Erratum: Effect of Dual-Task Training on Gait and Balance in Stroke Patients: An Updated Meta-analysis (American Journal of Physical Medicine & Rehabilitation December (2022) 101 (1148-1155) DOI: 10.1097/PHM.0000000000002016)

In the December 2022 issue of American Journal of Physical Medicine & Rehabilitation, in the article by Shu et al., "Effect of Dual-Task Training on Gait and Balance in Stroke Patients: An Updated Meta-analysis",1 there was an error in the statistical analysis that warranted correction. The authors wish to issue the following correction. In one of the articles included in the meta-analysis (Fishbein 2019), we made an error in the process of data extraction and mistakenly included the baseline data of the article. After re-checking and extracting the article data, we found that the article did not have specific variable values after intervention. The results presented in the article figure (Fishbein 2019) only indicate the differences obtained in actual walking variables and in walking confidence, however, the specific variable values after the intervention was missing. We wanted to get the specific variable values after the intervention in the article (Fishbein 2019) from the original author, but we did not receive a reply from the original author, so we eliminated the article (Fishbein 2019). The outcome indicators 10MWT and BBS in our meta-analysis both included this article. We deleted the article and re-analyzed the results, revised Fig. 3D and Fig. 4A, and the revised Fig. 3D and Fig. 4A are attached. Two studies13,20 took the 10MWT as the outcome indicator. After heterogeneity tests, I 2 =27% > 50%, = 0.24 > 0.1, indicated that there was no heterogeneity among the included studies. A fixed-effects model was used for merging effect size, and the results showed that the effect of the dual-task training group on the 10MWT in stroke patients was statistically significant compared with the control group (MD = 2.29, 95% CI = 3.64 to 0.94, = 0.0009; Fig. 3D). The error did not affect the major conclusions reported in the paper. Six studies11,13,19,23,24,26 took the BBS as the outcome indicator. After heterogeneity tests, I 2 = 88% 50%.