Immersive virtual reality upper-limb exercises in people with Parkinson’s disease: an observer-blinded randomised controlled trial

Upper-limb dysfunction is common in Parkinson’s disease (PD). Immersive virtual reality (iVR) offers an engaging rehabilitation approach, but its effectiveness versus conventional therapy remains unclear. This bi-centre observer-blinded randomised controlled trial compared iVR- and conventional upper-limb training (CON-ULT) on motor and non-motor outcomes in people with PD. Participants were randomised (1:1) to iVR upper-limb training (iVR-ULT) or CON-ULT for 30 min, 4 times/week, over 4 weeks of inpatient rehabilitation. Primary outcomes were arm function (Action Research Arm Test) and hand function (Jebsen-Taylor Hand Function Test) analysed by intention-to-treat. Secondary outcomes included dexterity, finger coordination, disease severity, motivation, health-related quality of life (HRQoL), adherence, acceptability, and iVR system usability (System Usability Scale, SUS). Outcomes were assessed pre- and post-intervention by blinded therapists. Of 97 patients screened, 58 were randomised: 29 to iVR-ULT and 29 to CON-ULT. No group-by-time interaction effects were observed in arm function. Both groups showed improved arm function (iVR-ULT: right, effect size r = 0.402, p = 0.034; left, r = 0.437, p = 0.021; CON-ULT: right, r = 0.640, p = 0.001; left, r = 0.567, p = 0.002; medium-to-large effects). For hand function, significant changes were found only after CON-ULT (right: r = 0.656, p = 0.0004; left, r = 0.628, p = 0.0007), reflecting a small between-group effect (r = 0.275, p = 0.038). For secondary outcomes, no group-by-time interaction effects were found. Across time, both groups improved in gross and fine dexterity (medium-to-large effects). Right-hand motor speed and coordination (Finger Tapping Test) improved significantly in only the iVR-ULT group (r = 0.516, p = 0.0064), while left-hand performance improved in both groups. Disease severity improved modestly in both groups (iVR-ULT: r = 0.431, p = 0.023; CON-ULT: r = 0.387, p = 0.037; medium effects), while Hoehn & Yahr stage remained stable. HRQoL analyses suggested changes in activities of daily living-related (r = 0.494, p = 0.009) and cognition-related domains (r = 0.391, p = 0.039) after iVR-ULT, stigma after CON-ULT (r = 0.386, p = 0.038) and emotional well-being after both interventions (r = 0.447, p = 0.018; r = 0.406, p = 0.029), with low-to-medium effect sizes. Adherence was high (median 16/16 sessions), acceptability ratings favourable (median 1–2/5), and device usability rated as good (median SUS 72.5). Both interventions produced within-group improvements across several outcomes, though between-group differences were limited. iVR-ULT appears feasible, acceptable, and potentially beneficial for motor function, warranting further evaluation in larger trials. German Clinical Trials Register: DRKS00034193. Date of registration: 16.05.2024.