Why Causation is the Fight You Can’t Avoid

In virtually every serious personal-injury mediation we do, causation sits at the centre of the dispute. I decided to do a deep dive into basic principles.

The central question always is this: did the defendant’s negligence actually cause this plaintiff’s injuries and losses, or are those problems explained by something else? Canadian law starts with a simple idea, that the plaintiff must prove, on a balance of probabilities, that the harm would not have occurred but for the defendant’s negligence.

The Supreme Court has repeated that the “global but-for” test is the default. Only in rare, truly exceptional cases does the law permit a looser “material contribution” route. Clements v. Clements, 2012 SCC 32^[1] reaffirmed this. In that case, a motorcyclist lost control and crashed while carrying his wife as a passenger because the bike was overloaded and a nail punctured the rear tire, and the wife sued alleging her husband’s negligent operation caused her injuries. The decision tightened the gate on material contribution, and reminded courts to resist treating evidentiary uncertainty as a reason to water down causation.

Inferences, Common Sense, and Medical Uncertainty

You rarely get perfect medical proof. That’s not fatal. In Snell v. Farrell, 1990 CanLII 70 (SCC)^[2], the plaintiff underwent cataract surgery and sustained optic nerve damage resulting in near-blindness in one eye. The exact mechanism of the injury was medically uncertain, and the defence ophthalmologist could not say definitively whether excessive pressure during the surgery caused the harm. The Supreme Court held that in such circumstances, the trial judge was entitled to draw a robust and common-sense inference of causation from the overall evidence, particularly given that the defendant was in control of the procedure and the plaintiff had no pre-existing visual impairment. That is so long as the inference was logical and the defendant had a fair opportunity to rebut it. Plaintiffs rely on Snell because it turns “not 100% certain” into “still provable,” while defendants emphasize that it does not permit speculation or reverse the burden of proof.

The Supreme Court later stressed that courts must stay disciplined about who bears what burden. In Benhaim v. St-Germain, 2016 SCC 48^[3]
, the plaintiff’s husband died of lung cancer after his physician failed to order diagnostic testing despite persistent symptoms. Two of the defendant doctors testified, but a third physician, whose involvement and knowledge could have been critical, was not called by the defence. The trial judge drew an adverse inference from that omission and found causation on the basis that earlier diagnosis would have materially increased the deceased’s chance of survival. On appeal, the Supreme Court held that while adverse inferences may be appropriate where a party withholds key evidence, they cannot be used to reverse the burden of proof, and causation must still be established on a balance of probabilities based on the full evidentiary record. The Court used the case to reaffirm that causation is an intensely factual exercise, highly resistant to appellate interference, meaning the trial record you build usually decides your fate.

Pre-existing Conditions: Thin Skull, Crumbling Skull, and Apportionment

Pre-existing vulnerability doesn’t excuse negligence. Under the thin skull rule, the defendant takes the plaintiff as found. But where a condition was already deteriorating, the crumbling skull principle limits recovery to the acceleration or aggravation caused by the tort. I recently wrote a blog on this topic, “The Thin Skull Meets the Crumbling Skull: Mediation Strategies for Chronic-Pain Litigation”.^[4]

The Supreme Court’s leading synthesis is Athey v. Leonati, 1996 CanLII 183 (SCC)^[5]. In that case, the plaintiff was involved in two minor motor vehicle accidents and later suffered a disc herniation while stretching at the gym. The defence argued that the gym incident, combined with his pre-existing degenerative back condition, broke the chain of causation. The Court rejected that argument and held that once the tort is shown to be a cause beyond the de minimis range, the defendant is liable for the full loss unless they can prove that a distinct, measurable, non-tortious deterioration would have occurred in any event. The plaintiff must still connect the dots with credible medical evidence showing the accident materially contributed to the injury, but defendants cannot escape liability by pointing vaguely to age, susceptibility, or degeneration. If they want to rely on a “crumbling skull” theory, they must particularize it with real evidence rather than argument.

A companion principle from Blackwater v. Plint, 2005 SCC 58^[6], often cited in damages analysis, underscores that courts carefully separate what the tort actually caused from what pre-existed or would have occurred in any event. The case arose from historical abuse at a residential school, where the plaintiff alleged psychological and spiritual harm resulting from physical and sexual abuse by a school employee. The Supreme Court accepted that the abuse materially contributed to his psychological injuries but also found evidence of significant pre-existing trauma and life circumstances that would have caused some degree of dysfunction regardless of the defendants’ conduct. The Court held that even when liability is established, damages must be apportioned so the defendant is only responsible for the harm they actually caused, not for the entirety of the plaintiff’s misfortune. This remains the core of the “divisible vs. indivisible injury” distinction and the restraint against over-compensation.

