B and G (Fact-Finding)

By separate appeal notices, a mother and father appeal against findings of fact made by a judge in care proceedings concerning their two children, a boy now aged 5 and a girl aged 2. In this judgment, I shall refer to the boy as B and the girl as G.

The relevant background is as follows. The mother and father met in 2018 and lived together for six years. In September 2020, the mother gave birth to B. He has been diagnosed as autistic and also suffers from a form of diabetes. He is said by his parents to be an active little boy who is accident prone. The medical records show that his mother had sought medical attention for B on several occasions but prior to the events which led to these proceedings there had been no suggestion that he was being abused or harmed.

In February 2024, the mother gave birth to G by Caesarean section, after a difficult labour during which forceps were used.

By this stage, the parents’ relationship was deteriorating. There were repeated arguments and incidents of domestic abuse. In the Spring of 2024, they agreed to separate and the father moved out of the family home, but they continued to share responsibility for looking after the children.

On 8 September 2024, the mother made an online request for a GP appointment for B. She wrote:

“B has a diagnosis of autism, sensory processing disorder and speech and language delay. B has no concept of danger and can be a clumsy child. I’ve noticed within the last few weeks he seems to be getting bruises on his legs and other areas very easily after a small bump and some bruises without injury which is causing me concern.”

On 11 September 2024, the mother took B for his appointment with the GP. B was referred to hospital where he was found to have sustained bruises which the treating doctors suspected had been sustained non-accidentally. Child protection medical examinations were performed on both children. They revealed that B had sustained bruises to his forehead, torso, back, pelvis and thigh, and that G had sustained a bruise over her ribcage. CT and MRI scans revealed that G had also sustained subdural haemorrhages in her frontal and parafalcine regions.

The parents were arrested, interviewed by the police and released on bail. Shortly afterwards, the father left the family homeand returned to live with his parents. After a strategy meeting, the children were discharged from hospital into the care of their maternal grandparents. The following day, B was readmitted to hospital after staff at his nursery observed further bruising. He was discharged again into the grandparents’ care. The following week, he was again admitted to hospital after the nursery observed bruising and then again, on 4 October 2024, returned to his grandparents.

Following a family group conference on 22 October 2024, the local authority agreed that the children could return to the mother’s care with the support of 24/7 cover by sessional workers. At that stage, the father was still living at his parents’ home but visiting the family home regularly, including staying overnight on occasions to assist in caring for the children. This arrangement continued for several months while the local authority carried out an assessment under section 47 of the Children Act 1989. The parents’ care of the children was recorded in detail by the support workers present in the home.

In December 2024, the parents took the children to a haematology clinic at Addenbrooke’s Hospital where they were assessed by a team under Dr Emmy Dickens, consultant paediatric haematologist. The haematological evidence in this case is complex and considered in detail below. Early tests revealed that both children had a mild Factor V deficiency, but at that stage Dr Dickens found no evidence of a tendency to spontaneous bleeding or bruising with normal handling.

During the period of 24/7 supervision, one incident occurred which featured in the evidence before the judge. It was alleged that, during the night of 9 January 2025, when B was awake and refusing to go back to bed, his father pinched his ear. Save for that incident, no concerns were noted about the care provided to the children by their parents. The sessional workers kept extensive notes of their observations during this period of 24/7 supervision, amounting to over 1,000 pages. In March 2025, the father moved back into the home to assist in caring for the children. The parents did not resume their relationship and they continued to live in the same property separately until early 2026 when the father moved out permanently.

At the conclusion of the section 47 assessment, a child protection conference concluded that care proceedings should be started. On 23 March 2025, the local authority duly filed an application under Part IV of the Children Act. At a contested hearing before HHJ Kushner on 16 April 2025, the children were made subject to interim care orders after which they were removed from the family home and placed with foster carers.

Case management directions were given for statements from the parents which were duly filed on 21 May in which the parents described their relationship, their personal circumstances (the mother’s mental health issues and the father’s history of substance abuse), their care of the children, the bruises seen on the children and their suggestions as to how they might have been caused. They described how G would roll onto her dummy which had a clip attached, and suggested that this might have caused the bruise to her chest.

Both parents also gave evidence in similar terms about G’s subdural bleeding. In her statement, the mother said:

“I was not aware that G had 2 subdural haematomas until the CP medical was sent to my legal representative. She has never presented with any symptoms. I do not recall any event, accident or time that G would have hurt her head. The only time that I can think of was when B threw his iPad and it accidently hit G on the side of her forehead. G was in her bouncy chair and my uncle and cousin were present. B does not have spatial awareness and threw out of nowhere the iPad which hit the right side of her face above the eyebrow before the hairline. G cried initially but nothing extensive, I knew that she was ok because she did not seem fazed by it. She was not sleepy nor distressed. There was no mark left either so I did not feel medical attention was needed. This occurred around July/August 2024.”

At a further case management hearing in May, a fact-finding hearing was listed in December 2025, and directions given for expert medical reports from Mr Jayaratnam Jayamohan, a consultant neurosurgeon, Dr Russell Keenan, a consultant haematologist, and Dr David Robinson, a consultant paediatrician.

Meanwhile, the children were seen again at the haematology clinic at Addenbrooke’s. Genetic testing revealed two abnormalities in both children (1) a heterozygous variant in the Factor V gene and (2) a heterozygous variant in the HPS1 gene related to a rare condition known as Hermansky-Pudlak Syndrome. The interpretation of these results was a significant aspect of the expert evidence before the judge.

On 26 September 2025, a professionals meeting was convened by the local authority. The minutes record that the meeting was told by a social worker that “the bruising has continued since the children have been in foster care” and that following a parenting assessment the local authority had prepared a 12-week transitional plan if the court decided that the children can return to the parents’ care. Dr Dickens attended the meeting and summarised the haematological findings. The minutes also record her as saying:

“They have been seen in Clinic with the Foster Carer and it is clear that both children still have bruises which are being body mapped regularly by the Foster Carer and Nursery. This is despite being in different circumstances to before.

There is sufficient evidence that B and G bruise disproportionately when they have had witnessed injuries. There is a genetically identified abnormality in one of the clotting factors present in both B and G. They are likely to bruise at a lower level than other children.”

Meanwhile, several reports were filed by the court-appointed experts.

The fact-finding hearing took place over four days between 6 and 9 January 2026. The local authority sought findings that the children’s injuries were inflicted by one or both of the parents, that the intracranial bleeding suffered by G had been caused by an act of shaking, and that the perpetrator of that act, knowing that G had been shaken, had failed to seek medical attention for her. In addition, the local authority sought other threshold findings arising out of its investigation, namely that the children had suffered harm, or were at risk of suffering harm, as a result of being exposed to the parents’ abusive and unstable relationship, the father’s drug use, and the mother’s poor mental health. Those additional findings were substantially conceded by the parents so that the focus of the hearing was on the cause of the children’s injuries. Oral evidence was given by the three experts, Dr Dickens, and the parents. At the end of the evidence, the judge adjourned for written submissions which were supplied by the parties on 19 January 2026.

On 20 February 2026, the judge delivered an oral judgment making findings which were subsequently recited in a schedule appended to the court order. In summary, the findings were as follows:

At the hearing on 20 February, after judgment had been delivered, the parents each made an application for permission to appeal which the judge refused. She then gave case management directions, including ordering the local authority to obtain a transcript of the judgment and listed the case for a further case management hearing on 9 March.

On 27 February, the parents each filed notices of appeal against the findings in respect of the children’s injuries.

Prior to the next case management hearing on 9 March, an unapproved transcript of the judgment was sent to the parties. At that hearing, the judge asked Mr Goodwin KC (leading counsel for the local authority) to read out his note of the questions she had asked the father at the conclusion of his evidence. She also asked to see a copy of the grounds of appeal submitted to this Court. The following day she emailed the parties saying that she had almost finished checking the transcript of her judgment and asking for a copy of the note of her questions to the father. There followed a series of emails from counsel, including from Ms Weston KC, leading counsel for the father:

“I am not aware that the court considered it necessary to address material omissions in light of the applications for permission to appeal that it refused. I am similarly unclear on the basis on which the court is now seeking to make additions to the judgment (although there are of course circumstances in which clarificatory addenda may be appropriate). In the circumstances, please could the court ensure that any additions to the transcript of judgment are clearly marked e.g. by the addition of an addendum paragraph rather than changes to the body of the judgment which the court gave.”

On 11 March, the local authority’s junior counsel sent to the judge a note of the judge’s questions to the father and the mother’s grounds of appeal. On 12 March, the judge sent to the parties an approved version of the transcript of the judgment (dated 10 March) and asked them to forward it to this Court. In the event, however, the version sent to the Civil Appeals Office was the unapproved version.

