Virtual reality (VR) is a frequently used intervention for the rehabilitation of individuals with neurological disorders. The aims of this review were to identify the short-term effect of VR on balance and to compare it with the effect of active interventions in individuals with Parkinson disease (PD). Searches for relevant articles available in English were conducted using the MEDLINE (via PubMed), EMBASE, CENTRAL, CINAHL, PsycINFO, and Physiotherapy Evidence Database databases from inception until March 2019. All randomized controlled trials comparing the effect of training with VR and the effect of training without VR on balance in individuals with PD were included. Two authors independently extracted data, assessed the methodological quality, and evaluated the evidence quality of the studies. Fourteen randomized controlled trials including 574 individuals were eligible for qualitative analyses, and 12 of the studies involving 481 individuals were identified as being eligible for meta-analyses. Compared with active interventions, the use of VR improved the Berg Balance Scale score (mean difference=1.23; 95% CI=0.15 to 2.31; I2=56%). The Dynamic Gait Index and Functional Gait Assessment results were also significant after the sensitivity analyses (mean difference=0.69; 95% CI=0.12 to 1.26; I2=0%). Both provided moderate statistical evidence. However, the Timed "Up & Go" Test and the Activities-Specific Balance Confidence Scale did not differ significantly. Publication bias and diversity in the interventions were the main limitations. Existing moderate evidence of the effectiveness of VR with the Berg Balance Scale, Dynamic Gait Index, and Functional Gait Assessment for individuals with PD was promising. Although the differences did not reach the clinically important change threshold, VR was comparable to active interventions and could be considered an adjuvant therapy for balance rehabilitation in individuals with PD.
Read full abstract