Nicholson v R.

This is an appeal against conviction brought with the leave of the single judge. The appeal raises issues as to the proper treatment of evidence said to be cross admissible as between one count and other counts in the indictment, and as to the judge’s decision to permit one of the counts in the indictment to go to the jury. The complainants are entitled to their anonymity and we shall use initials when necessary.

At a re-trial before HHJ Milford QC sitting at Newcastle-upon-Tyne Crown Court the appellant faced an indictment containing five counts as follows:

The appellant, now aged 59, qualified as a registered general nurse in 2000, and the following year was employed at the Spire private hospital in Washington, Tyne and Wear. At the time of the allegations brought by JU, LR and ES he was employed in the surgical recovery room looking after patients who were coming round after surgical procedure under general anaesthetic in theatre. It was the prosecution case that on several occasions the appellant sexually assaulted women under his care, some of whom were and some of whom were not recovering from anaesthesia.

The jury heard evidence concerning two further patients who were not the subject of any count in the indictment. The first was a woman whom we shall call SK who attended the Spire on 30 May 2006 for the removal of an ovary. She came forward after the commencement of the first trial. Her evidence was admitted at the first and second trials as ‘bad character’ evidence pursuant to section 101(1)(d) Criminal Justice Act 2003. It was suggested that her complaint fell into the same pattern of offending as that represented by the five count indictment and it was the first incident in time.

The witness gave evidence that once back in her room after surgery she became aware of her bed sheet being pulled back, her gown lifted and a hand touching her pubic area. She opened her eyes and saw a male nurse on the left side of her bed. She had been in his care on earlier occasions. His hand then squeezed her breast. She pretended to be asleep. He said, “Does that feel nice?” He replaced her bedclothes and left. She saw him going towards the door. She saw him again the following day when his behaviour was normal. She told no-one of her experience. In January 2010 she saw a picture of the appellant in a local newspaper in connection with the first trial. She then spoke to her brother’s partner. As a result of her encouragement she eventually told her brother what happened and, subsequently, the police. The appellant gave evidence that he did indeed attend to the patient in the recovery room but at no time did he touch her inappropriately.

Turning to count 1, SG was admitted on 12 March 2007 for breast augmentation. The complainant was not under the influence of anaesthetic. She complained that as she was getting ready for a visit from her mother after the operation, and the appellant was assisting her to put on a sports bra. He lifted her breasts into the bra before zipping the front. A care assistant, Susan Whittingham, gave evidence that it was not necessary for the appellant to handle the patient’s breasts which would have been hard and swollen. That task was for the patient herself, while the assistant would merely zip up the bra. SG made no complaint at the time. When interviewed by the police, who contacted her by telephone, she said there had been “a creepy old man”. She made her statement the following day. The defendant agreed that he had attended the patient. He could not recall the complainant but he would not have lifted her breasts.

As to count 2, JH was admitted on 8 October 2007 for breast reconstruction. It was necessary following the procedure to carry out a bladder scan for the purpose of ascertaining whether the patient was retaining liquid. The patient was not at that time recovering from the effects of anaesthesia. Gel was applied to the patient’s abdomen and a probe placed over it in order to produce the scan. It was necessary to lower the patient’s knickers to the level of her pubic hair. The complaint was that the male nurse conducting the scan had pulled down the patient’s knickers on the right side to thigh level, thus exposing part of her vagina. The complainant’s husband supported this account. The appellant accepted that he carried out the scan. He accepted that if he had lowered the complainant’s knickers as alleged it was unnecessary to do so. His evidence was that he had done only what was required. The issue was whether the appellant had exposed any part of the complainant’s vagina.

As we have said, the jury found the appellant not guilty of both counts 1 and 2. Neither of the complainants was, at the material time, recovering from the residual effects of anaesthesia when the alleged assaults occurred. In respect of each case there was a straightforward issue whether the appellant had behaved indecently towards the complainant.

