Comparison of Different Rehabilitation Techniques of Traditional Chinese and Western Medicine in the Treatment of Motor Dysfunction After Stroke Based on Frequency Method: A Network Meta-analysis.

Therapies are ranked using all the evidence in the network, so a therapy tested in many trials can rank highly without being significantly better than the alternatives; the authors state the ranking cannot settle which treatment is better and is only for clinical reference. Most of the included trials did not describe how patients were allocated to groups or whether anyone was blinded, which the authors list as a limitation of the evidence. Patients in the acute, subacute and chronic stages after stroke were pooled together and no analysis by stage was done, although the authors point out that different stages need different rehabilitation approaches. The authors say publication bias cannot be ruled out, and 12 studies sat outside the 95% confidence region on the funnel plot. Biofeedback, the top-ranked therapy for total and upper limb movement, was tested in only two trials that reported the total FMA score, and the authors say larger and higher-quality studies are needed. The paper's own account of how many trials were two-armed and how many were three-armed is unclear, and it reports differences in the number of cases before and after treatment in the included trials.

The authors state that financial disclosure statements were obtained and that no conflicts of interest were reported by them or by anyone in control of the article's content. The paper does not say who funded the work or whether any funder had a role in it.

Do not conclude from the rankings that biofeedback or virtual reality is the best treatment for your patient. Most comparisons between therapies in this network showed no significant difference, the order depends partly on how many trials happened to test each therapy, and the authors themselves say the ranking is only for clinical reference.