Accuracy of Machine Learning to Predict Upper-Limb Outcome Within the First 72 Hours Poststroke

The authors aimed to develop and validate a machine learning model that accurately predicts upper-limb motor outcomes in stroke patients within the first 72 hours post-stroke, using simple clinical assessments. They found that a model utilizing specific bedside tests, including Shoulder Abduction and voluntary finger extension, achieved a median absolute error of 5.9 points on the Action Research Arm Test, demonstrating both feasibility and predictive accuracy for rehabilitation planning.