As to count 3, on 16 June 2008 JU was admitted for the removal of her adenoids and the fitting of grommets. She was taken from surgery to the recovery room under the effects of general anaesthetic. The complainant said that she recalled being with a male nurse in the presence of a female. When the female left, an inference she drew from the sound of her footsteps, the male nurse placed his hand under her gown and caressed her right breast. As the female returned he removed his hand. When they were next alone he repeated his action. This happened repeatedly until the male nurse began to manipulate her nipple. Although conscious, the complainant was unable to speak or to move. JU made no complaint at the time. She said she had not wanted to tell anyone. However, she returned to the same hospital in May 2009 for cosmetic surgery and, before she did, she told her friend that last time there had been a man present who gave her the creeps. On her friend’s advice JU asked a female nurse at the Spire whether this time she could have a female chaperone. She gave a description of the male nurse who had previously attended her. The nurse responded by hugging her and telling her not to worry because the male member of staff had been suspended. The appellant said that he had no personal memory of the patient but could see from the records that both he and other males looked after the complainant. All sorts of people were going in and out. He denied impropriety.

As to count 4, LR was admitted for breast surgery on 23 June 2008. While in recovery after general anaesthetic she said she was conscious but unable to move or to open her eyes. A man with a Tyne and Wear accent was talking to her. He placed his hands between her legs and placed his fingers inside her vagina. She was shocked and confused. When she was able to open her eyes she saw a male nurse who spoke with the same voice and accent as the man who had assaulted her. Following her discharge, the complainant, on 29 June, sent an anonymous email to the hospital director. As a result of the follow up she agreed to speak to the police. The complainant said she did not think that she could identify her assailant and when she took part in an identification procedure she did not identify the appellant.

The other male who attended LR after her operation was a healthcare assistant, Michael Hume. He gave evidence that he could not have attended the patient alone because he was not, as an assistant, allowed to do so. He did not recall the patient and was not the person who assaulted her. Mr Hume agreed that the patient’s notes indicated that he had spent the first 15 - 20 minutes with the patient following her arrival at recovery from the operating theatre. She was recorded as being unresponsive at 4.25 pm and 4.30 pm. She was said to be responding to voice at 4.35 pm. At those times he would not have been alone with the patient. When asked to examine the notes again in re-examination he said that the times noted at 4.35 pm and 4.40 pm were not in his handwriting although the nursing entries were. From 4.45 pm onwards it was accepted that the handwriting in the notes was the appellant’s. The appellant denied that he was the patient’s assailant. The prosecution observed that it was clear when the appellant gave evidence that his voice was very different in tone from that of Mr Hume. That was not a material consideration at the time Mr MacDonald QC made a submission of no case to answer.

As to count 5, ES was in hospital on 11 September 2008 for an operation to her shoulder. She gave evidence that as she was coming around in the recovery room she felt a hand under her bedclothes. It reached inside her paper underwear and was placed on top of her vagina. It was removed and the same thing happened a second time. When she was able to open her eyes she saw a male nurse whom she described. She talked to him. She felt confused. When she was discharged home she made a complaint to her mother and informed the police. The theatre manager Janice Linley gave evidence. She said that she wheeled the patient from theatre into the recovery room where she was left in the care of the appellant. When she returned she saw that, for some reason, the curtains around the patient had been drawn across and she heard the appellant say from behind the curtains, “Sorry this is a bit personal”. When she drew the curtain back she found the appellant standing beside the patient with ECG leads in his hands. Mrs Linley said nothing to the appellant but left the curtain open and the following day spoke to the hospital manager. In cross examination Mrs Linley accepted that patients in recovery had ECG electrodes attached to their chests to monitor heart rate. It was the recovery nurse’s job to remove them when appropriate. She agreed that it may have been appropriate to draw the curtain around the bed for this purpose to protect the patient’s privacy, especially when there was a male patient in the next bed, as there was on this occasion. The appellant denied that he had behaved inappropriately towards this patient.

The jury also heard about a further incident which involved a patient to whom we shall refer as SH. She was admitted for a surgical procedure under general anaesthetic in March 2008. Immediately afterwards she said she was aware of a sensation in her vagina as though sexual intercourse had taken place. When she later made contact with the hospital it emerged that the appellant had not been on duty on that day. This was, the defence argued, an important incident since it demonstrated that a perfectly honest witness may make a genuine mistake about a sensation felt while recovering from general anaesthetic.

The possibility of false memories of sexual abuse by the counts 3, 4 and 5 complainants was the subject of evidence from three eminent expert witnesses. Professor Mark Forest gave evidence for the prosecution. He is a working hospital consultant in anaesthetics and critical care. Professor Ian Hindmarch also gave evidence for the prosecution. He is a professor in human psycho-pharmacology, the study of the effect of drugs on the mind. Professor Hindmarch had a particular interest in the side effects of anaesthetic drugs. It was the defence case that witnesses who described sensations of sexual interference may have been experiencing the unwanted side effects of recovery from anaesthesia. In support of that case they relied upon the evidence of Professor Alan Aitkenhead who had been senior lecturer in anaesthesia and honorary consultant in Leicester from 1979 until 1988. He became Professor of Anaesthesia at Nottingham and honorary consultant from 1989 until his retirement in July 2010. As honorary consultant he worked as a clinician for about half the time in clinical practice as would a NHS consultant anaesthetist. He was editor of a standard text book, ‘The Text Book of Anaesthesia’ which was in its sixth edition.

