Interventions for preventing falls in people post-stroke: A meta-analysis of randomized controlled trials

Background: Falls are a serious challenge facing individuals post-stroke. In the past decades, various fall prevention interventions have been developed. It remains unknown if any of these interventions are effective in reducing falls in this population. Such a knowledge gap could impede the effort of preventing falls in people post-stroke. Research questions: 1) Are there effective interventions to prevent falls among people in the post-acute and chronic stages of stroke? and 2) How do fall prevention interventions change three key fall risk factors in this population: balance, mobility, and lower limb strength? Methods: Eleven databases were searched for randomized controlled trials which included falls in people poststroke as an outcome measure. Information on the participants, training protocol, and outcome measures were collected for each study. The primary outcome is the number of falters and the explanatory variables included mean difference and standard deviation for fall risk factors. Studies were quality appraised using the Physiotherapy Evidence Database scale and the funnel plot. Results: Thirteen studies enrolling 1352 participants were identified. Effect size quantified by the odds ratio (OR) for falls and standardized mean difference (SMD) for fall risk factors were calculated. Overall no intervention appears to be significantly more effective in preventing falls than placebo training (OR = 0.88 with a range of [0.23 3.66]; 95 % confidence interval = [0.64 1.21], p = 0.44). All interventions showed little effect in improving the fall risk factors (SMD = -0.01 to 0.06 and p-value = 0.38-0.86), except one (the combined treadmill and overground walking) which significantly improved mobility. Significance: Currently no program is effective in reducing falls in people post-stroke. Future studies should measure falls as a primary outcome based on a consistent definition of falls and reliable approaches to collect falls data.