R V Clark

"Professor Meadow's evidence of the statistical probability of two SIDS deaths in one family undermined the safety of the convictions for the following reasons:

a) the figures cited were erroneous (application to call fresh evidence);
b) Professor Meadow's opinion as to the deaths being unnatural was founded in part on the statistical evidence cited in breach of the guidelines in R -v- Doheny & Adams [1997] 1 CAR 369;
c) the learned judge failed to warn the jury against the "prosecutor's fallacy" as referred to in R -v- Deen Times, 10th January 1994 CA. To the contrary, the learned judge appeared to endorse the prosecution's erroneous approach."
Background

102. An introduction is necessary. In 1992 the National Advisory Body set up CESDI (the Confidential Enquiry into Stillbirths and Deaths in Infancy - here the first year of life). In 1992 there had been a fall in the incidence of SIDS following the Back to Sleep campaign where parents were advised and encouraged to put their babies to sleep on their backs or sides, rather than prone. But SUDI remained the largest single group of deaths in the neo-natal range. Between 1993 and 1996 the CESDI SUDI Study was set up to identify possible risk factors and associations for such sudden unexpected infant deaths. The SUDI Study extended beyond SIDS. The classic definition for SIDS (from the American pathologist Beckwith) is:

"... the sudden death of a baby that is unexpected by history and in whom a thorough necropsy examination fails to demonstrate an adequate cause of death."

103. Clearly the accuracy of that definition depends on the pathologists' thoroughness in autopsy, and on his or her interpretation of the findings. The definition of SUDI is wider, and includes:

"• deaths that were unexpected, and unexplained at autopsy (ie those meeting the criteria for SIDS);

• deaths occurring in the course of an acute illness that was not recognised by carers and/or by health professionals as potentially life threatening;

• deaths occurring in the course of a sudden acute illness of less than 24 hours duration in a previously healthy infant, or a death that occurred after this if intensive care had been instituted within 24 hours of the onset of illness;

• deaths arising from a pre-existing condition that had not been previously recognised by health professionals;

It will be seen that all SUDI are potential SIDS, but further investigation may show it is not a true SIDS, which are unexplained and unsuspicious natural deaths.106. The CESDI Study report (on which we have relied for information in this paragraph, believing it to be uncontroversial) gives as a reason for the Study using the other category of SUDI rather than SIDS:

"The broader category of SUDI rather than SIDS was chosen because it is often not possible to distinguish between SIDS and other unexpected deaths until the first autopsy results become available, which may not be for some weeks after the death, and even then the distinction may not be clear-cut. In addition, it was expected that health professionals would more often have been involved in the care of babies whose deaths were explained, so that enquiry into those would be more likely to yield lessons in professional audit and consequent improvements in service."

107. The Study was a major work, covering 470,000 live births, of which 456 SUDIs were identified. Of these 93 were later fully explained leaving 363 finally classified as true SIDS.
108. The Crown took the view that neither Christopher's nor Harry's death was a SIDS death because both were attended by unusual/suspicious circumstances suggesting death caused by a parent because of (inter alia) the existence in the case of each boy of unexplained recent and old injuries found after death. But until cross-examination of the defence medical witnesses, the Crown believed that the defence experts at trial would be mounting a case that the deaths were SIDS deaths, ie unexplained natural deaths.
109. The trial started on Monday 12th October 1999, there were 13 days of evidence in all, and it was not until the defence expert medical evidence was called (ninth to twelfth days of evidence - 26th to 28th October) that it became clear that the medical experts called by the defence accepted that neither death was a true SIDS death. But though the precise measure of rarity was not a significant issue by the end of the trial, the principle of rarity was. The appellant and her husband had not abandoned her defence that both the deaths were natural and true SIDS.
110. The Crown had sought to meet the issue - true SIDS or not - by evidence of the rarity of true SIDS deaths - death by unexplained natural causes. On 5th June 1998 a witness statement by Professor Sir Roy Meadow was served. In paragraph 7 he stated:

Even when an infant dies suddenly and unexpectedly in early life and no cause is found at autopsy, and the reason for death is thought to be an unidentified natural cause (Sudden Infant Death Syndrome - SIDS), it is extremely rare for that to happen again within a family. For example, such a happening may occur 1:1,000 infants therefore the chance of it happening within a family is 1:1,000,000. Neither of these two deaths can be classed as SIDS. Each of the deaths was unusual and had the characteristics of a death caused by a parent."

111. From what we know now, it is clear that Professor Meadow was stating the approximate ratio of SIDS deaths to births (1:1,000) and squaring that to reflect the chance of a second SIDS death in a family that has already suffered one. That statistical practice makes the assumption that the chance of a second SIDS death in a family that has already suffered one is the same as the chance of the first SIDS death in a family that has not already experienced a SIDS death, an assumption that the defence challenge.
112. In arriving at those figures, Professor Meadow (who does not claim to be a statistician) was using the calculation above described on the global figures (see his evidence in chief, p 36 - 37). He was not (he did not have the material) relying on the effect on those figures of taking account of the "... three prenatal factors with the highest predictive value ... of an increased risk of SIDS" (see the second quoted sub-paragraph of paragraph 122 hereof) - factors which the CESDI Study showed affected the risk. It was this Study on which the ratio of 1:73 million was based, as will be seen.
113. The issue as to the theoretical probabilities of two true SIDS deaths in one family was explored in an old style committal at which Professor Meadow was cross-examined on propositions taken from the CESDI Study, and also the paper Professor Emery had prepared. For our present purposes we can go to the preliminary hearing one month before the trial as to severance and similar fact evidence. In that hearing, the Crown put forward a request for the mutual admissibility of the evidence between the two counts of murder, the issue on each count therefore whether the death was natural or unnatural. It was conceded that there was a prima facie case of unnatural death in Harry's case, but only suspicion short of a prima facie case in Christopher's case. To establish the mutual admissibility, the Crown relied on 13 similarities between the two deaths (and in particular the first six). The seventh similarity and its fate at that hearing is set out in the judge's ruling of 9th September 1999 at page 7:

"(7) The possibility of two babies dying from SIDS is, according to the evidence of Professor Meadow, extremely remote and estimated at 1 in a million. Mr Bevan [leading counsel for the defence] submitted that that evidence cannot assist the jury in deciding which death may have been natural and which may have been unnatural. He said if one was unnatural, the chances of there being one SIDS in the family remained at 1000:1. Mr Spencer [leading counsel for the Crown] did not press that aspect of the matter any further."

