Neonatal intensive care units: is the level of utilization still paralleled by infant mortality?

This research investigated the relationship of maternal place of residence with the utilization of neonatal intensive care units (NICU) and whether maternal residence was related to infant mortality in two contiguous Swiss cantons, Vaud and Valais, relying on the same tertiary NICU. Previous works have shown that infant mortality is close to the Swiss rate in Vaud, but elevated in Valais.

Analyses were based on linked birth and death certificates relating to the 57,962 single livebirths to mothers resident in Vaud or Valais delivered in hospitals over the 1979-1985 period. Data on utilization were retrospectively collected from NICU admissions' registers and linked to birth certificates.

Results of logistic regression models pointed to a large difference for all birthweight groups, in NICU utilization between Vaud and Valais after adjustment for risk factors abstracted from the birth certificate, when tertiary NICU beds were considered. When the definition of NICU utilization was enlarged to all beds of an identified NICU, irrespective of the location of the unit and of the intensity of care provided to admitted newborns, there was a clear reduction in geographical variations. Neonatal and infant mortality were significantly higher in the lower NICU utilization region only among newborns weighing > or = 2500 g at birth.

We found no difference in infant mortality (odds ratio 1.0) among < 2500 g newborns residing in regions characterized by a large difference in tertiary NICU utilization. Nevertheless, the limited sample size did not permit specific analyses of lower birthweight (e.g. < 1500 g) newborns. Replication of such research in other settings based on larger samples is needed.