Dr Forrest explained that modern anaesthetic drugs had three constituent parts and purposes, called the triad: sleep induction, muscle relaxant and anaesthesia. During recovery of consciousness, which was designed to be swift, the stages were (i) recovery of auditory speech or sound perception, followed (ii) by touch and movement sensation and, finally, (iii) visual awareness and alertness.

It was common ground between the experts that there were, in theory, two possible conclusions which the jury could reach upon the evidence of the witnesses SK, JU, LR and ES: either they were relating experiences accurately recalled or, honestly believing in the reality of their experience, their memory was in fact false having been induced by the effects of anaesthesia. Those false memories may have been created by (1) hallucination while awake, (2) dreaming while asleep or (3) misperception of an ‘innocent’ experience, such as an appropriate touching during a nursing procedure, while in a state between sleep and alertness.

The preponderance of the evidence was that if the witnesses’ memories were false, they were caused either by dreaming or by misperception.

Dr Forrest, having examined the medical records, including the recovery room nursing notes recording the patients’ stages of recovery from anaesthesia, and having considered the accounts given by the witnesses, expressed the opinion that the assertion of false memory was “considerably weakened” by the absence of amnesia and the impressive recall of the complainants. Professor Hindmarch considered that dreams and hallucinations could be excluded. The witnesses appeared contemporaneously to have been using their cognitive faculties both to decide upon and make sense of their experiences. They therefore appeared to be alert. It was possible that an experience could have been misperceived, but only if an innocent medical procedure had taken place which was capable of being misconstrued. Professor Hindmarch would not express an opinion as to the latter because post operative nursing procedure was outside his field of expertise. Professor Aitkenhead expressed the view that he could not exclude the possibility that false memory had been induced by sleep or misperception under the residual influence of anaesthetic drugs.

It was the common view that experiences similar to those described by the witnesses were rare.

The appellant argues that the judge should have acceded to a submission of no case to answer on count 3. First, the evidence was vague and insubstantial; secondly, the prosecution had failed to establish a prime facie case that the appellant had the opportunity to commit the offence. He had not been responsible for personal observation of JU until she was fully alert. Her evidence was that she had been assaulted during an earlier stage of recovery. If Mr MacDonald QC’s submissions upon count 3 were to succeed, they would have an impact upon other counts in the indictment having regard to the judge’s directions upon cross admissibility (see ground 2 below).

In our judgment, there was sufficient evidence of sexual interference by the appellant, which we have already summarised, to enable the jury to adjudicate upon the reliability of JU’s evidence. There was no dispute that the appellant was on duty throughout the complainant’s stay in the recovery room. Mr Hume’s evidence was that he was not permitted alone to attend to the patient. He was required to be under the supervision of the nurse, who was the appellant. The evidence was that staff would move in and out of the recovery room as and when circumstances required. While it was Mr Hume who made the initial observations of the patient at five minute intervals during her recovery, it was open to the jury to conclude that the appellant took advantage of Mr Hume’s absences to interfere sexually with the patient. We reject the argument that this count should not have been considered by the jury.

We observe that the term ‘cross admissibility’ is a useful label which is frequently used by counsel and trial judges considering evidence relevant to a number of counts in the indictment, or evidence upon one count which may be relevant in proof of another. However, it is important to remember that, strictly, cross admissibility is a label which identifies only whether evidence relating to one count is also relevant to the jury’s consideration of another. When it is ‘bad character’ evidence it must, by reason of the effect of section 112(2) Criminal Justice Act 2003, satisfy the requirements of a gateway under section 101(1), usually section 101(1)(d). When the issue arises whether evidence is ‘cross admissible’ between counts it is essential that thought is given to the questions: to what issue(s) may the evidence be relevant and for what purpose(s) may the jury properly consider that evidence upon one count is capable of supporting another? For reasons which we shall explain, we consider that this is what the judge did in the present case.