114. The judge, in making his order of mutual admissibility relied on the first six similarities. But it is clear that the Crown had not abandoned the seventh.
115. When the defence expert evidence was delivered a week before trial, it contained two statements relevant to the rarity of two unexplained sudden deaths in infancy (SIDS) in one family. First, Dr Rushton said:

"Therefore while the occurrence of two unexplained sudden infant deaths in a family always raises concerns as to whether or not there is an unnatural cause of the deaths, there are families in which such deaths do occur following unexplained but presumed natural causes."

The reason why the second such death in one family raises such concerns is because of the extreme rarity of true SIDS.

"The occurrence of repeat unexpected deaths is a topic on which I have been doing research over the last few years and have been able to do a confidential study involving approximately 100 such deaths. These have all been families who have presented as having two 'cot deaths'. My findings have been that approximately a third of the deaths have been due to a whole series of rare natural causes which had been missed by those doing the necropsy [ie, not true SIDS - see paragraphs 104 - 105]. A third of the deaths were associated with different forms of child abuse and having a whole number of features which indicated that they were unnatural deaths [ie, not true SIDS]. In the final third no suspicion of unnatural death was found and no natural cause was found and these children constitute what may be termed at the moment as 'true idiopathic unexpected child deaths' [ie, true SIDS]. The occurrence of two unexpected deaths in a family thus raises a definite suspicion of unnatural death which in my experience is confirmed only in a third of such cases".

ie that over the "... last few years ..." he has examined approximately 33 cases which he concluded were two genuine SIDS in one family. He was not called to give evidence at trial, but his report had been put to Professor Meadow in cross-examination at the old style commital proceedings and, as we shall see, Professor Meadow dealt with Professor Emery's report in his evidence at trial.

117. The Crown tell us and we accept that it was on the basis of that material that they concluded that the rarity of multiple true SIDS deaths in the same family was to be an issue which both prosecution and defence would explore at trial.
118. At this time counsel for the Crown learnt that Professor Meadow was writing the preface to the CESDI Study, and that Study included a table entitled "SIDS rates for different factors based on the data from the CESDI SUDI Study". This was the source of the 1:73 million figure and Professor Meadow delivered a Notice of Additional Evidence "further to his original statement" (ie that of 5th June 1998) on Thursday 7th October 1999 with the trial to start on Monday 11th October. This Notice reads:

"Further to my original statement, I have read the reports of the other medical experts.

Apart from non-accidental injury, no likely specific medical cause of death has been proposed. Thus it is suggested that the deaths of both children should be considered as examples of SIDS [that was how he understood the defence case].

The likelihood of SIDS varies with social circumstances. The most recent estimation of the incidence in England is that for a family in which the parents do not smoke, in which at least one has a waged income and in which the mother is over the age of 26 years, the risk is 1:8,543 live births. [the three prenatal factors referred to in paragraph 112 above]

Thus the chances of two infant deaths within such a family being SIDS is 1:73,000,000."[JA24]

119. Professor Meadow is a distinguished paediatric consultant. He is an acknowledged expert in the field of child abuse, and the discoverer of Munchausen Syndrome by Proxy, which led him to the difficult area of young children who died in unusual circumstances early in life. He is not a statistician (though as he made clear, he has to use statistics in his work). His original calculation of two true SIDS occurring in the same family was as set out in his original witness statement at paragraph 110 above. It is not surprising that, having seen the CESDI SUDI Study in draft, he referred to those figures in his Notice of Additional Evidence.
120. At that time, as the appellant's solicitor has deposed, there were a number of important and unavoidable last-minute matters to deal with. The trial therefore started without the defence having the assistance of a medical statistician, and in the event, no such witness was called. But the Crown had no statistical expert either.
121. The jump from 1:1,000,000 (Professor Meadows original figure) to 1:73,000,000 was not explained at the time of delivery of the Notice, but became clear later when the source of the figures was disclosed; a table prepared for the CESDI Study (together with supporting text, delivered soon after), which was then available in draft. The Table, as then delivered, was as follows:

Table 3.6.1: SIDS rates for different factors based on the data from the CESDI SUDI Study

SIDS Rate per 1000 livebirths*SIDS incidence in this
group*

Overall rate in the study populationin 1303

Rate for groups with different factors

Anybody smoke in the household
Nobody smokes in the household
in 737
in 5041

No waged income in the household
At least one waged income in the household
in 486
in 2088

Mother <27 years and parity
Mother > 26 years and parity
in 567
in 1882

None of these factors
One of these factors
Two of these factors
All three of these factors

in 8543
in 1616
in 596
in 214

* Based on the number of livebirths in each study region from 1993 to 1993 inclusive (OPCS)

106. 106. That table appears as Table 3.58 in the final published report. A number of points arise from it. First, as appears from the text of the Notice of Additional Evidence, the Crown assumed that the defence experts would be contending that these deaths were SIDS deaths. But after the cross-examination of Professor Berry (10th day of evidence, 26th October), the defence doctors were to agree that neither of these deaths was a SIDS death [JA25]because of the suspicious circumstances surrounding them. Second, the table gives rise for the first time to the 1:73 million statistic - see the explanatory paragraphs in the published version. For convenience, we put all the paragraphs relating to the Table (now numbered 3.58) together:

Overall, in the population included in this Study the SIDS rate was 0.768 per 1,000 live births, ie approximately one baby in 1,300 died as SIDS. From our data, it is possible to identify within the population a number of factors which are associated with an increased risk of SIDS. The identification of families at higher risk of SIDS is of importance in allowing the appropriate deployment of scarce health care resources, and in attempting to achieve changes in life style or patterns of child care that might reduce this risk. For families already at low risk, knowledge of the factors influencing risk may help to provide reassurance and encouragement in continuing appropriate patterns of care.

Table 3.58 shows the three prenatal factors with the highest predictive value (based on the Wald Score) of an increased risk of SIDS, and the likely effect of the presence or absence of each factor on the incidence of SIDS, along with the effect of combing those factors. [NB: we reproduce this paragraph in the final form in the Report, which the appellant's statisticians had, and not as originally delivered (in draft) to the defence, for which see Prosecution Appeal Bundle, p130.]

