Dance After Stroke: Feasibility Studies, Small Samples and What Would Need to Be Shown

Almost all the research on dance after stroke is feasibility work, and a feasibility study is designed to answer whether a full trial could be run rather than whether the intervention helps. That distinction decides how this whole page reads, and it is the thing most often lost when a small stroke and dance study reaches the news.

I have run a weekly singing session on a stroke rehabilitation unit for several years, so I am closer to this subject than to most on the site, and being close to it has mainly taught me caution. The physiotherapists on that unit work in a way I could not begin to imitate. They set a target, measure against it, progress the load, and write down what happened. I bring a guitar and some songs, and what I contribute is an hour that is not an assessment. Both things are worth doing and only one of them is rehabilitation. The most useful conversation I ever had there was with a physio who told me that the thing she could not prescribe was the reason to keep going after discharge, and that this was the gap she thought community activity might fill. She also said, in the same breath, that she had no idea whether it did.

Where a claim about effect appears below it carries a tier (Tier 1 supported, Tier 2 promising but limited, Tier 3 practice and anecdote), explained on reading arts in health research.

What a feasibility study is, and what it is not

It asks whether a proper trial is possible. It does not ask whether the thing works, and its numbers are rates rather than effects.

A feasibility or pilot study sets out to answer a specific list of questions. Can enough eligible people be found and recruited. Do they stay in, and how many drop out. Is the intervention safe, and did anything go wrong. Can it be delivered as written, by the people who are supposed to deliver it, in the settings available. Do the outcome measures behave sensibly in this population, and can the data be collected. Those are all genuinely important questions, and getting them wrong is how large trials fail expensively.

What a feasibility study is not powered to do is detect a difference between groups. Its sample is chosen to answer the process questions, not the effect question, so any change score it reports is uninterpretable as evidence of benefit. And feasibility studies almost always do report change scores, because the outcome measures are being trialled and the data exist. Those are the numbers that escape.

The tell is simple and worth memorising: if the paper’s own title or abstract contains the words feasibility, pilot, proof of concept or acceptability, the effect numbers in it are not the finding. The finding is the recruitment rate, the retention rate and the safety record. A press release that leads with the change score has inverted the paper.

Why a control group matters more after stroke than almost anywhere

Because recovery continues on its own, and a before and after design cannot tell the two apart.

Neurological recovery after stroke proceeds over months, steeply at first and then more gradually, and rehabilitation continues alongside it. Measure a group of people before an eight week class and again at the end, and they will nearly always look better at the end. Most of that improvement would have happened without the class. Some of it is ordinary therapy. Some of it is that the people well enough to attend all eight weeks are not the same people as those who enrolled.

This is the point at which the general warning about uncontrolled evaluations stops being general. The field as a whole produces a great many of them: the 2024 systematic review of arts on prescription screened 7,805 records, included 25, and found no randomised controlled trials at all, with the 8 quantitative studies all using uncontrolled before and after designs¹. In a stable population that design is weak. In a recovering population it is close to uninterpretable, because the untreated comparison, had it existed, would also have improved.

So the requirement here is not a stylistic preference for randomisation. It is that without a comparison group, a stroke study is measuring time.

The Tier 1 claim in this area belongs to music

There is one clearly supported rhythm and movement finding after stroke, and it is not a dance study.

Rhythmic auditory stimulation improved gait velocity by 11.34 m/min (95% CI 8.40 to 14.28), from 9 trials and 268 participants, at moderate quality². Tier 1, and the cleanest example on the site of what a supported claim looks like, with the effect size, the interval, the certainty rating, the trial count and the sample size all present and all from one review a reader can open.

Two things must be said next to it every time. First, it is a different intervention from a dance class. Rhythmic auditory stimulation is a rehabilitation technique in which a steady beat is used to cue the timing of gait, with the tempo set and progressed deliberately, usually by a therapist. It shares the rhythm with dance and it is not the same activity. Second, the claim is about gait velocity and nothing else. Not mobility in general, not independence, not confidence, not mood, not communication, not arm function, not quality of life. Letting that claim drift outward is the most common thing done with it, and this site’s rule is that a result about walking speed stays a result about walking speed. See rhythmic auditory stimulation and music therapy after stroke, which sets out where the rest of the stroke evidence sits and where it does not.

