To determine the effects of simultaneous dual-task training on cognitive function, physical function, and depression in older adults with mild cognitive impairment or dementia.

Comprehensive database searches were conducted in PubMed, Embase, the Cochrane Library, CINAHL, Ovid-Medline, Web of Science, and Scopus up to December 2022. Randomized controlled trials were included to assess the efficacy of simultaneous dual-task training for older adults with mild cognitive impairment or dementia. The analysis utilized Comprehensive Meta-Analysis version 3.0, presenting Hedges' g and the corresponding 95 % confidence interval (CI) for the pooled effect size and, applying a random-effects model. The I2 and Cochran's Q tests were employed to evaluate heterogeneity. The Cochrane Risk of Bias 2.0 tool was employed to assess study quality. The Copenhagen Trial Unit (version 0.9.5.10 Beta) was employed for trial sequential analysis, providing a rigorous methodology for evaluating cumulative evidence from multiple studies.

Of the 1676 studies identified, 20 studies involving 1477 older adults with cognitive impairment were included. Dual-task training significantly enhanced global cognition (0.477, 95 % CI: 0.282 to 0.671), executive function (− 0.310, 95 % CI: − 0.586 to − 0.035), working memory (0.714, 95 % CI: 0.072 to 1.355), gait (0.418, 95 % CI: 0.252 to 0.583), physical activity (0.586, 95 % CI: 0.012 to 1.16), and depression (− 0.703, 95 % CI: − 1.253 to − 0.153). Trial sequential analyses revealed the robustness of this meta-analysis, which was based on a sufficient sample size from the included studies. Moreover, dual-task training demonstrated beneficial effects on global cognition, executive function, working memory, and gait.

Dual-task training improved cognition, physical function, and depression among older adults with cognitive impairment. Accordingly, dual-task training should be considered a clinical nonpharmacological intervention for older adults with mild cognitive impairment or dementia. Nevertheless, the trial sequential analysis results were consistent with those of the pairwise meta-analysis but only global cognition reached significance by crossing the trial sequential analysis boundary. Future studies with higher-quality designs and larger sample sizes are required to obtain more conclusive results regarding other outcomes.

PROSPERO CRD42023418598.