Effects of rehabilitation exercise program types on dynamic balance in patients with stroke: a meta-analysis of randomized controlled trials

The purposes of meta-analysis are to evaluate evidence about the effects of Rehabilitation Exercise Program on the balance of post-stroke patients, evaluated by the Berg Balance Scale (BBS). The search was conducted ‘stroke,’ ‘rehabilitation,’ ‘dynamic balance,’ ‘Berg Balance Scale,’ ‘exercise’ and ‘randomized controlled trial’using MEDLINE (accessed by PubMed), Web of Science (WoS), ProQuest, and Google Scholar for journal studies published from January 2018 to October 2022. Two independent reviewers performed the article selection, data extraction, and methodological quality assessment. The main outcome was dynamic balance assessed by the Berg Balance scale. The review included 30 papers and a total of 540 patients. The overall effect size was 0.550, a medium effect size according to the Cohen’s standard. It was observed that gender has moderate effect size in male (0.551), female (0.458) and higher in male. Exercise type results showed large effect sizes in balance training (0.966), and aquatic activities (0.830), moderate effect sizes in virtual reality (0.762), moderate effect sizes in physically active (0.581), gait training (0.541), dual-task (0.478), trunk control (0.284), and small effect sizes in resistance training (0.128). Exercise programs are effective in improving dynamic balance in stroke patients. Especially, the meta-analysis showed higher Effect Size for balance training and virtual reality than for other programs making this relevant interventions for future head to head superiority studies that compare different balance interventions in stroke.

本元分析（meta-analysis）旨在通过伯格平衡量表（Berg Balance Scale, BBS）评估康复运动方案（Rehabilitation Exercise Program）对脑卒中患者平衡能力的影响效应。本研究以通过PubMed访问的MEDLINE数据库、Web of Science（WoS）、ProQuest及Google Scholar为检索平台，以“stroke”“rehabilitation”“动态平衡（dynamic balance）”“Berg Balance Scale”“exercise”及“随机对照试验（randomized controlled trial）”为检索关键词，纳入2018年1月至2022年10月发表的期刊类研究。由2名独立评审员分别完成文献筛选、数据提取及方法学质量评估。本研究的主要结局指标为经伯格平衡量表评估的动态平衡能力。本次综述共纳入30篇文献，合计540例脑卒中患者。合并效应量为0.550，依据科恩效应量判定标准（Cohen's standard）属于中等效应量。分析结果显示，性别维度的效应量表现为中等水平：男性患者效应量为0.551，女性患者为0.458，男性患者的效应量更高。运动类型亚组分析结果显示：平衡训练的效应量为0.966（大效应量），水上运动为0.830（大效应量）；虚拟现实（virtual reality）训练为0.762、体力活动干预为0.581、步态训练为0.541、双任务训练为0.478及躯干控制训练为0.284，上述干预的效应量均为中等效应量；抗阻训练的效应量为0.128，属于小效应量。研究表明，康复运动方案可有效改善脑卒中患者的动态平衡能力。尤其本元分析结果显示，平衡训练与虚拟现实训练的效应量高于其他运动方案，因此二者可作为未来开展脑卒中患者不同平衡干预措施头对头优效性比较研究的重点干预手段。