Robin Edward Jacobs v R

The provisions of the Sexual Offences (Amendment) Act 1992 apply to this offence. Under those provisions, where a sexual offence has been committed against a person, no matter relating to that person shall during that person’s lifetime be included in any publication if it is likely to lead members of the public to identify that person as the victim of that offence. This prohibition applies unless waived or lifted in accordance with s. 3 of the Act. We will refer to the victim in this case as L.

On 15 September 2022 in the Central Criminal Court before His Honour Judge Hillen, the applicant, then aged 39, was convicted of one count of rape. He was sentenced to four years’ imprisonment. Following refusal by the single judge, the applicant renewed his application for leave to appeal against conviction. On 20 September 2023 the full court adjourned that application, and an associated application for leave to rely upon fresh evidence, to a further hearing, which was to be treated as the hearing of the appeal, in the event of leave being granted. Both applications for leave (and, if appropriate, the appeal hearing itself) now come before us.

The applicant and L started dating in late August 2017. At their first meeting they had consensual vaginal sex in the applicant’s house. At their third meeting just over 2 weeks later, on 17 September 2017, they again had vaginal sex in his home. The prosecution’s case was that he withdrew and then raped L by forcing his penis into her anus without her consent. There was no dispute that anal penetration took place. The applicant’s primary case at trial was that L had consented to that act. Alternatively, if the jury were sure that she did not consent, he had reasonably believed that she had done so.

Accordingly, the prosecution had to prove that L did not consent to the anal penetration and also that the applicant did not believe that she had consented or, if he did, that that belief was unreasonable.

The applicant’s central position is that, in the light of what is said to be fresh psychiatric evidence not available at trial, the conviction is unsafe. Specifically, the jury were not directed to take the applicant’s autism spectrum condition (autism) into account when considering whether his belief that L consented to the anal penetration was, or may have been, reasonable. The applicant’s autism and ability to understand communication from L was a matter that the jury should have been directed to consider when assessing whether his belief in consent was, or may have been, reasonable.

Before us the applicant was represented by Ms Julia Smart KC and Mr David Lawson, neither of whom had appeared below. The prosecution was represented by Mr Jollyon Robertson, both at the second trial and in this court. We are grateful to all counsel for their helpful submissions, both written and oral.

The case took a long time to come to trial. Although L was interviewed in September 2017 and the applicant the following month, he was not charged with rape until June 2020, after the Covid pandemic had begun. The applicant remained on bail until his trial. In November 2020 the trial was fixed to begin on 4 January 2022.

In late 2021 the applicant referred himself to Dr Madeline Cawley, a psychologist, to be assessed for autism. In her report dated December 2021 she said that the applicant met the diagnostic criteria for autism. In a report dated 12 December 2021 Dr Caryl Marshall, a consultant psychiatrist and also instructed by the applicant, reached the same conclusion. She said that the applicant had “persistent deficits in social-emotional reciprocity” and had “made significant adaptations to compensate for this”, including “the use of strategies to mask deficits in cognitive empathy.” Neither report was prepared for the purposes of the trial and the question of belief in consent was not addressed.

On 22 December 2021 counsel then acting for the applicant successfully applied to break the trial date because of the recent diagnosis of autism. Counsel said that the defence wished to obtain a further report addressing two matters: first (and specifically), whether the applicant’s autism would or could have affected his belief in consent and secondly, whether his autism would or could explain his reactions immediately after the incident and the language he used to other people when discussing what had taken place. That first aspect speaks for itself; the second aspect sought to address the risk of adverse inferences being drawn from the applicant’s reaction (of laughter) following the incident and unpleasant jokes and comments about L by the applicant in texts to a friend. Dr Marshall provided a report dated 15 January 2022 dealing with those matters. We address the detail of that report below.

At the end of January 2022 the applicant instructed a new legal team. On 14 February 2022 the trial was fixed to begin on 16 May 2022, subject to the provision of certificates of readiness by 7 March 2022.

The applicant saw his new counsel in conference on 25 February 2022. In his “Observations on trial readiness” dated 4 March 2022 counsel told the court that the defence wished to call Dr Marshall to give evidence to explain the relevance of the applicant’s autism to his “words and behaviour in the immediate aftermath of the sexual activity, and the interpretation of the content of his communications with third parties concerning the incident.” Counsel then stated:

The prosecution decided that they needed to instruct a psychiatrist to consider the autism issue raised by the applicant and so the trial date in May 2022 was vacated. Dr Ian Cumming, a consultant forensic psychiatrist, produced his report dated 14 July 2022. It was served by the prosecution on the defence as unused material on 27 July 2022. Again, we will refer to the detail of that report in due course.

