CA-2025-001133 - [2025] EWCA Civ 993

This was, as Baker LJ observed when granting permission to appeal, in many ways a complex case. The Judge was presented with serious allegations of child abuse, in respect of which the evidence was in many respects unreliable, and the methods by which it was obtained fundamentally flawed. The Judge himself acknowledged that not one of the professionals investigating the allegations had complied with standard good practice in their work; this had the effect of excluding from judicial consideration potentially important material. Opportunities had been lost in the investigation to establish the truth.

In the final analysis, the pursuit of findings turned in large measure upon the evidence of an inexperienced foster mother, who had been, in the Judge’s finding “misguided” in her own personal investigation of what H alleged and was unskilled in much of her record keeping. Given his detailed assessment of the evidence, there was scope for the Judge to have concluded that she was over-zealous in this regard, suggestive perhaps of her inexperience in the role. It is well recognised that any initial questioning of a child who is alleging abuse should only ever be aimed at eliciting a brief account of what is alleged to have taken place; a more detailed account should not be pursued, but should be left to the formal (ABE) interview (see ABE Guidance, para.2.5). On the Judge’s account of the events of 21 April 2024, the foster mother asked a significantly greater number of questions of H than was necessary in the circumstances in order for her to know whether to take immediate action (ABE Guidance, para.2.7(d)).

To add to the complexity, the child complainant, H, was known to lie repeatedly; he had made serious allegations in the past against professionals and others which were patently not true. Moreover, the alleged perpetrators of abuse (the mother, father and CH) were all extremely vulnerable; each had a personal history of childhood sexual abuse, and each suffered from a degree of impaired functioning.

In a case in which there are multiple allegations, a Judge must always guard against the temptation to approach the evidence on the basis that something must have happened; the temptation in this case must have been all the greater given the reports of H’s sexually disinhibited behaviour and language, particularly once in foster care. In this case, the Judge had rightly been invited by counsel to consider the comments of Lord Hewart CJ in Bailey [1924] 2 KB 300 at 305, regarding the judicial approach required in cases in which the court is faced with determining a very large number of allegations:

“The risk, the danger, the logical fallacy is indeed quite manifest to those who are in the habit of thinking about such matters. It is so easy to derive from a series of unsatisfactory accusations, if there are enough of them, an accusation which at least appears satisfactory. It is so easy to collect from a mass of ingredients, not one of which is sufficient, a totality which will appear to contain what is missing. That of course is only another way of saying that when a person is dealing with a considerable mass of facts, in particular if those facts are of such a nature as to invite reprobation, nothing is easier than confusion of mind; and, therefore, if such charges are to be brought in a mass, it becomes essential that the method upon which guilt is to be ascertained should be stated with punctilious exactness” (Emphasis by underlining added).

The Judge was further taken to Macdonald J’s comments in Re P [2019] EWFC 27 at [272] where he said (having quoted the extract from Bailey above):

“The totalising approach must be avoided if the court is to steer safely clear of capitulating to suspicion and the beguiling adage that there is ‘no smoke without fire’” (Emphasis by underlining added).

The judicial advice from Bailey and Re P set out above was particularly apt to this case.

Notwithstanding the conscientious efforts of this judge in preparing his detailed judgment, this is a case in which the crucial analysis on the key facts would I suggest have been easier for him to undertake, and would altogether have been more coherent in its organisation and presentation, if he had adopted the approach advocated by Peter Jackson LJ in Re S [2023] EWCA Civ 346 at [33]; that is to say he could or should have identified and focussed on the “chapters of time” covered by the evidence, rather than structuring the judgment by reference to the sequence of witnesses and the individual allegations in isolation from each other. On the facts of this case, these ‘chapters’ might conveniently have been arranged under headings including: July 2023 (H’s report of being smacked); the first account of sexual abuse (December 2023: father); the second account of sexual abuse (January 2024: mother); other accounts of physical and sexual abuse (January – March 2024); the first ABE interview (5 April 2024); the allegations against CH on 21 April 2024; the second ABE interview. The Judge could then have concentrated on the first point in time at which H had ever given an account of alleged sexual abuse, the precise circumstances in which the account or accounts arose, whether and if so how they fitted into a context of lies, and how those were treated by those to whom they were made (Re P (Sexual Abuse – Finding of Fact hearing) [2019] EWFC 27). There was, thus, much force in Ground 1 of this appeal.

Turning to the specific findings under challenge, there are, in my judgment, obvious deficits in the Judge’s analysis of the evidence which led to finding 3, which drive me to the conclusion that it must be set aside. I identify altogether six flaws, or potential flaws, in the Judge’s approach.