On 27 March, Peter Jackson LJ granted both parents permission to appeal against findings 1 to 3. The other findings, relating to domestic abuse and the parties’ relationship, the father’s drug use, and the mother’s mental health, were not challenged. He directed the parents to consider “whether they wish to unite and refine their grounds of appeal so that there is one set of grounds with simple numbering.” Following this direction, the parents consolidated their arguments into five grounds of appeal set out below.

For reasons which remain unclear, it was not until 13 May, the day before the hearing of the appeal, that copies of the approved version of the transcript were sent to this Court. A comparison of the two versions shows that the judge, in addition to correcting typographical errors, also made a number of substantive changes, some of which relate to issues raised in the grounds of appeal which had been sent to her prior to final approval of the transcript.

One of the appellants’ main challenges to the judgment concerns the treatment of the expert medical evidence. Before turning to the judgment, it is therefore helpful to summarise the central aspects of the written medical reports. Unfortunately, the oral evidence given by the experts was not transcribed for the purposes of this appeal, and our knowledge of that evidence is confined to those extracts to which counsel referred in submissions.

In his first report, Mr Robinson set out a comprehensive summary of the relevant medical records for both children. He noted that at the medical examination on 13 September 2024, B had been found to have twelve bruises. Most had been sustained accidentally as a result of his increased activity. He noted:

“B has a diagnosis of autistic spectrum disorder. There are references to his increased activity and a tendency to injure himself from parents and professionals. Hyperactive/very active children often injure themselves more than others.”

However, Dr Robinson identified “concerns” about four large bruises (on B’s forehead, left mid axillary line, iliac spine and right pelvic bone) which were seen at that examination and for which there was no available account of an accidental event or its aftermath. It was his opinion that, in the absence of credible explanations, those bruises were most likely to have been inflicted. He noted that there were “haematological concerns” but said that “resolution of significant bruising does not support a bleeding tendency”. He also said:

“Resolution of significant bruising seen on 13.09.24 when re-examined on 29.09.24 in grandmother’s care is significant. Where inflicted injury is suspected substitution with sensitive care is usually followed by a resolution of bruising. This remains a strong indicator for previously inflicted injury.”

As to the bruise on G’s ribcage, Dr Robinson said that, although it was minor, it was not likely to be the result of her rolling onto a dummy clip or other object. It was more likely sustained by impact from a blunt instrument (hand or other), against a hard surface, or a grabbing action …. I have never once observed this in clinical practice, only in child protection cases as explanations provided by carers as a cause for a bruise.”

As to the subdural haematomas, Dr Robinson said:

“There was no history of encephalopathy with no retinal haemorrhages. In the absence of a medical cause, findings indicate a likely shaking acceleration/deceleration injury.

Following head trauma the point of deterioration is most likely to be close to the point of injury. No such history is provided. The force exerted would have been in excess of normal or rough handling. The perpetrator would have been aware of this. For some infants, symptoms of encephalopathy may be mild and resolve rapidly.”

Following the later reports filed by the haematologists that both children had two genetic variants (discussed below), and evidence that the children had continued to experience bruising, Dr Robinson modified his opinion in an addendum report, observing that given the haematological findings and reports of continued bruising both children were likely to have an increased susceptibility to bruising following minor, possibly unnoticed, injuries.

The haematological evidence included reports and a statement from Dr Dickens, the treating clinician, and Dr Keenan, the jointly instructed expert.

In her report enclosed in a letter to the local authority dated 4 November 2025, Dr Dickens said this about both children:

“Genetic testing confirms a single heterozygous variant in the Factor V gene that is likely to explain reduced levels of Factor V. Severe Factor V deficiency would be expected if a variant was found in both copies of the gene.

In addition, genetic testing has demonstrated a heterozygous pathogenic variant in the HPS1 gene. If both copies of the gene were affected this would be expected to lead to Hermansky Pudlak disorder, which can be associated with abnormal platelet function. Impedance platelet function testing in [the child] was normal. It is likely that [he/she] is a carrier of HPS1 but she does not appear to be affected.

In light of [his/her] ongoing bruising with minor trauma, the borderline Factor V level and the additional genetic findings, it is possible that [he/she] has a tendency to bruising with minor or no trauma.”

In G’s case, Dr Dickens added:

“The Factor V level could have been significantly lower at the time she presented with her intracranial bleed when she was much younger.”

The report recommended testing of family members “to clarify the clinical significance of the Factor V variant”. It should be noted at this point accuracy of reference to “ongoing bruising with minor trauma” was a matter of contention before the judge.

In an earlier letter to the children’s mother dated 15 October 2025, Dr Dickens described the Factor V variant as being “of uncertain significance”. She continued:

“Variants of uncertain significance are genetic changes that are present in genes of interest but have not previously been described as causing clinical problems

Factor V is one of the proteins involved in the coagulation cascade, important for normal blood clotting.

This change explains B and G’s low FV levels.

We can tell from our functional tests that the Factor V levels are not low enough to explain spontaneous bleeding in isolation but it is possible that this along with the HPS1 carrier status has a combined effect that has made B and G more likely to bruise or bleed with minor injuries.

Over the time we have known B and G, we have noted their tendency to bruise very easily and this has persisted in each of the settings where they are looked after (foster carer, nursery, supervised visits with mum and dad).

On this basis, I have updated my report for the safeguarding investigation to say that we cannot be sure G’s intracranial bleeding and B and G’s bruising could not be explained by these results.”

Dr Keenan prepared no fewer than four reports (with a very considerable degree of repetition), as well as responding to supplementary questions. In his initial report, he expressed “some concerns with the collation of the information provided” and asked to be provided with the primary source results from the laboratory. At that stage, he stated his opinion as follows. First as to B:

“The blood clotting investigations to date have identified a mild reduction in factor V but not a definitive blood clotting disorder that could cause bruising or likely to increase the risk of bruising. Not all blood clotting disorders have been tested for and therefore these disorders have not been diagnosed or excluded.”

With regard to G, he noted again that the Factor V was below the reference range but observed that this was an “even milder reduction” than in B’s case and concluded G’s level was normal for a girl of G’s age. In both cases, however, he recommended testing for further disorders, including testing family members.

This opinion was repeated in his first addendum report. In his second addendum report, however, dated 2 October 2025, drafted after the genetic test results became available to him, Dr Keenan observed that “this is not a straightforward case from the haematology perspective”. He said:

“The gene studies have identified 2 findings in both children.

1.

Both B and G have a variant of uncertain significance in the factor V gene. This means that the gene is not the standard common gene for factor V that is seen in the majority of the human population. We all have many variances in our genes. It is what makes us individuals and unique. Some variances (differences) in genes can cause disease and are pathological. Some differences do not cause disease and are not pathological.

The difference or variance found in B and G is of uncertain significance. This means that the pathological significance of this gene is not known as there is insufficient data in the literature.

It is possible that this is a variant that causes slightly lower factor V levels but no bleeding problems. It is possible that this variant causes a mild factor V deficiency and an increased risk of bruising and bleeding.

2.

In addition, both B and G have been found to have another gene abnormality in that they have one abnormal copy of a gene called HPS1. B and G both have one abnormal gene and one normal gene. This is considered to be a carrier state and not a disease. If both genes were abnormal this causes a disease called Hermansky Pudlak Syndrome. In Hermansky Pudlak syndrome with both genes abnormal there is a mild to moderate bleeding disorder. This is very rare. Hermansky Pudlak is a form of platelet storage pool disease. In the disease state it would be expected that platelet aggregation testing would be abnormal and the definitive testing of platelet nucleotide release would be abnormal. The platelet aggregation testing and platelet nucleotide release in B and G are both normal.

I consider taking all information together of results of the HPS1 gene test showing a single gene abnormality and the normal platelet aggregation testing and nucleotide release assay that this is most likely not to be of relevance. This is not certain. It is possible that the carrier state gives a mild weakness of platelet cell function. I would state that with the information as is current that ‘on the balance of probabilities’ the single copy of abnormal HPS1 gene, with normal platelet aggregation and normal platelet nucleotide release in B and G is not of significance and is not causative or contributory in the bruising or bleeding diagnosed.”

Under the heading “Possible Factor V deficiency”, he continued:

“The levels of factor V slightly below the reference range and the finding of a single F5 variant of uncertain significance (VUS) gives some uncertainty in this case.

As stated, it is possible that this gene is a variant that causes slightly lower factor V levels but no bleeding problems. It is possible that this variant causes a mild factor V deficiency and an increased risk of bruising and bleeding.

As the levels of factor V are very mildly reduced I consider that with the information as current and ‘on the balance of probabilities’ this is a safe level of factor V and probably not causative or contributory to the bleeding diagnosed.