The applicant’s first trial began on 22 August 2022. The jury were unable to agree on a verdict. The second trial, which resulted in the applicant’s conviction, began on 6 September 2022.

The parties agreed facts relating to autism which summarised parts of the psychiatrists’ reports (“Agreed Facts”). The Agreed Facts focused on the implications of the applicant’s behaviour after the incident in line with the approach taken by the defence as notified to the court in March 2022.

The Agreed Facts also referred to two of Dr Marshall’s conclusions on the consent issue. First, Dr Marshall stated that autism had not affected the applicant’s understanding of consent in sexual relationships. He demonstrated a clear understanding of both consent and ability to consent in that context. Secondly, while the applicant has some impairments in understanding “verbal and non-verbal cues” in social interactions as a result of autism, that had not been a significant factor in the incident in question.

In light of the Agreed Facts, neither Dr Marshall nor Dr Cumming was asked to give evidence at trial.

The applicant makes no complaint about the way in which the psychiatric opinions on the relevance of autism to the applicant’s conduct after the incident were summarised in the Agreed Facts and dealt with by the judge in his summing up.

Originally, the applicant sought to argue that the conviction was unsafe because insufficient adjustments were made to the trial process to address the applicant’s autism. That point was abandoned during the hearing before this court on 20 September 2023.

In summary, Ms Smart relies upon observations by this court in R v B (MA) [2013] EWCA Crim 3; [2013] 1 Cr. App. R 36 (R v B (MA)) and submits that the conviction is unsafe because:

The applicant seeks to advance those two grounds first, by reference to the material on autism before the jury. Ms Smart submits that the Agreed Facts were misleading. They were agreed between lawyers, not the experts. The latter did not see the document before it was used at trial. The document did not refer to Dr Cumming’s opinion that autism might have contributed to the applicant’s belief in consent.

Secondly, the applicant seeks leave to rely upon post-trial reports from Dr Marshall and Dr Cumming as evidence admissible under s. 23 of the Criminal Appeal Act 1968 (“the 1968 Act”). Ms Smart submits that this evidence shows that the agreed facts at trial were misleading and that the applicant’s autism was a relevant circumstance which the jury should have been directed to consider when assessing his belief in consent and the reasonableness of that belief. In particular, she refers to the possibility of the applicant missing or misunderstanding non-verbal cues from L. We note, however, that it has never been suggested that the applicant could have misunderstood the words spoken by L during the incident when she told him to stop (or “get out” or “get off”) (as it was common ground she did (at least once)).

L said that she had been looking for a new relationship and had met the applicant on Tinder. On 29 August 2017 they spent some time together getting to know each other and finding that they had things in common. They spent the night together. They had consensual sex several times. In cross-examination L agreed that the sex had been energetic and adventurous. Later text messages indicated that that was expected to continue.

The following week the applicant and L met again for a drink. They did not have sex on that occasion. In cross-examination L accepted that the applicant had been a “perfect gentleman” that evening.

On Sunday 17 September 2017 the applicant and L met for lunch. They went back to his house and had vaginal sex twice. The second time she was on all fours while the applicant entered her vagina from behind for about a minute or so. Then he withdrew and said, “wait there.” She thought that he might be repositioning himself. But then, without any discussion or warning, he forced his penis into her anus. There was no touching or foreplay. He used a huge amount of force. She told him twice to stop. He did not. L said he could not have misunderstood what she was saying. She then screamed “get off” and he did. Overall, this lasted for 20 to 30 seconds. They had never previously discussed having anal sex.

L said she was in a lot of pain. She went to the bathroom and found she was bleeding. She decided to leave. While she was getting dressed, the applicant was lying down, looking astonished or bewildered that she was so upset. He asked her whether she was going to report him to the police and whether he would see her again. She left and went to see a friend.

In cross-examination, L did not recall whether she had sucked the applicant’s testicles. But she accepted that if she had done so, she would not have asked him beforehand. She would have carried on if he did not object. However, anal sex had not been a progression. There had been no discussion about it. L thought that the applicant had been repositioning himself simply because he was not comfortable. She thought he was going to have vaginal sex with her again. L denied that he had moved her buttocks apart or placed his thumb near her anus. She disagreed that she had said “no” only once and that the applicant had withdrawn his penis immediately. When the applicant penetrated her it “hurt massively.” She felt the pain immediately.

The prosecution also relied upon evidence from four people to whom L made complaints, two on 17 September, one on 19 September and one on 20 September 2017. They also relied upon messages and a conversation between the applicant and two other people giving his reaction to the incident.

