Investigating the role of a mid-term clinical assessment in the prediction of motor recovery in post-stroke survivors

This research explores the impact of incorporating mid-term clinical assessments on predicting robotic rehabilitation outcomes for post-stroke patients using Random Forest models. While predicting motor recovery is complex and critical for personalizing therapies and reducing healthcare costs, adding mid-term assessments yielded limited improvements in prediction accuracy. The study compared predictions using only initial assessments versus both initial and mid-term assessments for three clinical scales: FMA, ARAT, and MI. Results showed that only ARAT predictions benefited from mid-term data, suggesting that further analysis is needed to fully understand the potential of mid-term assessments in enhancing predictive models.