Factual and Procedural Background
This opinion concerns a criminal appeal arising from the conviction of the Defendant for the deaths of two infants, Christopher and Harry. The case involved disputed medical and statistical evidence relating to whether the deaths were natural (Sudden Infant Death Syndrome - SIDS) or unnatural (homicide). The Crown's case relied on medical findings indicating inflicted injury and smothering, supported by statistical evidence purporting to show the extreme rarity of two natural unexplained infant deaths in one family. The defence disputed the interpretation of the statistics and the medical conclusions. Fresh expert evidence was later admitted on appeal, particularly regarding the medical evidence of haemorrhages in the eyes of one infant. The Court of Appeal considered multiple grounds of appeal, including the use and interpretation of statistical evidence, fresh medical evidence, and directions given to the jury regarding the Defendant’s "no comment" interview. The appeal was heard following conviction and involved detailed examination of expert reports and trial transcript.
Legal Issues Presented
- Whether the statistical evidence regarding the rarity of two SIDS deaths in one family was properly admitted and correctly explained to the jury, including whether the trial judge failed to warn against the prosecutor's fallacy.
- Whether the expert medical evidence, particularly concerning haemorrhages observed at the back of one infant's eyes, was correctly assessed and whether fresh evidence undermined the safety of the convictions.
- Whether the trial judge's directions regarding the Defendant's "no comment" interview were adequate and whether any adverse inference was properly disallowed.
- Whether any errors identified in the trial process rendered the convictions unsafe in light of the totality of evidence.
Arguments of the Parties
Appellant's Arguments
- The statistical evidence presented at trial, especially the 1 in 73 million figure for two SIDS deaths in one family, was misleading, overstated the rarity, and was treated improperly by the Crown and the trial judge.
- The calculation of the statistical risk failed to account for familial factors increasing recurrence risk, rendering the squaring of probabilities invalid and misleading.
- The trial judge did not adequately warn the jury against the prosecutor's fallacy, potentially allowing the jury to infer guilt from the rarity statistic improperly.
- Fresh medical evidence regarding haemorrhages at the back of the eyes would have undermined Dr Williams's reliability and should have been before the jury.
- The judge failed to direct the jury not to draw adverse inferences from the Defendant’s silence during the police interview, contrary to established legal principles.
- Errors in the use and explanation of statistics and medical evidence prejudiced the safety of the convictions.
Respondent's (Crown's) Arguments
- The CESDI Study, on which the statistical evidence was based, was a reputable government-supported study and its broad conclusions about the rarity of double SIDS deaths were admissible and relevant.
- The Crown did not rely on precise statistical probabilities but on the general rarity of two natural deaths to support their case.
- The trial judge gave appropriate cautionary directions about the statistics, emphasizing that convictions cannot be based on statistics alone and that individual circumstances must be considered.
- The medical evidence overwhelmingly supported the conclusion of unnatural death, independent of the statistical evidence.
- The fresh medical evidence on appeal did not undermine the convictions because the haemorrhages were a marginal issue and the fresh evidence would have affirmed rather than undermined Dr Williams's credibility.
- The judge’s directions regarding the Defendant’s silence were sufficient and agreed upon by defence counsel in advance, causing no prejudice.
