Abstract 16659: What Contributes to the Difference in Cost Between the CICU and NICU in Neonates Undergoing Congenital Heart Surgery?
Introduction: The mechanisms for previously reported increased costs of care for neonatal congenital heart surgery (CHS) in a neonatal intensive care unit (NICU) compared to a cardiac (CICU) or pediatric ICU (PICU) are not known. We leveraged decomposition analysis, commonly used in labor economics, to understand determinants of differences in cost by location of care.
Methods: Patient and center data were collected on neonates (<30 days old) who had CHS from 2003-13 using the Pediatric Health Information Systems. Differences in hospital costs adjusted for region and inflation based on care location (NICU vs. CICU/PICU) were modeled with gamma regression. A decomposition analysis with bootstrapping determined characteristic (explainable by differing covariate levels) and structural effects (if covariates are held constant). Covariates included hospital volume, gender, admission age, RACHS-1 score, prematurity, payor, race, admission year, syndrome, infection, ECMO and delayed sternal closure.
Results: For 20,595 neonates (NICU 47%, CICU/PICU 53%), NICU admission was associated with $22,222±3,430 higher unadjusted cost (p<0.001) compared to CICU/PICU admission. Characteristic effects accounted for $27,015±2,285 higher cost in the NICU (p<0.001). Contributing to this effect were higher rates of infection ($12,602±862, p<0.001), prematurity ($2,485±448, p<0.001), syndromes ($1,379±187, p<0.001), lower hospital volume ($10,587±731, p<0.001) and lower admission age ($3,979±823, p<0.001) for the NICU vs. CICU/PICU. Aggregate structural effects were not associated with cost differences (p=0.1), but were significantly different for some covariates examined individually, including higher NICU costs when admission age ($5,220±1361, p<0.001) and infection rates were held constant ($3,485±1334, p=0.009).
Conclusions: While many factors increasing cost of CHS care in the NICU may not be modifiable, the infection rates which contributed most to higher costs (compared to CICU/PICU) is a modifiable target. Higher costs in the NICU vs. CICU/PICU even when infection and admission age are constant may be secondary to differing overhead cost structures that require further delineation.
Author Disclosures: J.T. Johnson: None. K. Sullivan: None. X. Sheng: None. T. Greene: None. D. Bailly: None. A. Eckhauser: None. L. Minich: None. N. Pinto: None.
- © 2016 by American Heart Association, Inc.