Sir

Your News story “Quashed convictions reignite row over British cot deaths” (Nature 427, 384; 200410.1038/427384a) notes concern over the testimony of an expert witness, the distinguished paediatrician Roy Meadow, in several UK court cases in which parents were convicted of murdering their children.

Another aspect of his testimony that should be discussed more widely is the use of statistical arguments to conclude that the probability of two children in the same family dying a cot death was 1 in 73 million. Meadow based this testimony on statistics given in a government-commissioned report. The ‘1 in 73 million’ statistic was not the sole basis for Meadow's expert testimony, but sadly, neither the defence nor anybody else in court challenged the simple assumption behind the calculation that two cot deaths in the same family are unlikely to be related in any way.

A more convincing analysis by Ray Hill, a mathematician at Salford University, UK, reveals that, if there has already been one cot death in a family, the chance of a second one is 10 to 22 times higher. See http://pass.maths.org.uk/issue21/features/clark, where Helen Joyce uses Bayes' theorem to calculate the probability that the deaths arose from natural causes, by using plausible values for the alternative hypothesis that they were murdered. This analysis, which does not require an understanding of the underlying causes of cot death or child murder, leads to a probability of 0.625 that the deaths were natural (see B. Lewis, Math. Gaz. 87, 418–431; 2003).

Unhappily, the understanding that statistics is a difficult subject is not widespread, even among distinguished paediatricians.