Alexander v HCA International & Reffitt: £550,000 award after consultant failed to prescribe pre-ERCP suppository

Correction: An earlier version of this article carried the headline "hospital failed to administer pre-ERCP suppository." This was inaccurate. The claim against HCA International was dismissed; liability was found solely against the treating consultant, Dr David Reffitt, for failing to prescribe and communicate the instruction to nursing staff. The headline and standfirst have been updated accordingly.

A High Court judge has awarded £550,000 to a patient who developed life-threatening necrotising pancreatitis following an ERCP procedure at London Bridge Hospital, after finding that the treating consultant breached his duty of care by failing to instruct nurses to administer a Diclofenac suppository.

In Andrew Alexander v HCA International Limited & Dr David Reffitt [2026] EWHC 1284 (KB), HHJ Richard Roberts, sitting as a Deputy High Court Judge, found that Dr Reffitt had intended to prescribe 100mg rectal Diclofenac — a standard prophylactic measure against post-ERCP pancreatitis — but had failed to communicate this to the First Defendant's nursing staff. The claim against HCA International, the hospital operator, was dismissed.

A catastrophic chain of errors

Mr Alexander underwent an endoscopic retrograde cholangiopancreatography (ERCP) at London Bridge Hospital on 16 July 2019 to remove a stone from his common bile duct. Within hours of discharge, he presented at University Hospital Lewisham with severe abdominal pain. He was subsequently diagnosed with acute necrotising pancreatitis and spent nearly four months in hospital, undergoing multiple operations including a double-barrelled ileostomy, bowel resection and repeated surgical interventions for sepsis and haemorrhage.

He was left with permanently altered bowel function, insulin-dependent type 3c diabetes and a reduced life expectancy.

Consent and prescription failures

The judgement exposed compounding failures in both the consenting process and drug administration. Dr Reffitt had not ticked the box on the consent form indicating that the procedure would involve rectal administration of drugs — his signature overlapping that very section. All three expert gastroenterologists agreed this was below an acceptable standard of care.

Diclofenac was absent from the Drug Prescription and Administration Chart, the Short Stay Drug Chart and the Medication Discharge Summary. The only document in which it appeared was Dr Reffitt's own ERCP report, written immediately after the procedure — which the judge found could not constitute a valid prescription.

The judge found Dr Reffitt an unreliable witness. The consultant's detailed account of instructing Nurse Bouwer and observing her retrieve the suppository emerged for the first time during cross-examination at trial, more than six and a half years after the procedure, having been wholly absent from his witness statement.

Regulation 227 and widespread unawareness

A striking feature of the judgement was the court's finding that neither defendant displayed any awareness of Regulation 227 of the Human Medicines Regulations 2012/1916, which requires that directions for the supply of prescription-only medicines in hospital be given in writing. The hospital's Endoscopy Manager had characterised the absence of any prescription record as a "minor recording omission" — a characterisation the judge roundly rejected.

Causation and the statistical debate

On causation, the court accepted Professor Ian Gilmore's evidence that sub-group analysis of studies involving rectal Diclofenac specifically — rather than broader NSAID meta-analyses combining ineffective routes and less efficacious drugs — was the appropriate methodology. Across seven studies in the European Society of Gastrointestinal Endoscopy guidelines and five studies in Lyu et al., rectal Diclofenac consistently produced a risk ratio below 0.5, indicating a risk reduction exceeding 62%.

The court rejected the defendants' submission that the claimant was required to satisfy a 95% confidence interval, holding that the civil standard of balance of probabilities applied. The judge also found, in the alternative, that Diclofenac would have made a material contribution to reducing the severity of the pancreatitis.

In his closing remarks, the judge expressed concern that lessons had not been learned, noting that the systemic failures in record-keeping and prescribing practice at the hospital risked being dismissed as inconsequential when they were anything but.