Causation — factual and legal

Causation is the hinge between breach and liability. By Week 6 in Durham Tort, you should already be able to identify a duty of care, analyse breach, and avoid treating negligence as moral blame alone. The present topic asks a different question: assuming the defendant owed a duty and fell below the required standard, did that breach cause the damage for which the claimant seeks compensation?

The orthodox structure has two stages. First, factual causation asks whether the defendant’s breach made a difference to the occurrence of the damage. The ordinary test is the ‘but for’ test: would the damage have occurred but for the defendant’s breach? If the answer is yes, the breach is not a factual cause. If the answer is no, factual causation is established. Barnett v Chelsea and Kensington Hospital Management Committee remains the standard illustration: the casualty doctor was negligent in failing to examine the patient, but the patient would have died from arsenic poisoning in any event.

Secondly, legal causation asks whether responsibility should extend to the damage even though factual causation is established. This includes questions sometimes labelled remoteness, scope of liability, novus actus interveniens, and the effect of intervening natural events or third-party conduct. The point is not metaphysics. It is attribution. Tort law must decide whether this defendant should bear this loss.

The difficulty is that the two stages are not hermetically sealed. In many hard cases, especially industrial disease, medical uncertainty, loss of chance, and informed-consent litigation, the courts alter or relax the ordinary causal inquiry for reasons of policy and corrective justice. McGhee, Fairchild, Barker, Sienkiewicz and the Compensation Act 2006 form a distinct line. They do not abolish proof of causation generally. They respond to a particular evidential injustice: the claimant can prove negligent exposure to a causal agent, but medical science cannot identify which exposure produced the indivisible disease.

For Durham assessment, causation is a high-yield topic because it connects with nearly every earlier negligence theme. A public authority may have been in breach, but did its failure cause the harm? A doctor may have breached a duty to warn, but did the warning failure cause the injury? A defendant may have increased a risk, but is risk itself actionable damage? Good answers keep breach, factual causation, legal causation and damages analytically separate, while recognising that judges sometimes use causation language to control the reach of liability.

English tort law did not begin with a single, technical doctrine of causation. Older forms of action often treated the connection between wrong and harm as a matter of directness, common sense, and pleading. The modern law of negligence, however, required a more disciplined account. Once Donoghue v Stevenson generalised negligence around duty, breach and damage, courts needed a method for deciding whether a negligent act was sufficiently connected with the loss.

The first major modern movement was the separation of factual causation from remoteness. The Wagon Mound (No 1) replaced the older Re Polemis directness test with reasonable foreseeability as the governing test for remoteness in negligence. This did not answer the factual question whether the breach caused the damage, but it made clear that causation and remoteness are conceptually distinct: a defendant may cause damage which is too remote, and may breach a duty without causing any actionable damage.

Barnett gave the modern law its canonical factual test. The hospital casualty department negligently sent home three night-watchmen without examination. One died from arsenic poisoning. The claim failed because proper treatment would not have saved him. The case matters not because it is medically dramatic, but because it demonstrates that negligence is not liability insurance. Tort compensates damage caused by the breach, not damage merely coincident with negligent behaviour.

The second historical movement concerns evidential uncertainty. In McGhee, the claimant contracted dermatitis after working in brick kilns. The employer had negligently failed to provide washing facilities, so the claimant cycled home covered in brick dust. Medical science could not prove that the additional period of exposure caused the dermatitis. The House of Lords allowed recovery on the basis that the breach materially increased the risk. McGhee was later confined by Wilsher, where a premature baby’s blindness could have been caused by one of several agents, only one of which was connected with the defendant’s negligence. Wilsher insisted that increasing the risk was not normally equivalent to causing the damage.

Fairchild then revived and refined McGhee for mesothelioma. Mesothelioma is an indivisible disease caused by asbestos exposure, but science could not identify which employer’s exposure had triggered it. The House of Lords allowed recovery where each defendant had materially increased the risk of the very disease. Barker made liability several and proportionate to risk, but Parliament reversed that aspect for mesothelioma through section 3 of the Compensation Act 2006.

The third movement is the growing awareness that causal doctrine is doing more than science. Cases such as Chester v Afshar, Gregg v Scott, Bailey v Ministry of Defence and Corr v IBC Vehicles show courts confronting the limits of but-for causation in medical, evidential and responsibility-based settings. The result is not a single exception-ridden code. It is a structured common law in which the ordinary but-for test remains central, but exceptional doctrines are deployed where strict proof would defeat the law’s own protective purposes.

