England and Wales Court of Appeal (Civil Division) Decisions

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE FAMILY COURT AT BRISTOL
HH Judge Cronin
BS22C50071

		Royal Courts of Justice Strand, London, WC2A 2LL
		12 April 2024

This judgment was handed down remotely at 10.30am on 12 April 2024 by circulation to the parties or their representatives by e-mail and by release to the National Archives. .............................

LORD JUSTICE BAKER :

1. By separate notices of appeal, the parents of a little girl, hereafter called "A", challenge a finding made in care proceedings that, when she was a small baby, she suffered an abusive head injury inflicted by one or other of the parents.

Background

2. A, who is the first and so far the only child of her parents, was born on 19 December 2021. The family was unknown to social services and had no involvement with the courts before the event that triggered proceedings, which occurred when A was 2 months old.
3. The parents' account of the event, as recorded in the judgment, was as follows. A slept in a cot next to their bed on the mother's side. According to the mother, she was woken at about 4 o'clock by A in the early hours of 19 February 2022. Leaving A in her cot, she went downstairs to prepare a bottle of milk. She was away for about five minutes. As she was coming upstairs, she heard a scream, ran into the room, and found A in her father's arms, arching her back in pain. According to the father, when he woke he had gone to the bathroom and on returning he picked A up from the cot and took her into bed to settle her. She was arching her back and crying and, as the mother was coming upstairs, she screamed. A was unresponsive to her parents' attempts to rouse her so they called 111 and then 999.
4. A was taken to hospital by ambulance where, following an MRI scan, she was diagnosed with a subdural haematoma. She immediately underwent surgery, following which the treating surgeon reported that the bleed on her brain was the result of a ruptured cortical bridging vein. No other injuries to A's face or scalp were observed, save for a tiny scratch above her left ear which was not thought to have any relevance to the intracranial bleed. A skeletal survey confirmed that there was no evidence of acute or healing fractures and an ophthalmology assessment confirmed that there were no retinal haemorrhages or other damage to her eyes.
5. The opinion of the treating clinicians was that the injury could not be explained by a natural cause but rather had been caused by a traumatic event of some kind. The medical view was further that A's collapse due to the haematoma would have happened within minutes, up to one hour, after the damage to the cortical bridging vein and therefore whilst she was in the parents' care. The parents did not point to any such event and said that they did not know how the injuries were caused. They said that, at the time of A's collapse, she was in the care of the father upstairs at the family home while the mother was downstairs.
6. On 11 March 2022, the local authority started care proceedings and, on discharge from hospital, A was placed with her maternal grandparents under an interim care order. Over two years on, she remains living with her grandparents. With the local authority's approval, her parents have moved into the property and have supervised contact. Fortunately, A seems to have made a good recovery from her injury. There have been no concerns about the care she has received within the family or about her relationship with her parents. The parents worked closely with the local authority and with the medical professionals involved in A's care. Everyone accepts that, but for the injury sustained in February 2022, there would have been no need for the local authority to have any involvement with the family. We were told that there is no prospect of A being removed from her family's care. If the findings are upheld, there will be an assessment to determine whether A can be returned to the sole care of her parents. If it is concluded that she cannot be safely returned, it is currently agreed that she will remain in the care of her grandparents, possibly under a special guardianship order.
7. Although this was a case revolving round a single incident, the fact-finding hearing was not completed until twenty months after the start of the proceedings. At the hearing of the appeal, we did not explore the reasons for this, but the chronology included in the appeal papers sheds some light on what happened.
8. At a case management hearing on 29 March 2022, HH Judge Cronin, to whom the case had been allocated, gave permission for the instruction of three medical expert witnesses all regularly instructed in family court cases – Mr Jayaratnam Jayamohan, consultant paediatric neurosurgeon, Professor Stavros Stivaros, consultant paediatric neuroradiologist, and Dr George Rylance, a retired consultant paediatrician. Over the next few months, all three produced reports and in some cases addendum reports. A report was also filed by Dr Tim Hookway, consultant obstetrician and gynaecologist, following a further court direction.
9. On 31 August 2022, the judge listed the case for a 15-day fact finding hearing in February and March 2023. On 15 December 2022, an experts' meeting took place, attended by Mr Jayamohan, Professor Stivaros, Dr Rylance and Dr Tim Hookway. On 27 January 2023, the judge gave the mother permission to instruct a paediatric immunologist as an expert witness.
10. The fact-finding hearing started on 20 February 2023. The local authority sought a finding, in short, that A's head injury had been inflicted by one or other of her parents. Professor Stivaros gave evidence on 20 February, and Mr Jayamohan and Dr Rylance on 23 February. On 27 February, the report from the paediatric immunologist was filed. The mother gave evidence on 6 March and the father on the following day. On 9 March, the judge adjourned the hearing part heard. She also gave the mother permission to instruct a clinical geneticist as an expert witness. On 5 June 2023, the judge made an order with the consent of all parties authorising the instruction of Dr Russell Keenan, a consultant paediatric haematologist, as another expert witness. In the following weeks, the reports of the clinical geneticist and paediatric haematologist were filed. When the fact-finding hearing resumed on 17 July, the judge adjourned the hearing again and gave the parties permission to obtain an addendum report from Dr Keenan which was received in September. On 6 October, the fact-finding hearing resumed with Dr Keenan giving oral evidence and counsel giving closing submissions.
11. Judgment was handed down on 20 October, in which the judge found that the injury had been inflicted by either the mother or the father. At a hearing on 15 November 2023, the judge responded orally to a request for clarification, with corresponding changes made to the judgment in the form of track changes to the original draft. The track-changed version has been included in the appeal papers.
12. Following the hearing on 15 November, the judge's findings were recorded as follows:

