Ruth Del Río1, Marta Thió2, Mattia Bosio3, Josep Figueras4, Martín Iriondo5. 1. Departamento de Neonatología, Hospital Sant Joan de Déu, BCNatal-Hospital Clínic-Hospital Sant Joan de Déu, Universitat de Barcelona, Barcelona, España. Electronic address: rdelrio@hsjdbcn.org. 2. Newborn Research Centre, The Royal Women's Hospital, Melbourne, University of Melbourne, Melbourne, Australia. 3. Barcelona Supercomputing Center (BSC), Barcelona, España. 4. Departamento de Neonatología, Hospital Clínic, BCNatal-Hospital Clínic-Hospital Sant Joan de Déu, Universitat de Barcelona, España. 5. Departamento de Neonatología, Hospital Sant Joan de Déu, BCNatal-Hospital Clínic-Hospital Sant Joan de Déu, Universitat de Barcelona, Barcelona, España.
Abstract
INTRODUCTION: Extreme prematurity is associated with high mortality rates. The probability of death at different points in time is a priority for professionals and parents, and needs to be established on an individual basis. The aim of this study is to carry out a systematic review of predictive models of mortality in premature infants that have been published recently. METHODS: A double search was performed for article published in PubMed on models predicting mortality in premature neonates. The population studied were premature neonates with a gestational age of ≤30 weeks and / or a weight at birth of ≤1500g. Works published with new models from June 2010 to July 2019 after a systematic review by Medlock (2011) were included. An assessment was made of the population, characteristics of the model, variables used, measurements of functioning, and validation. RESULTS: Of the 7744 references (1st search) and 1435 (2nd search) found, 31 works were selected, with 8 new models finally being included. Five models (62.5%) were developed in North America and 2 (25%) in Europe. A sequential model (Ambalavanan) enables predictions of mortality to be made at birth, 7, 28 days of life, and 36 weeks post-menstrual. A multiple logistic regression analysis was performed on 87.5% of the models. The population discrimination was measured using Odds Ratio (75%) and the area under the curve (50%). "Internal Validation" had been carried out on 5 models. Three models can be accessed on-line. There are no predictive models validated in Spain. CONCLUSIONS: The making of decisions based on predictive models can lead to the care given to the premature infant being more individualised and with a better use of resources. Predictive models of mortality in premature neonates in Spain need to be developed. Crown
INTRODUCTION: Extreme prematurity is associated with high mortality rates. The probability of death at different points in time is a priority for professionals and parents, and needs to be established on an individual basis. The aim of this study is to carry out a systematic review of predictive models of mortality in premature infants that have been published recently. METHODS: A double search was performed for article published in PubMed on models predicting mortality in premature neonates. The population studied were premature neonates with a gestational age of ≤30 weeks and / or a weight at birth of ≤1500g. Works published with new models from June 2010 to July 2019 after a systematic review by Medlock (2011) were included. An assessment was made of the population, characteristics of the model, variables used, measurements of functioning, and validation. RESULTS: Of the 7744 references (1st search) and 1435 (2nd search) found, 31 works were selected, with 8 new models finally being included. Five models (62.5%) were developed in North America and 2 (25%) in Europe. A sequential model (Ambalavanan) enables predictions of mortality to be made at birth, 7, 28 days of life, and 36 weeks post-menstrual. A multiple logistic regression analysis was performed on 87.5% of the models. The population discrimination was measured using Odds Ratio (75%) and the area under the curve (50%). "Internal Validation" had been carried out on 5 models. Three models can be accessed on-line. There are no predictive models validated in Spain. CONCLUSIONS: The making of decisions based on predictive models can lead to the care given to the premature infant being more individualised and with a better use of resources. Predictive models of mortality in premature neonates in Spain need to be developed. Crown
Authors: Corsino Rey; Juan Mayordomo-Colunga; Roberts Gobergs; Reinis Balmaks; Ana Vivanco-Allende; Andrés Concha; Alberto Medina; Ana Colubi; Gil González-Rodríguez Journal: Biomed Res Int Date: 2022-03-03 Impact factor: 3.411