Díaz-Arribas et al (2019) systematic review on the effectiveness of Bobath therapy – Part 2: Interpretation and Critical Appraisal

//Díaz-Arribas et al (2019) systematic review on the effectiveness of Bobath therapy – Part 2: Interpretation and Critical Appraisal

Díaz-Arribas et al (2019) systematic review on the effectiveness of Bobath therapy – Part 2: Interpretation and Critical Appraisal

By |2019-06-11T20:43:22+00:00July 28th, 2019|Journal reviews|Comments Off on Díaz-Arribas et al (2019) systematic review on the effectiveness of Bobath therapy – Part 2: Interpretation and Critical Appraisal

Interpretation and Critical Appraisal

Methods: Search strategies and inclusions / exclusions: Searching four different databases and reference lists from prior Bobath reviews is a strength of this review. Searching at least two databases helps to ensure relevant trials are not missed. Another strength is the inclusion of non-English publications (ie German, French and Spanish). Many reviews exclude non-English studies because of the cost of translation.

Excluding non-English studies published in other languages is a limitation of this review. For example, one Lithuanian trial that was excluded had a large sample of 240 participants. Chinese language publications were also excluded. Multiple trials have been published in Chinese in the past 10 years, comparing Bobath with other treatments. The increasing number of trials conducted in China reflects the growing number of allied health professionals and researchers educated there, and the number of Bobath workshops conducted in the Asia Pacific region.  Exclusion of Chinese language studies represents a selection bias because important foreign language studies may be missed – perhaps as many as 15 or 20 studies.

Excluding trials that used Bobath combined with another treatment is a strength (eg task-specific training, forced use, constraint therapy). When evaluating a therapy it is important to define, publish and follow a specific therapy protocol (ie in this case, for Bobath therapy). Including other treatments with Bobath that might be a comparator/comparison treatment can confound the outcomes of a review. An equivalent in a systematic review of medical trials would be including a surgical treatment with a drug treatment, contaminating the treatment of interest.

Rehabilitation has always been a ‘black box” of multiple interventions. Over the past 20 years researchers have carefully started to unpack this ‘black box’ and define treatments such as task-specific training, constraint-induced movement therapy, mental practice and electrical stimulation to name a few.  In rehabilitation research we have progressed to naming, defining and describing the protocols for therapies so that their effectiveness can be investigated in trials.  It is important for researchers who evaluate the effects of Bobath therapy to describe their protocol and maintain fidelity to that protocol during a study.  Combining other therapies such as constraint-induced movement therapy or task-specific training into the intervention protocol complicates the intervention, which is no longer Bobath but something else.

Methodological quality of the 15 trials: Trials were of a moderate to high methodological quality, meeting between 4 and 8 of the PEDro criteria.  Most studies had good methodological quality (ie 12/15 met 6-8 PEDro criteria) and three studies had moderate methodological quality (ie 3/15 met 4-5 PEDro criteria).  No studies were rated as having low methodological quality.

Random allocation was not concealed in 12/15 trials, representing a significant methodological flaw. Researchers could have influenced allocation of individual patients to control or experimental groups. The person doing the allocation may have been aware that a consenting patient would be allocated to the control or experimental group when making their decision; they may have consciously or unconsciously influenced the group to which that participant was allocated, either by changing the order in which participants were enrolled, or the order in which treatments were provided. This process can produce systematic biases in an otherwise random allocation. Thus 12/15 trials could have systematic biases due to lack of concealed randomisation.

Assessors were not blinded to the group allocation in 4 trials, introducing a measurement bias. ‘Blinding’ means the person in question (participant, therapist or assessor) did not know to which group a patient participant had been allocated – an essential process in high quality randomised trials to minimise potential bias.  Assessor blinding ensures that effects of the intervention were not due to assessor biases (eg enthusiasm or lack of enthusiasm for a particular treatment when conducting a test such as the 6-minute walk test, a balance or upper limb test).

Grouping studies together according to outcome measures also influenced review outcomes in a way that is questionable. For example, how were studies separated that used “mobility’ measures (muscle tone, Modified Ashworth Scale, FIM, Motor Assessment Scale, Rivermead Motor Assessment) versus measures of ‘motor control of lower limb AND gait” (muscle tone, motor control, Fugl-Meyer Assessment, FIM, gait parameters, Motor Assessment Scale, Rivermead Motor Assessment). Measures of ‘gait’ could be considered a separate category but along with measures of ‘mobility’. Decisions about categorisation of outcome measures remain unclear to readers, and would be difficult to replicate.

Grouping study results together based on whether Bobath therapy was/was not effective against a comparison therapy is one of the most important weaknesses of this review. No between-group data were presented (in Table 3), consequently we do not know, cannot interpret or calculate the clinically important differences between intervention groups. Readers are only presented with summary statements about statistical significance (whether between- group differences were due to chance or not). For example, a difference of one point (between Bobath/another therapy approach) on the 66-point Fugl Meyer Scale or the 126-point FIM scale is NOT clinically important; a more thorough analysis of the data is needed. Categorising the effects of a treatment as being different (Y/N) or ‘better/worse’ than a comparator based on statistical significance alone perpetuates systematic biases in the original trials.

In conclusion, this review adds to knowledge by presenting more recent trials involving Bobath therapy, but should not change clinical practice.  Grouping comparisons together from studies with/without statistically significant differences is a major limitation or weakness of this review. Studies with more robust methods (ie higher PEDro score) should be allocated greater weighting in future reviews.  Furthermore, non-English studies, particularly Chinese studies, should ideally be translated and their results summarised in future reviews.  Researchers conducting randomised trials now have to publish their methods, including intervention protocol and outcome measures, in advance and in detail, before trial results can be published in a respected journal. Therefore, triallists should – in theory – be describing their version of Bobath therapy in sufficient detail to allow replication, enabling reviewers to determine what was done and to extract, report and compare raw data and outcomes.

Synopsis, interpretation and critical appraisal by Annie McCluskey

The StrokeEd Collaboration
E-mail: dranniemccluskey@gmail.com

  • Díaz-Arribas, M.J., Martín-Casas, P., Cano-de-la-Cuerda, R., & Plaza-Manzano, G. (2019 online early) Effectiveness of the Bobath concept in the treatment of stroke: A systematic review, Disability and Rehabilitation, DOI: 10.1080/09638288.2019.1590865