First, as I referenced in §30 above, the Judge opened his review of the evidence relevant to, and his discussion of, finding 3 by referencing the “backdrop of troubling inappropriate sexualised behaviour in CH’s past”. To recap, he stated that while CH’s background (involving sexual abuse) was “not determinative” of the allegation made by H, the court was “entitled to consider it”. Plainly, there were no circumstances in which it could be said that CH’s background was ‘determinative’ of the allegation that CH had abused H; given his particular mode of expression in this context, it was incumbent on the Judge to explain how he had indeed ‘considered’ CH’s background in his overall review. For instance, the Judge should have made clear in his judgment how he had ‘considered’ this background evidence, what if any probative value it had, and whether (for instance) the Judge regarded it as demonstrating CH’s propensity to abuse. He did not do so. These judicial comments, expressed by way of introduction to his review of the evidence of alleged abuse of H by CH caused Ms Cheetham to argue that the Judge may well have inadvertently approached his task by applying a reversed burden of proof. I, for my part, would not go that far.

Secondly, having found that the foster mother’s record-keeping had in material ways been defective when considering H’s December 2023 and January 2024 allegations against his parents, the Judge’s assessment that she had displayed “a significant improvement” in her record-keeping by 21 April required proper explanation. I regret that I do not find this explanation in the judgment. There was only the vaguest of references to the foster carers having received any, or any additional, training in record-keeping between January and April 2024. The Judge went on to offer confusing views as to why he felt able to rely on the record made on 21 April 2024 (in contrast to the earlier records); I have in mind that having recorded that the incident log did not in fact “follow the format of an interview record with a sequential note of the question asked and the answer provided”, he then contradicted this by stating that the foster mother had set out “verbatim, the questions she asked and in quotes, recorded the reply by H”. Ms Taylor was right to concede that the 21 April 2024 is not, contrary to the Judge’s finding, a ‘verbatim’ account. It is difficult to see how the Judge regarded himself in any better position to assess what H had been asked and precisely what H had said on 21 April 2024, than in the earlier recorded incidents (see §40 and §41 above).

Having earlier rightly recorded in his judgment that there are inherent forensic dangers in asking a child to repeat an allegation (see §42 above), the Judge appeared to overlook the fact that the foster mother had done precisely this on 21 April 2024, when she asked H to repeat the allegations for her daughter; the Judge recorded that the foster mother had “wondered whether he would tell the same story again”. It is clear from the incident log that the foster mother had gone on then deliberately to introduce the subject of Hallowe’en into a second conversation with H in the presence of her daughter, and in that context had drawn CH’s house. The irresistible inference is that this had been done by the foster mother in an unauthorised and misguided effort to investigate the allegations herself, and should have sounded a deafening alarm for the Judge about the reliability of all that followed. As the Judge had rightly earlier observed in his judgment, given H’s age, his “recall of events required care”; this level of care was, in my judgment, self-evidently lacking from much of the interaction between the foster mother and H in their discussions on 21 April 2024.

The Judge further failed to address discrepancies in the accounts between the foster mother and her daughter (see §43 above), yet he appeared to rely upon them both. I refer in particular to: (a) the account given by the foster mother’s daughter of the evening of 21 April 2024 (included in the judgment) did not make any reference to a gap of half an hour between two separate conversations, as recounted by the foster mother; (b) the foster mother’s daughter recorded that H had said that his penis was in “his (i.e. CH’s) mouth” (this is recorded within speech marks) whereas on the foster mother’s account H had alleged that CH’s penis was in H’s mouth; (c) the foster mother’s daughter recorded that she had said to H “something along the lines of ‘… no adult should be doing that’”, a comment of some significance, which was not recorded in any form at all by the foster mother.

Thirdly, although the Judge was right to reference in his judgment the many occasions on which H had made false accusations during his period in foster care, and although the Judge was also right to remind himself of the dangers of relying on the uncorroborated word of a child who has a history of lying, the Judge then offered, to my mind, an unconvincing rationale for why he had accepted H’s word for what he had said on 21 April 2024. In this regard, the Judge offers two reasons:

A helpful chronology of allegations was produced on behalf of the mother for this appeal; this starkly revealed the considerable number of false allegations made by H during his months in foster care. Had the Judge marshalled the evidence in this chronological way, rather than in compartments (see again §66 above), I suspect that he would have been able to see more clearly how the 21 April 2024 allegations sat amidst multiple and diverse false allegations (see again on this point Re: S (A child: Findings of Fact) above).

Fourthly, there was no support for finding 3 from the photographs taken on 31 October 2023. We were told that no comparison had been made at the hearing before the Judge between the photograph of H in his Hallowe’en costume and H’s description of it (in his account of abuse by CH). No photographs connected the children in their Hallowe’en costumes to the maternal grandparents’ home. No photographs connected the children in their Hallowe’en costumes with CH. It is difficult to see why the Judge regarded the mother’s answer as ‘evasive’ and ‘vague’ when she confirmed that on Hallowe’en 2023 she had been with the father, “[A], [H], [L] and the kids, we didn’t even go to my mum and dads”.

