DARE: Search Results

Document: 4
Electromyographic biofeedback for neuromuscular reeducation in the hemiplegic stroke patient - A meta-analysis

Schleenbaker R E, Mainous A G. Electromyographic biofeedback for neuromuscular reeducation in the hemiplegic stroke patient - A meta-analysis Archives of Physical Medicine & Rehabilitation, 1993, 74(12), pp. 1301-1304

Record status

This record is a structured abstract written by CRD reviewers. The review has met a set of quality criteria.

Author's objective

To determine whether electromyographic biofeedback improves function after stroke.

Type of intervention

Rehabilitation.

Specific interventions included in the review

Rehabilitation with electromyographic biofeedback, either on its own or in conjunction with physical therapy or gait training.

Participants included in the review

Subjects sustaining a cerebrovascular accident that resulted in hemiplegia. Time from cerebrovascular accident to rehabilitation varied from acute, 2-8 weeks, 2-3 months, more than 3 months, more than 6 months, to more than one year.

Outcomes assessed in the review

Functional outcome, defined as any measured outcome parameter that requires complex neuromuscular activity necessary for executing activities of daily living or ambulation. Isolated muscle strength, range of motion, or muscle electromyographic activity were excluded. Specific measures in the included studies: Upper Extremity Functional Test, time to trace circle with olecranon, improvement in Brunnstrom's stages of recovery, Action Research Armtest, Gait analysis grading system, change in need for ambulation aids and gait cycle time.

Study designs of evaluations included in the review

Original research studies with randomised or matched control group, which reported results such that an effect size statistic could be calculated.

What sources were searched to identify primary studies?

Searches for English language clinical studies of biofeedback and stroke or cerebral vascular disease between 1966 and 1991 were carried out using MEDLINE, PsycINFO, REHABDATA, and Dissertation Abstracts International databases. Bibliographies of retrieved articles and relevant textbooks were also searched. The search strategy is not given.

Criteria on which the validity (or quality) of studies was assessed

Not stated.

How were the judgements of validity (or quality) made?

Not stated.

How were decisions on the relevence of primary studies made?

Not stated.

How was the data extracted from primary studies?

A 'd' effect size was calculated for each study. This expresses the size of the effect in proportion to its standard error. The method for data extraction is not stated.

Number of studies included

8 studies, with a total sample size of 192 cases. It is not clear which studies were randomised and which were matched.

How were the studies combined?

Studies were combined using meta-analysis, although the specific method used is not clearly stated. The individual effect sizes were weighted by sample size. Analysis to estimate the number of unpublished studies required to make the meta-analysis results non- significant was also undertaken.

How were differences between studies investigated?

Chi-squared test for homogeneity of variance. Sub-group analysis was undertaken for the differing effects of electromyographic biofeedback on functional outcomes in upper and lower extremities.

Results of the review

The average effect size is 0.81 (95% CI: 0.50,1.12). The chi-squared test for homogeneity gave a non-significant value of 2.31 (p>0.05). The number of unpublished studies with null results required to bring the significance of the meta-analysis down to 5% is 56. For upper extremities studies the mean effect size is 0.77, and 0.89 for lower extremities studies, but no significance or confidence intervals are given.

Was any cost information reported?

No.

Author's conclusions

Electromyographic biofeedback appears to be a useful therapy for hemiplegic stroke patients and should be included in the therapeutic regimen.

CRD commentary

Not enough detail is given about the methods used to ensure the validity and relevance of the primary studies. No results of the individual studies are presented. The presentation of the statistical methods is complicated and unclear, making it difficult to assure the validity of the analysis. The interpretation of 'd' statistics in terms of a clinical effect is not possible from the data that are presented. No attention is given to the differing methodological rigour of the included studies, and little detail is given concerning the rigour of the review process. There is therefore concern about the strength of the authors' conclusion. No description of patient characteristics is given. No mention of follow-up is made.

What are the implications of the review?

Further research to determine the effect of electromyographic biofeedback in stroke rehabilitation is needed before any recommendations can be made. This treatment should be considered but a large multicentre randomised controlled trial assessing appropriate outcomes may provide more enlightenment.

Subject index terms

Subject indexing assigned by NLM: Biofeedback-Psychology/mt [Methods]; Cerebrovascular-Disorders/co [Complications]; Electromyography-; Hemiplegia/rh [Rehabilitation]; Activities-of-Daily-Living; Case-Control-Studies; Clinical-Protocols; Confidence-Intervals; Effect-Modifiers-Epidemiology; Hemiplegia/et [Etiology]; Hemiplegia/pp [Physiopathology]; Matched-Pair-Analysis; Randomized-Controlled-Trials; Research-Design; Treatment-Outcome; Human

Correspondence address

Randal E. Schleenbaker, MD, Department of Rehabilitation Medicine, Kentucky Clinic, University of Kentucky Medical Center, Lexington, KY 40536-0284, USA

Copyright

University of York, 1998
Database no.: DARE-940172