Thus, an infant living in a household in which nobody smoked had a risk of SIDS of around one in 5,000, whilst if anyone in the household smoked this risk rose to around one in 700. Similarly for an infant in a household in which there was no waged income, the risk was around one in 500, compared with one in 2,000 if there was a waged income.

The correlation between the factors was taken into account when more than one factor was used to calculate the rate, but because all three factors are independently significant in the multivariate analyses, the presence of more than one will have an increased effect.

Thus, it can be seen that for infants in families in which all three factors [ie, someone smokes in the household, there is waged income in the household, and the mother is 26 or under] are present the risk of SIDS was one in 214, compared with a risk of one in 8,543 for infants in families with none of the factors, ie a 40-fold difference in risk.

Since the factors will generally remain the same (with the possible exception of maternal age below 27 years) the risk of SIDS to a subsequent child in a family in which one infant has already died will range from one in 214 to one on 8,543. this does not take account of possible familial incidence of factors other than those included in Table 3.58.

For a family with none of these three factors, the risk of two infants dying as SIDS by chance alone will thus be one in (8,543 x 8,543), ie approximately one in 73 million. For a family with all three factors, the risk will be one in (214 x 214), ie approximately one in 46,000. Thus, for families with several known risk factors for SIDS, a second SIDS death, whilst uncommon, is 1,600 times more likely than for families with no such factors. Where additional adverse factors are present, the recurrence risk would correspondingly be greater still."

107. As we believe to be plain from the text, that figure went to show how unlikely it was [JA26]that a family with two children complying with the profile criteria would suffer one or two cot SIDS.

"One must always approach statistics with caution, but Professor Meadow has worked out the probability of there being two genuine unexplained natural deaths (SIDS) in a family such as this where material standards are high, the mother is over 26 years old and neither parent smokes[JA27]. The chances of one genuine unexplained natural death in such a family are about 1 in 8,500. The chances of two genuine unexplained natural deaths in such a family are about 1 in 73 million." [JA28]

109. There he dealt with the question of the rarity of genuine SIDS, which was a live issue. He returned to the question of whether, in all the circumstances, these could possibly be SIDS. "You don't find these sorts of recent unexplained injuries in a true case of SIDS".
110. The draft opening had been supplied to the defence. No complaint was made as to the inclusion of any of the relevant passages. The case proceeded on the basis that there would be an issue as to whether these were genuine unexplained natural deaths. In fact, as the expert witnesses called for the defence did not support the proposition that the deaths were SIDS, the actual issue was whether the Crown could satisfy the jury that neither death was natural, so that the jury could safely infer that each death was unnatural. No objection was taken to the Crown's case being opened in that way.
111. We then come to Professor Meadow's evidence. The importance of his evidence was: first, his identification of recurrent features commonly associated with unnatural infant death (eg episodes of unusual illness, inconsistent accounts by the parent of circumstances leading to death, the time of death (when mothers are under great stress), cases where the infant had just had a normal feed, cases where there are signs of overt physical abuse not directly causative of death); second, the seven similarities identified between the two deaths; and third, his identification of the physical findings on the bodies of these infants on which the Crown based their case. He dealt first with the case of Christopher (pp 12 to 25) and then with Harry (pp 25 to 32). Each summary concluded with his evidence that he could not think of any natural explanation for either death. That was the important part of his evidence.
112. Then he turned to the statistics (pp 32 - 39 of his evidence in chief). He prefaced his remarks with the need to approach statistics with caution, and then went on to describe the CESDI Study: a multi-disciplinary research study, government-supported, well -conducted: ... "it's the most reliable Study and easily the largest and in that sense the latest and the best." The Crown make the point that an expert medical witness, asked to give his opinion on the rarity of a medical condition can speak not only of his own experience, but also of what there is in the literature, including statistics. We agree with that submission. Objection may conceivably be taken that the statistics have to be proved, starting with the basic data, but absent any such objection (and none was taken here) the expert can rely on an up-to-date reputable study such as the CESDI SUDI Study.
113. That Study calculated the likelihood "... of a baby dying, and being labelled as sudden infant death syndrome" (a reference to the fact that that has been a registrable cause of death since 1971). The table reproduced above at paragraph 121 was put before the jury and explained to them. Professor Meadow was asked, in examination in chief:

"Q Just before we look at the figures, does this leave out of the equation deaths which showed suspicious features in any event?

A Yes, ... they did not include babies whose deaths were being investigated by the police or were the subject of a coroner's 'not ascertained', or it was thought to be ... murder or harm to the child. So these are babies that died suddenly and unexpectedly. The starting position was that these deaths were thought to be natural deaths."

"Q Later on in the paper does it go on to work out the risk of two infants dying of SIDS in that family by chance alone?

A Yes, you have to multiply one in 8,543 times 1 in 8,543 and I think it gives that in the penultimate paragraph, its points out that it's approximately a chance of 1 in 73 million .... It gives a chance of 1 in 73 million live births and in England, Wales and Scotland there are about, say, 700,000 live births a year, so it is saying by chance that happening will occur once every hundred years."

115. We have seen that the genesis of the 1:73 million figure was the CESDI Study, and not any individual calculation made by Professor Meadow. The grace note was that this would occur "... once in a hundred years". This evidence was given by Professor Meadow in examination in chief on the sixth day of the Crown evidence. While the "100 years" answer was new to the case, it was a straight mathematical calculation to anyone who knew that the birthrate over England, Scotland and Wales was approximately 700,000 a year. Mr Bevan submits that this answer must have greatly affected the jury. He suggested that its impact would have been "overwhelming". But he did not suggest the effect was such that the judge should have discharged the jury, and he did not so apply. Nor did the defence invite the judge to direct the jury to ignore the evidence relating to Table 3.58 of the Study, nor to give any special direction in relation to it.
116. Professor Meadow was then asked whether Christopher's death and Harry's death would have been included in the Study. Professor Meadow replied:

"They'd only be included if the pathologist had described them under the label Sudden Infant Death Syndrome, and I think with their findings at autopsy, fractures, bruises and things, they would not have described them as Sudden Infant Death Syndrome so they wouldn't have been .... I think these two cases would probably have been likely to have been excluded from the Study."

