Schembri v Marshall

In medical-negligence causation, a claimant need not establish the precise counterfactual mechanism by which timely treatment would have prevented death. After examining proposed mechanisms, the court may assess the evidence as a whole and decide whether survival was more likely than not. That assessment may draw on statistical evidence, expert evidence and the individual patient’s condition and likely treatment. Statistics are neither determinative nor to be rejected; they must be given appropriate weight and tested against particular evidence. The counterfactual cannot be assessed solely by assuming that events occurring at home, without monitoring or treatment, would have occurred in hospital. An appeal is not a retrial, and an appellate court should not undermine an overall trial assessment by selecting isolated passages from the evidence.

The appellant general practitioner admitted breaching his duty by failing to refer the deceased to hospital when she presented with symptoms of pulmonary embolism. The issue at trial was whether prompt referral would probably have led to her survival. Mr Justice Stewart found for the respondent and ordered payment of £260,000 plus costs: [2019] EWHC 283 (QB).

The appellant argued that the judge’s findings on the possible treatment mechanisms required dismissal of the claim and that general survival statistics could not replace individual analysis. The respondent defended the judgment and argued alternatively that the evidence should have led to a finding that treatment in hospital would probably have saved the deceased. The central issue was whether causation could be established from the evidence as a whole despite the absence of proof of a precise mechanism of survival.

1. Appeal dismissed. McCombe LJ gave the leading judgment, with Holroyde and Phillips LJJ agreeing.
2. The trial judge was entitled to separate the inquiry into possible specific mechanisms of survival from the final assessment of causation. A claimant need not prove the precise mechanism by which timely treatment would have prevented death. The question remains whether, on the balance of probabilities, the deceased would have survived had she been referred to hospital.
3. Where no specific mechanism can be positively established, the court may assess the evidence as a whole. Relevant matters include expert evidence, medical literature, the patient’s age and medical condition, likely haemodynamic status, and the monitoring, support and treatment that would have been available in hospital. This approach was consistent with the reasoning in Drake v Harbour [2008] EWCA Civ 25.
4. Statistical evidence has a legitimate role in medical-negligence causation. It is not determinative and must not obscure evidence placing the individual patient in a particular category. Its weight depends on its quality, comparability and clarity, together with the particular medical evidence. The treatment of statistics in Wardlaw v Farrar [2003] EWCA Civ 1719 did not require a different result, since that case concerned a patient whose response to treatment was available for assessment. The approach was also consistent with the reasoning of Lord Nicholls in his dissent in Gregg v Scott [2005] 2 AC 176.
5. The counterfactual hospital course could not be assessed solely by assuming that the events occurring at home, while the deceased was untreated and unmonitored, would have occurred in hospital. The judge was entitled to conclude that her prospects of survival in hospital were very high and that causation was proved.
6. The appeal was not a retrial. The appellate court should not engage in selective island hopping through the evidence where the trial judge had assessed the factual, expert and statistical material as a whole. The principle in Fage UK Ltd v Chobani Ltd [2014] EWCA Civ 5 applied. The respondent’s alternative argument concerning cardiogenic shock did not require determination because the trial judge had not been shown to have erred.

• Court of Appeal (Civil Division) — dismissed the appeal from the causation judgment.
• High Court of Justice, Queen’s Bench Division — Mr Justice Stewart found that the admitted breach caused the deceased’s death and ordered payment of £260,000 plus costs: [2019] EWHC 283 (QB).