Robot-assisted gait training for balance, mobility, and gait in Parkinson’s disease: a systematic review and meta-analysis

Abstract

Objective This systematic review and meta-analysis evaluated the effects of robot-assisted gait training (RAGT) on balance, functional mobility, walking endurance, and spatiotemporal gait parameters in patients with Parkinson’s disease. Methods PubMed, Embase, Web of Science, the Cochrane Library, CNKI, trial-registry sources, and supplementary search sources were searched from inception to the final search date. The protocol was retrospectively registered on the Open Science Framework (OSF; https://osf.io/njp39/) during manuscript revision. Randomized controlled trials comparing RAGT with conventional rehabilitation, treadmill training, balance training, medication, or usual care were included. In total, two reviewers independently screened records, extracted data, and assessed risk of bias. Hedges’ g and, in parallel, post-intervention mean differences (MDs) in the original clinical units were calculated using random-effects models. Subgroup, sensitivity, publication-bias, and GRADE analyses were conducted where data permitted; meta-regression analyses were treated as exploratory because each outcome included fewer than 10 studies and are reported only in the Supplementary Appendix. Results A total of 12 RCTs were included. RAGT was associated with better BBS scores (Hedges’ g = 1.08, 95% CI: 0.55 to 1.62; MD = 3.49 points, 95% CI: 1.85 to 5.12), shorter TUG time (Hedges’ g = −0.47, 95% CI: −0.77 to −0.16; MD = −1.82 s, 95% CI: −2.67 to −0.96), and a longer 6MWD (Hedges’ g = 2.21, 95% CI: 0.53 to 3.90; MD = 35.82 m, 95% CI: −0.94 to 72.58). Cadence increased by 4.14 steps/min (95% CI: 1.83 to 6.46), and step length increased by 5.70 cm (95% CI: 0.25 to 11.15). However, BBS, 6MWD, and step length showed substantial heterogeneity, and Egger’s tests suggested possible small-study effects or publication bias for several outcomes. Conclusion RAGT may be a promising adjunct to rehabilitation for PD, but the magnitude of benefit is uncertain, and the true effects may be smaller than the pooled estimates. Current evidence should be interpreted cautiously because of small trials, heterogeneous devices and comparators, underreported safety outcomes, risk of bias, and possible publication bias. Systematic review registration https://osf.io/njp39/, identifier (njp39).