Mr Alistair MacDonald QC recognised that evidence upon one count, or other ‘bad character’ evidence admitted under section 101(1)(d) Criminal Justice Act 2003, may be relevant in proof of a further count in the indictment. For example, a number of witnesses, unknown to one another, may make independent complaints of sexual assault by a defendant. Provided that the complaints were truly independent of one another, the fact that several similar complaints were made may be relevant to an important issue in the trial between the defendant and the prosecution. In the present case there were upon counts 3, 4 and 5 two issues for the jury to determine (1) whether the complainant’s memory was accurate, or false by reason of the effects of anaesthesia and (2) whether, if the complainant’s memory was accurate, the perpetrator of the assault was the appellant.

The learned judge directed the jury, in terms to which we shall return, that when assessing the evidence of each complainant in counts 3, 4 and 5 the jury could have regard to the evidence of any other count 3 - 5 complainant, and the evidence of SK. First, when deciding whether the complainant was the victim of false memory the jury could consider the unlikelihood of the coincidence that four women had suffered false memory of sexual interference in the same hospital in similar circumstances, in the case of three of them during June – September 2008, and in the case of the fourth, during May 2006. Secondly, the judge directed the jury that if they were sure that the appellant had sexually assaulted any one of the count 3 – 5 complainants, or SK, and, if they concluded in consequence that the appellant had a tendency towards committing sexual offences against female patients under his care, the jury was entitled to consider to what extent that propensity assisted them to resolve, in the case of any (or any other) count, the questions (1) whether the complainant’s memory was accurate or false and (2) whether any medical procedure was properly or inappropriately carried out.

Mr MacDonald conceded that if it was, in the factual circumstances of the present case, permissible for the judge to provide the jury with these ‘cross admissibility’ directions, he had no complaint to make about their content; the directions given were balanced, fair and contained the appropriate warnings. Furthermore, it was conceded that there was no evidence that the complaints made were other than independent, although in the case of SK there was a question whether she may have been influenced by the contents of the newspaper article in which she had recognised the appellant’s photograph. As to the latter, the jury received a specific direction.

Mr MacDonald submitted that on the facts of the present case it was dangerous to permit the jury to treat the evidence of any one complainant (or SK) as relevant to the issue whether a second complainant was recounting an accurate or a false memory. As the argument was succinctly expressed in writing:

For the first, and principal, of these submissions the appellant seeks to derive assistance from the decision and reasoning of this court in Norris[2009] EWCA Crim 2697. In Norris the appellant, a nurse, had been charged with the murder of four, and the attempted murder of one, elderly female patients under his care all of whom, it was the prosecution case, had died from hypoglycaemia induced by the appellant’s exogenous administration of insulin to his patient. The issues for the jury were whether, in the case of each patient, the prosecution (1) had excluded the possibility that the patient had died from an extremely rare condition of naturally occurring hypoglycaemia and (2) if so, had proved that the appellant had administered the fatal dose of insulin to the patient. It was the prosecution case that [12], “severe hypoglycaemia in non-diabetics was a very rare occurrence… [To] have five such cases, where four had led to the death of the patient, was quite extraordinary … [S]imilar patterns in each case made the five cases mutually supportive”. Quite apart from the evidence of pattern and opportunity the prosecution was able to point in the case of one of the victims to a particularly strong prime facie case against the appellant. This was the subject of a separate ‘propensity’ direction. Counsel for the respondent confirmed [68] that one of the reasons for adducing the evidence of rarity of cases of naturally occurring hypoglycaemia was “to demonstrate that because such a phenomenon was so rare, it would be extraordinary that all five cases in two hospitals in close proximity in a short period, were naturally occurring hypoglycaemia. It was not suggested at the trial that the jury could use the evidence … to show that all five cases must have been the result of malicious administration”. [original emphasis]

Counsel for the appellant conceded in argument [63] that as a general proposition the evidence in relation to one count was relevant and probative in relation to another on two crucial issues. As to the first:

Counsel for Mr Norris argued [65] that the jury should have received an explicit direction as to how they should approach the cross admissibility of the evidence upon the issue of causation of death. Further, they should have received assistance as to their approach to causation if there was a possibility that any one of the deaths was due to naturally occurring hypoglycaemia.

At paragraph 76 of his judgment, Aikens LJ posed two questions: was there a danger that the jury would deduce from the “extreme unlikelihood of five cases of naturally occurring hypoglycaemia” that “the cause of the hypoglycaemia in each of the five cases must therefore have been non-natural? Would such a line of reasoning have been wrong and should the judge have made a more specific warning against it?”

