B and G (Fact-Finding): Court of Appeal sets aside findings of inflicted injury over flawed reasoning

The Court of Appeal has set aside findings that two young children sustained non-accidental injuries, holding that the judge misinterpreted the neurosurgical evidence, left a crucial factual dispute unresolved, and failed to conduct any adequate evaluation of the parents' evidence.

In B and G (Fact-Finding) [2026] EWCA Civ 860, handed down remotely on 9 July 2026, Lord Justice Baker gave the leading judgement, with Lady Justice Andrews and Lord Justice Holgate agreeing. Joanne Brown KC and Daniel Sheridan appeared for the mother. Amanda Weston KC and Anna Hefford appeared for the father. Nicholas Goodwin KC and Lubna Rasul appeared for the local authority. Alison Moore appeared for the children's guardian, who supported the appeal on the perpetrator grounds.

Background

B, aged five and autistic, and his two year old sister G were examined in September 2024 after their mother sought a GP appointment because B seemed to be bruising easily. Both were found to have bruises, and imaging revealed chronic subdural haematomas in G's frontal and parafalcine regions.

Later genetic testing identified two heterozygous variants in both children, one in the Factor V gene and one in the HPS1 gene associated with Hermansky-Pudlak syndrome. Neither the treating haematologist nor the jointly instructed expert had encountered the combination. No research exists on its implications. Dr Keenan recommended genetic testing of the parents and the instruction of a geneticist. Neither step was taken.

HH Judge Kushner found in February 2026 that three bruises to B, the bruise to G's ribcage and the subdural bleeding were inflicted, that both parents fell within the pool of perpetrators, and that the perpetrator had failed to seek medical attention.

The judgement below

The judge reserved for six weeks, then delivered the judgement ex tempore from notes. After refusing permission to appeal she asked to see counsel's note of her own questions to the father and the grounds of appeal lodged with the Court of Appeal, and made substantive amendments to the transcript before approving it. The approved version reached the court the day before the hearing.

Baker LJ recorded considerable disquiet. Any amendment must represent a genuine expression of the judge's reasons rather than an ex post facto rationalisation. Having two versions risked confusion, as it did here, and extensive additions may obscure rather than clarify. Andrews LJ went further, endorsing the observation that some cases plainly call for a reserved written judgement circulated in draft under embargo, and describing the course taken as most unwise.

The reasoning

The central deficiency was identified by the guardian. The parents' evidence occupied half the hearing yet there was little mention of it in the judgement, no analysis of the impression the judge formed, and no indication whether she found either parent truthful. Saying that their evidence should be considered is not the same as considering it. That the judge found nearly all the protective factors in Re BR (Proof of Facts) [2015] EWFC 41 present made a careful analysis more, not less, necessary.

The judge also misinterpreted Mr Jayamohan, the neurosurgeon to whom the other experts deferred. She recorded him as concluding that the bleeding was non-accidental, and misattributed his "sitting on the fence" remark, which concerned the risk of subdural bleeding after forceps delivery through caesarean section. His actual position was that four aetiologies remained possible and he could not say which was more likely.

That did not preclude a finding of inflicted injury. As Charles J explained in A County Council v K, D and L [2005] EWHC 144 (Fam), a court may reach such a conclusion on the totality of the evidence even where nothing is diagnostic. But the analysis of the totality here was incomplete and flawed.

The judge accepted Dr Robinson's view that the September bruising was disproportionate, relying on foster carer logs which recorded only large bruises. She noted the criticism of those logs without explaining how she discounted it. She rejected Dr Dickens' clinical evidence as a snapshot without adequate reason, though Dr Dickens had examined the children. A twenty page chronology of bruising, and over a thousand pages of records from the period of round the clock supervision, went unaddressed. The foster carer's observation that B had bruises nobody could account for was arguably more significant than his incomplete logs.

Findings on inflicted injury and failure to seek medical attention were set aside. The case returns to Arbuthnot J to consider whether a rehearing is necessary.