Effects of robot-assisted gait training using virtual reality on balance and gait ability in stroke survivors: systematic review and meta-analysis of randomized controlled trials

Objectives To evaluate the effectiveness of robot-assisted gait training using virtual reality(VR) on balance and gait ability in persons with stroke.Design A systematic review and meta-analysis of randomised controlled trials.Methods Two reviewers independently screened studies, extracted data, and assessed risk of bias using the Cochrane Risk of Bias 2.0 tool. Random-effects meta-analyses were conducted using mean differences (MD) or standardised mean differences (SMD) with 95% confidence intervals (CI). Data sources were Pubmed, Embase, Cochrane Library and Web of Science searched between February 2013 and February 2025. The study eligibility criteria included randomised controlled trials (RCTs) involving stroke survivors. The intervention was robot-assisted gait training (RAGT) combined with VR. The control group included conventional physical therapy or RAGT without VR. Outcome measures included balance (BBS, TUG) and gait ability (10MWT).Results Eight randomised controlled trials involving 233 participants were included. Robot-assisted gait training using virtual reality significantly improved balance as measured by the Berg Balance Scale (MD = 4.05, 95% CI 1.72–6.37) and Timed Up and Go test (MD = −3.71, 95% CI −5.91 to −1.51), and gait ability assessed by the 10-Meter Walking Test (SMD = −1.86, 95% CI −2.53 to −1.20). Heterogeneity was moderate to high across outcomes.Conclusions This study confirmed that robot-assisted gait training using virtual reality is an effective intervention for improving balance and gait in stroke rehabilitation. However, findings should be interpreted with caution due to heterogeneity and the limited number of included studies.PROSPERO registration CRD420251138038