The 1958 British Perinatal Mortality Survey revisited

Abstract

The British Perinatal Mortality Survey is an important epidemiological study, for it provided detailed information of the medical and social background of virtually all births in Great Britain in the first week in March 1958. Writing in this journal twenty years ago Fedrick and Anderson concluded from their study of the Survey that the risk of preterm birth was related to a number of factors including low social class. Fedrick and Anderson used a definition of pretenn birth which is no longer current, and did not have the benefit of the statistical packages which are available today. On page 57 Hajo Wildschut and his colleagues re-examine the association between low social class and preterm birth, using the World Health Organization definition of preterm birth and more sophisticated statistical approaches than could be employed in the past. The authors randomly divided all fifteen thousand births into estimation and validation samples for primiparae and multiparae separately. The estimation samples were used to construct logistic regression equations for primiparae and multiparae which were then tested on the validation samples. They then derived a scoring system to predict preterm birth and estimated the Youden index, which measures the efficiency of the scoring system by combining the sensitivity and specificity into a simple formula. There was no relation between social class and preterm birth in both primiparae and multiparae. In primiparae only maternal age and maternal diastolic blood pressure were associated with pretenn birth; the associations in multiparae were bleeding before 28 weeks of gestation, maternal diastolic blood pressure, a previous low birthweight infant and place of booking. The scoring system performed poorly in both primiparae and multiparae, such that it cannot be used to predict risk of preterm birth in individual women. The paper by Wildschut and colleagues suggests that in strategies to prevent preterm birth there is little justification for a policy designed to reduce social disadvantage.