In his evidence the applicant described the various forms of energetic sex he and L had on 29 August 2017, including him licking her vagina and anus. She was a willing participant. L was confident and forward and there had been a lot of chemistry between them. But there was no discussion of anal sex. The applicant said that although anal sex has a stigma, he regarded it as “mainstream” in a relationship and had done it with the majority of his previous girlfriends.

The applicant said that on 17 September 2017 they had vaginal sex twice. The second time he asked L to go on all fours. They had sex in that position for a couple of minutes. He took his penis out. He thought the natural progression would be anal sex. He spread L’s buttocks to expose her anus and placed his thumb next to it as a marker. He told her to “hold steady”. He changed his position and angle and then inserted his penis just past the tip. Within two seconds L said, “get out” and he did so immediately. He was shocked that L’s attitude then changed completely. She became withdrawn. He tried to laugh it off. After she went to the bathroom he kept asking her if she was okay. She got dressed. He asked if he would see her again. Initially she said yes and then said that she was not sure. She left. He was worried that something would happen later because she was so upset.

The applicant said that when he spread L’s buttocks it would have been obvious to her what he was doing, and she had ample time to tell him to stop. He denied that the penetration lasted for anything like 20 or 30 seconds.

The applicant said “I was looking to use enough force to penetrate her. It was tight and I couldn’t go in immediately.” He said that he did not use lubricant. This was because he had previously had anal sex without lubrication and his penis was lubricated from the vaginal sex. He did not use a condom. The applicant agreed that anal sex is a “particularly intimate act” but he did not think it was different from other sexual activity so as to require “special consent.”

The defence relied upon the applicant’s good character. This was positively supported by a number of character witnesses.

A statement of facts agreed between medical experts instructed by the prosecution and the defence stated that on 19 September 2017 L had a fresh-looking laceration 2.5cm long at the anal margin with slight bruising. It was not possible to exclude a medical cause of that laceration, but it would have required the application of “at least moderate force”. It was not possible to tell from the medical evidence what had been the duration of the penetration or whether the injury had been caused by consensual or non-consensual activity. The injury was therefore consistent with the accounts of both L and the applicant.

In her report dated 15 January 2022 Dr Marshall set out the questions she had been asked to address. Question (i) was:

At paragraphs 37 to 38 of her report Dr Marshall said:

Question (ii) was:

Dr Marshall observed in the applicant impairments in social communication and interaction and difficulty in understanding some verbal and non-verbal cues (paragraph 39). The applicant said that he may miss minor verbal or non-verbal cues, but he did not believe that this affected his ability to determine whether L consented to sexual acts, whether before or during the incident (paragraph 41). He said that in previous sexual encounters he had been able to judge what a person might be willing to do. For example, he had not initiated anal sex where he could tell that someone was reserved or conservative. He gave examples of where he had understood a partner to consent to some acts but refuse anal sex (paragraph 42). He thought from the texts he had exchanged with L and the acts they had already performed that L was open to anal sex. He said his actions before the anal penetration gave L “ample opportunity to tell me to stop” (paragraph 43). When L had said “get out” he stopped immediately because he thought she had withdrawn her consent at that stage (paragraph 44).

On question (ii) Dr Marshall concluded at paragraph 45:

Question (iii) asked:

The applicant reported that he has developed strategies for meeting people for the first time and for managing social interactions (paragraphs 51 and 52). Dr Marshall referred to this as “social camouflaging”. It is used by some individuals with autism to disguise their differences and adapt socially (paragraph 54).

In his report dated 14 July 2022 Dr Cumming did not identify any questions that he had been asked to address.

Dr Cumming described in some detail the applicant’s account of the “rough and experimental sex” which he and L had enjoyed, leading the applicant to believe that she consented to anal sex (paragraphs 30 to 60).

Dr Cumming agreed with the diagnosis of autism (paragraph 98). Dealing with those with autism in general Dr Cumming said at paragraph 102:

Turning to the applicant, Dr Cumming said at paragraphs 103 to 104:

At paragraphs 108 and 110 Dr Cumming concluded:

The applicant waived legal privilege in relation to his legal team at trial. We commend trial counsel for his detailed and carefully prepared note on the grounds of appeal. The applicant has not raised any material issue about the note’s description of the defence case, its preparation and presentation.

At the conference on 25 February 2022 it was the joint view of counsel, the instructing solicitor and the applicant that Dr Marshall had excluded the applicant’s autism as having any relevance to the issue of consent or reasonable belief in consent. Counsel considered that paragraphs 37 and 38, 42 and 45 of Dr Marshall’s report represented her concluded opinion on those issues.