Given the uncertainty, I consider it reasonable both for clinical management and this court case that further investigation is performed. The children do not need further testing. It would be very useful to understand the significance of these genes by testing initially both parents and then potentially further family members. For example, if the parents or other family members are identified to have either of these genes then their own personal bleeding history and in particular any lack of bleeding history such as having surgical procedures with no issues would be very relevant.”

Dr Keenan was unable to attend the experts’ meeting but on 13 October 2025 replied in writing to the questions posed by the parties. In answer to a question about whether there was any helpful research to assist with the cumulative impact of the children’s genetic factors, he observed that “these conditions as true disease rather than carrier status are both very rare”. He continued:

“It is therefore unlikely that there will be a described co-inheritance case of factor V deficiency and Hermansky Pudlak syndrome and extremely unlikely to have sufficient cases to draw conclusions in terms of co-inheritance of disease. Co inheriting the carrier status of factor V deficiency and HPS1 will be more common but will still be very rare. I have not seen cases described of co-inheritance of the carrier status in the literature.”

On 19 November 2025, Dr Keenan filed a third addendum report. After quoting extensively from Dr Dickens’ report, he said that he did not disagree with any of her statements. He continued:

“My view is that it is possible that the combination of the 2 traits (heterozygous factor V variant of uncertain significance and HPS1 carrier state) that on their own are not likely to cause and bleeding or bruising problems may cause bleeding or bruising problems in combination.

The issue for the court is how likely this is.

As previously stated this case is complex from the haematological perspective and there is uncertainty.

With ongoing bruising in both children in different environments of nursery and foster parents it does seem that these children are experiencing bruising more than usual.

We cannot be certain with the current information of the subdural bleeding in G.

Subdural bleeding usually takes a significant injury or a severe abnormality of blood clotting to be present to have a spontaneous bleed.

No evidence of further subdural bleeding in either child is evidence that there was a significant traumatic event that caused the diagnosed subdural bleed.

It is my view that while the combination of mildly low factor V levels and a carrier state for HPS1 with normal laboratory platelet function could possibly cause a spontaneous subdural bleed on the balance of probability I consider they have not.

On the balance of probability I consider it probable i.e more than 50% than trauma was the cause rather than a spontaneous bleed with no trauma caused by only the haematological findings.”

Dr Keenan repeated his recommendation that both parents should undergo genetic assessment. He added that the instruction of a geneticist “remains reasonable in my view”.

In the course of submissions to this Court, Mr Goodwin for the local authority told us that, in his oral evidence. Dr Keenan had said that “on a balance of probabilities – not beyond reasonable doubt but very probably – the abnormal genes in combination were not causative or relevant to the bleeding in either child.”

Mr Jayamohan provided three reports. In the first, he set out his interpretation of the imaging of G’s head. Of a skeletal survey carried out on 17 September 2024, he reported no evidence of fracture or other abnormalities. Of a CT scan carried out on the same day, he reported no swelling or skull fracturing, no abnormalities in the brain itself and no changes in the subarachnoid spaces. He continued:

“There are intermediate density subdural collections seen over both frontal lobes. These contain some areas of higher density in keeping with fresh (re-)bleeding seen within them, although these may represent thrombosed veins. Looking posteriorly in the midline are smaller collections that are almost isodense to the brain, likely representing subdural collections also.”

He then considered MRI scans performed two days later on 19 September:

“… there is no appreciable difference in size in the subdural collections. MRI scanning does show that these are mostly proteinaceous fluid within the subdural collections, but there are membranes that can be seen, indicating that they are chronic subdural haematomas. The posteriorly placed bilateral subdural collections seen up at the vertex can still be seen and are not changed in size. There are a few smaller findings that are either veins or membranes in these areas.

I cannot convincingly see clear evidence of connection between the posterior vertex and the frontal collections, but regardless, there may a level of communication since they are in the same space. The posterior fossa and upper spine appear normal with no evidence of bleeding or injury that I can see within them.”

Responding to a question in his instructions about the timing of the injury, Mr Jayamohan wrote:

“The earliest mechanism by dating is birth. The studies that have that have looked at what is actually only a small number of the total number of births that take place, would suggest that elective caesarean section carries with it the lowest risk of asymptomatic subdural bleeding [he referred to three research papers commonly cited, called respectively the Whitby, Looney and Rooks papers, which he appended to the report]. In those babies that have had such, they would be expected in to reabsorb by one month old, and all to reabsorb by approximately three months old. However, one we note the small sample size, it is not safe to discount the possibility of birth related subdural occurring and remaining. It must be a very rare event overall given the lack of experience of this – I have had 2 clinical cases where I am convinced there was a high likelihood of such in 22 years as a consultant. The locations of these bleeds in G are seemingly separate – so this would in my view seem even more unlikely to be an explanation (as opposed to if there was just one area with a subdural collection).”

He then discussed the possibility of G’s subdural collections being caused at birth, but these observations were amended in his first addendum report considered below.

Continuing with the first report, Mr Jayamohan considered but discounted the possibility of a postnatal accidental trauma. As to a medical cause for the subdural bleeds, he said:

“I would defer [to] the other two experts in these regards but there seems to be none at the moment. The lack of any clinical ‘encephalopathy’ or radiological brain injury would, however, be in keeping with such an explanation, were one to be found.”

He continued:

“The last option is a non-accidental injury. Impact trauma would be a rather unlikely cause of these collections, given the symmetrical nature of the bleeds seen and the interhemispheric locations posteriorly – I cannot exclude it but would not favour it. However, these would be in keeping with a shaking injury. This would be the sort of mechanism that may be associated with a period of brain dysfunction also ….

Shaking injuries can be associated with mild brain dysfunction or sometimes none at all, and they can occur on multiple occasions … The forming of membranes (which only needs the original subdural bleed to occur previously) would then rebleeding to occur without any further traumatic events and can be associated with normal handling.

Therefore, the findings, if caused by a shaking event, may have only been caused by one event. The finding of multiple ages of blood completely confounds any ability to date from the radiology. There are no brain injuries that allow me to date any further either radiologically or from the clinical history.”

Mr Jayamohan’s second report was prepared after receiving Dr Dickens’ reports and Dr Keenan’s third addendum. As to the evidence of two genetic variants, he observed:

“While I understand that this may not be associated with spontaneous bleeding, it does raise the question of whether events that are within normal handling or perhaps not memorable to carers, may have provided sufficient energy to cause a traumatic bleed, although not a significant trauma. I also raise the question about the episode or episodes involving the iPad and whether this level of trauma in a child with a predisposition to bleeding or similar may have been sufficient to explain some of the subdural bleeding, or indeed the initial onset of such.”

He added that this was “clearly something that is highly specialised” and that he would defer to the opinion of the haematologists.

By this stage, Mr Jayamohan had also received the obstetric notes relating to G’s birth, which was described as “elective caesarean section, category four” and involved the use of forceps which were seemingly applied to the baby’s face. He was also shown a video of the birth, which, he said, showed evidence of forceps marks around the left cheekbone with talk of a similar mark on the other side. Mr Jayanohan continued:

“Assessing the risks of instrumented elective caesarean section is difficult from the literature to assess regarding subdural bleeding. There is evidence to suggest that elective caesarean section overall is related to the lowest risk of asymptomatic subdural bleeding. However, alongside this, is a clear suggestion, sometimes within the same papers, that instrumentation increases the risks regardless of mode of delivery. Therefore, the combination of these makes it difficult for me to give a clear sense of direction other than it is likely that even if an elective caesarean section has the lowest risk, the addition of forceps would increase this risk by some level. And therefore, perhaps it would be reasonable for me to "sit on the fence" and say there is to my understanding certainly a risk of subdural bleeding after forceps through a caesarean section. And of course this needs to be put alongside a potential increased bleeding tendency.

I will also note that there is no available history of an encephalopathy within the notes in keeping with a previous postnatal trauma. I am now less clear whether the combination of genetic changes in G may have put in place a risk of bleeding from events that are either minor enough to not have been noticed, especially if several occurred of differing types, or of the sort reported by parents.”

There was no additional opinion expressed in Mr Jayamohan’s third report.

We were not provided with a transcript or note of the experts’ oral evidence. But the following points were drawn to our attention. First, on the interpretation of the evidence of the subdural haematomas, the other experts deferred to Mr Jayamohan. Secondly, Mr Jayamohan said that “the court will be clear about my degree of uncertainty in this case”. Thirdly, he had identified four potential aetiologies of the subdural bleeding (birth trauma, accidental trauma, medical cause, inflicted injury) but was unable to say which was the most likely.

It is clear from reading the judgment that the judge was not reading out a written text but delivering the judgment from notes. Unsurprisingly, therefore, it is to a certain extent lacking in structure and written in a somewhat informal and discursive style. There is nothing inherently wrong with that, and indeed there was a time when most judgments were prepared and delivered in that fashion. It is, however, a hazardous course to take in a case, like this, where the evidence is complex. The risk is that something important is omitted. In this case, the parents’ arguments on appeal include that the judge failed to give proper consideration to their evidence and misconstrued aspects of the medical opinion evidence. In order to assess the merit of those arguments, it is necessary to go through the judgment in some detail.

