This post refers to two papers in the JAMA Journal of Internal Medicine: one is the report on a randomised controlled trial comparing equivalent times spent on mindfulness or disease-specific education in people with migraine; the other is a commentary on it, pointing out that although there was no effect on primary outcome (frequency of headaches), there were many and consistent statistically significant changes across a range of functional and quality of life measures. Whether or not migraine management falls within the scope of rehabilitation – and it is a very common long-term disabling condition, and so should – these papers have much to teach rehabilitation research.

The report on the original study illustrates many features that should be in a rehabilitation study:

• a plausible and potentially effective contrast intervention delivered in equivalent amounts, and similar in other ways
• a clear description of the interventions, with reference to a protocol describing at least one
• outcome measures covering several domains of interest: disability, quality of life, and emotional state. These are all of great importance to patients – all were helped.
• follow-up at 12, 24. and 36 weeks. (the treatment was eight weeks)
• a reasonable number of patients (96 randomised)
• very well presented and written up (would you expect otherwise? But useful to help other learn.)

The commentary discusses the dilemma from the point-of-view, for example, of a funding organisation. The study set out with a defined primary outcome which was not altered. Therefore one could argue that, by its own criteria, this has shown that mindfulness is ineffective, and all other data should be ignored. However a range of measures that are of great importance to patients were improved, and the pattern is consistent; this was not an isolated statistically significant difference in one measure. Therefore one should argue that, from the point-of-view of the patient, the benefits are real and important, and that, provided the patient understands that frequency of migraine will probably not be benefit (any more than it would be education alone), it is an effective intervention.

Daniel Cherkin, referring to trials investigating non-pharmacological interventions in people with chronic pain, concludes: “These trials, which reflect a shift from predominantly pharmacological approaches, require new ways of thinking about key aspects of study design, including choices of comparison groups and primary outcome measures.“