Counsel says that an important part of the context was that the applicant’s instructions did not indicate any mistaken understanding by him of the circumstances of the sexual activity. On the contrary, his instructions were that L was mistaken about the circumstances leading up to and after the anal penetration.

Consequently the primary case the applicant decided to advance at his trial was that L had consented to anal intercourse but changed her mind after it had begun. Given the applicant’s version of events, reasonable belief in consent was not advanced as the applicant’s primary case. The joint view in conference was that that alternative could dilute or undermine his primary defence that L had consented. This conclusion took into account the lack of any support in Dr Marshall’s report for a misunderstanding by the applicant of consent or signals associated with consent. In those circumstances it was considered unnecessary to seek any clarification or review of Dr Marshall’s opinion, given that it was substantially based upon her assessment of the applicant’s own account of events.

Trial counsel considered that the applicant had a clear understanding of Dr Marshall’s report and its implications for him personally. The applicant was keen to give evidence at the trial and did not wish to create the impression that he was “hiding” behind the diagnosis of autism.

In relation to Dr Cumming’s report, trial counsel said that there was concern that, although it contained some statements favourable to the applicant, it was not clear what he was saying. It was a “mixed bag” and the expert was “sitting on the fence.” Whilst the final sentence of paragraph 110 of the report appeared to support a defence of reasonable belief, that had to be balanced against the whole report. The prosecution had not been prepared to admit paragraph 110 because it was ambiguous.

A conference was held with the applicant on 12 August 2022. Three options were discussed:

The applicant had a strong preference for relying on the Agreed Facts because they established the autism diagnosis and went some way to assisting him to explain his conduct after the incident. Counsel favoured contacting Dr Cumming to ask him to clarify his opinion. Several attempts were made to reach him over the following days, but there was no response. In these circumstances, a joint decision was made to deal with the autism issue by relying upon the Agreed Facts.

Counsel said that that the Agreed Facts were not put to the experts for comment because it was thought unnecessary to do so. The document relied extensively upon Dr Marshall’s report. It did not misrepresent Dr Cumming’s opinion as reported in July 2022 and was not misleading.

The judge gave written directions to the jury which were reflected in his oral summing up. On the issue of whether the applicant reasonably believed that L had consented to anal sex, the direction was as follows:

No criticism is made of the judge’s directions in relation to the applicant’s case as advanced at trial. It is not suggested that the judge failed to summarise accurately or adequately the factors or cues relevant to reasonable belief in consent. We also note that the judge gave a very clear and fair direction that the jury should bear in mind the delay which had occurred and its possible impact on reliability of memory and whether the prosecution had made them sure of the applicant’s guilt. They were told to appreciate the risk of prejudice to the applicant.

Dr Cumming has prepared a further report dated 7 December 2022. Paragraph 1 states that this report should be read in conjunction with the earlier one dated 14 July 2022. Dr Cumming did not interview the applicant again, but was supplied with letters from him. Those letters are said to have included comments on a need for reasonable adjustments at the trial, a matter no longer relevant to this application. Nothing else is said about the content of those letters. Dr Cumming says that he had been requested to prepare the report by Mr Michael Charles, the applicant’s former employer. But he says nothing about what instructions or questions were put to him.

Dr Cumming states that his comments about the applicant in the first report remain pertinent and correct (paragraph 3). Paragraphs 4 to 8 then deal with the subject of reasonable adjustments which, as we have said, is no longer relevant.

Some of the report simply makes generalised statements about people who have autism (e.g. paragraph 9).

Dr Cumming says that an agreed statement of facts on a subject such as autism is best prepared by clinicians. A document prepared by non-clinicians, and not seen by the clinicians, runs the risk of misinterpretation and overlooking key issues regarded by the experts as important. Dr Cumming gave one example. High functioning individuals such as the applicant may have the same autism-related difficulties as other individuals with autism who are not high functioning, but their difficulties may be more “masked.” As time goes on individuals with autism develop strategies to mask and overcome their difficulties, giving the impression that autism is not present (paragraphs 13 to 14).

But the possibility of the applicant appearing not to have autism is not relevant to this application for leave. Instead, the issue is whether the applicant missed or misinterpreted non-verbal cues relevant to the issue of reasonable belief in L’s consent. On that matter Dr Cumming’s second report said this at paragraphs 15 to 17:

Dr Marshall produced an “addendum report” dated 19 January 2023. She states that that report should be read together with her report dated 15 January 2022. Dr Marshall has not interviewed the applicant again but has read letters from the applicant relating to the trial. We have not seen that material. The report sets out a series of questions put to the expert by Mr Charles. A number of the questions are of no relevance to the application before this court.

