Dropout rate in randomised controlled trials of balance and gait rehabilitation in multiple sclerosis: is it expected to be different for virtual reality-based interventions? A systematic review with meta-analysis and meta-regression.

Abstract

To assess and meta-analyse the pooled dropout rate from the randomised control trilas that use virtual reality for balance or gait rehabilitation in people with multiple sclerosis. A systematic review of randomised control trials with meta-analysis and meta-regressions was performed. A search was conducted in PubMed, Scopus, Web of Science, the Physiotherapy Evidence Database, the Cochrane Database, CINHAL, LILACS, ScienceDirect, and ProQuest. It was last updated in July 2022. After the selection of studies, a quality appraisal was carried out using the PEDro Scale and the Revised Cochrane risk-of-bias tool for randomised trials. A descriptive analysis of main characteristics and dropout information was performed. An overall proportion meta-analysis calculated the pooled dropout rate. Odds ratio meta-analysis compared the dropout likelihood between interventions. The meta-regression evaluated the influence of moderators related to dropout. Sixteen studies with 656 participants were included. The overall pooled dropout rate was 6.6% and 5.7% for virtual reality and 9.7% in control groups. The odds ratio (0.89, p = 0.46) indicated no differences in the probability of dropouts between the interventions. The number, duration, frequency, and weeks of sessions, intervention, sex, multiple sclerosis phenotype, Expanded Disability Status Scale score, and PEDro score were not moderators (p > 0.05). Adverse events were not reported and could not be analysed as moderators. Dropouts across the virtual reality and control comparators were similar without significant differences. Nonetheless, there is a slight trend that could favour virtual reality. Standardisation in reporting dropouts and adverse events is recommended for future trials.