Table 3.

Model 1: Multivariable Logistic Regression Analysis of the Primary Outcome Death or Neurodevelopmental Impairment with Two-Level Surgery Predictor Variable

Variablea	Adjusted Odds Ratio
	Estimate	95% CI
Surgery vs no surgery	1.29	1.08	1.55
Number of surgeries (for each additional surgery)	1.23	1.13	1.34
Birth weight (for each 250 g increase in weight)	0.70	0.60	0.81
Small-for-gestational-age	1.35	1.17	1.56
Male	1.63	1.42	1.88
Multiple birth cohort	1.30	1.15	1.48
Caucasian	0.75	0.67	0.84
5-min Apgar score ≤ 3	1.22	1.00	1.49
Antenatal corticosteroid exposure	0.84	0.72	0.98
Postnatal corticosteroid exposure	1.64	1.38	1.95
Seizures	2.93	2.31	3.71
Severe intracranial hemorrhage and/or cystic periventricular leukomalacia	2.12	1.78	2.54
Bronchopulmonary dysplasia (supplemental O2 at 36 wk)	1.47	1.25	1.72
Sepsis and/or meningitis	1.26	1.07	1.49
Necrotizing enterocolitis (Bell stage ≥ IIA)	1.26	0.83	1.92
Patent ductus arteriosus, excluding surgically closed patients	1.07	0.95	1.21
Caregiver highest educational level:
10–12 y vs ≤9 y	0.86	0.69	1.07
>12 y vs ≤9 y	0.62	0.51	0.76
Birth year:
2002–2005 vs 1998–2001	1.26	1.10	1.43
2006–2009 vs 1998–2001	0.64	0.55	0.76
Inborn	0.77	0.62	0.95
Propensity score- surgeryb AOR per 10% increase in the predicted probability	1.02	0.92	1.14

^aNeonatal Research Network center variable is also included in the model.

^bIncluded in the PS model were BW, SGA, sex, race, 5-minute Apgar score, ICH, BPD, sepsis and/or meningitis, NEC, PDA, multiple birth cohort, antenatal corticosteroid exposure (ANS), PNS, highest level of education attained by primary caregiver, birth year, and center.