When asked about the adequacy of the agreed facts prepared by trial counsel, Dr Marshall said at paragraph 30:

At paragraph 9 Dr Marshall was asked whether the applicant’s legal team had set out for her the legal ingredients of rape and the issues raised in R v B (MA) before she prepared her January 2022 report. She responded that that had not happened, but she had been aware of the ingredients of rape, including the absence of a reasonable belief by the defendant that the complainant had consented to the penetration. She added that she was not asked to consider whether the applicant’s belief in consent was reasonable. However, Ms Smart accepted that that ultimate issue was not a matter for an expert to opine on. We agree. It was solely a matter for the jury.

Dr Marshall was asked whether, if she had been advised about R v B (MA), she would have wanted to make any further comments relevant to the reasonableness of the applicant’s belief. She responded at paragraph 13:

Dr Marshall was asked whether the personal attributes of the applicant may have been a relevant factor as to the reasonableness of any belief he may have had in consent. The question was: “In what way, if any, was his belief affected by his autism?” At paragraph 20 Dr Marshall responded:

Ms Smart’s submissions fell into two parts. First, she argued that the errors in the process leading up to and including the trial rendered the conviction unsafe. Secondly, the additional expert evidence, if admitted by this court, demonstrates the inadequacy of the agreed facts placed before the jury and that the applicant’s autism should have been treated as relevant to the issue of reasonable belief in consent.

She pointed to the fact that s. 1(2) of the Sexual Offences Act 2003 (“the 2003 Act”) required the jury to have regard to all relevant circumstances. Applying observations by this court in R v B (MA), the reasonableness of the applicant’s belief in consent depended upon his reading of “subtle social signals” and so his impaired ability to do that was relevant to that issue. Ms Smart submitted that Dr Cumming’s first report at paragraph 110 engaged aspects of the applicant’s autism which were “potentially relevant” on that point. She sought to reinforce this submission by relying upon the views of the psychiatrists on “relevance” in their additional reports (see Dr Marshall at paragraphs 13 and 20 and Dr Cumming at paragraphs 15 to 17).

Ms Smart submitted that the errors began when the experts were instructed. They were not directed to consider the reasonableness issue, nor R v B (MA). Nevertheless, there was material in their initial reports sufficient to render the applicant’s autism relevant to that issue.

However, at the trial autism was dealt with solely in the Agreed Facts. That document was agreed by the parties’ legal teams who failed to refer it to the psychiatrists for their comment. As a result the Agreed Facts did not contain relevant material which the jury ought to have been allowed to consider, in particular Dr Cumming’s first report at paragraph 110. Furthermore, if the experts had been consulted, they would have provided further clarification of their opinions as set out in their additional reports.

Both the prosecution and the defence wrongly treated autism as being irrelevant to the issue of reasonable belief in consent. Consequently, the judge wrongly failed to give the jury a direction on the relevance of autism to that matter.

During their deliberations the jury sent a note to the judge asking, “can we have a legal definition of ‘reasonable’?”. The judge responded that there was no such definition, because it was for the jury to set the standard of what is reasonable, having regard to all the circumstances of the case. He then summarised again what the prosecution and defence said on this issue. Counsel did not ask for any further direction to be given. Ms Smart submits that the jury’s question is significant in relation to the failure to include a direction in the summing up on the relevance of the applicant’s autism.

In the second part of her submissions Ms Smart said that the additional expert evidence meets the requirements in s. 23 of the 1968 Act to be admitted by this court. It satisfies the overarching requirement in s. 23(1) that the receipt of the evidence be “necessary or expedient in the interests of justice” as well as the criteria in s. 23(2). The experts were directed to attend the hearing remotely in case the court decided that they should give live evidence. Ms Smart did not ask for the experts to give any further evidence. She was content for the application, and any appeal, to be determined on the material contained in the written reports. We agreed with that approach.

Ms Smart submits that the additional evidence demonstrates the misleading nature of the Agreed Facts provided to the jury. Both experts clearly state that autism is relevant to the applicant’s ability to appreciate L’s wishes and may have affected his ability to understand whether she was consenting to anal sex. The fact that a person with autism does not have the same level of perception of non-verbal cues as someone without autism, does not make the beliefs they form from another’s responses any less reasonable. Therefore, a jury “must be allowed to consider autism when assessing whether the belief formed by an accused was reasonable or not.”

To decide that autism is not a relevant circumstance, she says, would mean that those whose rational behaviour is influenced by autism will suffer discrimination in that they will be judged by “a standard that fails to take into account their reality.”