As noted above, the papers filed with this Court for the purposes of the appeal include two versions of the judgment – an unapproved version sent to the parties prior to the case management hearing on 9 March and included in the appeal bundle and an approved version sent to the parties in which the judge had made various amendments after seeing the note of questions she had put to the father at the end of his evidence and the grounds of appeal. On behalf of the mother, Ms Joanne Brown KC relied on some of the amendments as illuminating what she describes as failings in the judgment. In the following citations from the judgment, I have therefore underlined those passages inserted into the approved transcript which were not present in the original unapproved version.

The judgment started with a brief summary of the background, the precipitating event in September 2024, and the proceedings. In doing so, the judge stated (at paragraph 16):

“In the event, experts, whilst noting the blood disorder, particularly Dr Robinson, who had the role of drawing the various strands together in a paediatric overview, felt it was more likely than not that B’s and G’s injuries, each of them, that it was more likely than not that they were non-accidental injuries, even taking into account any blood disorder, and that the injuries were non-accidental injury.”

In the course of this introductory section of the judgment, the judge said (at paragraph 21):

“The parents say, and I stress they do not have to prove anything, but their evidence is important both on paper and in the witness box and should be considered, but they say this – and I summarise this – each one says ‘I wouldn’t do that. I would ask for help if there was a problem. If I had done something, I would tell. Moreover, knowing the other parent as I do, I cannot imagine he or she would do anything like that’.”

Next the judge recited the relevant legal principles, including the burden of proof, the approach to expert evidence, and the requirement to consider the totality of the evidence. The judge reminded herself that research has moved on over the years, adding that this was

“particularly important in this case, where much of the evidence has been about blood disorders, two of which were rare and unclear, and where the combination of the two blood disorder, and its effect, is unresearched and is regarded as uncharted territory.”

She added:

“neither medical nor non-medical evidence has precedence. Credibility assessments and proper exposition of the evidential canvas are central, and I have to be realistic but at the same time careful not to be selective.”

She proceeded to set out other legal principles, including the identification of a perpetrator of non-accidental injuries, the need to consider the possibility of an unknown cause, the correct approach to lies, and the assessment of allegations of failure to protect.

The judge said that she had read the papers, “many more than once and in fact several times”, including the experts’ reports and the parents’ statements. She started her consideration of the evidence with the incident in which the father was said to have pinched B’s ear in January 2025. Having done so, she said:

“I take the view that the burden of proof as it should be, lies with the local authority; and what I have heard on the evidence I take the view and I find that it is insufficient to make a finding on that particular injury.”

At this point, the approved version of the judgment contains the following paragraph (new paragraph 57) which was not in the original verbatim version:

“In respect of the other injuries which I will come to in turn, the local authority case is not based simply on frequency of bruising but rather the ‘disproportionality’ of the bruising, compared to other periods – disproportionality based primarily on location of the relevant bruise/injury and lack of explanation for it/them.”

The judge then considered B’s bruising in September 2024. She summarised the evidence of Dr Robinson, in particular about the four specific bruises on B which the local authority asserted had been inflicted non-accidentally. Dr Robinson had accepted in oral evidence that one of the four bruises – on the left mid-axillary line – could have been sustained in an accident. In this part of the judgment, the judge referred to part of the parents’ suggestions for how the injuries had been sustained. She noted that there was no suggestion of bruises being suffered during play-fighting. Regarding the bruise on G’s ribcage, she referred to the suggestion that it had been caused as a result of rolling on the dummy clip. She said that she did not think the parents were lying about this, but rather that it was a suggestion which “over the months they had firmed up on”. She continued:

“I appreciate that the parents do not have to prove anything, but as an explanation I am finding that difficulty lies within it. If it happened before and it did not result in a bruise in the past, then I am finding it difficult to understand why the parents are saying that rolling on to the dummy would have been a cause of the bruise on this occasion. So, why do they think it was the cause of the bruise on this occasion?”

The judge then considered Mr Jayamohan’s evidence, referring to various aspects of his reports and oral evidence. She noted that two elements of the triad of symptoms associated with a shaking injury – retinal haemorrhages and encephalopathy – were either not present or not reported, but that this “does not countervail the opinion of both Mr Jayamohan and Dr Robinson, who say that the subdural bleeding was entirely consistent with shaking”. She quoted passages from Mr Jayamohan’s first and second reports set out above. In doing so, however, she mistakenly said that he was “sitting on the fence on matters relating to the bleeding disorder”. In fact, as cited above, his comment about “sitting on the fence” related to “the risk of subdural bleeding after forceps delivery through a caesarean section” which “needed to be put alongside a potential increased bleeding tendency”.

Her summary of Mr Jayamohan’s oral evidence included his observation that “it is rare to see subdurals by fluke – in other words, if there were asymptomatic subdurals in the population, you would expect to see some of them when scanning.” She continued:

“114.

He said that a bleed from birth was extremely rare and he went further: the location of the bleeds frontal as opposed to birth bleeds or posterior. He then referred to the nuances and specifics for G and said that there was a slightly higher chance that it related to birth.

115.

He is later saying that if there were no concerns about the ability to clot then he would be fairly hawkish, saying that it was not birth related, but he could not exclude it. And he said if there was no clinically relevant bleeding disorder, then the birth related bleed would be exceptionally rare. If she had a disorder, then the chances would be elevated above exceptionally rare, but if a minimal increase in bleeding risk, it probably remains exceptionally rare. If there was a substantial risk, then it becomes more realistic, fifty per cent chance or more, but then indicated that that was not his understanding from the summary of the evidence.”

Next the judge considered the haematological evidence. She noted:

“there are two blood disorders, and the question is the effect not simply of each one separately, but in combination. And as I understand it both are relatively rare …. And it is common ground that the combination of the two blood disorders is completely uncharted territory.”

As to the HPS1 gene, she noted that Dr Keenan considered it “most likely not to be of relevance” but added that he had expressed caution and quoted him as saying (as I understand, in oral evidence):

“The carrier status might give a mild weakness of platelet cell function so should look at the clinical picture on the ground. But if there is no history of easy bruising then very unlikely that HPS1 was relevant.”

As to the Factor V variant, she recorded Dr Keenan’s evidence that “at its height each child might have a mild bleeding disorder, but possibly no bleeding disorder at all…”.

Turning to the differences between Dr Dickens and Dr Keenan, she noted first that Dr Dickens had at one point mistaken B’s results as G’s, which, she said, “troubles the court in the light of the other differences between the two haematologists”. She continued:

“130.

Under cross-examination on behalf of the mother, Dr Keenan was asked about the uncharted territory, the “unknown unknown”, and he said that even if the combination did create a bleeding clotting disorder, it would make a minor difference to the risk of bruising and would not cause spontaneous bleeding. He accepted he had only seen one significant Factor V case, but on that case it did produce unusual bruising; but nevertheless pointed out that he only has experience of one Factor V case.

131.

Each and every expert and each and every professional from their different perspectives accepted that the rare disorder taken singly, was something of an unknown quantity, and in combination there was no research at all. All accepted, and that included Dr Dickens, that you look at the clinical position on the ground to see if there was a disproportionate pattern of bruising because blood disorders, as they said, do not come and go.”

The judge then considered the evidence of the foster carer who had been asked to monitor the children’s bruising. She said:

“He said neither child had a pattern of disproportionate bruising. It was normal knocks and scrapes and the bruising was explained. He had not considered that the bruising had required medical attention, but also said that he logged only the large bruises; and the point has been taken up on behalf of the parents, to say that that is not a proper log and if you are looking for a pattern or checking for a pattern you look at it all, and you should undertake a proper analysis of all the bruises and you do not effectively just pick and choose, either based on size or possibly on location. I note that point. And they say that it is a point well made, and in fact better made, because Dr Dickens’ evidence is that on her appointment on 28 August she had a clear view that the foster carers were saying that the bruising had continued in their care and, as I understand it, she saw bruising and she considered from that that both children bruised more easily.”

She then quoted from the record of a professionals’ meeting at which Dr Dickens had observed that the genetic change in the Factor V gene and borderline low Factor V level

“could explain bruising that would be disproportionate to the injury but would not explain spontaneous bleeding …They have been seen in clinic with the foster carer and it is clear that both children still have bruises which are being body mapped regularly by the foster carer and nursery. This is despite being in different circumstances to before. There is sufficient evidence that B and G bruise disproportionately when they have witnessed injuries. There is a genetically identified abnormality on one of the clotting factors present in both B and G; they are likely to bruise at a lower level than other children.”

