OBJECTIVE:To evaluate feasibility of biofeedback therapy in China Medical Institutions to improve dysfunction following stroke. DATA SOURCE:A computer-based online search of publications was conducted using the Vip and PubMed Databases to identify publications that addressed biofeedback. The search key words included "electromyogram", "biofeedback", and "stroke". In total, 81 articles were retrieved. DATA SELECTION:Studies closely related to biofeedback, or studies with contents recently published in the same study field or in authorized journals, were included. Duplicated articles were excluded. Following full-text retrieval of selected articles, a total of 14 articles were collected, which addressed randomized, controlled trials of biofeedback therapy for dysfunction after stroke. Methodological quality was assessed for randomized, controlled trials using criteria from Cochrane reviewers handbook. Results were analyzed using Revman 4.2 software. MAIN OUTCOME MEASURES:Outcomes and evaluation indices were expressed by odds ratio (OR), weighted mean difference (WMD), and 95% confidence interval (95% CI). Potential publication bias was presented using the funnel plot. RESULTS:The study included 14 randomized, controlled trials of 1 147 patients. Following biofeedback therapy, meta-analysis results demonstrated that:(1) The total effective rate was significantly greater in the biofeedback therapy group compared with the control group [OR = 3.46, 95% CI (2.09, 5.73), P = 0.62]. (2) Electromyogram changes were better in biofeedback therapy patients compared to the control group [WMD = 22.31, 95% CI (17.19, 27.43), P < 0.001]. (3) Motor function was better in biofeedback therapy patients compared with the control group [WMD = 12.43, 95% CI (6.71, 18.16), P < 0.001]. (4) Daily living activities were better in biofeedback therapy patients compared with the control group [WMD = 18.11, 95% CI (15.77, 20.44), P = 0.36]. (5) Joint range of motion was better in biofeedback therapy patients compared with the control group [WMD = 6.43, 95% CI (4.44, 8.41), P = 0.77]. Sensitivity analysis also demonstrated similar results after eliminating articles that described unknown diagnostic criteria and statistical methods. CONCLUSION:Following stroke, biofeedback therapy for dysfunction was shown to result in significant and valid outcomes, increased motor function and electromyogram values, improved joint range of motion, and improved daily living activities.