Section 1(2) of the 2003 Act provides:

In R v B (MA) the issue was whether the trial judge had been correct to direct the jury that they should disregard the defendant’s paranoid schizophrenia when deciding whether he had a reasonable belief in the complainant’s consent for the purposes of s. 1(1) of the 2003 Act. Giving the judgment of the court, Hughes LJ said at [34] that the appeal should fail in any event because the appellant’s delusional beliefs had not led him to think that his victim was consenting when in fact she was not. But he went on to say that if the court was wrong about that, such delusional beliefs could not in law render reasonable a defendant’s belief in consent. The 2003 Act does not ask whether it was reasonable (in the sense of being understandable or not his fault) for the defendant to have his mental condition. It asks a different question: whether the belief in consent was a reasonable one. A delusional belief in consent would be, by definition, irrational and therefore not reasonable (at [35]). Hughes LJ stated at [40] that “Beliefs in consent arising from conditions such as delusional psychotic illness or personality disorders must be judged by objective standards of reasonableness and not by taking into account a mental disorder which induced a belief which could not reasonably arise without it”.

Under the previous law a genuine belief in consent, whether reasonable or not, was a complete defence to rape. The test was subjective (see Director of Public Prosecutions v Morgan [1976] AC 182). The 2003 Act deliberately departs from that model, and also from the approach taken in the law on self-defence, by requiring not only that a defendant’s belief must be genuinely held but also that it must be reasonable in all the circumstances of the case (see [36]).

The applicant’s grounds of appeal are based upon the following obiter observations of Hughes LJ at [41]:

We note the following points on this passage:

We agree with these observations. On this basis, depending on the evidence, the fact that a defendant accused of rape has autism may be relevant to the question of whether a belief in consent was reasonable, and the jury may need to be directed accordingly.

In this court the applicant’s arguments depend upon parts of the expert reports upon which he seeks to rely. Section 7.1 of the Criminal Practice Directions 2023 deals with the admissibility of expert evidence. Paragraph 7.1.1 provides that it is admissible if:

Part 19 of the Criminal Procedure Rules deals with expert evidence. By rule 19.3(3) a party wishing to introduce expert evidence, other than as “admitted fact”, must serve a report complying with rule 19.4 which specifies what a report is required to contain. By rule 19.4(h) a report must:

R v Dunleavy [2021] EWCA Crim 39 was concerned with the admissibility of a psychologist’s report diagnosing the applicant’s autism where he was alleged to have been involved in the preparation of terrorist acts. It was said that the applicant’s pursuit of information about guns was a compulsive, obsessive behaviour capable of being attributed to autism. This court upheld the judge’s decision not to admit the evidence. It decided that the justification for applying to introduce the evidence must be set out in the expert’s report. The judge had been correct not to hold a voir dire to enable the applicant to adduce evidence not contained in the report. In that case the issue on admissibility concerned relevance. The report had failed to focus sufficiently on the applicant rather than the range of effects that autism may have those who have it. There was a lack of evidence, for example, that the applicant had any compulsive, obsessive disorder.

In R v BRM [2022] EWCA Crim 385 the applicant argued that the trial judge had been wrong to exclude psychiatric evidence diagnosing him as having autism in relation to his case that he had stabbed the deceased in defence of another. The court held that it was for the trial judge to decide whether expert evidence on its face can properly be considered to be objective and reliable. If the expert evidence does not engage with the evidence in the case and does not tether any opinions expressed to that evidence, it will serve only to confuse the jury. The court (at [38]-[39]) rejected the suggestion that the evidence should have been admitted, leaving any deficiencies to be dealt with by cross-examination.

The court held that the judge had been entitled to conclude that the report was neither reliable (referring to Crim. PR 19.4 (h)) nor relevant, and therefore was inadmissible. The applicant’s case was that he produced and used his knife on the deceased because of his honest belief that the latter was about to pull out a knife. The applicant said he saw the deceased’s hand go to his waistband and he was known to carry knives. The psychiatrist’s opinions were predicated on how a person with the applicant’s characteristics would respond in a generalised fight situation. They did not relate the diagnosis of the applicant’s autism to what he said he had done (at [41]-[50]).