(1) The injury occurred in the early hours of the morning of the 19 February 2022 at or soon after 4am.

(2) A had a large torn cortical bridging vein which could not have been caused by any identified natural process.

(3) A suffered an abusive injury in the course of an undisclosed traumatic event at about 4am on 19 February 2022. This would have been by way of an acceleration or deceleration action with or without impact against a soft object and it caused a significant injury to her head.

(4) There is no evidence that there was an undisclosed accident which caused her injury.

(5) One of the parents has lied about not knowing what has happened. Both parents deny any knowledge of a traumatic event. Neither parent's evidence on this point is accepted.

(6) The court did not find that it is more probable that either parent was responsible and so both remain potential perpetrators.

(7) A suffered significant harm in the form of an abusive head injury whilst in the care of both of her parents and that harm was attributable to the care being given to her not being what it was reasonable to expect a parent to provide.

(8) There was no evidence that either of them should have recognised that the other presented a risk before this. Neither of them failed to protect her from injury before February 2022.

13. At a further case management hearing on 4 December 2023, the judge refused the parents' oral application for permission to appeal. Various directions were made as to the filing of evidence to progress the proceedings to the welfare stage, including a risk assessment of the parents in light of the court's findings and a special guardianship assessment of the maternal grandparents. The matter was listed for an Issues Resolution Hearing on 22 March 2024.
14. The parents each filed a notice of appeal to this Court. On 16 January 2024, I granted permission to appeal and gave various directions including detailed directions for the preparation of transcripts of part of the evidence. In the event, and not for the first time recently in relation to an appeal in a family case, the transcripts were not completed in accordance with the directions. Given the option of an adjournment of the appeal or proceeding without the transcripts, the parties elected for the latter course. In the event, we have been able to reach a conclusion on the merits of the appeal on the basis of the material put before us, although, as will become apparent, the absence of transcripts has affected the decision about what should happen next.

The issue between the medical experts

15. Before considering the judgment, I shall refer to an important feature in the medical evidence.
16. In most respects there was a consensus among the medical expert witnesses. All ruled out the possibility that the intracranial bleeding was attributable to a birth injury, or a blood disorder, or any genetic factor, or an aneurysm, or infection (although there had been evidence on admission to hospital that A might have been suffering from an infection). They were agreed that it was caused by a traumatic event. They thought that the event had involved an element of acceleration and deceleration. There was, however, some disagreement about the extent of the intracranial injuries. In his report on the medical records, Mr Jayamohan advised that A had sustained a large left-sided acute subdural haemorrhage caused by the avulsed vein with subsequent pressure effects on the brain and midline shifts. Professor Stivaros, however, interpreted the CT scans and MR imaging as showing that, in addition to the large left-sided bleed from the avulsed bridging vein, which was seen to be pushing the brain away from the skull and displacing the brain so as to cause a "midline shift", there was subdural bleeding in multiple locations overlaying the brain, between the two halves of the brain, beneath the brain and in the posterior fossa overlying her cerebellum. He also noted evidence of non-haemorrhagic contusional injuries in the frontal lobes and the back of the brain.
17. At the experts' meeting, Mr Jayamohan and Dr Rylance deferred to Professor Stivaros on the interpretation of the imaging. However, it is accepted by all parties that in the course of their oral evidence a divergence of views emerged. It would have been helpful to see the transcripts of evidence to get a complete and precise picture of what was said, but for our purposes the summary in the judgment is sufficient.
18. Professor Stivaros adhered to his opinion that there were multifocal areas of subdural bleeding of a type which did not occur with trauma such as a blow or a fall but required multiple changes of energy at high velocity, typically through shaking. It was his view that the bleed from the avulsed vein on the left side of the brain did not explain the other sites of subdural bleeding. It emerged in the hearing, however, that the hospital neuroradiologist who had initially reported on the imaging, Dr Likeman, had not seen multifocal bleeding nor any contusional injuries. Mr Jayamohan, while continuing to say that he deferred to Professor Stivaros, told the court that he did not consider that there was multifocal bleeding in the pre-operative imaging. In cross-examination, he agreed that blood can track from one hemisphere to another, was "very clear" that the bleeding had originated from a single avulsed bridging vein, noted that the treating neurosurgeon "had only found one bleed", and accepted that the saline wash used after the operation would promote the spread of blood. He agreed that the contusional injuries might be the product of pressure caused by the large haemorrhage. He also acknowledged that a number of the features seen in abusive shaking or shaking plus impact injuries were not present here – no thin film multifocal bleeds, no retinal haemorrhages, no grip marks, no evidence of blood in the spinal cord (although it must be noted that unlike most such cases these days no MRIs were taken of the spinal cord in this case.)
19. Our attention was also drawn to a document prepared by Professor Stivaros at counsel's request during the fact-finding hearing. He was asked to identify all cases in which he had given expert evidence involving a child who had a space-occupying subdural bleed that necessitated neurosurgical referral and had a midline shift where the findings were thought to possibly represent a shaking episode with or without associated impact injury. He identified twenty such cases. In all of those cases, there had been bleeding in multiple locations. In twelve out of twenty, there had been retinal haemorrhages.

