Introduction: Gait exercise assist robot (GEAR), a gait rehabilitation robot developed for poststroke gait disorder, has been shown to improve walking speed and to improve the poststroke gait pattern. However, the persistence of its beneficial effect has not been clarified. In this matched case–control study, we assessed the durability of the effectiveness of GEAR training in patients with subacute stroke on the basis of clinical evaluation and three-dimensional (3D) gait analysis. Methods: Gait data of 10 patients who underwent GEAR intervention program and 10 patients matched for age, height, sex, affected side, type of stroke, and initial gait ability who underwent conventional therapy were extracted from database. The outcome measures were walk score of Functional Independence Measure (FIM-walk), Stroke Impairment Assessment Set total lower limb motor function score (SIAS-L/E), and 3D gait analysis data (spatiotemporal factors and abnormal gait patter indices) at three time points: baseline, at the end of intervention, and within 1 week before discharge. Results: In the GEAR group, the FIM-walk score, SIAS-L/E score, cadence, and single stance time of paretic side at discharge were significantly higher than those at post-training (p < 0.05), whereas the stance time and double support time of the unaffected side, knee extensor thrust, insufficient knee flexion, and external rotated hip of the affected side were significantly lower (p < 005). However, no significant differences in these respects were observed in the control group between the corresponding evaluation time points. Conclusion: The results indicated significant improvement in the GEAR group after the training period, with respect to both clinical parameters and the gait pattern indices. This improvement was not evident in the control group after the training period. The results possibly support the effectiveness of GEAR training in conferring persistently efficient gait patterns in patients with poststroke gait disorder. Further studies should investigate the long-term effects of GEAR training in a larger sample.
All Science Journal Classification (ASJC) codes
- Biomedical Engineering
- Artificial Intelligence