The judge recorded that Dr Dickens had maintained that view in oral evidence, then added (at new paragraph 138):

“but I am unclear on the basis for this. She saw bruising on that day I understand, but there is no body map, no history, nothing to say whether they were explained or unexplained; nor the location; nor the size of what she saw; it seems the frequency was based on, what she says were the foster carer’s comments.. She has not analysed the full period”

The judge continued:

“I am grateful for the chronology which has been provided on behalf of the parents which sets out the bruising and the timings of the bruising and the dates …. But of course the analysis of the full period was done by Dr Robinson. He considered the logs of the foster carers and saw no pattern of disproportionate bruising. He accepted that there was one bruise which was unexplained on B to the left posterior thigh, but considered that one bruise did not make a pattern. On the other side, the parents say that because of the flawed logging by … the foster carer, that means, virtually by definition that it will be a flawed analysis by Dr Robinson. So, that is B.”

As for the bruising seen on G’s forehead in December 2024, the judge said that Dr Robinson “did not consider it unlikely at all” that it had occurred as a result a day to day fall in a toddler.

The judge then said that

“leaving aside for the moment the period of foster care what we know is this – in the period before September 2024 … there was nothing unusual in the pattern of bruising. The increase was noted by the mother, accepted by the father, and neither have resiled from that position that there was disproportionate bruising.”

The judge noted that there was no evidence of a disproportionate level of bruising in the periods when the children were with the grandmother or with the parents under 24/7 supervision. At this point in the approved version of the judgment (new paragraph 146), the judge inserted another paragraph that was not in the original verbatim version:

“I asked the father about the bruising when the children were in foster care. He said that ‘he did not think there was an increase in bruising but still large amount of bruising. About the same number as before the increased bruising’ (i.e. before September 24). ‘It was the same number as for the vast majority of his life. The appearance no different. As to etc it was not unusual in foster care.’”

The judge added (new paragraph 147) that, “in that regard the father’s evidence supported the foster carer’s”.

It was pointed out by the parents’ counsel on appeal that the new paragraphs 146 and 147 were a summary of the father’s answers to the judge’s questions at the end of his evidence, as recorded in counsel’s notes which the judge had asked to see at the case management hearing on 9 March 2026. I shall return to this issue later.

Under the heading “The general environment”, the judge then considered the circumstances in the home in the period leading to the admission to hospital. In doing so, she said that she placed “great store by the oral evidence that I heard from the parents”. She cited the judgment of Peter Jackson J (as he then was) in Re BR (Proof of Facts) [2015] EWFC 41 in which he set out and endorsed a list of risk and protective factors as “a helpful framework within which the evidence can be assessed and the facts established”. She noted that, with one exception, all of the protective factors identified in the list were present in this case, observing:

“These were well loved children and, it has to be said, very visible children. These were not children who were hidden away from health visitors or nursery or any of the like.”

She noted that some of the risk factors were also present, citing B’s disability, the history of domestic abuse, substance abuse, and the father’s stressful job, but added:

“on the flip side that notwithstanding the existence of those risk factors, that on the whole they had been able to weather those particular problems. But there were underlying problems, notably within the relationships. There were violent episodes and that is accepted by the parents ….”

The judge then considered the issue of lies allegedly told by the parents in their police interviews. In the approved version of the judgment, she said:

“I know that the local authority points to the interviews and say that the parties lied. I am not sure I would put it that way. It is rather that either they did not recognise it as a problem or at least the depth of the problem or rather more that having heard them in the witness box they just did not want to go there, and bottled it up. It is not so much that they were lying in the sense that that they knew the truth and consciously tried to conceal it; it’s a bit more subtle than that – it is rather more that they were dishonest with themselves, that they minimised their own problems.”

She reached the same view about the parents’ statements about the bruising, observing:

“Neither would or could contemplate that inflicted [sic] or could have been inflicted. They just did not want to go there. For most of the relationship they worked well both separately and as a team. With B they noticed a problem in his behaviour developing and they moved fast and early and did something. In respect of the bruising the mother moved fast and early and did something. She made an appointment with the GP. The exception is the bruise to G, the rib which nobody seems to have noticed or thought was particularly significant.”

The judge then observed that it was clear that from February 2024 there was a rise in tensions between the parties, including over a relationship between the mother and another man which “certainly affected the father”. The father “absented himself” which “caused B to be unsettled in his behaviour and more challenging”. The judge noted that the mother was unwell, adding “it is clear from her communications with the mental health team that she was feeling overwhelmed”. She then observed:

“it is one thing to have a stressful job when all is well at home; but stress at work and then stress at home, there is no respite, and that is an entirely different matter …. From what was said in the witness box it was unclear how much the father was at home and the amount of to-ing and fro-ing under the domestic roof. But it seems to me that when I heard their evidence it seems that when they were under the same roof they were not really speaking, and that you could on occasions cut the atmosphere with a knife.”

The judge referred to arguments and an incident of abuse in August 2024.

At paragraphs 164 to 168, the judge set out her conclusions about the causes of the injuries. This section of the approved judgment contains a number of additions or amendments to the original verbatim version which I have underlined in the citation below:

“164.

Taking all the matters together – the atmosphere in the home; the unusual pattern of bruising not seen, I find, before or since the pattern noticed by the mother at roughly the beginning the of September; the subdural haematoma, with no explanation and unlikely to be caused at birth because of the time since birth and because of the location and indeed the time since the birth; the bruise to G’s rib, which is one off, so a bleedings disorder resulting in a general pattern of odd bruising being unlikely; the expert opinion and analysis placing unlikely event or explanation on top of unlikely explanation; and I do not think that the haematology assists the parents when I consider the pattern of bruising over time including when events occur but with no bruising (e.g. G being hit with a tablet). Blood disorders do not come and go. Thus I have considered that possibility of unknown cause and I cannot see that I can make that finding.

165.

To summarise – I find that the bruises to B, three of them (but Dr Robinson was equivocal about one, and I give the benefit of the doubt to the parents on that) but three of them I find were non-accidental injuries, but also the subdural haemorrhage also not only non-accidental injury but likely to be a shaking injury; the bruise to the rib was unlikely to be caused by the dummy clip and so inflicted. It was all caused during a period of high tension within the family home and you might say it was something of a perfect storm, and I make those findings of non-accidental injury.

166.

In all probability on a balance of probabilities it is more likely that the injuries were caused by one parent. It is right that the father has flashes of anger and can be violent – not as I say a daily diet, but there are those flashes; and I also note his conduct in the hospital which caused a bystander to express their concerns.

167.

It is right that the mother took B to the GP as soon as she felt concerned, and it is said why would she do that if she inflicted the injuries? But equally there is no suggestion or evidence that the father stood in the way of that. Similarly, the hospital, both were surprised at the existence of the subdural, as they said to me. But to balance, primary care fell to the mother; so the circumstances were more open to her than for the father and as I understand it in the main when the father was present so was she. She had a high level of resentment for the lack of support and her frustration it has to be said was palpable – perhaps not unreasonably so, but nevertheless I observe that it was there.

168.

It is not clear to me the timeline about the father and when he was moving in and out, but I do not think that the mother moved out. She appears to have been on hand when the father was there and even with the benefit of hindsight there is no suggestion or evidence that the bruising occurred when she was not there or even coincidentally when the father was. And both were adamant that the other would not do it. I cannot say that one is more likely to have inflicted the injuries than the other.”

The judge addressed the local authority’s case that the parent who had not inflicted the injuries had failed to protect the children. She declined to make that finding, saying:

“I do not find that there was a failure to protect. There was of course violence in the relationship, but that does not mean that either party should consider that if there was bruising or injury to the child that it would necessarily flow from that on a balance of probabilities that the father had caused it. That does not mean that the violence within the relationship is not insignificant, and I agree with the guardian that it does need work, but that is not the same thing as linking it to a failure to protect.”

The judge then dealt briefly with the other matters, not challenged by the parents, on which the local authority relied in support of the threshold criteria. Finally, she added this observation:

“It is right that in September these children were halfway home and the local authority were considering the risks and considering that they could manage the risks, and of course it has now crystallised into findings and they are serious findings of serious injuries; but nevertheless it is also right that it does not take much to recognise the tensions in the home at that particular time; it is also clear that in general the parenting appears to have been good and if at all possible I would want the local authority to see if there is a route for these children to go back home, and I would want that to be considered and see whether it can be managed.”

Following this Court’s direction, the parents agreed to consolidate their case into five grounds of appeal.

At the appeal hearing, the parents’ cases were presented by Ms Brown KC leading Mr Daniel Sheridan for the mother and Ms Amanda Weston KC leading Ms Anna Hefford for the father. Their written submissions, prepared for the applications for permission to appeal and therefore based on the unapproved transcript of the judgment, were supplemented by oral submissions in which they focused substantially on the amendments to the approved version. Although it did not form a ground of appeal, they were critical of the way in which the judge had gone about correcting the transcript, asking to see one part of counsel’s note of evidence and the grounds of appeal submitted to this Court and then taking those matters into account in making the corrections. They accepted that the corrected version was the definitive one on which this Court should proceed. But their fundamental case was that, whichever version was relied on, the judge’s reasoning was fatally flawed.

