JMIR Publications

ABSTRACT

Background:

Agitation sedation is critical in ICU care, affecting patient safety. Nursing assessments suffer from low frequency and subjectivity. Automating these assessments can boost ICU efficiency, treatment capacity, and patient safety.

Objective:

The study addresses the critical need for enhancing patient safety and efficiency in Intensive Care Units (ICUs) by automating the evaluation of agitation and sedation using advanced machine learning techniques.

Methods:

Utilizing the Taichung Veterans General Hospital intensive care database (2020), an ensemble learning model was developed for classifying the levels of agitation and sedation. Different sequences of ensemble learning models were compared. In addition, an interpretable artificial intelligence approach (SHAP) was employed for explanatory analysis.

Results:

With 20 features and 121,303 data points, Random Forest achieved high AUC in all models (Sedation classification = 0.97, Agitation classification = 0.88). The ensemble learning model enhanced agitation sensitivity (0.82) while maintaining high AUC values for all categories (all above 0.82). The model explanation aligns with clinical experience.

Conclusions:

This study proposes ICU agitation-sedation assessment automation using ML, enhancing efficiency and safety. Ensemble learning improves agitation sensitivity while maintaining accuracy. Real-time monitoring and future digital integration promise ICU care advancements.