"Q Until this study was done and these figures were calculated was there a much lesser figure which didn't take account of these particular household features, smoking, age, wage earning and so on?

A Well in general if someone asks me what the risks of two babies within a family dying and being labelled as SIDS I would just say one in a thousand or 1,300 times 1 in 1,300. I'd just give the global rate but you then ... have to take into account as this paper has done and you have to in all individual cases of the circumstances. In other words, in a family where there's no wage earner, where they smoke, where it's a young mother, the chance of a sudden natural death is very, very much greater than the chance in a professional family, non-smokers and of a mature age."

118. He was then questioned about the study Professor Emery had played a part in, published in a paper called Recurrence of Unexpected Infant Death. This witness agreed that this was, in this country, the only other study of any scale on the recurrence of infant death within a family. The starting point for this study was families in which two or more infants had died suddenly and unexpectedly. The report starts:

"In England and Wales an estimated 50 families a year experience a second unexpected infant death."

"It was addressing a different issue. The research workers investigating families in which two or more children had died suddenly and unexpectedly, and they went back over all the records and had multi-disciplinary meetings and researched them very carefully to see if they could elucidate better the reasons for those deaths. They had in all 57 deaths to investigate in 24 families, most of families with two deaths. They couldn't get material on all of them but they were mainly two death families. Of those 57 deaths they, on more detailed and modern investigation were able to find a natural cause, that's a disease cause or a genuine accident for 30% of the cases, so their point there was saying nearly a third of those cases should never have been labelled as SIDS, they should have been identified as either a natural accident or a natural disease. Fifty-five percent of the cases they ended up by saying this child had been killed by the parents, and that was not observed first time around. In five percent of the cases they came to no conclusion, and in nine percent, that was in five children they ended up saying: 'I don't know what the matter is, we too would use the label SIDS'. So they ended up with five children where they couldn't decide upon a likely cause. They point out in their article that of all the families involved, 24 families, it was only in one family where there were two children who had died and in their opinion were completely unexplained ... in other words SIDS."

120. The points he made on this topic in his evidence in chief were to tell the jury of the nature of the CESDI Study, and to explain the calculations leading to the figure of 1:73 million, to confirm that the cases of Christopher and Harry would not be included because of the physical findings at autopsy; and that the CESDI Study was not undermined by the Recurrence of Unexpected Infant Death Study, as they dealt with different issues.

"Q Finally, coming back to these two babies, in your opinion, Professor, could either of these two deaths be classified as SIDS, Christopher and Harry.

122. Mr Bevan QC for the defence then cross-examined. He cross-examined the witness from the Care of Next Infant (CONI) Report, published in 1998. The point of the cross-examination (20th October pp 63 - 65) was to suggest that there were more families with two SIDS deaths than the CESDI Study figures suggested. Professor Meadow's response was that CONI's figures were not scientifically valid. Counsel put:

A Yes ... This is why you take what's happened to all the children into account, and that is why you end up saying the chance of two children dying naturally in these circumstances is very, very long odds indeed, one in 73 million. You know I mean ...

A I know, but I mean ... you have to say two unlikely events have happened and together it's very, very, very unlikely."

The defence called Professor Berry on Day 10 of the evidence. He was one of the four editors of the CESDI Study (though we were told this was more honorific than executive). He is a professor of Paediatric Pathology at Bristol University and a consultant paediatric pathologist to the Bristol Hospital for Sick Children since 1983. He gave evidence under cross-examination that he would not classify either of the deaths as SIDS, as the unexplained factors in both caused him concern (eg the injuries to the children were not sufficiently explained). He agreed that two SIDS deaths in one family would be "... unusual, but not very unusual ...", but "... we are talking about statistics here which generally speaking are not of great value in the individual case." This last point the judge came back to in his cautionary direction to the jury (see paragraph 145). He made the point that that he did not believe that statistics enables you to determine whether the death was natural. He agreed that SIDS did not usually repeat in families. While he accepted the 1 in 8,543 statistic in relation to the first SIDS death in low risk families, because this was an observed figure, he considered the squaring of that figure to calculate the risks of a second SIDS death to be an illegitimate oversimplification which a sentence of the Study warned against:

"This does not take account of possible familial incidence of factors other than those included in the Table." (see paragraph 122 above, the last sentence of the last paragraph quoted).

"There may be familial diseases which we are not clever enough to recognise but there may be other familial factors that predispose that family, for reasons we don't understand, to a second death. So this Study, this is a theoretical calculation. It is not the result of observation ... [but of] multiplying those figures together ... when you go out and look for families who have had ... second SIDS .... they are commoner than you would expect from this type of calculation."

124. The judge was specifically to remind the jury of this evidence that the risks were inherently greater in a family which had already had a SIDS death. The judge squarely left the issue to them, for them to decide. The defence put no specific figure on the probability of a second SIDS death in one family.
125. That then was the issue on the statistical evidence given in the trial. It was accepted by Professor Berry that to have two genuine SIDS deaths in the family was "unusual", and it seems to have been accepted that the higher risk families were those where someone smoked, where no-one was wage earning and the mother was 26 or under, and the lowest risk was for families such as the Clarks where the reverse of the above three factors applied. That should not have surprised anyone as the CESDI Study refers to international research and refers to (Chapter 1, page 4) "... a similar epidemiological profile ..." of high risk families in many foreign countries.
126. While to deal properly with this ground of appeal in its context in the trial it has necessary to consider the evidence and issues in detail, it was very much a side-show at trial. The experts were debating the incidence of genuine SIDS (unexplained deaths with no suspicious circumstances) in a case where both sides agreed that neither Christopher's death nor Harry's death qualified as such.
127. Just as the statistics had formed only a small part of the trial, they were only a small part of the summing-up. The main passage where the judge dealt with it introduced it on Day 2 of the summing-up, 37C to 39E. We quote it in its entirety for completeness:

"Well that, members of the jury, leads me to the conclusions of the various medical experts. I start with Sir Roy Meadow and I start with him on purpose because there was an aspect of his evidence which is not strictly speaking a conclusion but which is most conveniently dealt with at this stage. What I refer to is his statistical evidence because you may remember he dealt with statistics relating to two SIDS death occurring in one family.