The judgment

20. After a brief introduction, including a summary of the legal framework, the judge summarised the issue before the court in these terms:

"The haematoma is considered by the medical witnesses to have been caused by a torn bridging vein and the consensus of medical opinion is that that must have resulted from some traumatic event. This is the issue contested by the parents."

She then summarised aspects of the social services and police investigation, including citing from the parents' police interviews, and then described aspects of the medical treatment given to A.

21. Under the heading "The Court's Role", the judge summarised the legal principles applicable in fact-finding hearings. At paragraph 51 of the first draft judgment, this included the following passage:

"Many cases require the court to consider something which appears improbable and compare two unlikely scenarios. These might include the possibility that a child has a condition hitherto unknown to medicine or that the cause of an injury is simply unknown."

The judge supported this by citing two criminal authorities in a footnote – R v Harris and Others [2005] EWCA Crim 1980 and R v Henderson [2010] EWCA Crim 1269. Following a request for clarification, she added the following sentence to paragraph 51 (which appears as a track changed amendment in the version included in our papers):

"I was asked to identify the paragraphs in Harris and Henderson which support these principles: both are obvious and acknowledged by medical and legal thinking: in Henderson, paragraph 208, in Harris (less clearly) paragraphs 207, 208, 211 and 232."

22. The judge then set out in some detail a summary of the expert medical evidence. This included an account of the difference in opinion mentioned above as to the presence of multifocal areas of bleeding. She also recorded the experts' consensus that none of the other possible causes explored on behalf of the parents explained the injury or injuries sustained by A. She considered other points raised in the evidence of the parents. They said that A's left eye had been fully or partially closed after birth accompanied by puffiness on the left side of her face. Both signs were said to be visible on photos produced at the hearing and both to have disappeared after the surgery. They also referred to her apparent habit of shaking her head violently from side to side. Another issue explored was the fact that A's paternal grandmother had a ruptured cerebral aneurysm in 2020. The judge recorded that all these matters had been considered by Dr Rylance and ruled out as having any relevance. The judge referred to the investigations carried out by the expert geneticist and haematologist, who had ruled out any relevant genetic or clotting disorders. The judge recited a number of factors identified by the parents' counsel as being unusual if A's injuries had been inflicted – the absence of any external marks, or any retinal haemorrhages, or (if Dr Likeman and Mr Jayamohan were correct) any multifocal thin film subdural bleeds. She noted the apparent absence of any factors in the parents' background commonly found in cases of non-accidental injury, and the fact that professionals had found them co-operative and honest. She added:

"They were devoted parents in the hospital and [the mother] kept a diary to explain what had happened to A when she is old enough to want to know which speaks to her love and determination to protect her."

23. The judge then set out her findings as to the cause of the injury. Following the request for clarification of her judgment, she amended this section by deleting some words and inserting some additional passages. These were shown in track changes in the version included in our papers. In the citations below I have repeated the deletions and underlined the additional passages.

"112. The anomalies highlighted by the parents' case do not overcome the consensus of the medical experts, supported by the treating doctors' description of what they found in surgery.

113. I find that A had a large torn cortical bridging vein which could not have been caused by any identified natural process. I am not satisfied that she had multi-locational bleeding. I did not find contusions present on the CT scan. Professor Stivaros thought he could see them, none of the other doctors or experts did. What was seen on the MRI scan was not certainly contusion.

