Objective: To assess whether the volume of patients attended in neonatal intensive care units (NICU) is an independent predictor of 28-day mortality among very low birth weight (VLBW)/very low gestational age (VLGA) infants.
Methods: 54 NICU from 13 European countries submitting data for all admitted infants of a birth weight <1501 g or a gestational age <32 weeks in 2006 (N = 2301). Stepwise logistic regression was used to predict 28-day mortality rate (Akaike selection criteria). Variables included in the regression model were birth weight, gestational age, one and 5-minute Apgar scores, prenatal steroid use, major birth defects, multiple gestation, mode of delivery, gender and volume of VLBW/VLGA attended patients (<31, 31–96, >96 babies per NICU). Odds ratios (95% CI) were calculated and the performance of the logistic model was evaluated by the area under the receiver operator characteristic curve. The standardised mortality ratio was calculated for each NICU from the fitted model. A potential correlation of practices and outcomes at a hospital level was taken into account in the analysis.
Results: Mortality rates among VLBW/VLGA infants did not vary according to the volume of patients at the NICU (31–97 babies (OR 0.97; 95%CI 0.55 to 1.71; p = 0.92) and >97 babies (OR 0.88; 95% CI 0.53 to 1.47; p = 0.63)). Variables significantly associated with mortality were birth weight, gestational age, 5-minute Apgar score, major birth defects and gender.
In the 2006 EuroNeoStat cohort, the number of babies cared for in participating NICU does not seem to be associated with the 28-day mortality rate.