Bibliographic citations
Tovar, V., Zavala, M. (2024). Validación externa de dos modelos predictivos de displasia broncopulmonar en prematuros menores de 31 semanas de edad gestacional en una Unidad de Cuidados Intensivos Neonatal pública de nivel III en Lima, Perú [Universidad Peruana Cayetano Heredia]. https://hdl.handle.net/20.500.12866/16595
Tovar, V., Zavala, M. Validación externa de dos modelos predictivos de displasia broncopulmonar en prematuros menores de 31 semanas de edad gestacional en una Unidad de Cuidados Intensivos Neonatal pública de nivel III en Lima, Perú []. PE: Universidad Peruana Cayetano Heredia; 2024. https://hdl.handle.net/20.500.12866/16595
@misc{renati/1281994,
title = "Validación externa de dos modelos predictivos de displasia broncopulmonar en prematuros menores de 31 semanas de edad gestacional en una Unidad de Cuidados Intensivos Neonatal pública de nivel III en Lima, Perú",
author = "Zavala Romero, Marcelo",
publisher = "Universidad Peruana Cayetano Heredia",
year = "2024"
}
Background: Bronchopulmonary dysplasia (BPD) is associated with significant morbidity and mortality in the neonatal population. Multiple prognostic models for this disease have been developed, two of them of high quality, none of which have been externally validated in the Peruvian population. Objective: To externally validate two predictive models for the development of BPD, for postnatal days 7 and 14, in premature infants under 31 weeks of gestational age, in a Peruvian tertiary NICU. Materials and Methods: We validated the models by Laughon et al. and by Greenberg et al. with information from medical records of 368 patients under 31 weeks of gestational age hospitalized during 2012-2023. Discrimination was assessed using the M-index and the PDI; and calibration, with the O/E ratio. Results: The model by Laughon et al. presented a better overall discriminative ability than the model by Greenberg et al. on both measurement days (M index: 0.722 and 0.722 vs 0.662 and 0.682; PDI: 0.41 and 0.43 vs 0.34 and 0.37). The analysis of the pairwise C statistic and the specific PDI show large differences in the ability of both models to discriminate the different categories, being excellent for the non-BPD category and poor for the BPD severity grades. Both models show poor calibration due to the underestimation of some categories and the overestimation of others. Conclusions: Both models evaluated in this study did not have a good performance in our population, therefore we cannot recommend their use in Peruvian preterm infants.
This item is licensed under a Creative Commons License