Psychiatric Injury and Reasonable Foreseeability

Where the injury is psychiatric, Mustapha v. Culligan of Canada Ltd., 2008 SCC 27^[7] makes clear that foreseeability remains a distinct and necessary step in the analysis. In that case, the plaintiff developed a major depressive disorder, anxiety, and phobias after seeing a dead fly in a sealed bottle of Culligan drinking water he had purchased for his family. The Court accepted that the defendant’s negligence caused his psychiatric condition but held he could not recover because the type of harm he suffered was not reasonably foreseeable to a person of ordinary fortitude. Mustapha draws a sharp line between causation and remoteness. Causation asks whether the defendant’s conduct actually caused the injury, while remoteness asks whether that kind of injury was reasonably foreseeable in law. Plaintiffs still require reliable medical evidence linking their psychological symptoms to the triggering incident, while defendants will probe alternative life stressors, prior mental health history, or disproportionate reactions to argue that, even if causation is proven, the harm is too remote.

Ontario’s Threshold: Causation’s Gatekeeper in Auto Cases

For motor-vehicle torts in Ontario, the statutory threshold can quietly erase a plaintiff’s case even where causation is proven. The Court of Appeal’s “threshold trilogy” (Meyer v. Bright, Lento v. Castaldo, and Dalgleish v. Green^[8]) established that a plaintiff cannot recover general damages for pain and suffering unless they prove a permanent, serious impairment of an important physical, mental, or psychological function under s. 267.5 of the Insurance Act. In Meyer v. Bright, the plaintiffs suffered soft-tissue and psychological injuries after car accidents and the Court made clear that the threshold is not satisfied by pain alone. The impairment must be both caused by the accident and must seriously affect the person’s ability to work, care for themselves, or carry on daily activities in a meaningful way. In Lento v. Castaldo the Court stressed that intermittent symptoms and minimal functional interference do not meet the definition of “serious” or “important” even if some injury was caused by the crash. Dalgleish v. Green, reinforced that the trial judge, not the jury, decides the threshold motion after the verdict and is not bound by the jury’s implicit conclusions on causation or impairment. The trilogy confirms that causation runs through every branch of the test. The plaintiff may have genuine symptoms, but if they cannot prove those symptoms were caused by the collision and amount to a serious and permanent impairment of an important function, they recover nothing for non-pecuniary damages.

Kasap v. MacCallum, 2001 CanLII 7964 (ON CA)^[9] drives home two strategic points about threshold and causation in Ontario motor vehicle litigation. The plaintiff in Kasap claimed chronic pain and functional impairment following a rear-end collision, but her medical imaging was largely normal, her treatment was intermittent, and her evidence about daily functional limitations was inconsistent. The jury awarded damages, but on the post-trial threshold motion, the trial judge found she did not meet the statutory test for a permanent and serious impairment of an important function under s. 267.5 of the Insurance Act. The Court of Appeal upheld that decision, confirming that the trial judge may make independent findings on causation and impairment, even if they differ from the jury’s view, and that credibility, consistency, and functional impact matter more than diagnostic labels.

From my mediator’s chair, I see defence counsel rely on Kasap to show that minimal objective findings and weak functional evidence will not clear the threshold, especially where the plaintiff’s narrative is fragmented. Plaintiff’s counsel use the same case to argue that the judge must assess the entire evidentiary record, including lived experience and functional loss, rather than seizing on isolated “normal” test results.

The Statutory Deductible: When Causation “Wins” But the Money Vanishes

Ontario’s deductible for non-pecuniary damages can erase a jury’s number even if the plaintiff proves causation and clears the threshold. In Cobb v. Long Estate, 2017 ONCA 717^[10], the plaintiff was injured in a motor vehicle collision and the jury awarded him $50,000 in general damages, but after applying the statutory deductible under s. 267.5 of the Insurance Act, the net award for pain and suffering was reduced to zero. The Court of Appeal confirmed that deductibles must be applied to jury awards before any other adjustments and that prejudgment interest is calculated only on the net amount. In El-Khodr v. Lackie, 2017 ONCA 716^[11], which involved a catastrophic brain injury claim, the Court further clarified how statutory accident benefits (SABs) are to be deducted from the tort award, emphasizing that SABs offsets must be applied to each head of damage in a structured and methodical way. For plaintiffs, the message is brutal but practical. Even where causation is accepted, threshold is met, and a jury awards damages, a modest general damage figure can still be wiped out entirely by the deductible. For defendants, Cobb and El-Khodr provide a disciplined template for post-verdict arithmetic – combining deductibles, SABs deductions, and prejudgment interest – which makes marginal cases worth taking to trial when liability is strong but medical causation or severity of injury is weak.

