Arts in health: which creative therapies have the evidence, and which have the stories

Hospices employ music therapists for reasons that do not fit an outcome scale. What the work involves, and the measurement problem behind it.

A masters, supervised placements, personal therapy, and a very small profession at the end of it. What the UK and US routes each demand.

What counts as a good result has changed, and autistic adults have had a great deal to say about it. Why no effect size appears on this page.

Adult mental health is the use people assume is best evidenced. It is not, and the reasons are comparators, blinding and who reaches a trial.

I run two of these. What a good one looks like from the inside, and why the research behind them is thinner than the coverage suggests.

A tablet, a pair of headphones and songs from someone's twenties. What that is, how to build one carefully, and why it is not music therapy.

One outcome after stroke has strong trial evidence and the rest do not. Keeping gait velocity apart from the claims that travel alongside it.

Twenty six trials, 2,051 patients, anxiety 5.72 STAI-S units lower. Why this review carries no certainty rating, and what that omission means.

Thirty studies, 1,720 people randomised, a small effect on low mood and none on agitation. The 2025 review in full, including the row nobody quotes.

Gait velocity 11.34 m/min faster across 9 trials and 268 people. What the trials measured, who it suits, and why the unit must not be converted.

Start at a statutory register rather than a search engine. Routes country by country, NHS and school access, fees, and what to ask first.

Quieter and slower than most people expect. What an assessment covers, why improvisation dominates, and what the therapist writes down afterwards.

A registered music therapist is not a visiting singer. What the protected title means, and where the trial evidence is strong, thin, or actually null.

Both can help, only one is regulated, and the words get used interchangeably. Six checks that tell a clinical intervention from a good afternoon.

Community making is not therapy and does not need to be. What draws people in, what has been measured, and how the claims inflate.

Expressive writing and participatory photography are neither of them registered therapies. What they involve, and the consent trap in one.

Secure and inpatient settings employed dramatherapists long before the research arrived. What the work looks like there, and how thin the base is.

Role, story and improvisation used clinically, on the same statutory register as music therapy. What it involves and where it is actually offered.

Oncology has a large arts literature and very little of it is strong. What the work is used for alongside treatment, graded claim by claim.

Consent works differently when the client is eight. Where children are seen, who agrees to it, and what happens to the work a child makes.

A protected title on one side of the Atlantic and a patchwork of credentials on the other. What each route requires, and what it does not.

The claims here run well ahead of the trials. What is genuinely established, what is borrowed from adjacent fields, and what is unstudied.

Four hundred and seventeen people, three arms, no benefit on either primary outcome. What that trial settled, and the question it left open.

Nobody teaches you technique and nobody interprets your picture at you. The shape of the hour, and what happens to the work once it ends.

A painting group and a clinical intervention can look identical from the doorway. The checks that separate them, and why both are worth having.

Two protected titles cover one profession, and a craft afternoon covers none of them. What UK regulation guarantees, and why no fee figure is quoted.

Adapting downwards changes the room, not just the movement. Sightlines, physical assistance and the one thing a seated class honestly cannot train.

Three quite different activities share one heading. What separates them, why the title is not protected, and the figure this site will not print.

A relapsing condition, a self reported main outcome and heat that makes symptoms worse. Three reasons this research is unusually hard, and where it has got to.

A feasibility study answers whether a trial can be run, not whether the thing works. Why almost all the dance and stroke research is that first kind.

Three trials, 147 people, very low quality, and no firm conclusions. The number in circulation is not an effect size at all, and here is what it really is.

The claim travels much further than the study. Why no tango figure appears here, and the five stages by which a small result turns into a settled fact.

Best supported is a comparative statement, not a strong one. Why this corner has more trials than the rest, and why no effect size appears on this page.

Falls are counted, not scored on a questionnaire. Why that makes this the most testable dance claim, and why the strongest version is about exercise.

Postgraduate training and an accredited register, but no title protected in law. What the profession involves, and what checking a practitioner actually means.

No register exists on the community side, so every check is one you make yourself. Where these classes come from, and the question people forget to ask.

A policy report argues for something. A review measures it. Six checks that tell the two apart, and the citation error that outruns every correction.

Free at the point of use means somebody else is paying, usually on a grant with an end date. Where the money comes from and why schemes keep closing.

Nobody writes a prescription and nothing is dispensed. How a referred place on a creative course works, and where the widely quoted savings figures came from.

Most schemes run 8 to 10 weeks and follow the same arc. What happens in each part of it, including the week most people find hardest.

Four countries, four ways of organising the same idea, and one recurring trap: quoting a scheme's numbers with another phase's dates attached to them.

Object handling, closed sessions, and a building people will be seen entering. How museum schemes run, and where the self reported numbers part company.

Breath control, posture and a weekly group. What the programmes involve, why they are easier to test than most arts activity, and why no number appears here.

Follow the sums: £42,423 divided by 90 patients, set against £180,000 divided by 500. Where the savings figures came from and what they actually describe.

Vague requests disappear. The exact wording that gets a call back, who to say it to, and what to do when the honest answer is that nothing exists locally.

Link workers connect people to things that already exist locally. What that means in practice, and why arts is a small corner of a much wider job.

Eighty one trials and 5,576 people can still add up to very low certainty. The checks that separate a real arts in health finding from a repeated one.

Programmes fail on logistics, not on the quality of the work. Insurance, safeguarding, money, and who owns the thing once the enthusiast has left.

On an open ward the audience cannot leave. What that means for musicians, and the half minute rule that separates a good programme from a thoughtless one.

Making something that outlives its maker raises questions nobody asks until afterwards. Who owns it, who else is in it, and what was actually agreed.

Somebody who cannot consent cannot agree to being photographed, and a relative's permission does not reliably extend to publication. The rules that follow.

The fastest growing part of this field and the thinnest evidenced. Why staff wellbeing programmes are hard to test, and the question to ask before funding one.

One 1984 study of 23 matched pairs still underwrites a forty year claim. What hospital arts programmes do, and which of their claims hold up.

A ward art group and a registered therapy can look identical from the doorway. What separates them, and what the largest UK trial actually found.

Moderate certainty evidence of no effect on agitation, a small effect on low mood, and nothing lasting four weeks. What that means for a weekly session.

Activity sits within a home's wider duties, not clinical care. Who runs it, what good looks like, and the question families should ask about staffing.

Forty six patients, 23 matched pairs, one operation, one season, and no artwork in it at all. What Ulrich 1984 found, including the row nobody quotes.