Statistics from Altmetric.com
Practitioners of rehabilitation are often criticised for failing to evaluate their clinical work critically and assembling sound research evidence upon which to base it. The paper by Powell et al (this issue, pp 193–202)1 describes the results of an extremely ambitious and almost unique interventional study of community based rehabilitation in around 100 patients after moderate or severe traumatic brain injury, following the gold standard methodology of a randomised controlled trial.
At entry to the study, patients were allocated to either active multidisciplinary community rehabilitation, or a single therapist visit with the provision of written information about available resources, effectively a control group. Overall, despite the marked heterogeneity of the participants, the two comparison groups were well matched. The results showed significant improvements after active intervention for several hours a week for an average of 6 months, when followed up by a blinded assessor some 2 years later.
Considering the amount of work involved, both in providing the therapeutic input and gathering the outcome data, the therapeutic effect size may at first sight seem nevertheless disappointing. Randomised controlled trials are all very well in certain settings, but may not always be necessary or appropriate in evaluating rehabilitation, where benefits may be modest and multifactorial.2 One of the two primary outcome measures in the study was the original Barthel index. Although widely used in clinical practice as a rapid and easy measure of activities of daily living, it is notorious for its ceiling effect and has since been modified in an attempt to improve it. Both groups had a median score of 100%, with negative as well as positive changes at follow up. The improvement of one or two points at the upper limit of the index is of rather doubtful functional, as opposed to statistical, significance.
Fortunately, there was also significant improvement measured by the second primary outcome measure used by the authors, the brain injury community rehabilitation outcome (BICRO)-39 scale. This showed much more meaningful and impressive benefits of active intervention, with greater relevance, measuring what subjects really do, rather than can do, in their homes including such areas as social participation, productive employment and psychological wellbeing.
Powell et al1 acknowledge that they could only evaluate the overall framework for delivering rehabilitation rather than its specific components. They have offered important new evidence to support the wider development of community based multidisciplinary rehabilitation teams. Wilson3 has reminded us that it is wasteful spending hours delivering unevaluated therapy that may be of little or no benefit to a patient. The challenge now is to find evidence based ways of determining more precisely what these teams should actually do, when, and for how long.
Competing interestes: none declared
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.