It is a commonplace that the requirement of factual causation – as contained in the “but for” test – has a foundational role in negligence. The defendant is usually only obliged to make good harm to the claimant if his fault was a necessary condition for the same. Just as elementary is the principle that the burden of proof ordinarily rests upon the claimant to show causation on the balance of probabilities: it is he, not the defendant, who is impugning the status quo by seeking damages. Nonetheless, as is well known, there may sometimes be significant problems of causal uncertainty; in particular where – besides the risk posed by the defendant's faulty conduct – the claimant was exposed to other risks with the potential to cause the same harm. Here it may, after the event, be impossible for science to disentangle the different risks and say that, absent the conduct of the defendant, the claimant would probably have avoided injury. In this regard, a longstanding issue is how far the doctrine of “material contribution” may offer a legitimate alternative to the “but for” test, in permitting recovery on proof that the defendant's breach “materially contributed” to the harm (without the need to go further and show it would not otherwise have occurred). This question was at the heart of the decision of the Court of Appeal in the medical negligence case of Bailey v. Ministry of Defence [2008] EWCA Civ 883.

In Bailey, the claimant underwent an operation for a suspected gall stone at the Royal Haslar Hospital, run by the Ministry of Defence. Complications, including extensive bleeding, occurred, but afterwards she was simply returned to the ward and received little aftercare. Subsequently, her condition deteriorated, due both to continued bleeding and the development of pancreatitis. She was transferred in a critical condition to the intensive care unit of an NHS hospital, where she remained for ten days fighting for her life. Tragically, just when it appeared she was over the worst – and had been discharged from the ICU to a renal ward – she became nauseous and vomited after drinking some lemonade. Because of her weakened state she was unable to clear her air passages and choked. By the time she was resuscitated she had suffered brain damage.

Originally, the claimant joined the NHS hospital in proceedings, arguing that her care there was deficient. However, ultimately she abandoned that part of her claim and focused on events at the MoD hospital. Certainly, the lack of post-operative care had been negligent. The problem was causation. As noted, the immediate cause of the claimant's choking was her generally weakened condition. Nonetheless, this was contributed to not only by the inadequate aftercare, but by pancreatitis – a matter for which the defendant was not at fault. At trial, the experts were unable to say that “but for” the negligent care, she would probably have avoided the final catastrophe. The pancreatitis alone might have resulted in the same outcome.

In response, the Court of Appeal (upholding Foskett J.) held the claimant was nevertheless entitled to damages. Waller L.J., who gave the only judgment (Sedley and Smith L.JJ. concurring), categorised the case as one of “cumulative risk exposure”, in which the claimant was exposed to two sources of risk – inadequate aftercare, and pancreatitis – which combined to produce the harm. Here his Lordship, following dicta from Lord Rodger's speech in Fairchild v. Glenhaven Funeral Services [2002] UKHL 22, [2002] 1 A.C. 32, saw the decision in Bonnington Castings v. Wardlaw [1956] A.C. 613, as relaxing “but for” causation in favour of a test of “material contribution”. The latter permitted recovery “where medical science cannot establish the probability that ‘but for’ an act of negligence the injury would not have happened but can establish that the contribution of the negligent cause was more than negligible”, at [46]. In contrast, his Lordship, at [44], distinguished the earlier medical negligence case of Wilsher v. Essex Area Health Authority [1988] 1 A.C. 1074 (where the House of Lords had required “but for” causation) on the basis that it involved “alternative” sources of risk, operating mutually exclusively. At the same time, Waller L.J. disclaimed any suggestion that “policy factors” might have a role to play (as the implicit justification for liability in industrial disease cases such as Bonnington Castings and Fairchild, and its denial in Wilsher) asserting, at [46], that: “In my view one cannot draw a distinction between medical negligence cases and others”.

There are difficulties with the decision in Bailey. We may first note the decision's potential breadth of application: it appears to give free rein to claims against hospitals when a patient suffers harm bound up with their weakened physical state, provided this was borne upon by earlier negligence (see now also Canning-Kishver v. Sandwell and West Birmingham NHS Trust [2008] EWHC 2384 (QB)). As noted, the Court of Appeal justified this outcome by invoking the doctrine of “material contribution” as governing “cumulative risk” cases. Admittedly, there are dicta, particularly in Bonnington Castings and McGhee v. National Coal Board [1973] 1 W.L.R. 1 that support such an approach. Nonetheless, in an area where the House of Lords has notoriously vacillated in the line of cases from Bonnington Castings through McGhee and Wilsher to Fairchild and Barker v. Corus UK Ltd. [2006] UKHL 20, [2006] 2 A.C. 572, it cannot claim authoritative status. Indeed, by the time of Barker – a decision not referred to in Bailey – Lord Rodger was arguably alone in holding to it.

Secondly, the need to delimit the doctrine's application has encouraged a host of arcane distinctions in the case law, including that relied on in Bailey between “cumulative” and “alternative” risk cases. It is unclear, at least in indivisible harm cases, why this distinction matters. In both cases the underlying problem is identical – the impossibility of establishing that the defendant's breach was necessary for the harm. Arguably, the issues are better seen in evidential terms as going to the proper distribution of the burden of proof: should this remain on the claimant, or exceptionally be shifted to the defendant (to show the risk he created did not materialise)? It is true that an analysis in terms of proof reversals was rejected in Bonnington Castings, and that Lord Wilberforce's attempt to revive it in McGhee was disparaged in Wilsher. Nevertheless, this analysis emphasises that these cases do not touch the underlying substantive law requirement for causation (which remains the “but for” test), but concern how, on policy grounds, the courts may accommodate factual uncertainties in its application.

This brings us to the other main shortcoming in the decision in Bailey, which was the Court's failure to engage with policy concerns. Thus, despite Waller L.J.'s assertion to the contrary, there are, with respect, significant differences between industrial disease and medical negligence claims, justifying a more claimant-friendly approach in the former. There, typically, the claimant is exposed to risk factors that, even if the defendant is only at fault for one, all ultimately derive from the workplace environment. By contrast, in medical cases, the doctor intervenes on behalf of the patient to ward off natural risks (stemming from illness), and the treatment itself usually adds to the risks in play; furthermore (as Lord Hoffmann noted in Fairchild) in the case of NHS care, allowing recovery in doubtful causation cases will affect the resources available for other patients. Even though the claim in Bailey was not ultimately against the NHS, and notwithstanding that the case was a very sad one, it is submitted that the Court of Appeal's approach to resolving it is not sustainable.