The validity of the variable "NICU admission" as an outcome measure for neonatal morbidity: a retrospective study.

Objective: To determine whether "neonatal intensive care unit (NICU) admission" is a valid surrogate outcome measure to assess neonatal condition in clinical studies.

Design: Retrospective study.

Setting: Tertiary hospital in the Netherlands.

Population: Neonates admitted to NICU during a 10-year period. Inclusion was restricted to singletons born beyond 37 weeks of gestation, and admitted to NICU in the first 24 h for delivery-related morbidity.

Methods: Patient characteristics and admission data were compared for four groups based on the line of care during delivery, i.e. home birth (Ia), midwife-led hospital delivery (Ib), secondary care (II), tertiary care (III).

Main Outcome Measures: Percentage of neonates/infants that died during NICU admission, diagnosis on admission, treatment received and a Neonatal Therapeutic Intervention Score System (NTISS).

Results: We studied 776 newborns (Ia 52, Ib 25, II 160, III 512, 27 unknown). The mortality rate differed significantly (Ia 15%, Ib 12%, II 22%, III 1%, p < 0.01), as did the NTISS morbidity scores at admission [Ia 12.0 (6.0-23.0), Ib 8.5 (6.3-10.0), II 21.0 (15.0-30.0), III 6.0 (4.0-9.0); p < 0.01], diagnosis at admission, received treatment and the duration of admission.

Conclusions: The severity of neonatal illness after 37 weeks of gestation differed depending on the line of care in which they were born, with neonates born in secondary care consistently having the highest morbidity, and those born in tertiary care having the lowest. NICU admission should not be used as an outcome measure for neonatal morbidity, specifically not when comparing different birth settings.