In our recent paper we reported that signs of asphyxia at birth were associated with the subsequent development of schizophrenia (Reference Dalman, Thomas and DavidDalman et al, 2001). Crow (Reference Crow2001), in his invited commentary, suggested that the birth records were assessed by midwives who were not ‘blind’ as to case—control status. As stated in the paper, we took care to eliminate this possibility and think it highly unlikely that the midwives became unblinded. We should add that, following the Vancouver agreement (International Committee of Medical Journal Editors, 1997), the midwives were not listed as authors as they only contributed to data gathering. We understand that Professor Crow has also adopted this policy in relation to the National Child Development Study interviews (Reference Done, Johnstone and FrithDoneet al, 1991).
Why were our findings so clear-cut in relation to asphyxia? There are at least two possible reasons. First, we took care to adjust for confounders and also adjusted for the association between different pregnancy and delivery complications in order to examine for an association independent of other complications. Second, by using paediatricians to examine birth records we may have been measuring birth asphyxia more accurately than with the Apgar index, which is only poorly related to asphyxia (Reference Sykes, Molloy and JohnsonSykes et al, 1982). Most of the other large studies carried out recently have relied upon routinely available data on pregnancy and birth complications. This might have introduced a random measurement error and could have obscured important associations.
Finally, the paper by Thomas et al (Reference Thomas, Dalman and David2001) does not contradict that of Dalman et al (Reference Dalman, Thomas and David2001). Thomas et al (Reference Thomas, Dalman and David2001) were concerned only with the possibility that pregnancy and delivery complications were more strongly associated with schizophrenia in certain subgroups. The results indicated that there were no statistically significant interactions so the association between asphyxia and schizophrenia was apparent in the whole sample.
eLetters
No eLetters have been published for this article.