North Yorkshire County Council v SA & Ors

Where non-accidental injury is proved but the perpetrator cannot be identified on the balance of probabilities, the court must determine whether there is a real possibility that each proposed carer was a perpetrator. A test requiring proof that there is no possibility of a person’s responsibility is too wide, because it may include anyone with even fleeting access to the child.

The court should assess all relevant circumstances, not opportunity alone. It should identify possible perpetrators and relevant issues at an early stage, particularly before directing a split trial, so that affected persons can participate fairly.

An 11-week-old child suffered serious non-accidental injuries. The court also found that he had suffered an earlier non-accidental injury. In the first stage of split care proceedings under the Children Act 1989, Charles J found that the section 31 threshold was met but, applying a test of no possibility, declined to exclude either parent, the maternal grandmother or the night nanny from the pool of possible perpetrators of the later injuries.

All parties agreed on appeal that this was the wrong test. The central issues were the correct test for identifying possible perpetrators in an uncertain-perpetrator case and the proper disposal of the appeal, including whether the night nanny and grandmother should remain within the pool.

1. Appeal allowed unanimously. The judge’s finding as to the pool of possible perpetrators could not stand because it rested on the wrong test. The section 31 threshold itself remained satisfied: the child had suffered significant harm through two non-accidental incidents.

2. Where the evidence permits a positive identification of a perpetrator, the ordinary civil standard applies. Where it does not, the question is whether there is a likelihood or real possibility that each identified person was a perpetrator. In this context, likelihood and real possibility express the same test. A no-possibility test is patently too broad because it may encompass a person with only a fleeting opportunity to injure the child.

3. The court applied the approach in Re O and N (Minors); Re B (Minors) [2003] UKHL 18. Findings at the preliminary stage about possible perpetrators, and any judicial assessment of comparative likelihood, may properly inform the welfare stage. A person cannot be included or excluded solely by an elimination exercise; all relevant facts and circumstances must be assessed.

4. On the existing evidence, there was no real possibility that the night nanny caused the later injuries. She could not have caused the earlier injury; the judge’s assessment of her evidence was favourable; and it was highly improbable that two different people had separately inflicted such injuries in the child’s first 11 weeks. There was also no real possibility that the maternal grandmother caused either incident. Relevant matters included the medical timing evidence, the absence of any substantive allegation against her, and the procedural unfairness caused by her not having been a party at the first hearing.

5. Both parents remained possible perpetrators of both incidents. The mother’s evidence could not exculpate the father while the injuries remained unexplained. The court added that early case management should identify all relevant issues and possible perpetrators, and should consider whether a split trial will produce unfairness or require later reconsideration.

• Court of Appeal (Civil Division): Allowed the local authority’s appeal. It set aside the trial judge’s no-possibility approach and applied the real-possibility test to the night nanny and maternal grandmother.
• High Court, Family Division (Charles J): At the first stage of a split trial, found two non-accidental injuries and that the threshold under section 31 of the Children Act 1989 was met. The judge did not identify a perpetrator and declined to exclude the parents, grandmother or night nanny from the pool.