BIRTH-WEIGHT-SPECIFIC TRENDS IN PERINATAL-MORTALITY BY HOSPITAL CATEGORY IN SOUTH-AUSTRALIA, 1985-1990

Objective: To investigate differences by birthweight in risk of perina tal death between level 3 hospitals (which provide care for high risk pregnancies and neonatal intensive care) and other hospitals in South Australia, using perinatal data for the 1985-1990 period. Design: Anal ysis of birthweight-specific trends in risk of perinatal death by hosp ital category for singleton births, adjusting for risk factors. Subjec ts: 114 725 singleton births of at least 400 g birthweight (or at leas t 20 weeks' gestation) born in hospitals in the 1985-1990 period and n otified to the perinatal data collection. Main outcome measure: The re lative odds of a perinatal death, as opposed to a live birth which sur vived the neonatal period. Results: Births at level 3 hospitals had a higher crude risk of perinatal death than those at other hospitals, bu t this was due to the higher frequency of low birthweights at level 3 hospitals. For birthweights under 2000 g, and especially for the very low birthweights, there was a higher risk at non-level-3 than level 3 hospitals. There was also the unexpected finding that births at level 3 hospitals in the 2500-2999 g range had a comparatively high risk of perinatal death. There was little difference in risk for births of hig her birthweight. Conclusions: The greatly reduced risk of perinatal de ath in level 3 hospitals for babies with birthweights under 2000 g see ms likely to be due to the specialist services in these hospitals. Fur ther investigation is required to determine why babies in the 2500-299 9 g range of birthweights had a comparatively high risk of perinatal d eath at these hospitals. This appears to be due, at least in part, to an excess contribution of deaths from congenital abnormalities. Also, it seems that the higher prevalence of complications in pregnancy in l evel 3 hospitals, and the transfers for induction of labour after intr auterine fetal death, would have made a contribution. These same facto rs may also have affected the risk in level 3 hospitals for higher bir thweight births.