His figures were derived from a recent and very thorough research study, Government funded report, call CESDI. You have, members of the jury, a table containing those statistics in section 8 of the folder. In short, in a case were nobody smokes in the household and there is at least one waged income in the household and the mother is more than 26 years old, the probability of one SIDS death in the family is 1 in 8,543 and the probability of two SIDS in the same family is 1 in 73 million live births. That means there is a chance of two SIDS in the same family happening one every hundred years. In addition to that, in these two cases, there are features which are suspicious in any event. So that was the evidence that he gave in evidence in chief about those statistics.

In cross-examination figures from a report by CONI - Care of the Next Infant programme - which was published in October 1998 were put to Sir Roy Meadow. In that report, 8 [deaths] out of [5,000] babies in their programme were treated as true cot deaths after confidential enquiries. ... It was suggested that using the three high risk factors in the table from the CESDI Study, the risk of a SIDS death would be 1 in 214 which would be a risk of a double death of 1 in 46,000, and applying that correlating it to the figure of 5,000 babies, the risk it was put at that time would be 72 in every 46,000. I think Mr Bevan in his final submissions referred to a figure of 54 rather than 72 but the point still remains the same.

Sir Roy Meadow's response to this was that that was not a valid way of assessing the evidence. The CONI report was nothing like the same standard of the CESDI report, he said, and the research was not scientifically valid. He said that it does not have any value statistically and cannot be compared to the CESDI Study."

128. The judge then gave the jury a cautionary direction as to their use of statistics:

"I should I think members of the jury just sound a note of caution about the statistics. However compelling you may find those statistics to be, we do not convict people in these courts on statistics. It would be a terrible day if that were so. If there is one SIDS death in a family it does not mean that there cannot be another one in the same family. That part of the evidence relating to statistics is nothing more than that. It is a part of the evidence for you to consider. Although it may be part of the evidence to which you attach some significance, it is of course necessary for you to have regard to the individual circumstances relating to each of these two deaths before you reach your conclusion on the two counts on this indictment.

Having said that, members of the jury, I turn then to what truly were the conclusions of the relevant experts relating to Harry. The first one is Sir Roy Meadow and his conclusion was that Harry's death could not be classified as a SIDS death and in his opinion Harry did not die a natural death. He also said that the further information that he had received since he made his witness statement in June 1998 sadly increased the strength with which he felt that these two deaths were not natural."

"He said that he would not classify either of these deaths at SIDS deaths because they have features which are not typical of SIDS deaths. Both deaths, he said, gave him considerable concern. He thought the timing of the deaths was unusual but not very unusual for a SIDS death.

So far as the statistics are concerned relating to the likelihood of there being two SIDS deaths in one family, that figure of 1 in 73 million, he thought it was wrong to multiply the two figures of 1 in 8,543 together because if there was one SIDS in the family, then there maybe other factors present in the family which might make it more likely that another death would occur."

130. The judge then, on the last day of his summing-up, summarised how both sides put their cases. Dealing first with how the Crown put their case, he reminded the jury that the main issue was whether the defendant had done it. Neither death could be considered a SIDS death because of the existence of the recent and old injuries that were found in each case. The defendant and her husband had not been truthful, she had lied as to Harry's position in the bouncy chair, he had lied as to the time he got back from the office on the night of Harry's death, lying out of blind loyalty to her. She was getting tired, he was late, and going to Glasgow the next day. He then set out (as we have heard) the important findings of Dr Williams which had been confirmed, the seven findings, showing that Harry must have been the subject of shaking or some violent movement. As to Christopher, his nosebleed was consistent with prior attempted smothering (haemosiderosis was a non-starter) and his torn frenulum diagnostic of deliberately inflicted injuries indicating abuse shortly before death, consistent with smothering. Then there were the bruises, and the fresh bleeding in the lungs, the significance of which Dr Keeling had dealt with.
131. He then concluded his summary of the Crown case with this passage:

"It was therefore submitted by the prosecution all of those finding on their own were ... too powerful to amount to anything else but a deliberate killing of Christopher. All Professor Meadow's features for an unnatural death rather than a SIDS death were applicable in the case of these two babies, it was submitted by the prosecution, namely previous unusual episodes, inconsistent accounts between parents, both events had happened in the evening, both babies had just taken a feed, [JA29]and there was, say the prosecution, smothering combined with some other form of abuse. Professor Meadow's opinion you were reminded was that neither death was natural.

Reliance was also placed by the prosecution on the statistics mentioned by Professor Meadow for the probability of two SIDS deaths within the family, namely one in 73 million and even longer odds, it was said, if you take into account the existence of the old and fresh injuries, and reliance was also placed on the seven similarities between the two deaths which I mentioned to you early in my summing up, and which the prosecution suggest make it beyond coincidence that these two deaths were natural deaths."

132. He then summarised the defence case over five pages, culminating with a matching passage in relation to the statistics:

"So far as the statistics are concerned relating to the possibility of 2 SIDS deaths in one family, reliance was placed on the CONI figures, those were the figures which Professor Meadow had dismissed in his evidence, and reliance was also placed on Professor Berry's point that the risks were inherently greater in any event in a family which had already had a SIDS death.

You were told, quite correctly, that what you must not do is to conclude that if the defendant killed one baby she must have killed the other. That would be quite wrong. It was suggested that the only safe approach was to look at the death of each child independently and only if you reach the conclusion that the defendant killed one child should you ask yourselves whether that helps you in relation to the other child.

It was submitted that you cannot say with any degree of certainty that the defendant deliberately killed either of these children. She had done nothing untoward to either child and whilst there may be suspicious features, the prosecution have not proved this case so that you can be sure that she had killed either of them. Well, members of the jury, that is I hope a reasonably fair summary of the way in which both sides put their cases to you."

It was against that evidential background that the Full Court gave leave to the defence to adduce expert statistical evidence. It made clear that it did so having "... no preconceived view of the likelihood of the evidence being accepted ..." and granted leave: "without prejudice to any matter which the subsequent Court may decide."