Fifthly, it is clear from the incident log that almost as soon as H had made the allegation against CH he withdrew it, telling the foster mother that he was “joking”; the Judge did not, in his final analysis, comment upon this at all. This retraction (partial or complete) required some examination.

Finally, the Judge did not attempt to analyse the evidence relevant to the five different ways in which H said he had been abused by CH at Hallowe’en 2023. Nor was there any explanation in the judgment for how the Judge had converted H’s allegation of “he put his hand up my bum” and “grabbed up his bottom hard” (see §19 above) into a finding that CH had “penetrat[ed] his anus with his fingers/hand and/or plac[ed] his hand between his buttocks” (see §51 above). There was no medical evidence. Such a serious finding warranted explanation.

For these reasons in combination, I am satisfied that finding 3 should be set aside.

I see nothing inherently wrong with finding 4. The parents accept that H displayed the behaviour described, and accept that it was ‘inappropriate’; they dispute the term ‘sexualised’. For my part, I am satisfied that at least some of the behaviours listed (particularly at (iv) and (v)) were appropriately described by the Judge as ‘sexualised’. As it happens, and as I have already mentioned (see §52), finding 4 does not in fact contain a complete list of quite extreme sexualised behaviours as revealed by the evidence.

If finding 3 is, as I propose, to be set aside, the words “as a result of the sexual abuse that H has suffered” in finding 5 would obviously need to be removed. That leaves the conclusion that the inappropriate and/or sexualised behaviours were attributable to ‘inappropriate boundaries in the home’; it is difficult to ascertain from the judgment how the Judge could reach that conclusion. The father made concessions about bad language used in the home but that would not explain H’s behaviours. Suspicion of loose or inappropriate sexual boundaries in the home, based on the maternal family history and/or on the basis of H’s apparent knowledge of sexual matters, would not be enough to support a factual finding. Therefore, albeit with some hesitation, I am inclined to accept the submission of Mr Feehan that the finding of sexualised behaviour being attributed to ‘inappropriate boundaries within the family home’ is too evidentially insecure to stand. This leaves the facts in finding 4 without any attribution; while in ‘threshold’ (section 31(2) CA 1989) terms this is not helpful, the facts in finding 4 are likely nonetheless properly to inform the welfare stage.

Finding 7 should in my view stand. The Judge reviewed with care and at some length the evidence about the relationship between the mother and her family, and the mother’s ambivalence about the risks posed by them. The Judge was entitled to reach the conclusion, based on mobile phone records, nursery records, sightings and admissions, that the mother and father had not been frank with the court about the level of contact between the children and the maternal grandparents over the years. Although the Judge had formed a more favourable assessment of the father, the Judge had also found the father someone who found it difficult to stand up to the mother and they had argued about the level of maternal family contact. The Judge’s conclusion on finding 10 further informs finding 7; for the reasons set out below (§83) I find no proper basis for disturbing finding 10.

It is fair to point out that the maternal grandparents had not been found to be perpetrators of abuse by HHJ Lopez, and after the 2017 proceedings, the mother had of course been permitted to continue living with them at the conclusion of those proceedings. It seems to me from the uncontroverted evidence that the Local Authority had in some respects acquiesced in contact taking place between the parents and the wider maternal family over the years; as recently as summer 2023 the Local Authority had indeed permitted the maternal grandmother to reside with the mother and children for a short period, while the father was excluded from the home (following H’s allegation that the father had smacked him). There was, therefore, some substance, in my view, to the mother’s evidence (recorded in the judgment) that she felt she had received ‘mixed messages’ about the risks posed by the maternal family. So although finding 7 is essentially sound, it needs to be seen in this context.

Given that finding 3 should be set aside, I am of the view that findings 8 and 9 should also fall away.

Finding 10 turned upon a straightforward dispute of fact. The Appellants presented cogent arguments to the Judge about the quality of the recognition evidence and the Judge’s lack of consideration of the gaps in the evidence (i.e., that the paternal grandmother had not been approached for a statement). They repeated these points before us. Ms Taylor argues that the Judge was entitled to prefer the evidence of the nursery worker; she knew R and knew the maternal grandmother. The nursery worker was “95% sure” that she had seen the maternal grandmother with R. The Judge preferred that evidence to the evidence of the parents whose evidence on other aspects he had had cause to doubt, and who had denied this. In my judgment, it was for the Judge to form his own evaluation of the reliability of the witness evidence on this allegation; he had seen and heard the witnesses. It is of course irrelevant whether I would necessarily have reached the same conclusion, but I find it impossible to say that that the Judge was plainly wrong in his assessment (Volpi v Volpi [2022] EWCA Civ 464 at [2] and [65]), and would not therefore interfere with this finding.