Using machine learning models to predict oxygen saturation following ventilator support adjustment in critically ill children: a single center pilot study

Abstract

Clinicians experts in mechanical ventilation are not continuously at each patients bedside in an intensive care unit to adjust mechanical ventilation settings and to analyze the impact of ventilator settings adjustments on gas exchange. The development of clinical decision support systems analyzing patients data in real time offers an opportunity to fill this gap. The objective of this study was to determine whether a machine learning predictive model could be trained on a set of clinical data and used to predict hemoglobin oxygen saturation 5 min after a ventilator setting change. Data of mechanically ventilated children admitted between May 2015 and April 2017 were included and extracted from a high-resolution research database. More than 7.105 rows of data were obtained from 610 patients, discretized into 3 class labels. Due to data imbalance, four different data balancing process were applied and two machine learning models (artificial neural network and Bootstrap aggregation of complex decision trees) were trained and tested on these four different balanced datasets. The best model predicted SpO2 with accuracies of 76%, 62% and 96% for the SpO2 class \"< 84%\", \"85 to 91%\" and \"> 92%\", respectively. This pilot study using machine learning predictive model resulted in an algorithm with good accuracy. To obtain a robust algorithm, more data are needed, suggesting the need of multicenter pediatric intensive care high resolution databases.