Abstract
OBJECTIVE: There are several meta-analyses of randomized controlled trials (RCTs) demonstrating the benefits of virtual reality (VR) training as an intervention for motor performance, activity of daily living (ADL) and quality of life (QoL) outcomes in patients with Parkinson's disease (PD). However, the aggregate evidence collected to date has not been thoroughly evaluated for strength, quality, and reproducibility. An umbrella review from published meta-analyses of RCTs was conducted to evaluate the strength and quality of existing evidence regarding the efficacy of VR training in improving the motor performance, ADL and QoL outcomes of patients with PD. METHODS: PubMed, PsychInfo, Web of Science, and Scopus were searched to identify relevant meta-analysis of RCTs examining the effects of VR training on motor performance and quality of life outcomes in PD patients. We recalculated the effect sizes (Hedges'g) for VR training using DerSimonian and Laird (DL) random effects models. We further assessed between-study heterogeneity, prediction interval (PI), publication bias, small-size studies, and whether the results of the observed positive studies were better than would be expected by chance. Based on these calculations, the quality of evidence for each outcome was assessed by using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) criteria. RESULTS: Four meta-analysis with eight outcomes included in the umbrella review was recalculated effect size. Pooled results found VR training can large improve the basic balance ability, moderate improve the overall balance capacity and moderate improve the stride length in PD patients. For ADL and QoL, the effect sizes were pooled that suggested VR training can moderate improve ADL and QoL for PD patients. However, no statistically clear evidence was found in walking speed, motor function and gait function during VR training. The analyzed meta-analyses showed low-to-moderate methodological quality (AMSTAR2) as well as presented evidence of moderate-to-very low quality (GRADE). Tow adverse reactions were reported in the included meta-analyses. CONCLUSIONS: In this umbrella review, a beneficial correlation between VR and balance ability, stride length, ADL and QoL in PD patients was discovered, especially for the very positive effect of VR on balance because of two of the eight outcomes related to balance ability showed large effect size. The observations were accompanied by moderate- to very low-quality rating evidence, supporting VR training as a practical approach to rehabilitation.