134. It was clearly both sensible and expedient for the Full Court to deal with those questions as they did.
135. The evidence adduced by the appellant consists of reports by Dr Evett, a chartered statistician and member of the Forensic Science Service who specialises in the application of probability theory and statistics to the interpretation of scientific evidence; and by Professor Dawid, Professor of Statistics at the Department of Statistical Science, University College, London. No additional evidence has been put in by the Crown.
136. Because of pressure of time and the limited utility of what would have been necessarily argumentative cross-examination of the statisticians, we took the defence new evidence (from Doctor Evett and Professor Dawid) as read, and with the agreement of the parties tested that evidence by the submissions made on each side, and did not require the statisticians to give evidence.
137. Mr Bevan submits that there were serious errors of approach towards the statistical material at trial, both in evidence and in the Crown submissions and in the judge's summing-up. He contends that the errors were highly significant since the statistics - in particular, the figures of 1:73 million or once in 100 years - had a "devastating" impact on the jury. Mr Bevan accepts that he did not complain at the time. He did not object to the opening of the evidence or to the admission of the evidence or to the passages in counsel's speeches or the judge's summing-up of which complaint is now made. His eyes were opened by the evidence received after trial from the experts now called on behalf of the appellant. But on such an issue, he submits, the appellant should not be prejudiced by counsel's failings at trial.
138. As to Ground 3(a), Professor Dawid's first criticism relates to the calculation of the figure of 1 in 73 million. He concedes that the CESDI Study appeared to have been carefully planned and executed. But he says the report did not contain sufficient detail or transparency and he concludes:

"Whereas such methods may be appropriate for suggesting broad general hypotheses about the relationship between the outcome studied (here SIDS) and possible explanatory factors, it can be very misleading to use them to construct precise numerical formulae for predictive or explanatory purposes."

He also raises the question of the appropriateness of the calculation to the specific case of Sally Clark, and whether it should not have had built in all her individual characteristics. None of those matters, however, provides great assistance to the appellant's case, since the Crown was using the CESDI Study not for its precise figures but for a very broad point, namely the rarity of a double SIDS.

139. Professor Dawid goes on to question the appropriateness of squaring 1 in 8,543 in order to produce the figure of 1 in 73 million. He notes the fact that the relevant table:

"... does not take account of possible familial incidence of factors" [see paragraph 122 and the penultimate quoted sub-paragraph]

and emphasises the possibility that there might be factors predisposing a family with one cot death to another. He suggests, not unreasonably, that the 1:73 million must be regarded as a "ball park" estimate. We do not imagine that the Crown would demur from that suggestion. In any event it is common ground, we believe, that the statistical justification for squaring exists only where the risk of a single SIDS death in a family that has already had a SIDS death is the same as the chance of a single SIDS death in a family that has not previously suffered one. The existence of arguments against squaring was known to the jury at the trial. Professor Berry made the points to which we have already referred, and the judge reminded the jury about these in his summing-up. But again the precise figures are not important, since the Crown was making the broad point that repeat SIDS deaths were very unusual, in which exercise the number of noughts separating the lower risk households from higher risk households did not matter once the overall point was made, as here it was.

140. Similar points apply to the brief observations made by Dr Evett on the method of calculation of 1 in 73 million. Dr Evett makes two points on this. He accepts the statistical validity of the assumption justifying the squaring, but states that the Study does not show the supporting evidence for that assumption, and "In the absence of such evidence the calculation is invalid". He also complains that the appellant's alcoholism and depression were not taken into account. That point was presumably taken in ignorance of the fact that the defence had successfully applied for the evidence of Mrs Clark's alcoholism not to be taken into account, as there was no evidence that she was under the influence at the time of the children's deaths. No specific submissions were made to us on the question of depression.
141. He also makes the obvious point that the evidential material in Table 3.58 tells us nothing whatsoever as to the guilt or innocence of the appellant.
142. Thus we do not think that the matters raised under Ground 3(a) are capable of affecting the safety of the convictions. They do not undermine what was put before the jury or cast a fundamentally different light on it. Even if they had been raised at trial, the most that could be expected to have resulted would be a direction to the jury that the issue was the broad one of rarity, to which the precise degree of probability was unnecessary.
143. As to Ground 3(b), objection is taken to the alleged reliance by Professor Meadow on the statistical evidence in expressing his view that the deaths were unnatural (see the transcript of his evidence at paragraph 137 hereof). No objection is taken to his reaching that conclusion on the medical and circumstantial evidence. But in R v. Doheny and Adams [1997] Cr App R 369 a number of observations were made about the role of the expert in cases involving DNA evidence, stressing the limits of the evidence that the forensic scientist can properly give (which will include evidence about the number of people with matching characteristics who are likely to be found in the United Kingdom or in a more limited relevant sub-group). The court stated (at 374G):

"The scientist should not be asked his opinion on the likelihood that it was the defendant who left the crime stain, nor when giving evidence should he use terminology which may lead the jury to believe that he is expressing such an opinion."

144. It is unnecessary to examine in greater detail the history of the common law rule that expert witnesses should not give opinion evidence on the "ultimate issue" which the court has to decide. Suffice it to say that in our judgment Professor Meadow did not overstep the line between the expert's role and the task of the jury when he gave it as his opinion that Christopher and Harry did not die natural deaths. Mr Bevan's submission proceeds on the basis that Professor Meadow's opinion was founded both on the medical and circumstantial evidence and on the statistical evidence, and that it was in founding himself on the statistical evidence that Professor Meadow fell into error. In our judgment, however, Professor Meadow's opinion was based on his expert assessment of the medical and circumstantial evidence, not on the statistical material. Most of his examination in chief was concerned with the medical issues. He nowhere suggests that Table 3.58 (which did not deal with deaths such as these) provides any evidence that these deaths were unnatural, only that true SIDS were rare[JA30]. No-one would know better than Professor Meadow that the important evidence as to whether these deaths were unnatural lay in the physical finding port-mortem, in the account of the last hours of the infants, and in the evidence and credibility of the parents - it certainly did not lie in the statistics. And it is clear from reading his evidence that his conclusion was firmly based on that medical and circumstantial evidence, as one would expect. He then dealt briefly with the statistical material towards the end of the examination in chief, before being brought back in conclusion to "these two babies" for the purpose of expressing an opinion on whether the deaths were natural or not. As we read the transcript, that involved a move away from the subject of statistics and back to the medical and circumstantial evidence relating specifically to Christopher and Harry.
145. For that reason we reject Ground 3(b). But there remains the separate question whether Professor Meadow's statistical evidence and the way in which it was dealt with at the trial involved an "... error of approach" as counsel for the appellant describes it.
146. For convenience we repeat Ground 3(c) of the Grounds of Appeal:

"The learned judge failed to warn the jury against the 'prosecutor's fallacy' as referred to in R -v- Deen, The Times, 10th January 1994. To the contrary, the learned trial judge appeared to endorse the prosecutor's erroneous approach."

