Re A, B and C: Court of Appeal upholds mother as perpetrator of non-accidental fracture in care proceedings

The Court of Appeal has dismissed a mother's appeal against a finding that she inflicted a non-accidental fracture to her infant son's arm, affirming that the trial judge's assessment of witness credibility was both carefully reasoned and properly open to her on the evidence.

The proceedings concerned three young children whose parents had a prolonged and volatile relationship marked by repeated separations and reconciliations. The injury in question — an incomplete fracture of the proximate left humerus sustained by C, then aged approximately 15 weeks — was identified following an overnight contact visit with the father. Medical evidence, which neither parent challenged, established that the fracture was non-accidental, required significant force, and would have caused immediate and continuing signs of distress. The only realistic perpetrators were the parents.

HH Judge Magee, sitting in the Family Court at Portsmouth, found on the balance of probabilities that the mother had caused the fracture. That conclusion rested substantially on her assessment of the witnesses. The mother was found to be a poor witness, with numerous inconsistencies in her account of when she first noticed something was wrong with C's arm. A friend, LP, gave evidence that the mother had mentioned C's arm being floppy within minutes of arriving that morning — hours before the mother claimed to have noticed anything untoward. The mother failed to follow LP's advice to seek medical attention, did not mention that advice to the maternal grandmother when she called in the evening, and never contacted the father to ask whether anything had happened. This was particularly striking given that, only days earlier, she had promptly notified the social worker when C sustained a bruise during a previous contact visit.

In contrast, the judge found the father a more reliable witness, characterising differences between his written and oral evidence as an evolution of insight rather than inconsistency. She accepted the evidence of both the father and the paternal grandmother that nothing untoward was noticed with C's arm during the contact weekend, and was satisfied that, had the father harmed C, the paternal grandmother would have spoken out.

The mother advanced four grounds of appeal. It was argued that the judge had given excessive weight to inconsistencies in the maternal grandmother's evidence, insufficient weight to the paternal grandmother's minimisation of a violent incident involving the father, and failed to consider the wider canvas of the father's propensity to anger and aggression — including incidents involving a child, a family member and a traffic warden. A further ground alleged a failure to evaluate the evidence as a whole and to explain why the outcome selected was preferred over alternatives.

Baker LJ, with whom Arnold and Dove LJJ agreed, rejected each ground. The appellate court reaffirmed the well-established principle that assessment of witness credibility is quintessentially a matter for the trial judge. The judge had been fully aware of the father's history of anger and aggression and had not overlooked it; she had weighed it alongside his candid acknowledgment of wrongdoing and his demonstrated remorse. On the question of propensity, the Court noted that the authorities on similar fact evidence cited on the mother's behalf — including R v Mitchell [2017] AC 571 and R v P (Children: Similar Fact Evidence) [2021] 1 FLR 652 — were primarily directed at admissibility rather than weight, and had no direct bearing on the exercise performed by the judge. The judge's reasoning was found to meet the standard required: detailed enough to show the basis on which she acted, without being required to address every point raised by counsel.

Arnold LJ added an observation that, whilst the judge's witness-by-witness approach was fully open to her, a chronological analysis of events between 14 and 18 November 2024 would have been advantageous. Such an approach could have assisted in identifying both the probable perpetrator and the probable timing of the injury, which he noted are inter-related questions.

The appeal was dismissed.

Copyright & permissions