under which the adverse inference is most powerful and most likely to be drawn.

First, the plaintiff must establish some affirmative evidence of causation, not merely the theoretical possibility that earlier diagnosis would have helped. Population-level statistics about staging probabilities for cancers discovered incidentally are relevant and legitimate, but Benhaim confirms they must be assessed against the individual clinical picture. Where nothing about the individual case makes it more likely that the patient was in the unfavourable minority, the statistical inference is stronger. Where individual clinical evidence is inconsistent with the statistical generalization, as the majority found here, the statistic alone is insufficient.

The plaintiff benefits most from the adverse inference when the defendant has adduced no compelling affirmative contrary evidence of alternative staging. Where the defence expert’s opinion about staging is highly speculative, dependent on retrospective reading of imaging, or rests on assumptions that are clinically uncertain, the absence of solid contrary evidence allows the adverse inference greater weight. The dissent in Benhaim emphasizes this: the three justices found that what the defence offered was hypothesis and speculation dressed as clinical expertise.

Individual survival evidence, namely the specific patient’s actual clinical course, how long they survived, when symptoms appeared, can be powerful corroborating evidence. As Fournier J.A. found, and as Côté J. agreed, Mr. Émond’s 31-month survival was fundamentally inconsistent with the staging hypothesis the defence was advancing. Evidence of this kind bridges the gap between population statistics and individual clinical reality in a way that can anchor the adverse inference to specific, verifiable facts about the plaintiff.

The Significance of Expert Evidence Quality

Benhaim confirms the irreducible importance of expert evidence specifically addressed to the question of staging at the time of the alleged negligence. The question is not what stage the cancer was at when actually diagnosed; that is known. The question is what stage it probably was at when it should have been diagnosed. Expert evidence that reconstructs, from tumour growth rates and elapsed time, what the cancer was probably doing during the period of negligent delay, and what probable staging at the earlier time would have meant for treatment and prognosis, is the most direct way of supporting the adverse inference with affirmative evidence. The Benhaim dissent illustrates the risks of relying on a defence expert whose reasoning, on close examination, was characterized as speculative throughout.

The Outcome and Its Human Dimension

The Supreme Court’s decision in Benhaim v St-Germain meant that Cathie St-Germain did not recover death damages for the loss of Marc Émond. The trial judge’s $70,000 damages award, for the anguish of the negligent handling of the treatment process, not for the death, was restored. The physicians who failed to investigate a shadow on a chest X-ray for over a year, who did not consult prior films as the radiologist had recommended, and who did not arrange the CT scan their colleague had suggested, were found negligent but not found to have caused the death that followed.

Four justices concluded that the evidence did not compel the inference that earlier action would probably have saved Marc Émond’s life. Three justices concluded the opposite, finding the trial judge had misapprehended the expert evidence and ignored compelling evidence of Mr. Émond’s survival trajectory. The proximity of the vote reflects genuine uncertainty at the heart of the case, an uncertainty that Benhaim does not resolve, but which, by articulating the legal framework precisely, gives future courts, practitioners, and patients the tools to navigate.

Conclusion: A Tool, Not a Rule

Benhaim v St-Germain is, at its core, a decision about institutional roles. It says that the trier of fact, the trial judge who hears the witnesses, assesses the experts, weighs the statistical evidence against the clinical evidence, and reaches a judgment about what the evidence most probably shows, is the right person to decide whether the adverse inference should be drawn. Not an appellate court reviewing a record. Not a rule that mandates the inference whenever its preconditions are met.

This respect for trial-level fact-finding is consistent with everything Snell v Farrell established about causation as a practical question of fact and everything Clements v Clements confirmed about the robust and pragmatic approach. The adverse inference is part of that framework, a tool that acknowledges what negligence does to evidence and allows courts to address it fairly, without converting fairness into a mechanical rule that bypasses judgment.

For patients and families in delayed cancer diagnosis cases: the adverse inference exists and may be powerful. It may be the difference between establishing causation and failing. Its application depends on the quality of the affirmative evidence supporting causation, the strength of the contrary evidence the defence produces, and the individual clinical picture as it relates to the population-level statistics. For lawyers: the work of building a strong causal case, namely specific staging evidence, expert evidence on tumour behaviour, evidence connecting the staging probability to the treatment that would have been available, remains essential. The adverse inference supplements that work in the specific circumstances where negligence has made complete evidence impossible to gather. It does not replace it.

Quick Reference: Benhaim v St-Germain

The Decision

• Citation: Benhaim v St-Germain, 2016 SCC 48 (CanLII), [2016] 2 SCR 352 — https://canlii.ca/t/gvk2h
• Decided: November 10, 2016 | Heard: April 28, 2016 | File No.: 36291
• Majority (4): Wagner J. (McLachlin CJ, Karakatsanis and Gascon JJ. concurring)
• Dissent (3): Côté J. (Abella and Brown JJ. concurring); agreed on the law; found palpable and overriding error in the trial judge’s factual assessment
• Outcome: Appeal allowed; trial judge’s decision restored; $70,000 damages for anguish only; no death damages; no causation established
• Appeal from: Court of Appeal for Quebec (Kasirer, Fournier and Bélanger JJ.A., 2014 QCCA 2207)
• Trial: Marcotte J., Quebec Superior Court, 2011 QCCS 4755