Finally, the statutory test for the admission of fresh evidence on appeal, or an application for leave to appeal, is set out in s. 23 of the 1968 Act as follows. The Court of Appeal may, if it thinks it “necessary or expedient in the interest of justice” receive any evidence which was not adduced below (s. 23(1)). By s. 23(2), in considering that question, the Court of Appeal shall have regard in particular to:

We reject at the outset the suggestion that trial counsel did not have in mind the possibility that the applicant’s autism might be relevant to the reasonableness of his belief in L’s consent. The very purpose of breaking the trial date of 4 January 2022 was to enable Dr Marshall to consider whether the applicant’s autism could have affected his belief in consent. The questions put to Dr Marshall were relevant to that issue. When new trial counsel saw the applicant in conference on 25 February 2022 he advised that Dr Marshall’s report provided no support for the applicant having had a mistaken understanding of consent or signals associated with consent. Plainly that was in the context of the applicant’s autism. Although counsel’s note does not refer to R v B (MA) expressly, the advice he gave plainly shows either that he had the observations of Hughes LJ in mind or that he acted in accordance with those observations. That line of argument was investigated. We have already referred to the reasons why the applicant chose not to pursue that approach.

Furthermore, the matter was reconsidered when Dr Cumming’s report was served by the prosecution towards the end of July 2022. Trial counsel focused on paragraph 110 of that report, the very paragraph upon which Ms Smart has relied in her submissions before us. He rightly advised that Dr Cumming’s report was ambiguous and needed to be clarified with the expert. At his instigation the instructing solicitor made efforts to contact Dr Cumming, albeit unsuccessfully. This course of action was plainly referable to the reasonable belief in consent issue and the approach suggested in R v B (MA) at [41]. Here we should note that there is no suggestion that counsel should have made an application to adjourn the trial so that Dr Cumming could be approached by the defence and called to give evidence, if that was judged to be advisable.

We also reject the contention that the Agreed Facts put before the jury were misleading or failed to set out the position of the psychiatrists adequately. As trial counsel rightly advised, Dr Marshall’s report excluded the applicant’s autism as having any relevance to the applicant’s reasonable belief in consent. Dr Marshall’s report in January 2023 does not show that advice to have been incorrect on the material available by the time of the trial.

We do not consider that any additional parts of Dr Cumming’s report in July 2022 could or should have been included in the Agreed Facts. Part of Dr Cumming’s report merely discussed in general terms difficulties which people with autism may have (see e.g. paragraph 102). At paragraph 103 Dr Cumming said that it was difficult to link the applicant’s autism and the incident in question. The connection was “complicated” and “theoretical.” Paragraph 104 of the report does not assist the applicant. It begins with a generalisation. Dr Cumming then suggests that a difficulty in reading cues in social situations may well apply to the applicant “in terms of the offence”, but that was counter-balanced by the absence of such difficulties in the applicant’s previous relationships. Likewise the applicant gained no assistance from paragraph 108.

The applicant’s argument essentially depends on paragraph 110 of the report. Here Dr Cumming merely suggested that the applicant’s autism might have had some bearing on the interpretation of non-verbal cues, but the issue was subtle and affected by hindsight. Although it was not a dominant theme in the incident, it might have contributed to how he perceived L and what was permissible. At best this was a very tentative proposition. It appears to suggest the possibility that the applicant’s autism might have made a relatively minor contribution to belief in consent. In that situation, it was all the more important for the expert to explain how that opinion related to the circumstances of the incident. But here that opinion was not tethered to the evidence about the circumstances of the incident, or to what had happened previously between the applicant and L or in previous relationships.

In any event, the applicant faces the further difficulty that the tentative suggestion in paragraph 110 is undermined by the end of paragraph 104. There, Dr Cumming stated that the absence of any difficulty in the applicant being able to read social cues in previous relationships and encounters weighed against the suggestion that that ability was impaired by his autism during the incident in question. It was therefore necessary for the expert to explain why he had suggested that the applicant’s autism was a material factor in relation to that incident. Although, the defence tried unsuccessfully to contact Dr Cumming before the trial, the inconsistency between paragraph 110 and the earlier part of his report was not resolved.

In these circumstances, Dr Cumming’s suggestion in paragraph 110 that the applicant’s autism could have had a small part to play in his reading of social cues was unreliable. It would also have been misleading to put paragraph 110 before the jury without paragraph 104 (and the respondent would clearly not have agreed to such a course). The report would then have confused rather than assisted the jury. The report did not show that the applicant’s autism was, or might have been, relevant to his belief in L’s consent or the reasonableness of that belief by reference to the circumstances of the case.

On this analysis, it follows that there was no need for any direction to have been given to the jury about the applicant’s autism in relation to the issue of his reasonable belief in consent. The fact that the jury asked for a direction on the meaning of “reasonable” does not alter the position.