Table of Precedents Cited
| Precedent |
Rule or Principle Cited For |
Application by the Court |
| R v. Doheny and Adams [1997] Cr App R 369 |
Limits on expert evidence, especially regarding the ultimate issue and the proper use of statistical evidence. |
Used to reject the argument that the expert improperly gave opinion on guilt based on statistics; confirmed that Professor Meadow did not overstep expert role. |
| R v. Deen, The Times, 10th January 1994 |
Warning against the prosecutor's fallacy in interpreting statistical evidence. |
Referenced in relation to the failure to warn jury adequately about the misuse of statistical probabilities. |
| R v. McGarry [1999] 1 Cr App R 377 |
Requirement for judges to direct juries not to draw adverse inferences from silence when s.34 Criminal Justice and Public Order Act 1994 direction is not given. |
Applied to assess whether the judge should have given a direction on the Defendant’s "no comment" interview. |
| Condron v. United Kingdom (2000) ECHR |
Fair trial rights under Article 6 regarding use of silence and inferences. |
Considered in relation to the fairness of the trial judge’s approach to the Defendant’s silence. |
| R v. Denis Adams (No 2) [1998] 1 Cr App R 377 |
Legal impracticality of performing statistical exercises comparing probabilities of guilt and innocence. |
Referenced to reject the notion that statistical probabilities could realistically assess competing causes such as double infant murder. |
| Stafford and Luvaglio v. DPP [1974] AC 878 |
Approach to appeals involving fresh evidence. |
Applied to frame the court’s approach to fresh evidence and assessment of conviction safety. |
| R v. Callaghan (1989) 88 Cr App R 40 |
Standard for assessing safety of conviction with fresh evidence. |
Applied to determine whether convictions remain safe after considering fresh evidence and trial errors. |
| R v. Jones [1997] 1 Cr App R 86 |
Obligation of appellate courts to exercise independent judgment on safety of conviction in light of new evidence. |
Reaffirmed approach to safety of conviction on appeal with fresh evidence. |
| R -v- Davies & Others (2000) |
Relationship between safety of conviction, unfairness, and misdirection at trial. |
Used to explain that a conviction may be unsafe due to trial irregularities even if guilt is not in doubt. |
Court's Reasoning and Analysis
The Court carefully examined the statistical evidence and its presentation at trial. It acknowledged that the CESDI Study was a reputable, large, government-supported epidemiological study providing broad estimates of SIDS risk in different family contexts. The Crown used it to illustrate the rarity of two genuine SIDS deaths in a low-risk family, quantified as approximately 1 in 73 million. The Court noted that the defence experts criticised the precision and assumptions underlying this figure, including the invalidity of squaring the probability to estimate recurrence risk due to unaccounted familial factors.
Despite these criticisms, the Court found that the statistical evidence was intended only to show the rarity of two SIDS deaths and was not conclusive proof of guilt. The trial judge gave clear cautionary directions emphasizing that convictions must not be based on statistics alone and that individual circumstances must be considered. The Court rejected the contention that Professor Meadow improperly relied on statistics in forming his opinion that the deaths were unnatural, finding his opinion was grounded on medical and circumstantial evidence.
The Court addressed the alleged prosecutor's fallacy, concluding that there was no evidence the Crown argued that the odds against innocence were 73 million to one. While the judge’s summing-up included a passage that could be read as endorsing such an approach, the Court found no indication that the jury were misled to that extent, especially given the overall focus on medical and circumstantial evidence.
Regarding the fresh medical evidence about haemorrhages at the back of one infant's eyes, the Court reviewed detailed expert reports and testimony. The new evidence suggested that the haemorrhages were likely real and consistent with asphyxial death rather than artefacts from autopsy procedures. However, the Court found this issue was peripheral and minor compared to the overall medical evidence. The fresh evidence would have strengthened the prosecution’s case rather than undermined it.
On the issue of the Defendant’s "no comment" interview, the Court acknowledged that while the judge did not give an explicit direction forbidding adverse inferences, the jury were clearly told that the Defendant was entitled to follow solicitor’s advice and that no adverse inference should be drawn. The approach was agreed by defence counsel and did not cause prejudice or unfairness.
In assessing the safety of the convictions, the Court found overwhelming medical and circumstantial evidence of guilt, including evidence of inflicted injuries, inconsistent accounts, and the suspicious timing and circumstances of both deaths. The Court concluded that the statistical evidence was a minor part of the trial and that even if properly handled, the convictions would have been the same. The identified error in the treatment of statistical evidence did not render the convictions unsafe.
Holding and Implications
The Court held the appeal against conviction DISMISSED.
The convictions for the deaths of the two infants were upheld as safe and satisfactory. While the Court accepted that there was an error in the approach to the statistical evidence at trial, this did not undermine the overwhelming medical and circumstantial evidence supporting guilt. The fresh medical evidence did not weaken but rather could have strengthened the prosecution case on a peripheral issue. The directions regarding the Defendant’s silence were adequate and did not prejudice the trial.
No new legal precedent was established. The decision confirms the importance of careful use and explanation of statistical evidence in criminal trials, the limits of expert opinion on ultimate issues, and the requirement for clear jury directions on the use of silence. The case underscores that even where statistical evidence is imperfectly handled, convictions may remain safe if supported by strong independent evidence.