Robot-assisted gait training in stroke based on ambulatory status: A systematic review and meta-analysis
Eun-Ji Kim, Yixin Wang, Hye-Seon JeonBackground
Robot-Assisted Gait Training (RAGT) can improve gait and balance after stroke; however, its efficacy depends on ambulatory level, robot type, and outcome measure.
Objective
This meta-analysis evaluated the overall and subgroup-specific effects of RAGT in stroke patients based on ambulation level, robot type, and outcome measures to guide individualized rehabilitation strategies.
Methods
Six databases (PubMed, Scopus, MEDLINE, Embase, Web of Science, and CINAHL) were searched through December 1, 2024. Twenty RCTs involving patients with stroke who underwent exoskeletal or end-effector RAGT were included. Subgroup analyses examined ambulation classification (independent or assisted ambulators), robot type (exoskeletal or end-effector), and outcome measures, including gait speed (m/s), the timed up-and-go (TUG) test, the Berg Balance Scale (BBS), and the FMA-LE.
Results
RAGT resulted in a small but statistically significant improvement in the gait and balance of patients with stroke. Small effect sizes were identified across the ambulatory levels. The exoskeletal robot group demonstrated a small effect size, whereas the end-effector group showed no significant changes. Subgroup analysis revealed RAGT was most effective in improving gait speed and TUG test performance among independent ambulators, whereas improvements in BBS scores were more prominent among those requiring assistance. The FMA-LE outcomes showed small but statistically significant improvements overall, with improvements observed primarily among assisted ambulators, not among independent ambulators.
Conclusion
This meta-analysis supports the effectiveness of RAGT for improving gait and balance in patients with stroke. Findings suggest that tailoring robotic interventions should be based on ambulation classification and rehabilitation goals.