The Parties

• Appellants: Dr. Albert Benhaim (general practitioner) and Dr. Michael O’Donovan (radiologist)
• Respondent: Cathie St-Germain (in her own name, as tutor to her minor son, and as universal legatee of Marc Émond)
• The patient: Marc Émond, a 44-year-old non-smoker in excellent health in November 2005; died June 6, 2008, age 47; survived by Ms. St-Germain and their eight-year-old son

Key Facts

• November 9, 2005: Routine chest X-ray found a 1.5–2 cm ill-defined opacity in Mr. Émond’s right lung; Dr. O’Donovan recommended consulting prior X-rays and suggested a CT scan; Dr. Benhaim did neither, ordering only a follow-up X-ray
• January 17, 2006: Follow-up X-ray showed no change; Dr. O’Donovan suspected chronic lesion and recommended a third X-ray in four months; that third X-ray was never taken
• December 4, 2006: Annual physical; X-ray showed lesion had grown to approximately 2.5 cm; cancer now suspected; further tests ordered
• December 2006 / January 2007: CT scan and PET scan confirmed stage IV lung cancer, incurable; diagnosis January 2007; Mr. Émond still asymptomatic at that time
• February 2007 onward: Palliative chemotherapy; suspended June–November 2007 (Mr. Émond feeling well; biked over 5,000 km that summer); resumed November 2007; stopped February 2008 as unsuccessful
• June 6, 2008: Marc Émond died, age 47
• Trial damages: $70,000 to Ms. St-Germain personally and as universal legatee for anguish from negligent handling of treatment, not death damages
• Trial judge’s staging finding: cancer likely already at Stage III in November 2005 and likely incurable at that time
• Key statistical evidence: 78% of incidentally discovered lung cancers are at Stage I; Stage I has a cure rate of approximately 70%
• Key survival evidence: 31-month survival (14 months without treatment, 17 months with) vs. 8–12 month life expectancy for untreated Stage III/IV lung cancer

The Central Legal Question and Answer

• Question: Where a defendant’s negligence undermines the plaintiff’s ability to prove causation, and the plaintiff adduces some affirmative evidence of causation, is the trier of fact required to draw an adverse inference of causation?
• Answer (unanimous, 7-0 on the law): No. The adverse inference is available but not mandatory. It is a discretionary tool for the trier of fact, not a rule of law triggered by defined preconditions.
• The 4-3 split was entirely on the factual question of whether the trial judge’s refusal to draw the inference constituted a palpable and overriding error.

The Governing Principles

• Adverse inference of causation may be drawn where: (a) the defendant’s negligence undermined the plaintiff’s ability to prove causation; and (b) the plaintiff has adduced some affirmative evidence of causation
• The inference may discharge the plaintiff’s burden on the balance of probabilities, but is not triggered automatically by those conditions
• Whether the inference is warranted and how to weigh it are matters for the trier of fact, assessed against all of the evidence
• The legal burden of proof remains with the plaintiff throughout; the adverse inference does not shift it
• A trial judge who declines to draw the inference has made a finding of fact, entitled to appellate deference; only a palpable and overriding error justifies appellate reversal
• Population-level statistics are relevant but not determinative in individual cases; they must be weighed against individual clinical evidence; the trier of fact’s assessment of statistical evidence attracts appellate deference
• In Quebec: the adverse inference is a presumption of fact under art. 2849 of the Civil Code, requiring serious, precise, and concordant presumptions; this is substantively equivalent to the Snell approach in common law provinces

The Dissent (Côté J., Abella and Brown JJ.) on Palpable and Overriding Error

• Three specific errors in the trial judge’s factual assessment: (1) misapprehension of Dr. Ferraro’s evidence, in that she described his position as “impossible” progression when he said “possible but unlikely,” and she found he identified mediastinal involvement when he did so only once, in re-examination, which was a mistake; (2) omission of Mr. Émond’s 31-month survival trajectory, entirely inconsistent with Stage III/IV disease given the 8-12 month untreated life expectancy; (3) resulting failure to draw a serious, precise and concordant presumption under art. 2849 that Mr. Émond’s cancer was not at Stage III or higher in November 2005
• The dissent agreed fully on the legal framework; its disagreement was entirely about whether these three specific errors reached the palpable and overriding threshold

Relationship to Prior Causation Decisions

• Snell v Farrell, 1990 CanLII 70 (SCC) https://canlii.ca/t/1fstw — source of the adverse inference principle; causation as a practical question of fact; inference without scientific proof; burden on plaintiff
• Clements v Clements, 2012 SCC 32 https://canlii.ca/t/frvld — “but for” as the primary test, applied robustly and pragmatically; adverse inference is one way of satisfying “but for,” not an exception to it; material contribution to risk is a narrow multi-tortfeasor exception
• St-Jean v Mercier, 2002 SCC 15 — Quebec civil law application of Snell; causation under the Civil Code; presumptions of causation require serious, precise, and concordant foundation
• Laferrière v Lawson, 1991 CanLII 87 (SCC) — legal and scientific causation are distinct; loss of chance not recognized in Canada; factual causation is for the trier of fact
• Ediger v Johnston, 2013 SCC 18 — confirmed that the trier of fact may draw an inference against a defendant who does not introduce sufficient contrary evidence; recently affirmed the Snell approach in an obstetrical malpractice context