1. Damage is an element of negligence, not merely a consequence. Negligence is not complete on breach alone. A claimant must prove actionable damage and a causal connection between breach and that damage. This is why causation should be addressed after breach, not fused with it. In a problem question, write: duty, breach, factual causation, legal causation and remoteness. Do not say that because the defendant was negligent, the defendant is liable.

1. The ordinary factual test is ‘but for’ causation. Ask whether the damage would have happened in any event if the defendant had taken reasonable care. The inquiry is counterfactual. The court constructs the hypothetical world in which the defendant was not negligent and compares it with the actual world. If the claimant would have suffered the same injury at the same time, factual causation fails. Barnett is the model. The defendant’s negligence was real, but it did not cause the death.

1. The claimant bears the burden of proof. In civil litigation the standard is the balance of probabilities. The claimant must show that it is more likely than not that the breach caused the damage. A 49 per cent chance is insufficient; a 51 per cent probability is enough. This explains the harshness of loss-of-chance cases. In Hotson, the claimant’s hip injury had probably already made avascular necrosis inevitable before negligent diagnosis. In Gregg v Scott, a reduced chance of cure from cancer was not recoverable as such in an ordinary clinical negligence claim.

1. Multiple sufficient causes create difficulty. The but-for test can under-identify causes where two independent sufficient causes operate together. If two fires merge and destroy a house, neither fire may be a but-for cause if each would have destroyed the house alone. English law deals with such cases pragmatically, often through material contribution to damage, concurrent causes, or the language of substantial cause. The but-for test is a tool, not a metaphysical definition.

Barnett is the starting point because it demonstrates the negative function of causation. The hospital’s failure to examine the patient was negligent, but it did not cause the death. The case prevents negligence from becoming liability for all adverse events occurring after breach. It is also a reminder that factual causation is usually a medical or evidential question, not a moral one.

Bonnington Castings introduced the language of material contribution to damage. The claimant’s pneumoconiosis resulted from inhaling dust from both negligent and non-negligent sources within the defendant’s factory. Because the negligent dust materially contributed to the disease, causation was established. This is not a risk case: the negligent exposure formed part of the actual causal process.

McGhee is more controversial. The absence of washing facilities lengthened the claimant’s exposure to brick dust. The court could not say that the extra exposure caused dermatitis, but held that a material increase in risk was enough. Later courts struggled to define whether McGhee was a true causation exception or merely an evidential inference. Wilsher rejected any broad reading. The premature baby’s blindness could have resulted from several different causal agents. Negligent excess oxygen was only one candidate. The claim failed because the defendant had not been shown, on the balance of probabilities, to have caused the injury.

Fairchild is the central modern exception. Several employers had negligently exposed employees to asbestos. The claimants developed mesothelioma. Medical science could not identify which exposure triggered the disease. The House of Lords held that a defendant who materially increased the risk of mesothelioma could be liable. The reasoning is explicitly exceptional. It responds to a combination of negligent exposure, a single noxious agent, indivisible disease, and evidential impossibility.

Barker altered the remedial consequence by holding that liability should be proportionate to the risk created. Parliament then reversed that result for mesothelioma through the Compensation Act 2006. Sienkiewicz confirmed that, in mesothelioma cases, the Fairchild exception survives even where the tortious exposure is small compared with environmental exposure, provided the increase in risk is material rather than de minimis.

Hotson and Gregg are the main loss-of-chance controls. In Hotson the claimant could not recover for losing a 25 per cent chance of avoiding avascular necrosis because the injury had probably already made that outcome inevitable. In Gregg the House of Lords refused to compensate a reduction in the chance of survival in ordinary clinical negligence. These cases show that English tort generally requires proof of causation of damage, not merely proof of reduced prospects.

Chester v Afshar is a separate informed-consent exception. The surgeon negligently failed to warn of a small inherent risk. The risk materialised. The claimant could not show she would never have had the surgery, only that she would have deferred it. A majority allowed recovery to vindicate the duty to warn. The case is important, but dangerous: it should not be generalised into a broad principle that breach of an autonomy-based duty always supplies causation.

Corr v IBC Vehicles illustrates legal causation. The defendant’s breach caused severe physical and psychiatric injury; the claimant later died by suicide. The House of Lords held that the suicide did not break the chain. It was a consequence of the depressive illness caused by the defendant. The case shows that intervening acts are not automatically novus actus merely because they are deliberate.

See practice questions section below.

See further reading section below.