TITLE:
Machine Learning-Guided Intervention in Excited Delirium Syndrome: A Case Report of Multimodal AI Integration
AUTHORS:
Rocco de Filippis, Abdullah Al Foysal
KEYWORDS:
Excited Delirium Syndrome (ExDS), Machine Learning in Healthcare, Gradient Boosting Classifier, Predictive Analytics in Emergency Medicine, Clinical Decision Support Systems
JOURNAL NAME:
Open Access Library Journal,
Vol.12 No.1,
January
30,
2025
ABSTRACT: Excited Delirium Syndrome (ExDS) is a critical medical emergency characterized by severe autonomic dysregulation, agitation, and a high risk of fatal complications if not managed promptly. Effective treatment relies on early and accurate identification of high-risk patients, yet clinical decisions are often subjective and time-sensitive. This study explores the application of a Gradient Boosting Classifier (GBC) to predict patient outcomes and support treatment prioritization for ExDS. A synthetic dataset of 500 patient records, including demographic, physiological, and behavioral data, was used to train and evaluate the model. The GBC achieved an accuracy of 55% and a ROC-AUC of 0.54, demonstrating its potential in distinguishing stabilized and critical cases. Feature importance analysis identified body temperature, stabilization time, and heart rate as the most significant predictors, aligning with clinical insights. Visualizations, including a confusion matrix, prediction probability distributions, and cumulative gain and lift charts, provided valuable insights into the model’s performance and areas for improvement. While the model showed promise in triaging high-risk patients and prioritizing resources effectively, its performance for critical cases requires further refinement to reduce false negatives. This study highlights the potential of machine learning in augmenting clinical decision-making for ExDS, offering a foundation for future research and real-world applications to improve patient outcomes in emergency care settings.