In her submissions on behalf of the father, Ms Weston cited the judgment of Peter Jackson LJ in Re B (A Child) (Adequacy of Reasons) [2022] EWCA Civ 407, in particular his observations at paragraphs 57 to 59 about the structure of a judgment, the eight components of a good judgment and the ultimate need in particular to “evaluate the evidence as a whole, making clear why more or less weight is to be given to key features relied on by the parties” and “give the court's decision, explaining why one outcome has been selected in preference to other possible outcomes.” At paragraph 60, Peter Jackson LJ continued:

“The last two processes – evaluation and explanation – are the critical elements of any judgment. As the culmination of a process of reasoning, they tend to come at the end, but they are the engine that drives the decision, and as such they need the most attention. A judgment that is weighed down with superfluous citation of authority or lengthy recitation of inessential evidence at the expense of this essential reasoning may well be flawed. At the same time, a judgment that does not fairly set out a party's case and give adequate reasons for rejecting it is bound to be vulnerable.”

Ms Weston submitted that the judgment in this case fell well short of what was required. It was deficient in its evaluation of the evidence as a whole and in its explanation for why one outcome had been preferred over another. Ms Weston described the judgment as “a meandering stream of consciousness” from which it was impossible to discern how the judge reached her conclusions. The deficiencies were on a scale which could not properly be repaired through a process of seeking clarification: see Re C, D and E (Care Proceedings: Adequacy of Reasons) [2023] EWCA Civ 334.

In the course of Ms Brown’s submissions, we were taken through all the corrections made to the approved transcript. Ms Brown sought to demonstrate that the corrections made had only served to highlight the flaws in the judge’s reasoning. In developing this argument, she focused in particular on those changes which, she submitted, had been made to counter the accusation that she had failed to assess the reliability of the parents’ evidence and failed to explain on what basis their evidence in relation to the children’s injuries was rejected.

In going through the amendments, Ms Brown alighted on the judge’s additions to paragraph 21. To the original words that the parents’ evidence “is important both on paper and in the witness box”, the judge had added “and should be considered”. Ms Brown submitted that the fact that the judge had inserted these words did not mean that she had given sufficient consideration to their evidence and only served to illustrate that she had not in fact done so. Ms Brown also drew attention to the new paragraph 57 inserted in the approved version of the judgment in which the judge had summarised the local authority’s case as to the features of the bruising which pointed to a non-accidental cause. Ms Brown submitted that this amendment served only to highlight the judge’s failure to engage with the parents’ case and evidence about the bruising. Ms Brown was particularly critical of the judge’s addition of new paragraphs 146 and 147 to the approved version of the judgment. She submitted that it was wrong for the judge to ask for a note of one part of the father’s evidence to buttress the findings she had made when there had been no proper evaluation of the parents’ evidence as a whole.

Ms Brown further submitted that important parts of the foster carer’s evidence about bruising were not cited in the judgment. She acknowledged that the judge had noted that he had “logged only the large bruises”. In addition, however, the foster carer had said that B sustained other bruising which had not been included in the log, that he had not recorded any small bruises to the back, arms, backs of legs, or thighs, and that B had sustained bruises for which he could not provide an explanation (“B has bruises that no one can say where they come from”). Ms Brown submitted that this was important evidence which, alongside the evidence in the chronology of bruising prepared on the parents’ behalf, supported Dr Dickens’ opinion that the two children were liable to bruise easily. Ms Weston also cited parts of the evidence given by the foster carer which were omitted from the judgment. In particular, she relied on his comment that the bruising in both children was 'more visible' and that there appeared to be a mismatch between the degree of seriousness of any reported incident giving rise to the bruising and the appearance of the bruise and/or bump.

In addition, Ms Brown and Ms Weston submitted that the judge’s analysis of the expert evidence was flawed. A major element of Ms Brown’s argument on appeal was that the judge had misinterpreted Mr Jayamohan’s opinion. Contrary to what the judge had said at paragraph 16, it was not his view that the subdural bleeding was non-accidental. Ms Brown emphasised the judge’s error about his reference to “sitting on the fence” set out at paragraph 53 above. In addition, the judge had erred in assessing the significance of the genetic variants and her conclusions about the bruising were flawed. She wrongly rejected Dr Dickens’ clinical evidence about the disproportionality of the bruises, and instead relied on Dr Robinson’s analysis which in turn had been based on the foster carer’s incomplete recording.

Ms Brown further submitted that the very limited analysis leading to the judge’s “pool” finding was inadequate. It had been the mother’s primary case that the injuries had not been inflicted but if the court reached a contrary conclusion it was inherently improbable that the mother was the perpetrator. Ms Brown acknowledged that there was evidence of tensions in the family home during the months prior to September 2024, but pointed to the evidence of the subsequent period when the mother was the primary carer under 24/7 supervision in which there had been no evidence of any risk to the children from the mother.

Finally, Ms Brown drew attention to the amendment made to the final paragraph in the judgment after she had made her findings. In expressing her wish for the local authority to “see if there is a route home for these children”, the judge had inserted the observation “it is also clear that in general the parenting appears to have been good”. Ms Brown submitted that this positive observation about the quality of parenting ought to have been taken into consideration as part of the analysis of whether the injuries were inflicted and, if they were, the identity of the perpetrator. Extensive submissions had been made on the mother’s behalf about how often the children were taken by the mother to be examined by medical professionals. The judge did not engage with those submissions at all.

The guardian invited the Court to allow the appeal on grounds 3(e) and (f) relating to the identity of the perpetrator, whilst remaining neutral on the other grounds. On her behalf, Ms Alison Moore acknowledged the significant shortcoming in the judgment that there was little if any mention of the parents’ oral evidence (which, she observed, had taken up half of the time of the hearing) and no analysis as to what impression she gained of them that would then inform the findings made against them. There was no indication or consideration of whether the judge found that either parent lied or told the truth. Ms Moore also submitted that the judgment lacked clarity as to why, in light of all of the evidence as part of the “wider canvas” and given the presence of many of the protective factors identified in Re BR (Proof of Facts), the judge nevertheless found both parents to be in the pool of perpetrators.

On behalf of the local authority, Mr Goodwin put up a robust defence of the judgment. This was a case in which two young children had sustained injuries in the context of a history of parental dysfunction, described by the judge as “a perfect storm”, spilling over on occasions into domestic abuse. The judge had to navigate her way through the evidence, surveying a wide canvas, from the parents, foster carer and various experts. Mr Goodwin described the judge as being “steeped in the evidence”, adding that no judge is going to reflect every nuance of evidence or submissions presented to her. The question was whether this judgment was good enough. The local authority submitted that it was.

Mr Goodwin submitted that there was nothing inherently wrong with the course taken by the judge in correcting the transcript. Her reasoning was not invalidated by the fact that she made corrections, even if some of them had been prompted by reading the grounds of appeal. There was no basis upon which this Court could properly conclude that amendments made to the transcript did not reflect the judge’s thinking at the time the judgment was delivered. He accepted there had been a number of amendments but contended that none of them was determinative of the appeal. In every case, the amendment was an expansion of the judge’s thinking already discernible in the unapproved draft. For example, the issues about disproportionality of the bruising were, in Mr Goodwin’s phrase, plainly on the judge’s radar. In paragraph 164 to 168, the judge had been drawing the threads of the totality of the evidence and none of the alterations made when correcting the transcript undermined that ultimate analysis.

Mr Goodwin submitted that the judge’s references to Re BR (Proof of Facts) and the lists of risk and protective factors indicated that she had in mind the positive aspects of the parents’ case. He pointed to her comment that “these were well loved children and, it has to be said, very visible children.”

Mr Goodwin accepted that Mr Jayamohan had not expressed any conclusion on the ultimate issue as to whether the subdural haematomas were inflicted non-accidentally or attributable to a birth injury. He conceded that the judge had therefore been wrong, at paragraph 16 of her judgment, to include Mr Jayamohan as being amongst the experts who had concluded that the injuries were non-accidental. He submitted, however, that this was not a fatal error because elsewhere in the judgment she had recorded the nuances of Mr Jayamohan’s position, including his opinion as to the rarity of subdural bleeding being caused at birth, in particular such bleeding extending for seven months after birth, in the context of a possible bleeding disorder. Mr Goodwin pointed to paragraphs 114-5 in the judgment as examples of the judge correctly citing Mr Jayamohan’s opinion of those matters. In considering the possibility of a birth-related injury, Mr Goodwin emphasised the evidence given by Mr Jayamohan that subdural haematomas were almost exclusively posterior, not frontal.