147. We say by way of introduction that we find the forensic history of this point curious and untidy, because the introduction and use of statistics was never canvassed with the judge at any stage. It seems to us obvious, as we have already stated, that the evidential value of the contents of Table 3.58 tells us nothing whatsoever as to the guilt or innocence of the appellant. It would simply have told all within the relevant category of the population that their risk of having unexplained infant deaths with no suspicious circumstances is very small.
148. So far as we are aware, the judge was never referred to the prosecutor's fallacy, nor was any objection taken to the admission of any evidence at trial. (For the "prosecutor's fallacy" reference can be made to R -v- Doheny & Adams [1997] 1 Cr App R 369, at 372-373.)
149. The complaint is now made as to the admission as evidence of Table 3.58 and its contents in relation to the chances of any family (sharing the broad characteristics of the Clarks) having a first and then a second unexplained infant death with no suspicious circumstances being very rare: 1 in 8,543 for the first, and 1 in 73 million for the second. The rarity of such events was not seriously disputed at trial, though there was evidence, fairly summed-up to the jury, that the table overstated the rarity of a second SIDS death in one family.
150. That evidence was clearly relevant and admissible for the reasons set out above.
151. The complaint is now made that the prosecution reversed the evidence in Table 3.58 eventually using it to suggest that

"... the odds of the defendant being innocent are greater than 73 million to 1 against."

That would be quite impermissible, because Table 3.58 simply deals with the odds against any family with the same broad socio-economic characteristics suffering one or two SIDS deaths, and tells us nothing at all about any matter going to the guilt or innocence of any of the parents within that category should they be the unlucky family to suffer that death.

152. The defence complain (the source of the quotations is the defence skeleton argument):
a) that the judge did not direct the jury "... to reject the Crown's erroneous reasoning" - but they did not draw the judge's attention to that reasoning and its errors;
b) that the new evidence of the statisticians "... demonstrates that the prosecution invited the jury to adopt the figure of 73 million as having a significance in itself, without reference to the competing possibility ["...the incidence of double infant murder by a mother"] ... the figure has no significance or relevance. Consequently the jury were led to believe that this evidence demonstrated that the chances of the appellant being innocent was at least 73 million to 1 against."
153. So far as the first point goes, the Crown unsurprisingly deny that they ever put the case on statistics that way. It would surprise us if they had, because the point is thoroughly bad. It would also surprise us if the defence had not noticed so dramatic a change of position. If they had noticed, they would certainly have complained. If they had complained, the judge would have ruled in their favour.
154. In other words, it is stating the obvious to say that the statement "In families with two infants, the chance that both will suffer true SIDS deaths is 1 in 73 million" is not the same as saying "If in a family there have been two infant deaths, then the chance that they were both unexplained deaths with no suspicious circumstances is 1 in 73 million". You do not need the label "the prosecutor's fallacy" for that to be clear. It is clear that the second statement does not follow from the first, nor does it tell you anything about the children or their parents other than there were no smokers in the household, there was one waged income, and the mother was 27 or over - all being factors which put the Clarks in the lowest of all risk categories.
155. It is suggested by Dr Evett that the fact that the second statement does not follow from the first needs to be carefully explained to the jury. As a generalisation, we agree, but it all depends on just what was said. He also suggests that Professor Meadow contributed to the danger of misinterpretation. We do not agree that he did. Dr Evett quotes Professor Meadow's evidence under cross examination at 65A and 65D (see the passage quoted in paragraph 139 above):

"... you take what's happened to all the children into account, and that is why you end up saying the chance of two children dying naturally in the circumstances is very, very long odds indeed, one in 73 million" (emphasis added)

"You have to say two unlikely things have happened, and together it is very, very, very unlikely."

At 65A Dr Evett suggests that "... in these circumstances ..." refers to the odds against Christopher and Harry having died naturally. It is clear from the context that that is not what Professor Meadows was saying. He was dealing with the CESDI Study, and "... in these circumstances" clearly referred to the category of family the Clarks fell into. Equally, in the second passage relied on, Professor Meadow is clearly referring to the CESDI Study, which shows that for a family liked the Clarks a double SIDS death would be "... very, very, very unlikely". If Mr Bevan QC, for the defence, had understood him to be saying that the odds against both of these deaths being a SIDS death were 73 million to 1 that is a point which would certainly have been brought out in cross-examination and not left where it was, with the remark "Lies, damned lies and statistics". We will come back to this point.

156. Dr Evett rightly identifies the Crown's point that so long as the probability of two children dying of SIDS is very small, its exact size is of little relevance. But he inferred that:

"The 'logic' implicitly applied at the trial was as follows: a certain event (the deaths of two babies in one family) has occurred. We are unsure of the cause. One possible cause is that both babies died of SIDS. However, the probability of two babies in the same family both dying of SIDS is extremely tiny. Therefore we can exclude that possibility and, in consequence, accept that the babies were murdered - if that is the only alternative."

We do not find support for such "implicit" application of that logic in the evidence. That was not how the case was put in the evidence. The transcript supports the explanation given by the Crown: to establish rarity when there was an issue as to whether the deaths were natural and true SIDS in one family.

157. Dr Evett's alternative source for his theory is an item in The Times, published after the conviction, (which we have not seen) stating:

"... the prosecution said that it was beyond coincidence that both children could have died naturally. The probability was one in 73 million".