At first sight it may appear that the court was posing the question whether the unlikelihood of coincidence was inadmissible in consideration of causation of any one death. However, it is clear from the succeeding paragraphs that this was not the question the court was asking. At paragraph 77, Aikens LJ said:

At paragraph 78 Aikens LJ observed that the trial judge had sensibly avoided the risk of false reasoning by directing them that they must decide separately upon each count whether naturally occurring hypoglycaemia could be excluded. Aikens LJ continued:

Mr MacDonald’s argument is that the rarity of false memory in recovering patients, and the unlikelihood of coincidence of false memory in these four patients, was or should have been treated by the judge as inadmissible to support either and both of the following propositions:

Any one of the four complainants was not the victim of false memory; and

All of the four complainants were not victims of false memory.

It is necessary to recall that Mr Norris was not arguing that the evidence in support of one count was inadmissible upon another in support of the issue of causation. The risk which the Court was considering was that the jury might have jumped to the conclusion that, because a cluster of deaths caused by naturally occurring hypoglycaemia would be quite extraordinary, all of the five deaths were caused by the exogenous administration of insulin. That line of reasoning was forbidden since the statistical chance that any one of the deaths was non-natural remained constant.

It seems to us that Mr MacDonald’s argument not only lacks support in authority but it also defies experience. Coincidence may be unlikely in a variety of circumstances. One is that four independent complainants are lying about sexual assault. Another is that four independent complainants are suffering from the rare phenomenon of false memory of sexual assault. The jury was just as entitled, when considering the evidence of any one of the complainants, to have regard to the unlikelihood of a cluster of false memories as they would have been entitled, if the defendant’s assertion had been that they were lying, to have regard of the unlikely coincidence that they were all liars. However, before the jury could be entitled to have regard to the unlikelihood of a cluster of victims of false memory, there would have to be expert evidence that the coincidence was indeed unlikely because this was not a matter about which the jury would have experience. In Norris the jury could not have paid any regard to the ‘extraordinary’ coincidence of allegedly natural deaths from hypoglycaemia without the expert evidence of Professor Ferner and Dr Kroker. There was such evidence in the present case. What the jury could not do was, on the basis of coincidence, leap to the conclusion that all four of the complainants were recounting true rather than false memories. What was required was an examination of the evidence relevant to each count separately and a separate conclusion upon the reliability of each complainant’s memory. In reaching their conclusion upon the evidence relevant to each count the jury was entitled to have in mind the rarity of false memory of sexual assault and, if they so concluded, the unlikelihood of the coincidence upon which the defence relied.

HHJ Milford QC explained to the jury (summing up, page 2/7) the defence case that four of the complainants were coming round from anaesthesia. The defence was not saying that the women were telling lies. They believed they were speaking the truth and the defence made “an utterly fair point” that for this reason they would appear to be far more convincing witnesses than those who know perfectly well they were not telling the truth. Separate verdicts must be returned. They did not stand or fall together (summing up, page 3/19). In the case of counts 3, 4 and 5 the issue in each case was the same: “… [h]ave the prosecution proved that the memory of the complainant is a true memory or is it possible that the memory is a false memory caused by the anaesthetic”. As to the evidence of the four complainants the judge said (summing up, page 7/12):

The judge pointed out (summing up, page 8/1) that the count 3 - 5 complainants all made similar complaints:

The judge was in these passages making quite clear to the jury that, while the evidence of one witness might be treated as supportive of another, they were not permitted to lump the evidence together to reach a blanket conclusion upon counts 3 – 5.

He proceeded to draw the jury’s attention to the complaint of SH (summing up, page 8/22) in respect of which it was established that the appellant was not even on duty at the relevant time. He continued at page 9/17:

The judge proceeded to explain (summing up, page 10/3) the second way in which the evidence might be ‘cross admissible’. If in respect of any one complaint the jury was sure of the appellant’s guilt it was open to them to conclude that the appellant had a tendency to commit sexual offences in the circumstances alleged. That was a factor which the jury could consider as relevant to the issues, first, whether another complainant had described a true or a false memory and, second, whether she may have misperceived a normal and appropriate nursing procedure. As in the case of the coincidence direction, Mr MacDonald had no complaint to make about the terms of the judge’s propensity direction provided that it was appropriate to give such a direction at all. The judge gave a conventional warning as to the limitations of propensity evidence and concluded (summing up, page 11/17):

