Machine Learning-Based Clusters of Vital Signs and Lactate Levels Predict Vasopressor Use in Sepsis.

Key Points

• To explore how machine learning clusters of vital signs and lactate levels can predict vasopressor use and associated mortality in sepsis.
• Utilized machine learning algorithms to cluster initial vital signs and lactate levels.
• Analyzed patterns of vasopressor use across different clusters.
• Evaluated associations with mortality rates.
• Identified distinct patterns of vasopressor use linked to specific clusters of vital signs.
• Demonstrated variations in mortality rates based on these patterns.
• Found potential for machine learning models to inform timely vasopressor therapy.