114. I find, on the balance of probabilities, that A suffered an abusive injury in the course of an undisclosed traumatic event at about 4:00 AM on 19th February 2022. This would have been by way of an acceleration or deceleration action with or without impact against the soft object and it caused a significant injury to her head.

115. I need next to consider whether I can decide who is most likely to have caused the injury or whether both parents remain in the pool. The combined evidence of the parents is that A became unwell whilst her mother was downstairs, and her father was carrying her back to bed with him. The medical evidence is that her injury would have occurred at the time or very shortly before she was observed to be unwell.

116. The Court has to consider all realistic possibilities even if not raised on any party's case and Courts are always anxious to find an explanation that will result in the lowest level of interference with the family life of the parties. I have considered whether Tthere may have been an accident – she may might have been dropped if she wriggled when picked up by a tired parent, she may might have fallen from the "next to me" if it was not against the bed (the photograph shows it at an angle to the bed). On the other hand, either parent may might have shaken her or thrown her onto the bed in frustration – but I cannot speculate. The best evidence I have is the very experienced medical consensus that there must have been a traumatic event.

117. It might be more likely that an accident occurred than that either parent lost control and did something dangerously, but it is also more likely that a parent who made a mistake or had an accident would admit it and that a parent whose action was shameful, even if regretted, would try to conceal it by claiming not to know what caused the injury. On the balance of probabilities, I exclude undisclosed accident as a cause of her injury. A parent who injured a child accidentally or came upon a child who had had an accidental fall, would, on the balance of probabilities, acknowledge what had happened rather than continue to lie."

24. The judge recorded that each parent had ruled out the possibility that the other could have caused the injury. She continued (at paragraph 120):

"I have to remember that witnesses tell lies for many reasons. I have not found that either parent has told a specific lie, but I do find that one of them, at least, knows what happened and the other is at best naïve not to think that the other did something now that we have examined all the evidence. One of the parents has lied about not knowing what has happened. The evidence is that there was a traumatic event. Both parents deny any knowledge of a traumatic event. They do not have to prove anything. I have made a positive finding on other evidence which has the effect that I do not accept either parent's evidence on this point but I have not approached the finding from an assessment of their denials. I have considered their evidence very carefully but I have been satisfied on the balance of probabilities that the local authority has proved its case."

25. The judge then (at paragraph 122) considered but dismissed submissions made on behalf of the parents about the timing of the injury:

"Arguments for both parents draw my attention to the timing they describe: for the mother, the submission is that A became unwell when she was downstairs, for the father that he was only alone with A for five minutes in which he had to wake, get out of bed, go to the bathroom, come back and pick A up. The medical consensus was that she would have collapsed within minutes, at most an hour, of the injury being inflicted. Timing alone does not exclude either parent: the event may have been an accident that took moments only. The father was able to sleep through a feed and could have slept through something that happened just before he woke, the scream of pain may not have coincided with the impact but have been a reaction to the pain caused by the pressure of the developing haemorrhage. The mother's interview … says that she went downstairs, took a bottle out of the fridge, put it in hot water (did she boil a kettle?), and sat at her computer (to fill the time or because that is her seat?) and waited, once she had the bottle she turned the lights off, closed the door and as she landed on the second step heard the scream. Five minutes may have been a very optimistic estimate. Since at least one of them has not told me what happened I cannot found a conclusion about what happened on their evidence about timing."

26. The judge set out her conclusion as to the perpetrator of the injury as follows:

"123. I cannot distinguish between the parents unless I accept their evidence that the mother was downstairs when A became unwell and I can be satisfied that her injury was inflicted at the moment when she screamed and began to collapse. I found both parents apparently to be trying hard to help me as they gave their evidence. They were both likeable people who spoke about their good fortune in their baby and each other in glowing terms. I accept that their efforts to understand the evidence and mother's diary keeping are as consistent with anxious innocence as with trying to find a defence. I thought this was an unusual behaviour in the circumstances of being in hospital with a baby who was so ill and I had to consider whether the mother was trying too hard to make a case for herself as a good parent. The effect of this evidence on my finding is neutral. Counsel for the Local Authority pointed to minor uncertainties in Father's evidence and his own counsel recognised that the history the parents gave put Father with A when she became unwell. Without direct evidence of the traumatic event that must have taken place I cannot find that it is more probable that either parent was responsible and so both remain potential perpetrators. I was asked if there were other reasons for my finding that mother should remain in the pool of potential perpetrators other than opportunity and my finding that it was not more probable that one or other parent was responsible for determining that the mother remains as a potential perpetrator: there were none.

