Abstract
Abstract
Introduction:
Neonatal intensive care unit (NICU) evacuation poses distinct operational and ethical challenges because neonates often depend on continuous life-sustaining technologies and specialized care. Although several neonatal or pediatric disaster triage systems have been proposed, actual NICU evacuation cannot be ethically or practically tested under controlled disaster conditions. A reproducible framework comparing evacuation-order strategies is therefore needed.
Study Objective:
This study aimed to develop and apply a reproducible simulation framework for NICU evacuation-order evaluation and, within that framework, to compare structured triage with evacuation without triage and quantitative with categorical neonatal triage approaches.
Methods:
A retrospective simulation study was conducted using routinely recorded clinical triage assessments from a tertiary NICU, comprising 1,079 patient evaluations in 162 neonates across 49 triage days. During clinical practice, clinical severity and triage status had been recorded using the Neonatal Therapeutic Intervention Scoring System (NTISS), Neonatal Extrication Triage (NEXT), and Simple Triage and Rapid Treatment for Neonates, Revised (START-Neo-R); NEXT was a quantitative six-domain resource-dependency score ranging from zero to 12, whereas START-Neo-R was a five-level categorical neonatal triage system. Three evacuation strategies were evaluated: descending-order evacuation using NEXT, descending-order evacuation using START-Neo-R, and evacuation without triage. Evacuation-order performance was quantified using the Total Evacuation Score (TES), defined as the time integral of cumulative NTISS scores among patients remaining in the NICU across sequential evacuation steps. The TES distributions were compared descriptively across strategies rather than inferentially.
Results:
Within the 49 observed triage days, both structured triage strategies generated TES distributions shifted lower than those from evacuation without triage. In the direct comparison between NEXT and START-Neo-R, the worst-case TES under NEXT was lower than more than one-half of the START-Neo-R TES distribution on 34 of the 49 observed triage days (69.4%).
Conclusion:
In this simulation study using real-world NICU data, TES provided a reproducible framework for comparing neonatal evacuation-order strategies under explicit assumptions. Within this TES-based framework, structured triage generated TES distributions shifted lower than those from evacuation without triage, and NEXT showed a more favorable distribution-based pattern than START-Neo-R on most observed days. These findings suggest that TES-based evaluation may help identify triage strategies that support more transparent and testable NICU evacuation planning.