Cross-validated prediction model for severe adverse neonatal outcomes in a term, non-anomalous, singleton cohort

Abstract

ObjectiveThe aim of this study was to develop a predictive model using maternal, intrapartum and ultrasound variables for a composite of severe adverse neonatal outcomes (SANO) in term infants.DesignProspectively collected observational study. Mixed effects generalised linear models were used for modelling. Internal validation was performed using the K-fold cross-validation technique.SettingThis was a study of women that birthed at the Mater Mother’s Hospital in Brisbane, Australia between January 2010 and April 2017.PatientsWe included all term, non-anomalous singleton pregnancies that had an ultrasound performed between 36 and 38 weeks gestation and had recordings for the umbilical artery pulsatility index, middle cerebral artery pulsatility index and the estimated fetal weight (EFW).Main outcome measuresThe components of the SANO were: severe acidosis arterial, admission to the neonatal intensive care unit, Apgar score of ≤3 at 5 min or perinatal death.ResultsThere were 5439 women identified during the study period that met the inclusion criteria, with 11.7% of this cohort having SANO. The final generalised linear mixed model consisted of the following variables: maternal ethnicity, socioeconomic score, nulliparity, induction of labour, method of birth and z-scores for EFW and cerebroplacental ratio. The final model had an area under the receiver operating characteristic curve of 0.71.ConclusionsThe results of this study demonstrate it is possible to predict infants that are at risk of SANO at term with moderate accuracy using a combination of maternal, intrapartum and ultrasound variables. Cross-validation analysis suggests a high calibration of the model.

Preventing The Human Factor: Organisational Aspects Linked to Foetal Asphyxia.

“…An extra step forward could be screening to identify children that are at higher risk of experiencing neonatal asphyxia from a low risk population. Some studies have been conducted and have found that it is possible to predict infants that are at risk of severe adverse neonatal outcomes at term with moderate accuracy using an algorithm that combines maternal, intrapartum and ultrasound variables (28).…”

Protocol for the development and validation of a risk prediction model for stillbirths from 35 weeks gestation in Australia

“…The performance of development and validation datasets will be assessed via overall performance (R 2 ), calibration, discrimination, and clinical performance will be assessed through positive predictive value (PPV) and negative predictive value (NPV). A fixed false-positive cutoff of 10% will be used for PPV and NPV [38]. Calibration characterizes model performance in terms of agreement between predicted (expected) risk and observed risk and is reported using a calibration plot [39].…”

Protocol for the development and validation of a risk prediction model for stillbirths from 35 weeks gestation in Australia

“…Backward stepwise elimination in a multivariable logistic regression model will be applied to remove non-signi cant factors with p-values greater than 0.100 in line with Akaike's Information Criterion (30). Finally, the risk prediction model will be applied and fully validated for each week gestation from 35 weeks (six total models: 35,36,37,38,40, and 41+ weeks).…”

“…The performance of development and validation datasets will be assessed via overall performance (R 2 ), calibration, discrimination and clinical performance will be assessed through positive predictive value (PPV), and negative predictive value (NPV). A xed false positive cut-off of 10% will be used for PPV and NPV (37).…”