In our judgment, there is no prospect either that the jury would have misunderstood the judge’s direction, or that they might have fallen into the trap of reasoning improperly to the conclusion that the rarity of the coincidence upon which the defence relied disproved the defence upon all three counts, without examining the state of the evidence, including the defence evidence, upon each count separately. We do wish to point out that when the jury is being invited to consider the evidence of several complainants for an assessment of the unlikelihood of coincidence, care needs to be taken by the judge before also giving the propensity direction. A conclusion, partly based upon the unlikelihood of coincidence, that the defendant is guilty upon one count (and therefore has a propensity to commit such offences) may enhance the probability of guilt upon other counts, but the jury should be aware of the risk of overvaluing the accumulation of inference. In the present case the jury received the appropriate warning.

Secondly, Mr MacDonald advanced the bold submission that the jury should not have been permitted to consider the unlikelihood of the coincidence that the complainants were suffering from false memory without reliable statistical proof of the degree of unlikelihood of the coincidence. In the course of evidence given by the expert medical witnesses the results of surveys and research were explored, but none of them was asked to place a statistical probability value upon the cluster of cases the jury was considering. Their evidence concentrated upon each complainant, the anaesthesia administered, the stage of the patients’ recovery when the alleged assault was experienced, and the quality of the witness’ recall. Mr MacDonald’s argument was that without an appreciation of the statistical probability of coincidence the jury could have had no sound basis for reaching a conclusion based, even in part, upon that coincidence. He does not rely, in support of his argument, upon any principle of law or evidence approved in the cases but seeks to draw an analogy with the admission of DNA evidence. Juries, he submits, are permitted to consider DNA evidence only because the evidence is given meaning by the value of the probability that more than one person in the population may be found to have an identical profile.

In our judgment, Mr MacDonald’s analogy is a false one. Any evidence capable of narrowing a range of relevant possibilities is likely to be admissible, e.g. the offender had dark hair, was left handed and walked with a limp. The evidence may establish circumstances which, when considered as a whole, have the effect of proving guilt. It is not the law that a statistical value must be placed upon any coincidence on the unlikelihood of which one of the parties to a criminal trial relies. DNA evidence is capable of being, together with other evidence in the case, such a potent source of identification that the prosecution is required to tender evidence of statistical probability (properly explained to the jury) so that it can be evaluated fairly. In some circumstances, even the absence of statistical precision will not prevent the jury considering DNA evidence provided that they understand its probative relevance and its limitations (see, e.g. Bates[2006] EWCA Crim 1395, particularly at paragraphs 29 - 31). The use of statistical evidence by expert forensic scientists does not imply that every time the prosecution relies upon the remote chance of coincidence it must prove the statistical probability of that coincidence. If that were the case the admission of such evidence, approved in Freeman and Crawford [2008] EWCA Crim 1863, [2009] 1 Cr. App R 15, would be impermissible in the overwhelming number of prosecutions relying on circumstantial evidence for their potency, including the prosecution in Norris.

We recognise that there will be occasions on which the nature of the evidence is such that either the evidence will be excluded on the grounds of fairness or it will be the subject of warnings to the jury as to its limitations. The probative value of the evidence may be tenuous and for that reason its effect unfairly prejudicial or, while the evidence may have an enhanced probative value upon one or more issues, it may require a warning that it should not be overvalued. Such warnings are commonplace, for example, when propensity evidence is admitted. If the evidence is admitted, the requirements for directions in each case must depend upon the judgment of the trial judge as to the nature and effect of the evidence and the issues which the jury is being asked to resolve. These problems should always, we think, be the subject of discussion before speeches. It may be necessary for the judge to warn the jury against using the evidence for a purpose which would be unfair.

In the present case, the nature of the evidence was such that no statistical evaluation could realistically be attempted, not least because the precise circumstances of the complainants were not replicated in the research papers to which the experts referred. The experts were, in the main, reporting the clinical experience of themselves and their colleagues and comparing the available research with the present cases. As the judge pointed out these were circumstances which the jury was entitled to consider subject to the warnings given in his summing up. Since Mr MacDonald has no complaint to make of the judge’s directions to the jury upon their approach to the evidence, it does not appear to us that the risk of unfair prejudice to the appellant’s case is made out.

For these reasons we conclude that the verdicts of the jury are safe and the appeal is dismissed.