Article by Graeme Bentley, Principal Solicitor and Accredited Member of the Law Society’s Children Panel
The Court of Appeal in G (A Child: Scope of Fact-finding) [2025] EWCA Civ 1044 dismissed an appeal by the Local Authority and the Children’s Guardian against a refusal to hold a fact-finding hearing into the death of a mother’s first child, Z, six and a half years earlier. The case has immediate significance for social workers, lawyers, guardians and local authorities because it highlights the difficulty of persuading courts to investigate historic allegations where threshold and welfare planning can be supported on other grounds.
The decision was split, with Baker LJ delivering a powerful dissent. The divergence illustrates the continuing uncertainty around when fact-finding will be permitted and highlights the importance of tailoring submissions to the overriding objective and the Oxfordshire factors.
Background
The mother, 15 at first assessment, had a history of emotional dysregulation, aggression, and erratic behaviour. She gave birth to Z shortly after turning 16. Early assessments in a residential setting raised concerns about her focus on Z and possible drug use. A week after moving in with her grandmother, Z was found unresponsive and died aged five months.
Medical experts broadly agreed the cause of death was shaking, but the coroner returned an open verdict, and the CPS declined to prosecute.
When the mother gave birth to her second child, X, in January 2025, concerns remained about drug use, hostility to professionals, and her history with social services. Proceedings were issued at birth. Threshold included positive drug testing, mental health concerns and the history of Z’s death. The Guardian sought a paediatric overview of the earlier medical evidence and the Local Authority applied to expand threshold to seek findings about Z’s death.
Assessments during proceedings showed X was thriving in her mother’s care, with a strong bond, though support would be needed for at least 12 months and there were doubts about maternal engagement.
The judge at first instance concluded that fact-finding into Z’s death was unnecessary. He emphasised that safety planning and risk assessment could proceed without it; that delay, cost, and the evidential limitations of revisiting a 2018 death were powerful counterweights; and that focus should remain on present circumstances.
The majority decision
Peter Jackson LJ, delivering the lead judgment, described the decision as a core case management choice. Referring to Re H-D-H (Children) and Re C (A Child) [2021], he confirmed that the long-standing Oxfordshire approach (A County Council v DP [2005]) remains valid: the court should consider the child’s interests, time, cost, evidential yield, necessity, relevance to care planning, impact on parties, prospects of a fair trial, and overall justice.
These factors must be applied flexibly in light of the overriding objective in FPR 2010 r.1.1: dealing with cases justly, proportionately, and efficiently, with welfare at the forefront.
The majority held that the trial judge had correctly directed himself, was immersed in the evidence, and had conducted five case management hearings. His conclusion – that threshold and welfare assessments could proceed without resolving responsibility for Z’s death – was therefore one the Court of Appeal would not disturb unless “badly wrong.”
On risk assessment, the court referred to Re T (Risk Assessment) [2025] EWCA Civ 93. The judge could answer the four key questions – what harm might arise, how likely, consequences if it did, and how risks could be managed – by reference to the mother’s long-standing traits and her present parenting of X. It was not necessary to determine her role in Z’s death.
Proportionality was emphasised. Seven years had passed, making fresh fact-finding unlikely to shed new light. The process would be lengthy and costly, potentially hundreds of thousands of pounds, and would impose emotional strain on the mother and consequential impact on X. Meanwhile, progressive care planning had continued, with no present intention to separate X from her mother.
Although the judge accepted that fact-finding “might be helpful,” he was entitled to conclude it was not necessary.
The dissent
Baker LJ saw the case differently, regarding fact-finding as required. Drawing on Re H [1996], Re B [2008], and Re S-B [2009], he emphasised the binary principle: the court cannot proceed on suspicion or quasi-findings; the likelihood of harm must be based on proven fact.
For Baker LJ, the central question was whether X faced a real possibility of serious head injury in her mother’s care. That depended on whether the mother inflicted Z’s fatal injuries. Without such a finding, risk assessors would have to disregard Z’s death, leaving harm assessments incomplete and potentially unsafe.
He was critical of the idea that a finding of “impulsivity” sufficed. Many parents act impulsively; very few fatally shake a child. The risk category changes fundamentally if it is established that the mother had previously done so.
The dissent also applied the Oxfordshire factors differently: where the potential outcome of fact-finding is “of the utmost importance” to care planning, other factors such as delay or cost should rarely outweigh it. In his view, this was such a case.
Contrasting approaches
Bean LJ, concurring with Peter Jackson LJ, stressed that the trial judge had applied the law correctly and that an appeal court should not intervene where the balancing exercise was properly undertaken.
The split is instructive. The majority placed weight on proportionality, judicial discretion, and the adequacy of risk assessment without findings. The dissent insisted that without proven facts about past harm, welfare planning risked being unsafe.
The case contrasts with Re P and E (Care Proceedings: Whether to Hold Fact-Finding Hearing) [2024] EWCA Civ 403, where an appeal succeeded because the first-instance judge was seen to have pre-judged that injuries were accidental. There, evidence could support deliberate infliction; in G, the “top end” of possible findings was a momentary loss of control.
Practical guidance for practitioners
The case underlines that:
- Fact-finding is not automatic. Even serious allegations will not be investigated unless findings are necessary and relevant to threshold or welfare planning.
- Case management discretion is wide. Appeals rarely succeed unless the judge has gone “badly wrong.”
- Tailored submissions are essential. Practitioners must:
- Show specifically how the absence of findings would hinder risk assessment or care planning.
- Frame submissions within the overriding objective – proportionality, fairness, efficiency and child welfare.
- Address the Oxfordshire factors directly, particularly necessity and relevance.
- Single-issue cases differ. Where threshold depends on one disputed allegation, fact-finding is more likely. Where other concerns independently support threshold, courts may assess welfare without revisiting history.
- Expect judicial divergence. The split between Peter Jackson LJ and Baker LJ shows outcomes may turn on judicial perspective, especially where historic harm is balanced against proportionality.
Conclusion
G (A Child) reminds practitioners that the court’s task is not to resolve every disputed allegation but to do justice efficiently and proportionately. For social workers, lawyers, guardians and local authorities, the lesson is clear: success in arguing for (or against) fact-finding depends on linking the investigation directly to the child’s future welfare and demonstrating why it is, or is not, necessary to secure a safe and just plan.
Disclaimer: This article does not constitute legal advice and is provided for general information purposes only. Published 9 September 2025.
