Comparative Effectiveness of Digital Exercise-Based Interventions for Improving Balance in Older Adults: Systematic Review and Network Meta-Analysis With GRADE Certainty of Evidence

Background Technology-based rehabilitation is increasingly used for balance training in older adults. A 2023 network meta-analysis (NMA) found only exergaming with motion capture superior to no treatment; however, the evidence has since expanded, and no GRADE (Grading of Recommendations Assessment, Development, and Evaluation) appraisal has been applied. Objective This study aimed to update the comparative effectiveness of 5 technology categories—exergaming with balance platforms (EBP), exergaming with motion capture (EMC), other serious games, interventions with wearables, and telerehabilitation (Tele)—versus conventional exercise (CE), and no treatment (NT), applying GRADE adapted for NMA. Methods A systematic search of 4 databases from inception through April 2025, with an identical supplementary update search to January 31, 2026, identified 60 randomized controlled trials (N=3469; mean age 73.1, SD 5.2; range 63.3-86.2 y). Fifty-seven trials (2758 analyzed participants) contributed to 4 prespecified networks (all balance, functional scales, Timed Up and Go [TUG] test, and gait speed). A random-effects NMA was fitted using multivariate meta-regression. Treatment hierarchies were summarized by the surface under the cumulative ranking curve (SUCRA) and the probability of being best; certainty of evidence was rated with GRADE adapted for NMA. Results In the all-balance network, Tele (standardized mean difference [SMD] 0.69, 95% CI 0.08‐1.29; P =.03), EBP (SMD 0.63, 95% CI 0.31‐0.95; P <.001), EMC (SMD 0.58, 95% CI 0.22‐0.95; P =.002), and CE (SMD 0.35, 95% CI 0.03‐0.68; P =.03) were superior to NT, with consistent findings in the TUG network (EBP 0.81, 95% CI 0.42‐1.20; Tele 0.82, 95% CI 0.16‐1.49; EMC 0.79, 95% CI 0.36‐1.21; all P <.05). Tele ranked first by SUCRA (76.3%), but its probability of being best was only approximately 41%, and the estimate rested on 5 trials (n=198). No technology-based intervention was superior to CE in any network. Certainty of evidence was moderate for EBP versus NT and EMC versus NT in the TUG network only; the 37 other comparisons were low or very low. Dropout did not differ between technology-based and control arms (odds ratio 1.10, 95% CI 0.87‐1.40; P =.41). Conclusions Tele, EBP, and EMC were each associated with improved balance versus NT, although none exceeded CE. The first-place SUCRA ranking of Tele is preliminary. Certainty was moderate only for EBP and EMC versus NT on the TUG test; the remaining evidence is too uncertain for firm clinical recommendations. As no included trial reported falls, these findings are not direct evidence of fall prevention. Larger trials with standardized comparator intensity and posttraining follow-up are required.