Predicting the need for intubation within 3 h in the neonatal intensive care unit using a multimodal deep neural network

Abstract

Respiratory distress is a common chief complaint in neonates admitted to the neonatal intensive care unit. Despite the increasing use of non-invasive ventilation in neonates with respiratory difficulty, some of them require advanced airway support. Delayed intubation is associated with increased morbidity, particularly in urgent unplanned cases. Early and accurate prediction of the need for intubation may provide more time for preparation and increase safety margins by avoiding the late intubation at high-risk infants. This study aimed to predict the need for intubation within 3 h in neonates initially managed with non-invasive ventilation for respiratory distress during the first 48 h of life using a multimodal deep neural network. We developed a multimodal deep neural network model to simultaneously analyze four time-series data collected at 1-h intervals and 19 variables including demographic, physiological and laboratory parameters. Evaluating the dataset of 128 neonates with respiratory distress who underwent non-invasive ventilation, our model achieved an area under the curve of 0.917, sensitivity of 85.2%, and specificity of 89.2%. These findings demonstrate promising results for the multimodal model in predicting neonatal intubation within 3 h.

Clinical Limitations of Non-pharmacological Comfort Tools for Needle Sticks in Neonatal Intensive Care Units (NICU)

“…39 During the time an infant is intubated, they may require needle sticks for blood gas monitoring, routine lab draws, etc. 40,41 Intubation covers the oral cavity of the neonate eliminating the option of a pacifier or gloved finger as a comfort tool during needle sticks.…”

Explainable artificial intelligence (XAI) for predicting the need for intubation in methanol-poisoned patients: a study comparing deep and machine learning models

“…Additionally, repeated intubation attempts are linked to a heightened risk of complications including cardiac arrest, hypoxemia, arrhythmia, regurgitation, and airway trauma 4 . Various Artificial intelligence (AI) techniques such as machine learning (ML) and deep learning (DL) can accurately predict the need for intubation in people and prevent many unwanted complications and even death 5 , 7 , 8 .…”

“…The findings from the study by Im et al 7 , demonstrate that utilizing a multimodal deep neural network (MDNN) model integrating clinical data and time-series variables accurately predicted the need for intubation within the next 3 h in neonates with respiratory distress, achieving an accuracy of 88.2%. This suggested model could assist in guiding decision-making for neonates experiencing respiratory distress necessitating endotracheal intubation 7 .…”

“…The findings from the study by Im et al 7 , demonstrate that utilizing a multimodal deep neural network (MDNN) model integrating clinical data and time-series variables accurately predicted the need for intubation within the next 3 h in neonates with respiratory distress, achieving an accuracy of 88.2%. This suggested model could assist in guiding decision-making for neonates experiencing respiratory distress necessitating endotracheal intubation 7 . Additonally, the findings from the study conducted by Siu et al demonstrated that machine learning techniques could forecast the requirement for intubation in critically ill patients by leveraging routinely gathered bedside clinical parameters and laboratory findings.…”

“…To our knowledge, no study has utilized DL and ML techniques to predict the need for intubation in methanol-poisoned patients. The studies conducted cover predicting intubation needs in various contexts: among COVID-19 patients 8 , in the emergency department 4 , within 3 h in neonatal intensive care units using a multimodal deep neural network 7 , and within the first 24 h after critical care admission using machine learning approaches 5 . In these studies, only ML or DL techniques have been used; these two techniques have not been compared, and they have not focused on the need for intubation in methanol-poisoned patients.…”