Harnessing machine learning for predicting successful weaning from mechanical ventilation: A systematic review.

Area of Science:

• Critical Care Medicine
• Biomedical Informatics
• Computational Biology

Background:

• Machine learning (ML) models are increasingly utilized for predicting mechanical ventilation (MV) weaning outcomes in adult patients.
• The application of ML to MV weaning is a relatively recent but rapidly evolving field.
• This systematic review evaluates the efficacy of ML in predicting successful weaning from MV.

Purpose of the Study:

• To systematically review and assess the performance of machine learning models in predicting successful weaning from mechanical ventilation.
• To identify the most effective ML algorithms and methodologies for MV weaning prediction.
• To highlight gaps and future directions in ML applications for respiratory care.

Main Methods:

• Comprehensive literature search across PubMed, EMBASE, Scopus, Web of Science, Google Scholar, ACM Digital Library, and IEEE Xplore up to May 2024.
• Inclusion of peer-reviewed studies focusing on ML models for predicting successful MV weaning in adult patients.
• Quality assessment of included studies using a modified Joanna Briggs Institute checklist.

Main Results:

• Eleven studies involving 18,336 patients were included.
• Boosting algorithms like XGBoost and Light Gradient-Boosting Machine were most common, followed by Random Forest and Neural Networks.
• XGBoost generally showed superior performance in Area Under the Curve comparisons, with high-quality studies predominating.

Conclusions:

• Machine learning models are effective tools for predicting successful weaning from mechanical ventilation in adult patients.
• XGBoost emerged as a top-performing algorithm for this predictive task.
• Future research should explore advanced architectures like transformer models to further enhance prediction accuracy.