Comparative effectiveness of non-pharmacological interventions for post-stroke upper limb motor dysfunction: a systematic review and network meta-analysis of randomized controlled trials.

Background: Post-stroke upper limb motor dysfunction (PS-ULMD) is a common and disabling consequence of stroke. Although multiple non-pharmacological interventions are used to treat PS-ULMD, the optimal treatment remains unclear. Thus, this study aimed to identify effective non-pharmacological interventions for improving upper limb motor function in patients with PS-ULMD. Methods: Eight databases were searched from database inception to May 18, 2026. Randomized controlled trials (RCTs) evaluating 15 non-pharmacological interventions for PS-ULMD were included based on multiple guidelines. Risk of bias was assessed using the Cochrane Risk of Bias tool (RoB 2.0). Primary outcome was the improvement in the Fugl-Meyer Assessment for Upper Extremity (FMA-UE). Pairwise meta-analyses were conducted using Review Manager (RevMan, version 5.4). Network meta-analyses were performed using STATA (version 15.0) and ADDIS (version 1.16.8). The quality of evidence for outcomes was evaluated using the Confidence in Network Meta-Analysis (CINeMA) online tool. Results: A total of 89 RCTs, involving 5129 participants, were included. The RoB 2.0 tool indicated that the majority of included studies presented some concerns. Pairwise meta-analyses showed that, compared with conventional therapy, non-pharmacological interventions significantly improved FMA-UE scores [mean difference (MD) = 4.95, 95% confidence interval (CI): 3.74-6.17]. Compared with sham control, the overall effect remained statistically significant (MD = 5.12, 95% CI: 3.43-6.81). Network meta-analysis further indicated that BCI, CIMT, BF, and PES were associated with relatively larger improvements in FMA-UE scores compared with conventional therapy and sham control. However, according to the CINeMA framework, the overall certainty of evidence for most comparisons was rated as low to very low. Conclusion: Overall, non-pharmacological interventions appear to be effective in improving post-stroke upper limb function. Among them, BCI, CIMT, BF, and PES were associated with relatively greater improvements in upper extremity motor recovery. However, the certainty of the evidence remains limited, and further high-quality RCTs are warranted to confirm these findings and to better define the relative effectiveness of non-pharmacological interventions. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/, identifier CRD420251243062.