Abstract

Background: Previous studies suggest that high risk and low birthweight babies have better outcomes if born in hospitals with level III neonatal intensive care units. Relations between obstetric care, particularly intrapartum interventions and perinatal outcomes, are less well understood, however.

Objective: To investigate effects of obstetric, paediatric, and demographic factors on rates of hospital stillbirths and neonatal mortality.

Methods: Cross sectional data on all 65 maternity units in all Thames Regions, 1994–1996, covering 540 834 live births and stillbirths. Hospital level analyses investigated associations between staffing rates (consultant/junior paediatricians, consultant/junior obstetricians, midwives), facilities (consultant obstetrician/anaesthetist sessions, delivery beds, special care baby unit, neonatal intensive care unit cots, etc), interventions (vaginal births, caesarean sections, forceps, epidurals, inductions, general anaesthetic), parental data (parity, maternal age, social class, deprivation, multiple births), and birthweight standardised stillbirth rates and neonatal mortality.

Results: Unifactorial analyses showed consistent negative associations between measures of obstetric intervention and stillbirth rates. Some measures of staffing, facilities, and parental data also showed significant associations. Scores for interventional, organisational, and parental variables were derived for multifactorial analysis to overcome the statistical problems caused by high intercorrelations between variables. A higher intervention score and higher number of consultant obstetricians per 1000 births were both independently and significantly associated with lower stillbirth rates. Organisational and parental factors were not significant after adjustment. Only Townsend deprivation score was significantly associated with neonatal mortality (positive correlation).

Conclusions: Birthweight adjusted stillbirth rates were significantly lower in units that took a more interventionalist approach and in those with higher levels of consultant obstetric staffing. There were no apparent associations between neonatal death rates and the hospital factors measured here.