In a witness statement dated 14 September 2023 the applicant appears to criticise his trial counsel for advice that he gave. He says that he was told that the reasonableness of his belief in consent was an objective rather than a subjective test, and that although his autism was relevant to the way he had spoken and acted after the incident, it was not relevant to the issue of reasonable belief in consent. For the reasons we have given, however, the additional evidence did not assist the applicant on the question of reasonableness of belief. Counsel was correct to say that the reasonableness test is objective rather than subjective. That flows from s. 1(1) and (2) of the 2003 Act. It is an objective test applied to all the relevant circumstances of the case. In some instances that might include specific characteristics of the defendant attributable to autism (see [79]-[81] and [84]-[86] above). But in this case the applicant’s specific characteristic of autism was not relevant to reasonable belief in consent on the available evidence.

The remaining issue is whether the additional psychiatric evidence prepared after trial would alter the above analysis, assuming that it would be admissible under s. 23 of the 1968 Act. We have considered the evidence de bene esse.

Only paragraphs 13 and 20 of Dr Marshall’s report in January 2023 are potentially relevant to the question of reasonableness of belief. She says that in view of the applicant’s previous experiences of consensual sexual activity and his deficits in social interactions attributable to autism, it is possible that he missed cues specific to the incident. On this basis it is said that his autism was an “important consideration” when considering whether his belief in consent was reasonable. In paragraph 20 Dr Marshall said that this was “theoretically possible.”

There are several difficulties with this evidence.

First, Dr Marshall asks that her reports in January 2022 and January 2023 be read together. Both are based upon the same account given by the applicant. There has been no material change. But she does not acknowledge that the opinion she now puts forward is contradicted by her earlier report, notably paragraph 45. Of course, an expert is entitled to change their mind, but if they do so they must clearly set out what the change of opinion is and the reasons for it. The lack of explanation for the apparent change of opinion in the most recent report is striking.

Secondly, Dr Marshall’s additional report does not tie her revised opinion to the evidence about the circumstances of the incident, or to what had happened previously between the applicant and L or in previous relationships and why she considers the applicant’s autism made a material difference to those matters. For example, there is no explanation as to what non-verbal cues may have been missed or misinterpreted, or how the missing or misinterpretation of those non-verbal cues was, or might be, relevant to whether any belief by the applicant in L’s consent was reasonable.

Thirdly, Dr Marshall describes any contribution of the applicant’s autism to his belief in L’s consent as a theoretical possibility. Dr Cumming made essentially the same point in his first report at paragraph 103. We do not see how a theoretical possibility of the kind described by the experts could assist a jury meaningfully. Rather it is likely to confuse proper decision-making.

In relation to Dr Cumming’s second report, as we have explained, the applicant’s case rests on paragraphs 15 to 17. Here again there are serious flaws in the evidence presented.

First, Dr Cumming does not acknowledge the inconsistency in his first report (see [93]-[94] above). Although the applicant’s legal team at trial sought to obtain clarification from Dr Cumming, there has been no further explanation for the purpose of these applications.

Secondly, the general difficulties which a person with autism, including the applicant, may have in interpreting non-verbal cues is not related to the evidence in the present case as regards the incident and the applicant’s previous experiences with L and others.

Thirdly, paragraph 16 of the second report does not acknowledge that L gave a very different account of the incident. In his first report Dr Cumming referred to the difficulty of relying upon the applicant’s autism to explain the differences between the two accounts (paragraphs 103 and 110).

Fourthly, paragraph 17 of the second report describes factors which might often be relied upon by a defendant who does not have autism. The report does not explain how the applicant’s autism would, or might, make a difference in the specific circumstances of this case. Given that Dr Cumming, like Dr Marshall, has described the connection between the incident and the applicant’s autism as “theoretical” in his first report (paragraph 103), that omission is significant.

There are therefore significant difficulties in terms of admissibility of the fresh expert evidence for the purpose of s. 23(2)(c) of the 1968 Act, and reliability more generally.

In addition, there has been no reasonable explanation for the failure to adduce the expert evidence at trial. On the contrary, the applicant chose to make the tactical decision described by trial counsel. The applicant is an intelligent person – indeed was a qualified practising lawyer - who, as his counsel at the trial has made clear, well understood the psychiatric evidence and the issues that he was facing.

For all these reasons, it is not necessary or expedient in the interests of justice for the additional expert evidence to be admitted. This is a case which demonstrates why finality in litigation is an important consideration (see e.g. R v Foy [2020] EWCA Crim 270).

For these reasons, and having considered all the submissions and materials in this case fully, we reach the conclusion that it is not arguable that the applicant’s conviction is unsafe. The application for leave to appeal against conviction, and the associated application for leave to rely upon additional evidence, are refused.