Education and debate

A case of murder and the BMJ

Roy Meadow, emeritus professor of paediatrics and child health. 

Leeds LS17

Sally Clark, a 34 year old mother, was convicted in 1999 of the murder of her two sons Christopher, who died at the age of 11 weeks in December 1996, and Harry, who died at the age of 8 weeks in January 1998. Both the children had been previously healthy; they died suddenly in her care, in the evening, at home. Postmortem examinations of both children showed multiple abnormal findings.

After the verdict, the media reported the family's claim that both children had "died of cot death" and that incorrect statistical evidence given at the trial had greatly underestimated the likelihood of recurrence of cot death. The BMJ published an editorial as deficient in its sources as it was sensational in its title.¹ It was called "Conviction by mathematical error?" and it ignored the fact that at the trial neither the defence nor any of the expert witnesses advanced the claim that the children's deaths were examples of sudden infant death syndrome. The risk of recurrent sudden infant death syndrome was irrelevant to the conviction.

	What was the evidence?

The trial, at Chester Crown Court, was long, and the fact that both parents were solicitors led to more publicity than usual. Many medical experts were called by both prosecution and defence, including eight pathologists with Home Office, forensic, paediatric, ophthalmic, and neurological expertise. There were many days of pathological evidence. Opinion differed about the extent, nature, and cause of the unusual findings identified on the dead infantsfor example, the exact number of petechial haemorrhages on the face, the location of bruises, the cause of the torn and bruised frenulum, the degree of haemorrhage at the back of the eye, whether the spinal cord had been damaged at more than one level, the timing of the old rib fracture, the likelihood of the recent rib dislocation being caused by resuscitation, the proportion of alveoli showing evidence of previous bleeding into the lungs, and the exact timing of the hypoxic damage identified in the brain. This heavy morphological evidence was supplemented by that from experts seeking, and failing to find, natural conditions, disorders, or events that could account for these abnormalities, which are generally associated with trauma.

At the end of the trial none of the pathologists or clinicians had described or classified the death of either child as an example of sudden infant death syndrome. Several attributed the findings to physical abuse, with smothering and shaking as the probable causes of death.

	Misleading BMJ editorial

The BMJ editorial, which suggested that "the conviction is unsafe," did not cite from the trial. The writer did not seem to know the substantive issues of Mrs Clark's trial but instead relied on newspaper reports. The editorial was circulated and publicised by the family as indicating miscarriage of justice.

Sally Clark

The headline was irresponsible because it was unrelated to the facts of the case. There had been days of detailed pathological evidence, clinical opinions, and clinical discussion but only a few minutes of statistical discussion that had come relatively early in the trial. At the committal proceedings five months before the trial, at which medical evidence for the defence was not submitted, I referred to the frequency with which physical abuse recurred within families compared with the rarity of recurrence of sudden infant death syndrome. At the full trial, the detailed information concerning the likelihood of sudden infant death syndrome in successive children came from the recently published book reporting the results of the confidential inquiry into sudden unexpected deaths in infancy.² The court had been provided with pages 91 and 92 of that book, which states that for a family in which the mother is over the age of 27, does not smoke, and has a waged income, "the risk of two infants dying as SIDS by chance alone will thus be one in (8543×8543), i.e. approximately 1:73 million." I had been asked to provide the text.

I testified that in my opinion neither child's death was an example of sudden infant death syndrome. As it quickly became clear that none of the other clinical or pathological experts believed that the deaths were examples of sudden infant death syndrome, discussion of its recurrence rate was irrelevant. In the judge's final summing up, which extended to about 170 pages, there were only a few paragraphs about statistics. In these the judge advised the jury to treat the statistics with caution.³

	Grounds of appeal

An appeal against the conviction was launched on five grounds, two of which concerned medical evidence. The first related to the eye haemorrhages, which five ophthalmic pathologists were summoned to review. The opinion of the appeal court was that the expert review strengthened the evidence concerning the extent and nature of the haemorrhages.⁴

