Factors Influencing Walking Improvement After Immersive Virtual Reality Training with a Bodyweight-Supporting System in Poststroke Patients: A Prospective Single-Arm Study.

Objectives: We developed an original protocol for standing balance and walking function training using immersive virtual reality (IVR) combined with a bodyweight-supporting (BWS) system. The purpose of this study was to evaluate the feasibility of IVR training in poststroke patients and to identify clinical characteristics associated with walking improvement. Methods: A total of 22 poststroke patients admitted to a rehabilitation ward were enrolled. All participants received conventional physical therapy daily, and IVR training was initiated when balance recovery plateaued. The original IVR training was conducted for 10 consecutive days using a BWS system, enabling safe and repetitive practice of standing weight shifting and advanced stepping tasks, with task difficulty individually adjusted. Primary outcomes were the Berg Balance Scale (BBS), Functional Reaching Test (FRT), and walking time on the 10-m walk test (10 MWT). Changes in 10 MWT (Δ10 MWT) were calculated, and participants were dichotomized into good and poor walking improvement groups based on Δ10 MWT to compare baseline clinical characteristics (e.g., BBS, FRT, and Functional Ambulation Categories [FAC]) between the groups. Receiver operating characteristic (ROC) curve analyses were conducted to identify cutoff values associated with greater improvements in walking. Results: BBS, FRT, and 10 MWT improved significantly after the 10-day IVR training (p < 0.01). ROC analyses identified FAC ≤ 3 and BBS ≤ 42 before training as cutoff values associated with greater improvement in walking performance. Conclusions: Our proposed IVR training approach using a BWS system could be safely implemented and might improve standing balance and walking function in poststroke patients with a balance disability.