124. I therefore find that A suffered significant harm in the form of an abusive head injury whilst in the care of both her parents and that harm was attributable to the care being given to her not being what it was reasonable to expect a parent to provide."

27. Next, the judge rejected the local authority's argument that she should make a finding that the other parent failed to protect A, observing that there was no evidence that either of them should have recognised any risk before the incident that led to her collapse. Finally, the judge explained her reasons for rejecting earlier submissions that it was not in A's interests to continue the fact-finding hearing. It is unnecessary to consider this issue in the context of this appeal.

The Appeal

28. On behalf of the mother, Mr Charles Hyde KC leading Mr Jonathan Wilkinson put forward six grounds of appeal:

(1) The judge failed to consider the evidence in a holistic way but instead adopted a linear approach and concluded that A had suffered an abusive injury before giving any consideration to the parents' evidence or evidence of others about the care afforded to A/background generally.  

(2) The judge wrongly concluded that the medical evidence established that there must have been a traumatic event.

(3) The judge failed to consider the impact of her own findings on the medical evidence, namely that she was not satisfied that 2 out of 3 injuries identified by some medical experts were present.

(4) The judge attached too much weight to the medical evidence and failed to attach any, or sufficient, weight to the background evidence/evidence of the parents.

(5) The judge failed to set out or apply the proper test when concluding that the mother was in the pool of perpetrators.  

(6) The judgment was inconsistent in that the judge, on the one hand, expressly did not find that the parents told a specific lie but, on the other, found that one of the parents had lied about knowing what had happened.  Moreover, this finding, coupled with the erroneous conclusion that the medical evidence confirmed there must have been an abusive injury, formed the bedrock of the judge's conclusions that the mother should remain in the pool of perpetrators.

29. On behalf of the father, Mr James Tillyard KC and Mr Nathan Jones put forward eleven grounds of appeal:

(1) On the basis of the judge's overall conclusions in the initial judgment, she should have found that the causative event may have been an accident and she was unable to find that it was a traumatic event by way of an acceleration or deceleration injury with or without impact contrary to her finding at paragraph 114.

(2) The judge was asked for clarification as to the apparent inconsistency referred to in ground one and instead of clarification of the judgment, she changed its meaning.

(3) The judge was in error in determining without qualification that it was wrong to rule any particular mechanism or injury in or out on the basis of the presence or absence of symptoms seen in other cases (see paragraph 47 judgment).

(4) The judge erred when summarising Professor Stivaros' review of cases that he had been involved in where there had been a space-occupying subdural bleed that necessitated neurosurgical referral and had a midline shift.

(5) Having found that she was not satisfied A had multilocational bleeding (paragraph 113) or that there were contusions the judge should have considered the impact that had on her consideration of the medical evidence and the weight she should have attached to it.

(6) The judge fell into error in deciding that, as none of the medical experts could provide an explanation for A's presentation at hospital other than an inflicted traumatic event involving shaking (paragraphs 84 and 112), on the balance of probabilities that must be the cause, an error which Mr Tillyard characterised as the "Sherlock Holmes" approach.

(7) The judgment was not, as it should have been, an evaluation of all the evidence in a holistic manner but an exercise in considering the medical evidence alone and determining, on that evidence, that there had been an undisclosed traumatic event involving an acceleration or deceleration action with or without impact (paragraph 114) and so one of the parents must have lied about not knowing what had happened (paragraph 120).

(8) The judge was wrong to decide what had happened before deciding the credibility of the parents. This led to the erroneous and inevitable conclusion that they had not been truthful.

(9) On the evidence, the judge should have found that both the mother and father were credible witnesses who were telling the truth and that there was no reason to believe they had lied and she should then have placed that in the balance when making her findings.

(10) The judge was wrong to dismiss the parents' evidence that they had been concerned that there was something wrong with A's left eye on the basis that neither she nor Dr Rylance could see anything was wrong from the photographs.

(11) The judge has failed to engage with or consider the wider canvas and credibility issues which point towards this being an accident and point away from this being an inflicted injury.

30. There was plainly a considerable degree of overlap between these various grounds advanced by the two appellants, so in granting permission to appeal I asked their representatives to liaise to refine the grounds and arguments in support. That led Mr Hyde and Mr Wilkinson to summarise their primary criticisms in a way which broadly covers the gravamen of principal arguments put forward by both appellants:

(1) The judge erroneously considered the medical evidence in isolation.

(2) She converted the collective medical opinion into a "finding".

(3) Her conclusion that the medical evidence demonstrated that there must have been an abusive event was not a reflection of the medical opinion that she had received.

(4) In her analysis of the medical evidence, she failed to undertake any assessment of the impact of her not being satisfied of the existence of 2 out of 3 purported injuries, identified by some of the medical experts and assumed (out of deference) by others to exist.