The Court later clarified “silo” deductibility for overlapping statutory benefits in Cadieux v. Cloutier, 2018 ONCA 903, which often becomes part of the negotiation on heads of damage where causation is conceded but the quantum case is thin. Plaintiffs should model the set-offs early. Defendants should press precise silos and documentary proof of paid/owing benefits.

Evidence That Moves the Needle: Medical, Collateral, and Surveillance

Causation cases are won with functional evidence blended with medicine. Treating-practitioner notes that track work tolerance, daily activity, and symptom trajectories are often more persuasive than one-off IMEs. Plaintiffs should tie complaints to time-stamped functional change (before/after). Defendants will mine family doctor charts for pre-accident complaints, gaps in care, and inconsistent reporting.

Surveillance is a double-edged sword. When properly disclosed and used for impeachment, it can undermine a plaintiff’s causation narrative. When mishandled, it backfires. Iannarella v. Corbett, 2015 ONCA 110^[12] warns that undisclosed surveillance and trial-by-ambush will not be tolerated, and trial judges must police the limited purpose of such evidence. If you want surveillance to do heavy lifting on causation, comply strictly with disclosure and be ready to anchor short clips to clinical evidence rather than innuendo. See my Substack, “Cat and Mouse: Surveillance, Surprise, and the Ethics of Withholding Evidence”.^[13]

Material Contribution: The Trap for the Hurried

Counsel still try to plead “material contribution” as a fallback when experts are cagey. Clements makes clear that material contribution is a narrow exception for situations of circular causation and evidentiary impossibility where but-for proof is impossible and the defendant’s negligence created the very risk that materialized. If a standard but-for analysis is available, even if less than perfect, the exception does not apply. Plaintiffs should resist the siren song of the exception unless the facts truly fit. Defendants should be ready to show why ordinary but-for analysis is workable on the record.

Mediation Posture: How Causation Drives Settlement Ranges

Mediation is diagnostics, not just horse-trading.

On plaintiff’s side, come with a coherent chronology: pre-accident baseline, the index event, early complaints, treatment adherence, objective findings (where available), and a functional arc. Pin your theory to Snell and Athey. The Court can infer causation from a persuasive whole, and pre-existing fragility is not a defence. It only limits quantum through crumbling-skull analysis.

On defence, lead with Clements. Insist on but-for proof, highlight alternative causes (degeneration, subsequent incidents, occupational stressors), and quantify how Cobb/El-Khodr/Cadieux math curtails recovery even if causation is found.

Threshold and deductible should be priced explicitly. For borderline impairment cases, the defence will stress Meyer/Kasap to argue the judge can and will say “no threshold,” while plaintiffs should marshal functional losses tied to important life domains (work, self-care, family roles) and credible permanence opinions. Numbers should reflect the realistic odds that a judge either accepts or rejects threshold causation, and what the deductible does to a middling jury award.

Strategy Checklists

From the plaintiff’s chair, the story must be linear and documented. Use treating records to show before/after function. Lock your expert to Athey principles (thin vs. crumbling) and have them address alternative explanations head-on. Be explicit about permanence and importance for threshold and model the deductible impact so your demand isn’t fantasy-land if the jury returns a modest number. If psychiatric injury is in play, your expert should speak to Mustapha’s ordinary-fortitude foreseeability as well as medical causation.

From the defence side, press the timeline and the gaps. If but-for is your gate, don’t let the discussion drift into “possible contribution.” Tie the plaintiff’s complaints to documented pre-accident issues, ordinary life stressors, and post-accident non-compliance. If surveillance will be used, follow Iannarella. Disclose it properly, use it to impeach specific testimony, and resist turning it into a standalone medical opinion. Keep Cobb/El-Khodr/Cadieux in the foreground so that even if causation is conceded on some heads, the economics still make sense to settle low.

The Bottom Line

Causation in personal injury is not a metaphysical exercise. It’s a disciplined, evidence-driven narrative. The orthodox rule is but-for. Snell allows common-sense inference on a solid record. Athey keeps pre-existing conditions in their proper box. Mustapha reminds us that foreseeability matters for psychiatric injury, and Clements closes the backdoor when parties try to stretch material contribution beyond its narrow role. In Ontario auto cases, Meyer and Kasap keep judges in the driver’s seat on threshold, while Cobb, El-Khodr, and Cadieux tell you how much any win is actually worth after deductibles and benefits set-offs.

If you build your file with those realities in mind, mediation becomes a rational pricing exercise rather than a referendum on hope.