SIZE OF THE MATERNITY
HOSPITAL AND PERINATAL OUTCOME
Devlieger H, Van Maele G, Martens G,
Bekaert A, Spitz B
Study Centre for Perinatal Epidemiology, Brussels, Belgium
Introduction: Since more than 10 years, all deliveries
in the Flanders (Northern part of Belgium) are registered on a free basis
by the Study Centre of Perinatal Epidemiology (SPE) together with 39
maternal and 24 neonatal items. The registration concerns yearly ¡À 62000
infants in 64 maternity hospitals.
Aim of the study: Within the scope of
this registration, risk stratification for each delivery was performed in
the different maternity hospitals to evaluate interhospital differences in
foetal and neonatal mortality.
First by
logistic regression analysis odds ratios were calculated for each
identification- and process variable according to the outcome (perinatal
mortality). Next a risk index score was derived from the odds ratios of
each of the significant variables. This risk index score allows classifying
deliveries according to the outcome. As a result, for each delivery a sum
of risk indices could be calculated, each score corresponding with a given
risk of perinatal mortality. Finally, based on the sum of risk scores of
all deliveries the corresponding mortality risk in individual maternity
hospitals was calculated and compared with the observed mortality.
Maternity hospitals with discordance between expected and observed
mortality were identified.
Results: 9 out of 24 small maternity hospitals (< 800
deliveries yearly) had a significantly higher observed perinatal mortality
than expected from the sum of risk index scores. Only one hospital with
more than 2000 deliveries a year had a worse than expected perinatal
outcome.
Conclusions: The risk stratification based on
logistic regression analysis and determination of a risk index score allows
for correct assessment of risk in a different case mix. Based on this model
of risk stratification it results that a consistent number of small sized
maternity hospitals in the Flanders have a higher risk of perinatal
mortality than the larger ones.