If the reporter who heard that did, as Dr Evett thinks, understand that to mean that there was only one chance in 73 million that the children died naturally, we agree with Dr Evett that that would be a profoundly incorrect inference. But we do not know what the reporter heard, nor what he understood. The Crown say they did not put their case that way. Certainly if the Crown did say that in the course of the trial, we would have expected an immediate challenge from the defence, and a ruling from the judge favourable to the defence. But that did not happen. We conclude that at no time in the evidence was the case put that the odds against innocence were 73 million to 1. [JA31]

158. There is clearly firmer ground in the passage in the judge's summing-up and this is what the appellant relies on (see the skeleton argument, page 27). There the judge summarises his understanding of the Crown case:

"Reliance was also placed by the prosecution on the statistics mentioned by Professor Meadow for the probability of two SIDS deaths within the family, namely one in 73 million and even longer odds, it was said, if you take into account the existence of the old and fresh injuries, and reliance was also placed on the seven similarities between the two deaths which I mentioned to you early in my summing up, and which the prosecution suggest make it beyond coincidence that these two deaths were natural deaths." [for the avoidance of doubt, these "similarities" do not include the seventh similarity referred to at paragraphs 113-114 above]

159. As is generally the case in the Crown Court, no record has been made of counsels' closing speeches, and it would seem that the judge was here referring to Crown counsel's speech. Mr Bevan complains that in that passage the Crown suggests that "... the odds of the defendant being innocent are greater than 73 million to 1" and then goes on to analyse the impact of the judge's warning as to the statistics on that passage. We will take those points in turn.
160. The appellant's case is developed as per their skeleton:

"The fresh evidence of Dr Evett and Professor Dawid demonstrates that the prosecution invited the jury to adopt the figure of 73 million as having a significance in itself when in truth, without reference to the likelihood of a competing possibility, the figure has no relevance or significance at all".[JA32]

The competing possibility identified is a double infant murder by a mother. That may be capable of being expressed in terms of a statistical probability, but legally speaking the exercise is not realistic - see R -v- Denis Adams (No 2) [1998] 1 Cr App R 377, which shows that it is not an exercise the courts would perform.

161. The history of the matter, set out at length at the beginning of this section, shows why the rarity of SIDS deaths, in particular double deaths in the same family, was - or was properly thought to be - relevant and that Table 3.58 was legitimately introduced in that connection. We have made it clear that these two deaths were not SIDS because of the unexplained, or insufficiently explained, previous injuries to the children. We have made the point that the evidential value of Table 3.58 to that Study only gives a general indication of the rarity of a double SIDS. That general indication cannot be increased by the facts of a particular case, because Table 3.58 addresses the chance of any family being so afflicted and does not help us as to the likelihood that a specific parent or parents abused their child, because it tells you nothing relevant to the question of guilt or innocence. That is a different question the answer to which cannot affect the general Table 3.58 question: namely what is the risk of a two child family suffering a double SIDS?
162. Therefore we accept that when one is looking ex post at whether two deaths were natural or unnatural, the 1:73 million figure is no help. It is merely a distraction. All that matters for the jury is that when your child is born, you are at a very low risk of a true SIDS death, and at even lower risk with the second child.
163. Professor Meadow did not misuse the figure in his evidence, though he did not help to explain their limited significance.
164. We return to the sentence of the summing-up referred to in paragraph 175. In our judgment, counsel for the Crown should not have said that the existing injuries led to "... even longer odds ..." than the 73 million to one. The existing injuries to the infants went to guilt, the odds went to rarity, and it was a mistake to add them together. Mr Spencer QC explained this on the basis of an "advocate's flourish", which does not impress us, but we are not persuaded that counsel for the appellant or the judge then understood the Crown to have submitted to the jury that the odds against the appellant being innocent were (because of the statistics in Table 3.58) 73 million to one against. That submission would in our judgment have been obviously fallacious, and had it been made, we would have expected Mr Bevan for the defence to have objected, the judge to have upheld the objection, and the 1 in 73 million figure would have gone as an unnecessary distraction. That there was no such application suggests the lack of impact of "... 1 in 73 million and even longer odds ..." on the third day of the summing-up of this long trial. But we must and do assume that counsel said what the judge reported him as having said. Might the jury have focused on that to the exclusion of the real and compelling evidence in this case?
165. We come then to the adequacy of the warning as to the use of statistics which the judge gave when first introducing them before his summary of Professor Meadow's evidence. For convenience we repeat that warning:

166. We have made clear what the judge should have told the jury: that it was the prosecution's case that to have one unexplained infant's death with no suspicious circumstances in the family was rare, and for there to be two such in the same family would be rarer still. That was the only relevance of Table 3.58, and the statistics were capable of showing that, but nothing more. They could not help as to whether the defendant was guilty or not guilty. Against this, we examine the caution that the judge gave the jury, found in the summing-up at Day 2, page 39. He was sounding "... a note of caution ...". However compelling the jury found the statistics "... we do not convict people on the statistics". That was a clear warning. He was not saying he found the statistics "compelling". He gave a limited purpose to them: "If there is one SIDS death in a family, it does not mean there cannot be another one". His passage was introduced as sounding a note of "caution" about the statistics. He conceded that they could attach "some significance" to them, but did not say what. But he led them back to the necessity of "... having regard to the individual circumstances of these deaths". He followed the cautionary passage with what "truly" were the conclusions of the experts relating to the deaths, with no mention of statistics. And that reflected the reality and focus of the trial. And it was that evidence that the trial had concentrated on. The difficulty we feel here is that by the time of the speeches, rarity was largely accepted, so the measure of rarity, the CESDI Study, was not important. The 73 million figure should have been cleared away as a distraction. Instead the judge considered that the statistics could be considered. Might the jury have been misled into attributing to those statistics a significance they did not have, i.e. as lengthening the odds against the deaths being natural?
167. The argument that the jury might have been so misled relies entirely on the wording of the passage. That has to be read in context. The context is a substantial trial properly concentrating on the medical issues and the circumstantial evidence, including the parents' evidence and their credibility. The contrary argument seeks to put the passage into context. The trial was not about statistics. The summing-up was detailed (170 pages approximately), careful and fair. Only two or three pages, in scattered passages, dealt with statistics. And there is the difficulty that the point identified in the grounds of appeal seems not to have been noticed in the trial. Had it been, it is difficult to see how it escaped judicial attention.
168. Despite those objections, we conclude that there is some substance to the criticism that the judge appeared to endorse the prosecution's erroneous approach in this particular. The ultimate question for us is whether the error of approach rendered the conviction unsafe. That is something best considered after examining the remaining grounds of appeal, when we can look at the case in the round.