AGREE: an upper limb motorized exoskeleton for restoring arm functions: a single-blinded randomized controlled trial.

Objectives: Technology-assisted neurorehabilitation is a promising strategy to address the healthcare system overload we are witnessing in recent years. Given the lack of concordance on the effectiveness of upper-limb exoskeletons, it is important to clinically test innovative prototypes. This randomized controlled trial evaluated AGREE, a motorized exoskeleton assisting shoulder and elbow movements with a novel unified control system, against standard care.

Methods: The study involved 32 neurological patients with upper limb motor deficits at Casa di Cura del Policlinico (Italy). The randomization list was generated electronically to allocate participants into two groups. Both groups received 15 training sessions (45 min, three times weekly). The outcome measures considered were: (i) usability (System Usability Scale), (ii) clinical changes in Fugl-Meyer Assessment, Action Research Arm Test, Motricity Index, and Box and Block Test from baseline to post-treatment, and iii) treatment dosage.

Results: The usability of the system was evaluated at the end of the intervention by participants with a resulting median SUS score of 68.7 (IQR = 31.8). Both groups showed comparable significant clinical improvements in all clinical outcome measures, with experimental group patients improving both according to the Fugl-Meyer ( $\Delta$  FM = 2, IQR =4.5 ) and the Action Research Arm Test ( $\Delta$  ARAT = 2, IQR = 4.5). The experimental group received significantly less actual treatment time than the control group, while the total number of sessions and therapy duration were equivalent for the two groups.

Discussion and conclusions: Therapists operated the AGREE system independently, and no adverse events occurred. Patients reported a moderate usability of the system. While outcomes slightly favoured the control group for FM and ARAT scales, the obtained results support literature evidence, which indicates that the two treatments are globally comparable. Notably, the experimental group achieved comparable improvements with reduced treatment time, suggesting therapy consistency and frequency may be more impactful than session duration. Future research should further examine optimal rehabilitation dosing strategies.