|Biofeedback treatment for headache disorders: a comprehensive efficacy review
|Nestoriuc Y, Martin A, Rief W, Andrasik F
Biofeedback for migraine was supported as an effective treatment and biofeedback for tension-type headache was effective and specific. The strong conclusions of this review should be treated with caution due to the poor reporting of the analyses.
To systematically review the efficacy of biofeedback for headache disorders: migraine and tension-type headaches (an update).
The following databases were searched for articles from inception to March 2008: MEDLINE, PsycINFO, Cochrane Central Register of Controlled Trials (CENTRAL), and PSYNDEX. Search terms were reported and two previous reviews (Nestoriuc, et al. 2007 and 2008, see Other Publications of Related Interest) were checked for eligible studies. Published papers in English or German were considered.
Controlled and uncontrolled studies were eligible provided that they reported on individually administered biofeedback or biofeedback in combination with other behavioural therapies for the treatment of migraine or tension-type headache (TTH). Studies of patients diagnosed with both migraine and TTH were excluded. Eligible biofeedback treatments were: peripheral skin temperature feedback, electromyographic feedback, blood-volume-pulse feedback, electroencephalographic feedback, and galvanic skin response training. The primary outcome was frequency of pain, measured with a structured headache diary. Other outcomes of interest were the intensity and duration of pain; headache or medication indices; anxiety; depression; self-efficacy; and physiological outcomes. Studies were required to report sufficient data to allow the calculation of effect sizes. Samples with less than four patients per treatment arm and case studies were excluded.
The included studies were either controlled studies, some of which were randomised, or pre-post single-arm studies. Control groups included no treatment, placebo using pseudo feedback or similar, and alternative relaxation treatments. Studies were largely conducted in the USA, but they also took place in European countries, Canada, and India. The mean patient age was similar for both migraine (37 years) and TTH (38 years), as was the gender split of 88% female in the migraine group and 73% female in the TTH group. The TTH studies included a few focusing on biofeedback for children and adolescents. In adult patients the average period since diagnosis was 17.1 years (migraine) and 14.8 years (TTH). A standard diagnostic system was used in 80% of migraine and 51% of TTH studies. The number of biofeedback sessions ranged from three to 24 with an average of 11 sessions for both migraine and TTH studies. Treatment manuals were described in 79% of migraine and 80% of TTH studies.
It was not clear how many reviewers performed the study selection.
Assessment of study quality
A comprehensive validity assessment was performed using a validated scale, which included 12 items grouped under internal validity, external validity, construct validity, and statistical conclusion validity. Validity assessment was performed by the first author and two independent reviewers (graduate students) after training on the coding system. A random 20% sample of the migraine and all the TTH studies were double-coded.
Effect sizes were calculated using Hedges' g for controlled trials and the pre-post equivalent for uncontrolled studies, for each outcome variable, treatment group, and time point. A correction for small sample sizes was applied. Multiple effect sizes from single studies were averaged with covariance adjustment. Data extraction was performed by the first author and two independent reviewers (graduate students) after training on the extraction system. A random 20% sample of the migraine and all the TTH papers were double-coded.
Methods of synthesis
Random-effects and fixed-effect meta-analyses were used to pool the studies and calculate weighted mean effect sizes (ESs), with 95% confidence intervals (CIs), for treatment comparisons, feedback modalities, outcomes, and time points. Outcome variables were integrated into three symptom categories of headache pain (frequency, intensity, and duration), and physiological and psychological or behavioural variables. For analysis of general efficacy, the variables headache frequency, duration, and intensity were integrated. Pre-post effect sizes were pooled from all studies, and used to assess the effect size after an average 14-month follow-up, and the efficacy of different biofeedback modalities. The moderating effects of patient, treatment, and study characteristics were tested with planned contrasts and weighted multiple regression analysis. The Cochran Q statistic was used to assess statistical heterogeneity and the fail-safe N was calculated for publication bias.
Results of the review
A total of 94 studies, with over 3,500 patients, were included; 56 studies on migraine (mean 40 patients per study) and 45 studies on TTH (mean 29 patients per study), nine of these were excluded from meta-analysis as they were on children. In seven studies both migraine and TTH patients were treated. Controls were used in 78% of migraine and 58% of TTH studies, the others were uncontrolled pre-post test studies. Within the controlled studies, 26% of migraine and 22% of TTH used either single or double-blinding.
Migraine (40 controlled studies): biofeedback was shown to have a significant small-to-medium effect size in comparison with waiting-list control for headache reduction (ES 0.46, 95% CI 0.27 to 0.64; 15 studies), a non-significant small-to-medium effect size compared with placebo control (ES 0.25, 95% CI 0.00 to 0.49; 12 studies), and a small non-significant effect size compared with relaxation (ES 0.10, 95% CI -0.39 to 0.50; five studies).
Tension-type headache (28 controlled studies): biofeedback was shown to have a significant medium-to-large effect in comparison with waiting list control for headache reduction (ES 0.79, 95% CI 0.40 to 1.17; eight studies), a significant medium effect size compared with placebo control (ES 0.50, 95% CI 0.26 to 0.75; eight studies), and a significant small effect compared with relaxation (ES 0.18, 95% CI 0.06 to 0.30; 14 studies).
Results over an average 14-month follow-up (pre-treatment versus follow-up for all studies) indicated a significant medium-to-large effect size in favour of biofeedback treatment for both migraine and TTH patients. The highest treatment gains were shown for blood-volume-pulse feedback and differences between the modalities were not significant. Further results were reported. The fail-safe N calculations indicated that publication bias was unlikely to be present.
Biofeedback for migraine was supported as an effective treatment and biofeedback for TTH was effective and specific.
This review addressed a clear question with relatively specific inclusion criteria. A minimum follow-up period of six months for migraine and three months for TTH patients was specified in the two original reviews, but it was unclear if this was maintained in this update. The searches did not incorporate unpublished literature and language restrictions were applied, but no evidence of publication bias was reported. The processes were not clearly reported for study selection, but quality assessment and data extraction appear to have been performed rigorously. A validated quality assessment tool was used, but the results were not reported, which makes it difficult to assess the reliability of the primary studies. The analysis was not clearly described nor justified, which is problematic given the large number of heterogeneous studies. Statistical and clinical heterogeneity were not clearly reported nor explored and sensitivity analyses were not reported. Some of the conclusions were drawn from a combination of controlled and uncontrolled study results without addressing issues of validity or reliability.
The strong conclusions of this review should be treated with caution due to the poor reporting of the analyses.
Implications of the review for practice and research
Practice: The authors stated that for the treatment of TTH, the combination of biofeedback with relaxation training, and biofeedback alone could be recommended, especially for juvenile headache patients.
Research: The authors did not state any recommendations for research.
Nestoriuc Y, Martin A, Rief W, Andrasik F. Biofeedback treatment for headache disorders: a comprehensive efficacy review. Applied Psychophysiology and Biofeedback 2008; 33(3): 125-140
Other publications of related interest
Nestoriuc Y, Rief W, Martin A. Meta-analysis of biofeedback for tension-type headache: efficacy, specificity, and treatment moderators. Journal of Consulting and Clinical Psychology 2008; 76(3): 379-396.
Nestoriuc Y, Martin A. Efficacy of biofeedback for migraine: a meta-analysis. Pain 2007; 128: 111-127.
Subject indexing assigned by NLM
Biofeedback, Psychology; Electromyography; Evidence-Based Medicine; Humans; Migraine Disorders /therapy; Randomized Controlled Trials as Topic; Tension-Type Headache /therapy; Treatment Outcome
Database entry date
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.