Background: Recovery of hand function after stroke represents the hardest target for clinicians. Robot-assisted therapy has been proved to be effective for hand recovery. Nevertheless, studies aimed to refer patients to the best therapy are missing.
Methods: With the aim to identify which clinical features are predictive for referring to robot-assisted hand therapy, 174 stroke patients were assessed with: Fugl-Meyer Assessment (FMA), Functional Independence Measure (FIM), Reaching Performance Scale (RPS), Box and Block Test (BBT), Modified Ashworth Scale (MAS), Nine Hole Pegboard Test (NHPT). Moreover, patients ability to control the robot with residual force and surface EMG (sEMG) independently, was checked. ROC curves were calculated to determine which of the measures were the predictors of the event.
Results: sEMG control (AUC = 0.925) was significantly determined by FMA upper extremity (FMUE) (>24/66) and sensation (>23/24) sections, MAS at Flexor Carpi (<3/4) and total MAS (>4/20). Force control (AUC = 0.928) was correlated only with FMUE (>24/66).
Conclusions: FMUE and MAS were the best predictors of preserved ability to control the device by two different modalities. This finding opens the possibility to plan specific therapies aimed at maximizing the highest functional outcome achievable after stroke.
Keywords: Prediction; Recovery; Robotic; Stroke; Surface electromyography.
Copyright © 2021 Elsevier Ltd. All rights reserved.