Feasibility, Acceptability, and Efficacy of Mindfulness Training in People With Upper Motor Neuron Disorders: A Systematic Review.

Only studies published in English from 2001 onwards were included, and no unpublished or non-peer-reviewed work, so relevant work in other languages or from earlier years is missing. Very few studies were in people with spinal cord injury (1 study) or ALS (2 studies), so the findings may not carry over to those groups. Among the trials used for the outcome analyses, 19% of the depression and anxiety studies, 29% of the fatigue studies and 10% of the quality-of-life studies were rated high risk of bias, and the authors say this should make readers cautious about those results. The outcome figures are medians of a relative change calculation across studies using different questionnaires, not pooled effect sizes, so the size of any benefit is approximate and should not be read as a precise treatment effect. Most comparisons were against usual care or a waiting list rather than another active treatment, which tends to make any active intervention look better. Fewer than half of the studies (19 of 44) measured anything after the programme ended, and only four followed people out to 12 months. Only 9 of the 32 studies that recommended home practice reported whether participants did it, so how much home practice people actually manage is uncertain. Acceptability data came from only 8 studies, which the authors say is too few to compare in-person with online delivery or live with self-paced delivery.

Do not read the percentage reductions as proof that mindfulness beats other active treatments. The 27% figure for depression is a median of relative changes across studies that used different questionnaires, not a pooled effect size, and against people who received another active treatment such as exercise or psychoeducation the reduction was 14%. The authors also note that 19% of the trials in the depression and anxiety analysis and 29% of those in the fatigue analysis were rated high risk of bias.