(5) Her finding led her to the conclusion that one or other of the parents must be lying.

(6) In reaching that conclusion, she failed to undertake an analysis of the broad canvas or the specific evidence that each of the parents gave.

(7) By undertaking her analysis in this way, she reversed the burden of proof and/or put the onus on the appellant mother to exclude herself from being in the pool of possible perpetrators.

(8) The judgment lacked analytical rigour, was contradictory, and contained impermissible speculation.

31. On behalf of the local authority, Ms Libby Harris unsurprisingly emphasised the high hurdle faced by an appellant seeking to challenge findings of fact in this Court. Her primary argument was that the findings made by the judge were open to her having heard the evidence of the expert witnesses and the parents. This was a case where the "wider canvas" evidence was not in dispute. In particular, the high level of care provided by both parents was accepted on all sides and recognised by the judge at the start of her judgment. Ms Harris acknowledged the differences of opinion between the experts but submitted that their important overall opinion – that trauma was the likeliest explanation for the injury – remained unchanged. The judge had proceeded to consider the parents' evidence and their various explanations or suggestions as to how A's injury had been sustained. She had been entitled to form the view that she could not find them to be credible witnesses about the events leading to A's hospital admission. The local authority did not accept that the judge failed to apply the proper test when determining whether the mother was in the pool of perpetrators. Ms Harris submitted that the analysis in paragraphs 123-4 of the judgment was in compliance with the established principles. In short, the judge's findings were within her discretion and accordingly this Court should refrain from interfering.
32. In written submissions, counsel for A adopted a neutral position on the outcome of the appeal whilst stressing the guardian's concern that, whatever the outcome of the appeal, there should be no unnecessary delay in making decisions about the future care of A.

Discussion and conclusions

33. Before considering the merits of the appeal, there are three general observations about the hearing and the judgment which must be made.
34. First, it is highly unusual and in principle unsatisfactory for a fact-finding hearing to be adjourned for several months in the way that occurred in this case. The hearing started in February, was adjourned in March, adjourned again in July, and only concluded in October. I understand why the judge agreed to the adjournments. She wanted to leave no stone unturned in trying to find a medical explanation for the injury which A had sustained and acceded to applications for further expert assessment. We did not explore whether the judge was right to allow the additional assessments or the reason why the applications were made at such a late stage in the proceedings. It is, however, highly regrettable that the need for these additional experts was not identified at an earlier stage.
35. Secondly, most of the oral evidence was heard in February and March. So far as I am aware, no evidence was given in July and only one witness – the expert haematologist – in October. Thus the judgment was delivered over eight months after the key evidence had been given. On any view, this must have presented a considerable challenge to the judge. With the benefit of hindsight, given the fact that the hearing was being adjourned for several months, the very technical nature of the expert evidence, the disagreement between the experts which only became apparent during the hearing, and the crucial importance of assessing each piece of evidence in the context of all the other evidence, it would have been advisable to obtain transcripts of the evidence given in February and March by the experts and the parents. In passing, I would advise any judge in a family case who finds themselves in a similar position adjourning a fact-finding hearing for any length of time to order transcripts of the evidence. As it is, the judge did not do so and therefore had to make do with her own notes and recollections, supplemented by counsel's submissions.
36. Thirdly, whilst appreciating the judge's good intentions in amending the judgment in response to the requests for clarification, the final track changed version does not fully clarify ambiguities in the judge's reasoning but instead in some respects causes further uncertainty. The better course would have been to respond to the request for clarification in a separate document in so far as a response was merited in accordance with principles established in case law – see most recently Re YM (Care Proceedings) (Clarification of Reasons) [2024] EWCA Civ 71.
37. Having made those general observations, I turn to the matters raised on behalf of the parents. It is clear that, during a series of hearings extending over many months, the judge devoted a great deal of care and attention to this troubling case. It is therefore very unfortunate that in her final judgment there are a number of significant flaws in the judge's reasoning.
38. First, the judge made a clear finding that there was no multifocal subdural bleeding, nor any contusions present, and that the haemorrhaging was all attributable to the rupture of the cortical bridging vein. But the consequence of that finding, as noted in Mr Jayamohan's initial report, was that the likely cause of the ruptured vein was impact rather than shaking. Mr Jayamohan reported that he had seen large draining veins torn in the context of a shaking injury but only where "associated with clear findings of a shaking injury (i.e. injuries to other parts of the brain and accompanying changes within the posterior fossa and also described in the eyes)". That interpretation was supported by Professor Stivaros's review of twenty cases in which he had given expert evidence. In every case involving a child who had a space-occupying subdural bleed requiring surgical intervention which was found to possibly represent a shaking episode (with or without associated impact injury), the child had subdural bleeding in multiple locations. In twelve out of the twenty cases, the child also had retinal haemorrhages. The judge's finding that the injury was caused "by way of an acceleration or deceleration action with or without impact against a soft object" was therefore contrary to the weight of the evidence.
39. Furthermore, the consequence of the judge's finding was that she was left with a highly unusual medical picture. The child suffered a ruptured cortical bridging vein for which the likeliest cause was said to be impact. Yet it is striking that an impact sufficient to tear this major vein left no other sign of injury. I am not satisfied that the judge gave sufficient consideration to this conundrum before reaching her conclusion that the injury "must" have resulted from a traumatic event.
40. Secondly, having concluded (at paragraph 113) that A "had a large torn cortical bridging vein which could not have been caused by any identified natural process" but that she was "not satisfied that she had multi-locational bleeding", the judge then immediately concluded (at paragraph 114) that A "suffered an abusive injury in the course of an undisclosed traumatic event". She reached the conclusion that the injury was (a) as a result of a traumatic event and (b) inflicted abusively without giving due consideration to the totality of the evidence, including the evidence of and about the parents.
41. In the final, track changed version of paragraph 120, the judge wrote:

"The evidence is that there was a traumatic event. Both parents deny any knowledge of a traumatic event. They do not have to prove anything. I have made a positive finding on other evidence which has the effect that I do not accept either parent's evidence on this point but I have not approached the finding from an assessment of their denials."

It is not entirely clear what the judge meant by this passage, but it seems to confirm that the judge made her finding that the injury was both traumatic and abusive on the basis of the medical evidence without assessing the parents' denials.

42. I agree with the appellants' submission that this amounted to a "linear" approach. As Dame Elizabeth Butler-Sloss P observed in Re T (Children) [2004] EWCA Civ 558 at paragraph 33:

"…evidence cannot be evaluated and assessed in separate compartments. A judge in these difficult cases has to have regard to the relevance of each piece of evidence to other evidence and to exercise an overview of the totality of the evidence in order to come to the conclusion whether the case put forward by the local authority has been made out to the appropriate standard of proof."

This applies to expert medical evidence as to any other evidence. In A County Council v K D & L [2005] EWHC 144 (Fam) at paragraph 39, Charles J observed:

"It is important to remember (1) that the roles of the court and the expert are distinct and (2) it is the court that is in the position to weigh up the expert evidence against its findings on the other evidence."

Later in the same judgment, Charles J added at paragraph 49,

"In a case where the medical evidence is to the effect that the likely cause is non-accidental and thus human agency, a court can reach a finding on the totality of the evidence either (a) that on the balance of probability an injury has a natural cause, or is not a non-accidental injury, or (b) that a local authority has not established the existence of the threshold to the civil standard of proof … The other side of the coin is that in a case where the medical evidence is that there is nothing diagnostic of a non-accidental injury (or human agency) and the clinical observations of the child, although consistent with non-accidental injury (or human agency) of the type asserted, [are]more usually associated with accidental injury or infection, a court can reach a finding on the totality of the evidence that on the balance of probability there has been a non-accidental injury (or human agency) as asserted and the threshold is established."

43. In this case, the medical evidence pointed to the injury having been sustained in a traumatic event. But before reaching a conclusion on that point – and even more importantly, before finding that the injury was inflicted abusively – it was incumbent on the judge to consider the totality of the evidence, including, as Mr Tillyard submitted, the wider canvas evidence relating to these parents and how likely it was that they would injure their child, how possible it would have been for either parent to have injured the child under the circumstances without the other knowing, and the overall credibility of the parents and their account of what had happened. Instead, she found that, on the basis of the expert evidence, the injury occurred as a result of a traumatic event and that it was abusive.
44. Thirdly, the judge's treatment of the possibility that the injury was sustained accidentally was unsatisfactory. She only considered the possibility of an accident after expressing her conclusion that the injury was abusive and when she was considering whether she could identify the perpetrator. In the original draft of paragraphs 116 and 117, the judge raised the possibility that "there may have been an accident" without expressing a conclusion. In the final, track changed, version, she excluded undisclosed accident on the balance of probabilities, on the basis that "a parent who injured a child accidentally or came upon a child who had had an accidental fall, would, on the balance of probabilities, acknowledge what had happened rather than continue to lie." No basis for this assertion was advanced. Then a few paragraphs later, when considering submissions about the timing of the injury, she reiterated the possibility the event may have been an accident.
45. Fourthly, there is a flaw in the judge's treatment of the issue of timing. In paragraph 122, she briefly summarised the parents' case that the child collapsed during a five-minute period when the mother was downstairs and the father woke up, went to the bathroom, and, on returning, picked up the child. The judge discounted this point by making several speculative observations ("the event may have been an accident that took moments only"; the father "could have slept through something that happened just before he woke"; "the scream of pain may not have coincided with the impact but [may] have been a reaction to the pain caused by the pressure of the developing haemorrhage"; while the mother was downstairs, "did she boil a kettle?"; did she sit at her computer "to fill the time or because that is her seat?"; "five minutes may have been a very optimistic estimate"). She concluded by saying: "since at least one of them has not told me what happened I cannot found a conclusion about what happened on their evidence about timing." This was putting the cart before the horse. She ought not to have reached a conclusion that one or both of the parents had withheld information without considering all of the evidence, including their evidence about timing.
46. Finally, there is a flaw in the judge's reasoning in concluding that both parents remain in the pool of possible perpetrators. The correct approach to be followed by a judge seeking to identify the perpetrator of injuries to a child was summarised by Peter Jackson LJ in Re B (Children: Uncertain Perpetrator) [2019] EWCA Civ 575 at paragraph 49:

"The court should first consider whether there is a 'list' of people who had the opportunity to cause the injury. It should then consider whether it can identify the actual perpetrator on the balance of probability … Only if it cannot identify the perpetrator to the civil standard of proof should it go on to ask in respect of those on the list: "Is there a likelihood or real possibility that A or B or C was the perpetrator or a perpetrator of the inflicted injuries?" Only if there is should A or B or C be placed into the 'pool'."

In Re A (Children) (Pool of Perpetrators) [2022] EWCA Civ 1348, King LJ said (at paragraph 43):

"In my view the proper approach is not to seek to distinguish as between the possible perpetrators in order to see which one inflicted the injuries. Rather the proper approach is to consider each individual separately in order to determine whether that individual can be found on the balance of probabilities, to be the perpetrator."

47. At paragraph 120, the judge said:

"I have not found that either parent has told a specific lie, but I do find that one of them, at least, knows what happened and the other is at best naïve not to think that the other did something now that we have examined all the evidence. One of the parents has lied about not knowing what has happened."

At paragraph 123, she said:

"I cannot distinguish between the parents unless I accept their evidence that the mother was downstairs when A became unwell and I can be satisfied that her injury was inflicted at the moment when she screamed and began to collapse."

She concluded:

"Without direct evidence of the traumatic event that must have taken place I cannot find that it is more probable that either parent was responsible and so both remain potential perpetrators."

48. In my judgment, the judge did not consider each parent separately as she was required to do. She did not carry out a sufficiently detailed analysis of their evidence to form a clear view of their credibility and reliability She concluded that one of them had lied and the other was "at best naïve". Contrary to the course prescribed by King LJ in Re A, she asked herself whether she could distinguish between the parents. She concluded she could not "without direct evidence of the traumatic event that must have taken place". I am reminded of the observation of Peter Jackson LJ in Re A (A Child: Adequacy of Reasoning) [2019] EWCA Civ 1845. In that case, the only reasons given by the trial judge for the inability to identify which parent was the perpetrator of a child's injuries were that

"They were both in the flat with him at the time the injuries occurred. They have both lied. They are both protecting each other."

Peter Jackson LJ observed:

"That takes one nowhere. What was required was an analysis of the factors that pointed towards and away from each adult as being the perpetrator. If the result was an inability to identify, so be it, but the attempt had to be made."

I regret to say that I do not consider that the judge made a sufficient attempt to carry out that analysis in this case.

49. For those reasons, I conclude that both appeals must be allowed.
50. Ideally this Court should try to arrive at an outcome that enables the proceedings to come to an end today. Regrettably, I do not think we are in a position to do so. Without transcripts of the evidence, we cannot reach a conclusion as to whether the local authority case is proved. Even if transcripts were available, we would have to decide whether to proceed on the basis of the judge's finding that there was no multifocal bleeding. For my part, I feel uneasy about one court proceeding on the basis of some findings made by another judge while rejecting others.
51. In the circumstances, if My Lord and My Lady agree, I would remit the proceedings to the Family Presiding Judge for the Western Circuit, Judd J, to decide what course of action should now be followed. It may be that one or other party will propose that a rehearing of the fact-finding hearing is unnecessary and disproportionate in the circumstances. Alternatively, it may be suggested that the rehearing should go ahead before Judd J or another experienced family judge on a curtailed basis, relying on transcripts of some of the evidence given at the first hearing and confining the oral evidence to a limited number of witnesses. I make no comment as to which course should be followed, save to say that I hope that a way can be found to avoid a full rehearing which, even if it did not require the fifteen days originally allocated, would undoubtedly involve significant resources and lead to further delay before final decisions can be taken about A's future.

LORD JUSTICE NUGEE

52. I agree.

LADY JUSTICE FALK

53. I also agree.