The other medical ground for appeal related to misinformation concerning the statistical probability of two deaths from sudden infant death syndrome in one family. The appeal judges accepted that statistics relating to cot death had been handled inappropriately. They stated that "Professor Meadow did not misuse the figure in his evidence, though he did not help to explain their limited significance."⁵ They noted that there had been little questioning of experts about recurrence rates of sudden infant death syndrome because none of the experts considered that either child's death was an example of the syndrome. The appeal court judges concluded that "In the context of the trial as a whole, the point on statistics was of minimal significance and there is no possibility of the jury having been misled so as to reach verdicts that they might not otherwise have reached."⁶

After considering the two medical and the three non-medical grounds for appeal, the judges unanimously dismissed the appeal and gave eight reasons why they concluded "that the evidence was overwhelming." Two of those reasons were that "in each case there was evidence of previous abuse," and "in each case there was evidence of deliberate injury recently inflicted."⁷

The family will continue to fight their corner. Some journalists will continue to support them, and very few journalists reporting this case, or the subsequent appeal, have mentioned that both children showed signs of both recent and past abuse, or that none of the medical experts considered that they had died of cot death. In today's world it is inevitable that, despite the appeal judges' finding that both Mr and Mrs Clark gave untrue evidence, formal letters of complaint from the family to the Police Complaints Authority, the General Medical Council, the royal colleges, or other statutory bodies will be treated with respect and will consume considerable resources.

	Balance and accuracy

What part did the BMJ play? There will be different opinions. My main concern is that the BMJ did not check the facts and verify the issues of the Clark case before using the title "Conviction by mathematical error?" The three appeal court judges pointed out the fallacy of opinions based on newspaper reports where "we do not know what the reporter heard, nor what he understood."⁸ Although the BMJ published an editorial suggesting a miscarriage of justice, it did not subsequently publish the fact that the Court of Appeal pronounced the evidence for murder overwhelming.

All doctors recognise the conundrum of the BMJ's relationship with the medical profession. The BMA is a prestigious representative and respected professional organisation. The BMJ is its journal. Inevitably, those outside medicine believe that the BMJ represents the British medical establishment; and they probably pay more attention to its leading articles than do medical readers who are attuned to its idiosyncrasies. Yet at a time when the rest of the media are so fast to criticise doctors, diminish their work, and point to their failings, there is an increased need for the BMJ to be scrupulous in checking the facts and to be prepared to speak out on behalf of doctors who, in difficult and unsupported circumstances, commit themselves to unpopular areas of work.

Recently the BMJ has published, alongside thoughtful letters from experienced scientists and clinicians, letters from the mad, the bad, and the misinformed. Presumably that has been done "to achieve balance." But the appearance of balance is not the same as balance itself. When the rest of the media are reporting preferential information against doctors, the best way for the BMJ to achieve balance is to respect the facts and present the medical case clearly and accurately.

	Footnotes

   Competing interests: RM was an expert witness for the prosecution.

	References

1.	Watkins SJ. Conviction by mathematical error? BMJ 2000; 320: 2-3[Free Full Text].
2.	Fleming P, Blair P, Bacon C, Berry J, eds. Sudden unexpected deaths in infancy. London: Stationery Office, 2000.
3.	Court of Appeal Judgment. R v Sally Clark. Case No 1999/07495/Y3. London: Royal Courts of Justice, October 2000:para 181.
4.	Court of Appeal Judgment. R v Sally Clark. Case No 1999/07495/Y3. London: Royal Courts of Justice, October 2000:para 227.
5.	Court of Appeal Judgment. R v Sally Clark. Case No 1999/07495/Y3. London: Royal Courts of Justice, October 2000:para 179.
6.	Court of Appeal Judgment. R v Sally Clark. Case No 1999/07495/Y3. London: Royal Courts of Justice, October 2000:para 272.
7.	Court of Appeal Judgment. R v Sally Clark. Case No 1999/07495/Y3. London: Royal Courts of Justice, October 2000:para 271.
8.	Court of Appeal Judgment. R v Sally Clark. Case No 1999/07495/Y3. London: Royal Courts of Justice, October 2000:para 173.