Mindfulness: what the trials support, and where claims outrun the evidence

Mindfulness is one of the areas where the gap between what is claimed and what has been shown is widest. Both halves of that sentence matter: there is real evidence, and there is a great deal of overclaiming built on top of it.

What is reasonably well supported

Relapse prevention in recurrent depression. This is the strongest and most specific finding. NICE’s 2022 guideline on depression in adults concluded there was good evidence that group CBT and mindfulness-based cognitive therapy were effective, and on average cost-effective, for people at high risk of relapse, with data covering up to two years. It is recommended as one option among several.

Moderate effects on anxiety, depressive symptoms and pain, in structured eight-week programmes. The effect sizes reported in the meta-analytic literature are moderate for those outcomes and smaller for general stress reduction and quality of life.

Note what that evidence is of: eight-week, twenty-plus-hour, teacher-led group programmes with daily home practice. It is not evidence about apps, ten-minute recordings, or a mindfulness slot in a staff away day.

Where the claims run ahead

Claims that mindfulness treats or alleviates cancer, HIV, multiple sclerosis, lupus, ME, addictions, eating disorders, obsessive-compulsive disorder or trauma — or that it extends lifespan by protecting telomeres — are not supported by the evidence base. Some have been examined and found wanting; most have simply never been tested at a standard that would justify the claim. Assertions about grey-matter growth in eight weeks rest on small neuroimaging studies whose findings have been difficult to replicate.

Claims about improved creativity, innovation, happiness, focus and workplace productivity are mostly extrapolation. Where mindfulness has been trialled for sleep or for positive mood, the evidence has been weak.

This is worth stating flatly because the archived page at this domain made several of these claims, and they were typical of how mindfulness was described in the UK in the mid-2010s. They were not supportable then either.

Adverse effects: real, and under-reported

The idea that meditation is risk-free is an assumption, not a finding.

• A systematic review of adverse events in meditation practices and meditation-based therapies put the overall prevalence at around 8%, with a striking split by study design: about 4% in experimental studies and 33% in observational ones — a gap that itself indicates a measurement problem rather than a real difference.
• A large study of regular meditators found 22% reported unpleasant meditation-related experiences.
• An international cross-sectional study of regular meditators found unpleasant and adverse effects were more likely in people with pre-existing mental health difficulties.
• The commonest reported effects were anxiety, depression and cognitive anomalies.
• By contrast, a review of randomised trials of MBSR and MBCT found only about 1% of 4,031 participants across 36 trials reported adverse events — and the great majority of mindfulness trials have not systematically assessed or reported harms at all.

The honest reading is that the low figures from trials reflect under-measurement, not safety. For most people, most of the time, structured mindfulness practice is uneventful. For a minority — concentrated among people with existing difficulties, and among those doing intensive practice without a teacher — it is not, and a well-run programme screens for that and says so beforehand.

How to read a mindfulness claim

Four questions do most of the work:

1. Which programme? MBSR and MBCT are specific and studied. “Mindfulness” alone is not a protocol.
2. What dose? Eight weeks and daily practice, or ten minutes on a phone? The evidence is about the first.
3. Which outcome, in which population? Relapse prevention in recurrent depression is a very different claim from stress at work.
4. Compared with what? Against a waiting list, almost anything looks effective. Against an active control, much less so — and much of the older literature used waiting-list controls.

What this does not mean

It does not mean mindfulness is useless, and it does not mean nobody should do it. People find structured practice genuinely valuable, NICE recommends one form of it for a specific purpose, and Parliament’s own inquiry took it seriously enough to publish a report.

It means the claims should be sized to the evidence — and that anyone selling a course that promises to cure an illness is making a claim the literature does not carry.