Development of a Predictive Model for Neonatal Near Miss at a High-Complexity Referral Hospital in Peru during 2024

Marquina Mencía A.F.; Neira Mendoza M.P. · Zeitschrift fur Geburtshilfe und Neonatologie (2026) · DOI: 10.1055/a-2792-3454

Resumen

Objective To develop a multivariable predictive model for neonatal near miss (NNM) in a high-complexity hospital in Lima to improve early identification of high-risk newborns and optimize NICU resource allocation. Study Design Retrospective cohort of 264 neonates admitted at birth to the NICU in 2024. Maternal, sociodemographic, clinical, and neonatal management variables were analyzed. Multivariable logistic regression developed the model, compared against existing international pragmatic and combined criteria approaches. Results 82.2% experienced NNM and 17.8% died. Maternal factors were not associated. Key predictors included Apgar<7, CPR, CPAP use, surfactant therapy, and phototherapy. The model achieved AUC=0.94, sensitivity=94.01%, specificity=87.23%, outperforming reference models. Conclusion A model integrating NICU intervention indicators demonstrated high discriminatory power. Its use may enhance early detection of vulnerable neonates, guide care priorities, and rationalize resources. Further research should assess implementation feasibility and impact on outcomes across diverse healthcare settings.

Instituciones: Universidad Peruana de Ciencias Aplicadas