Mr Goodwin invited this Court to focus on what the judge had done rather than what she had left undone. She had conducted an extensive analysis of the evidence on all the relevant issues, drawing on all strands, and not considering any piece of it in isolation. He took the Court through the judge’s analysis of G’s bruised ribcage and her subdural bleeding. He referred to the judge’s analysis of B’s bruising, including the parents’ explanations, her detailed consideration of the individual bruises, leading ultimately to Dr Robinson’s assessment and her own conclusion. Mr Goodwin pointed out that Dr Robinson had been given an opportunity to update his view in the light of updated disclosure during a break in his evidence, none of which caused him to alter his opinion. In cross-examination, Dr Dickens had accepted Mr Goodwin’s description of her assessment as a “snapshot”. It had been Dr Robinson who had carried out the totality of the evidence about the bruising. Both Dr Robinson and the judge had taken into consideration the criticisms of the reliability to the foster carer’s logs. It was correct that the foster carer had only logged the bigger bruises, but those were the bruises about which Dr Robinson was concerned. The judge had been entitled to accept Dr Robinson’s assessment and regard Dr Dickins’ clinical assessment as a snapshot. Mr Goodwin emphasised Dr Keenan’s view that the children had only carrier status for the HPS1 gene and that it was most likely not to be of relevance, but one should look at the clinical picture; that the Factor V levels were (for B) mildly reduced and (for G) either normal or slightly below the reference levels; that on a balance of probabilities the Factor V levels would not cause or contribute to the bleeding disorders; and that the combination was unlikely to make a difference.

Mr Goodwin submitted that the judge had been careful to warn herself about the limits of the expert evidence. She referred to the legal principles about the treatment of such evidence, the small cohorts in the research studies about birth-related subdural haematomas, the fact that Mr Jayamohan had very limited experience of such cases, and the fact that the combination of genetic anomalies was “uncharted territory”.

Mr Goodwin acknowledged that the judge had not dealt expressly in a separate section of her judgment with the parents’ evidence and credibility. But he pointed to a number of places where she had referred to their evidence and submitted that, on that basis, it could not be said that the judge ignored it. As to the pool finding, Mr Goodwin acknowledged that the analysis had been brief (three short paragraphs, 166 to 168, quoted at paragraph 64 above) but submitted that this Court should be slow to interfere, given the extent of the judge’s knowledge of the evidence in what was a particularly fact-sensitive evaluation. Issues relevant to the identity of perpetrator had been alluded to earlier in the judgment which did not require repetition at this point. He also submitted that, given the nature of the injuries – a chronic subdural haematoma which could have been inflicted at any point over a period of six months, with no evidence of encephalopathy to help pinpoint the timing, and a pattern of bruises also inflicted over an period of time – a finding as to the perpetrator might in the circumstances have been beyond the reach of the court.

Before turning to the grounds of appeal, it is necessary to say something about the way in which the judgment was prepared. No party asserted that the judge’s substantive amendments to the transcript amounted to a serious procedural irregularity so as to justify setting aside the findings, but there was considerable criticism of the judge’s conduct which needs to be addressed.

It is well established that a judge is entitled to correct the transcript of an oral judgment before the final version is approved and that such corrections may in certain circumstances extend to substantive elements as well as typographical errors. The extent to which a judge may correct a judgment has been considered in a number of reported cases. We were not, however, referred to any of those authorities and for my part, not having heard legal argument on the issue, I am reluctant to embark on an analysis of the case law.

I do, however, have considerable disquiet about the course taken by the judge in this case. A substantial proportion of the hearing before us involved a comparison of the two versions of the judgment. Although they did not say so explicitly, it was implied by the parents’ counsel that the judge had included points in the approved version of her judgment that were not part of her reasoning at the time she was delivering it and only added after she saw the grounds of appeal. Mr Goodwin submitted that it is difficult if not impossible for this Court to say what was in a judge’s mind at any point. But the purpose of requiring a reasoned judgment is to demonstrate what was in the judge’s mind when the decision was made. It is an element of the system of open justice.

In this case, the course taken by the judge ran the risk of causing considerable problems. At the very least, having two versions of the judgment may cause confusion – as here, where the appellants’ skeleton arguments on the appeal were seemingly drafted on the basis of the unapproved transcript and the local authority’s skeleton on the basis of the approved version. Another risk is that extensive deletions and additions may obscure rather than clarify the reasoning. Furthermore, as Ms Brown argues has happened here, in amending the judgment the judge may end up illuminating its failings. Most importantly, any amendments to a judge’s reasons must represent a genuine expression of the judge’s reasons for reaching her decision and not an ex post facto rationalisation.

The real problem with the judgment in this case is the way it was prepared and delivered. The judge reserved judgment for six weeks. But it is clear from reading it – and confirmed by the parties – that it was then delivered ex tempore from notes rather than by reading out a settled text. With respect to the judge, this was an unwise course to take in a case of this complexity, with difficult medical evidence which had to be considered in the context of all the other evidence, in particular the accounts given by the parents. As noted above, the risk is that something important is omitted.

As I observed to Ms Brown in the course of submissions, it is not for this Court to dictate to judges precisely how they should write judgments. As Peter Jackson LJ said when identifying the necessary components of a judgment in Re B (Adequacy of Reasons) at paragraph 59, “judgments reflect the thinking of the individual judge and there is no room for dogma”. But where a reader – including an appellate judge but more importantly the parties – cannot discern from the judgment the judge’s assessment of crucial aspects of the evidence (including in this case the parents’ own evidence), the judgment is likely to be open to challenge.

In this case, an assessment of the parents’ evidence, their credibility and reliability ought to have been a central feature of the judge’s reasoning. As has been stated on many occasions (see for example Devon County Council v EB & Ors (Minors) [2013] EWHC 968 (Fam), paragraph 59), the evidence of the parents and any other carers is of the utmost importance and it is essential that the court forms a clear assessment of their credibility and reliability. In this case, it was a key aspect of the evidence about the cause of the injuries and, if they were inflicted, the identity of the perpetrator.

At several points in the judgment (in particular in the approved version), the judge referred to the parents not having to prove anything. For example, at an early stage in the judgment, the judge said of the parents: “I stress they do not have to prove anything, but their evidence is important both on paper and in the witness box”. In the approved version, she added at this point “and should be considered”. But it is not enough just to say that. The judge has to go on to consider it. In this case, although the judge referred to aspects of the parents’ evidence at some points in her judgment, there was insufficient consideration given to their evidence.

It was Ms Moore in her written submissions who pinpointed the central deficiency in the judgment. Although the parents’ evidence had taken up half of the time of the hearing, there was little mention of it in the judgment and no analysis as to what impression she gained of them that would then inform the findings made against them. There was no indication or consideration of whether the judge found that either parent lied or told the truth. I was unpersuaded by Mr Goodwin’s attempt to demonstrate that the various references to their evidence was sufficient engagement with it. I accept Ms Weston’s submission that there is nothing to indicate the impression the judge formed of the parents, or their evidence, or the basis on which she reached her conclusion that one of them had deliberately harmed their children. As noted above, at one point the judge said that she placed “great store by the oral evidence that I heard from the parents”. But she then failed to provide any or any adequate analysis of it.

It is true that the judge did address the list of protective and risk factors identified in Re BR (Proof of Facts) [2015] EWFC 41 and identified which were present or absent in the family environment. But given her assessment that, with one exception, all of the protective factors identified in Re BR were present in the family home in this case, it was even more important to carry out a careful analysis of the parents’ evidence.

It is an established principle that in care proceedings involving allegations of child abuse, whilst appropriate attention must be paid to the opinion of medical experts, those opinions need to be considered in the context of all the other evidence. The observations of Charles J in A County Council v K, D and L [2005] EWHC 144 (Fam) at paragraph 49 have been cited many times but bear repetition:

“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, are the type asserted is 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.”

As Ryder J observed in A County Council v A Mother and others [2005] EWHC Fam 31,

“A factual decision must be based on all available materials, i.e. be judged in context and not just upon medical or scientific materials, no matter how cogent they may in isolation seem to be.”

The present case was a paradigm example of the importance of these principles. Taken by itself, elements of the medical evidence pointed to non-accidental injury. But even in isolation the “medical and scientific materials” put before the judge did not conclusively point to the injuries being inflicted. One reason for setting out the medical evidence in such detail in this appeal judgment is to illustrate the complexities of it and the degree of uncertainty. The interpretation of bruising is never straightforward and in this case no photographs were taken at the child protection examination in September 2024. The children had two genetic variants which individually may have increased the propensity to bleeding and to easy bruising. There were no reported cases of the two variants being identified together and no research material to assist the experts or the court as to the implications of the combination. Dr Keenan recommended that the parents undergo genetic testing. He also recommended that an expert opinion should be obtained from a geneticist. In the event, neither step was taken. The fact that two out of the three elements of the triad of signs of a shaking injury were absent did not undermine the expert opinion that the subdural bleeding identified in G was consistent with shaking. But the absence of the two elements was a factor to be taken into account in assessing whether on a balance of probabilities, having regard to the totality of the evidence, the bleeding had occurred non-accidentally.

