91 Determinants of Death or Palliative Care in Children's Hospital NICUs

Abstract

Previous studies suggest that mortality is higher in neonates admitted to children's hospitals compared to perinatal centres, even after correction for severity of illness. We examined 625 admissions to two children's hospital neonatal intensive care units (NICUs) to ascertain if admission source (another NICU versus non-NICU) is a predictor of outcome after correction for severity of illess [Score for Neonatal Acute Physiology (SNAP-II) and SNAP Perinatal Extension (SNAPPE-II)]. Cases included all acute NICU admissions to Hospital for Sick Children from July to Dec. 2002 (n=338) and Montreal Children's Hospital in the year 2002 (n=287). Mean (SD) birth weight was 2654 g (1062), gestational age 36 wks (5), and 5-min Apgar 8 (2). Median (range) SNAP-II was 0 (0–65), SNAPPE-II 15 (0–84), age of admission 1 day (0–244), and length of stay 8 days (0–312). 183 (29%) admissions were transferred from another NICU, 254 (41%) had congenital anomalies that required admission, 234 (37%) required surgical intervention, and 63 (10%) died before discharge or were discharged on palliative care. Admissions from other NICUs had increased risk of death or palliative care [RR 2.20 (95% CI 1.38, 3.49)], congenital anomalies [1.30 (1.08, 1.58)], surgery [2.46 (2.03, 2.97)], and longer length of stay [mean (SD) 24 (48) vs 14 (33) days; p=0.0006)]. Binary analyses showed that admission from another NICU, SNAP-II, SNAPPE-II, and congenital anomalies were predictors of death or palliative care (all p's<0.01), while surgery was not (p=0.91). On logistic regression, independent predictors of death or palliative care were NICU source of admission [OR 2.85 (95% CI 1.53, 5.31); p=0.0010], SNAP-II [1.09 per point (1.06, 1.11); p<0.0001], and congenital anomalies [4.68 (2.46, 8.92); p<0.0001], while surgery was inversely associated [0.33 (0.17, 0.66); p=0.0015]. Results for SNAPPE-II were similar. These results suggest that comparisons of outcomes between NICUs should include source of admission and associated casemix (congenital anomalies, surgeries) as well as severity of illness for neonates admitted to children's hospitals.