Correction of a statistical error: The data does not show Letby's presence at the baby deaths could have occurred by chance

Dear editor,

Lucy Letby was found guilty of killing seven neonatal babies, in part because of her consistent association with the deaths. However, O’Quigley ¹ believes he has shown this association could have occurred by chance, and this has been widely quoted in support of the need for a retrial. I now show the underlying calculation is erroneous.

The error, which is often seen in statistical applications in biomedicine, is a failure to appreciate biological differences. Had there been nothing singular about the deaths of the seven neonates killed by Letby, the frequency of neonatal deaths could, as O’Quigley suggests, have allowed her presence by chance.

But the murdered babies didn’t die ordinary deaths: they were in a stable condition, and the deaths were sudden and unexpected; and as a post-mortem study of 1516 neonatal deaths showed, such unexpected deaths account for only about 6% of the total. ² It is this relatively small number of unexpected deaths, not the total number of neonatal deaths, that should have been used in O’Quigly's calculations; and that would have led to the opposite conclusion, namely, that Letby's presence at the baby deaths could not have been by chance.

O’Quigly's paper should be withdrawn or modified accordingly.