Editor,—The series of reprints that have arisen from the New Zealand cot death study, which took place around 1990, such as the article by Williams and colleagues1 are replicating work that we carried out in Sheffield in the l950s.2-4 In the Sheffield studies, as well as noting the significance of prone sleeping, we also saw an increase in children presenting as unexpected deaths at the weekend. In addition, we found a relative increase on the night after the day when family doctors took their traditional half day off and did not hold afternoon or evening surgeries. This feature became most apparent when the cot death rates were seen as part of the total pattern of deaths—that is, a diminution in acute deaths following admission to hospital was replaced by greater numbers of home cot deaths.

This pattern of deaths in Sheffield changed after we introduced the prevention programme identifying children at increased risk of unexpected death.5 We found that we had largely eliminated the partially explained group of cot deaths, and the total infant and cot death rates in the city fell considerably.

The point that needs to be made is that such sociopathological studies on child deaths should always be carried out in relation to the pattern and site of deaths, and to the total infant death rate in the local community. Much of the confusion related to risk discriminants results from the false assumption that with cot deaths one is dealing with a single cause. Particular causes can be increased, reduced, or eliminated. This has been particularly striking during the past 50 years relating to what are almost certainly accidental suffocation deaths. An increase in the unexpected death rate occurred following propaganda arising from neonatologists recommending prone sleeping in children. The recent complete reversal of that policy seems to have resulted in the elimination of that group and hence the fall in “cot death” rates to their original level.