Validation of a Costing Algorithm and Cost Drivers for Neonates Admitted to the Neonatal Intensive Care Unit

Objective Neonatal intensive care units (NICUs) account for over 35% of pediatric in-hospital costs. A better understanding of NICU expenditures may help identify areas of improvements. This study aimed to validate the Canadian Neonatal Network (CNN) costing algorithm for seven case-mix groups with actual costs incurred in a tertiary NICU and explore drivers of cost. Study Design A retrospective cohort study of infants admitted within 24hours of birth to a Level-3 NICU from 2016 to 2019. Patient data and predicted costs were obtained from the CNN database and were compared to actual obtained from the hospital accounting system ( Cout par Parcours de Soins et de Services ). Cost estimates (adjusted to 2017 Canadian Dollars) were compared using Spearman correlation coefficient (rho). Results Among 1,795 infants included, 169 (9%) had major congenital anomalies, 164 (9%) with <29 weeks' gestational age (GA), 189 (11%) with 29 to 32 weeks' GA, and 452 (25%) with 33 to 36 weeks' GA. The rest were term infants: 86 (5%) with hypoxic-ischemic encephalopathy treated with therapeutic hypothermia, 194 (11%) requiring respiratory support, and 541 (30%) admitted for other reasons. Median total NICU costs varied from $6,267 (term infants admitted for other reasons) to $211,103 (infants born with <29 weeks' GA). Median daily costs ranged from $1,613 to $2,238. Predicted costs correlated with actual costs across all case-mix groups (rho range 0.78-0.98, p <0.01) with physician and nursing representing the largest proportion of total costs (65-82%). Conclusion The CNN algorithm accurately predicts NICU total costs for seven case-mix groups. Personnel costs account for three-fourths of in-hospital total costs of all infants in the NICU. Key Points Very preterm infants born below 33 weeks of gestation account for most of NICU resource use. Human resources providing direct patient care represented the largest portion of costs. The algorithm strongly predicted total costs for all case-mix groups.