On the interpretation of the evidence about G’s subdural bleeding, Mr Jayamohan was the key witness. I accept Ms Brown’s submission that the judge misinterpreted his opinion. Reading the judgment as a whole, the judge seems to have concluded that he thought on balance that the bleeding was attributable to an inflicted injury. But Mr Jayamohan’s ultimate view was that he could not say which of the possible aetiologies was more likely than not. As I read his addendum report, his uncertainty as to the cause of the subdural bleeding was based on a combination of the risk of subdural bleeding after forceps delivery through a caesarean section, the “potential increased bleeding tendency”, and the absence of any history of encephalopathy.

Mr Jayamohan’s final position was that there were four potential aetiologies of the subdural bleeding (birth trauma, accidental trauma, medical cause, inflicted injury) but he was unable to say which was the most likely. Ms Brown submitted to the judge that in those circumstances it was not possible for the local authority to discharge the burden of proof. I do not accept that submission. For the reasons set out by Charles J in K, D and L, where the medical evidence is that there is nothing diagnostic of a non-accidental injury, it is nonetheless open to the court to reach a finding that the injury was sustained non-accidentally on the totality of the evidence. In this case, however, the judge misinterpreted the opinion of Mr Jayamohan, to whom the other experts deferred on the issue of the subdural haematomas, and her analysis of the totality of the evidence was incomplete and flawed. As Ms Brown observed in reply, there is no getting round the fact that in her summary the judge mistakenly said it was Mr Jayamohan’s view that on a balance of probabilities the subdural haematoma was sustained non-accidentally.

In their closing submissions to the judge, Ms Brown and Mr Sheridan had put forward a series of points which they said were not in dispute as to the aetiology of G’s subdural bleeding. In particular,

Turning to the bruising, I accept the parents’ submission that there was a factual dispute as to whether the children continued to sustain injuries in foster care which was left unresolved by the court in its judgment. Analysis of bruising is always difficult and here the evidence was complex and, in some respects, sub-optimal. The judge’s conclusion, based on Dr Robinson’s opinion, that the bruising identified in September 2024 was disproportionate when compared to other periods in the children’s lives was based on the foster carer’s logs of bruises which, as the parents maintained, were flawed because they did not register every bruise seen by the carer but only recorded the large bruises. The foster carer was asked to keep a record of “significant” bruising. But of course he would have no idea of what would be significant in the context of a case of a suspected bleeding disorder.

In contrast, Dr Dickens’ assessment that the children bruised more easily was based not only on her understanding of the foster carer’s analysis but also on her own experience of the children in clinic. As already noted, in her written report she said:

“Over the time we have known B and G, we have noted their tendency to bruise very easily and this has persisted in each of the settings where they are looked after (foster carer, nursery, supervised visits with mum and dad).”

I agree with Ms Weston’s comment that, notwithstanding Dr Dickens’ apparent concession to Mr Goodwin in cross-examination, her evidence was more than a snapshot. It is true that Dr Dickens only saw the children on one day, but the fact is that she saw bruising on that day which, in the context of the history, she considered was disproportionate. Overall, I accept the submission that the judge failed to give any or any adequate explanation of her reasons for discounting the evidence of Dr Dickens, who had seen the children in the course of a clinical examination, over the opinion of Dr Robinson, who had not.

The judge seems not to have addressed the important point raised on behalf of the parents that the foster carer’s logs of the bruises only recorded the large bruising and that, when looking for a pattern, it is necessary to consider all bruises, large and small, whatever the location. The judge referred to this point but did not explain how she discounted it, as she would have had to do, when accepting Dr Robinson’s evidence that the foster carer’s logs disclosed no pattern of disproportionate bruising.

In discounting Dr Dickens’ evidence on the bruising, the judge observed that no body maps had been produced nor any history to support Dr Dickens’ account of what she saw. But there was a 20-page chronology of bruising prepared on behalf of the parents. It indicated that both children had continued to sustain bruises at various stages during the proceedings, in a variety of settings (at nursery, at home during the 24/7 supervision period, in foster care). It included detailed descriptions of the bruises and in some instances photographs, some of which showed significant marks. The judge referred to it in passing but gave no further consideration to it in the judgment.

There was in fact considerable evidence that B was liable to bruise more easily than other children. That was the reason for the mother’s original request for a GP appointment on 8 September. The child protection examination on 13 September identified 12 bruises, of which 8 were considered to have been sustained accidentally. The evidence of the foster carer included the striking observation that B had “bruises that no one can say where they come from”. It might be thought that this single observation by the foster carer was more significant than his incomplete logs.

I also accept the submission that the judge failed to attach sufficient weight to the fact that these children have a combination of genetic variants, each of which may dispose them to bleed and/or bruise more easily, which neither Dr Dickens nor Dr Keenan had seen in combination before and about which there is no research evidence. Given the uncertainty over the significance of the children’s genetic abnormalities, the clinical picture assumed an even greater degree of importance. The judge failed to resolve the issues about the extent of the children’s bruising over the full period. That had to be resolved before the judge attempted to reach a conclusion about the haematological evidence. The paragraph added by the judge to the approved version of the judgment (at paragraph 146) did not in my view resolve the issue. There was no admission by the parents that the bruises in September 2024 were materially different from those seen at other times.

Finally, I accept the parents’ case that no or no sufficient consideration was given by the judge to the positive features about the care given to the children by the parents, in particular the mother, in the extended period of 24/7 supervised care, which continued over several months up to their removal into foster care. We were told that the records kept by those supervisors amounted to over a thousand pages. The judge said that she had read all of the papers but, so far as I can see, made no further reference to those records at all. They amounted to a significant body of evidence about the parents’ overall care of the children which provided an important part of the context in which the evidence had to be evaluated.

For those reasons, I conclude that the judge’s finding that the injuries sustained by the children were inflicted cannot stand. It follows that the findings that both parents were in the pool of perpetrators, and that the perpetrator of G’s head injury failed to seek medical attention, fall away. On the perpetrator issue I agree with the parents’ submission, supported by the guardian, that the judge’s analysis was perfunctory, although there is force in Mr Goodwin’s submission that the time period over which these injuries were sustained meant that, if they were inflicted, it may have been beyond the reach of the court to identify the perpetrator.

In granting permission to appeal, Peter Jackson LJ presciently observed that

“This was a case where the medical evidence was complex and less than diagnostic, and where the non-medical evidence did not contain particularly strong pointers towards the injuries being inflicted ones. In those circumstances it is arguable that the judge’s analysis of the evidence and process of reasoning was insufficiently rigorous, and in particular that she did not engage with the parents’ evidence and explain her conclusions about it.”

Drawing the threads together, this was a case in which the complex medical evidence was far from conclusive and the judge’s analysis of that evidence was flawed. She failed to resolve a crucial issue as to the extent of the bruising sustained by the children in foster care, and above all failed to conduct a proper evaluation of the parents’ evidence. This was a paradigm example of a complex and difficult fact-finding hearing which called for an intense degree of judicial scrutiny. Unfortunately, the judgment contained no adequate analysis of the totality of the evidence nor a satisfactory explanation as to why the judge came to the conclusion that the injuries were inflicted.

For those reasons, I would allow the appeal and set aside findings (2) and (3) as summarised in paragraph 18 above. The remainder of the findings remain in place.

If my Lady and my Lord agree, I would remit the case to the Family Presiding Judge for the South Eastern Circuit, Arbuthnot J. Some months have now passed since the previous fact-finding hearing. I anticipate that Arbuthnot J will want to consider first whether in all the circumstances a rehearing of the fact-finding hearing relating to the cause of the injuries is necessary for decisions to be taken about the children’s future care and, if so, the scope of such a hearing.

I agree. I particularly wish to endorse the observations made by Baker LJ at [89]. Whilst judges are, and should be, free to decide whether to deliver an oral judgment or hand down a written one, there are some cases in which it should be readily apparent that the latter course should be followed unless there are good reasons not to (e.g. extreme urgency). In my view, because of the nature and complexity of the issues, this case fell within that category, and it was most unwise of the judge to embark on the course which she did. The practice of circulating a draft of a reserved written judgment under embargo gives the parties’ legal representatives the opportunity to draw the judge’s attention to omissions or mistakes, which can then be addressed before the judgment is handed down. This case provides a graphic illustration of what can go wrong if that course is not followed.

I agree with both judgments.