Why no dance and stroke figure appears here

Because none is locked in this site’s source document, and quoting one from a summary would be the exact practice the site exists to stop.

That is a short answer to what may feel like a frustrating absence, so here is the longer version. The rule is that a figure is printed only after somebody has opened the study or review it came from and recorded the design, the sample size, the outcome measure, the effect size, the interval, the certainty rating and the year. For dance and stroke, that has not been done here. The gap is named rather than filled.

What would change this page: an adequately powered randomised controlled trial, or a systematic review of several, reporting a named outcome with an effect size, a confidence interval, a certainty rating, a sample size and a date. Better still if the comparator were another supervised group activity of similar length and sociability rather than usual care alone, so that dancing could be separated from attention, structure and company. Better again if outcome assessment were blinded, which is achievable for physical measures even though the participants cannot be.

For a sense of what a traced dance figure looks like when one does exist, the only dance Cochrane review sourced on this site covers depression: 3 studies, 147 participants, whole sample SMD -0.67 (95% CI -1.40 to 0.05), very low quality, with its authors declining to draw firm conclusions³. That is on dance for depression, along with the widely quoted number from the same review that turns out not to be an effect size at all.

What dance might plausibly be for after a stroke, stated as a hypothesis

There is a coherent case for it, and calling it a hypothesis rather than a finding is the honest framing.

The case runs like this. Formal rehabilitation ends, often at a point that has more to do with service capacity than with the end of recovery. What follows is frequently a cliff: fewer appointments, less structure, no one expecting you anywhere, and a slow contraction of activity that has its own consequences for strength, balance, mood and isolation. A weekly class involves standing, weight transfer, sequencing, timing, memory for movement, and being in a room with people. All of those are things somebody is otherwise likely to do less of.

Every element of that is plausible. Plausibility is not evidence, and this field’s central failure is treating the two as interchangeable. The proper response to a plausible mechanism is a trial, which is precisely what the feasibility studies are trying to make possible.

The clinical version, and the community version

Two quite different things are available and they get called by similar names.

Dance movement psychotherapy is a psychological therapy delivered by a practitioner trained to postgraduate level, with assessment, agreed goals, notes, clinical supervision and a planned ending. In the UK its title is not protected in law; the four protected arts therapy titles are art psychotherapist, art therapist, dramatherapist and music therapist, and the Health and Care Professions Council register listed 6,103 arts therapists across those four in July 2026⁴. The Association for Dance Movement Psychotherapy UK runs its own accredited register instead⁵. See dance movement psychotherapy.

A community dance class is led by a dance artist, is open ended, and is there for the movement and the company. No register, no protected title, no assessment. Most dance after stroke, in practice, is this. Finding a dance for health class covers how these are organised and what to ask, and seated and chair based dance covers the adapted versions that suit many people after a stroke better than a standing class does.

Neither is physiotherapy, and neither is a substitute for it. Why the three get conflated in the first place is the subject of dance and health.

Practical questions before starting

Ask the stroke team first, then ask the class.

The clinical questions worth raising with a physiotherapist or the stroke team: standing tolerance, whether a hand needs to be free to hold on, any visual field loss that makes a moving room difficult, fatigue and how suddenly it arrives, blood pressure on standing, and whether there is anything about the pattern of weakness that makes particular movements unwise. These are specific enough that a general answer is no use.

The class questions: what training the leader has and whether they have worked with people after stroke, whether there is a seated version of everything, what happens if somebody becomes unsteady, whether the venue is reachable by whatever transport is available, whether they are insured, and how long the class is funded for. That last one catches people out, because provision runs on short grants and a class recommended six months ago may already have ended: who pays for arts on prescription explains the pattern.

And the general framing, which I would rather state twice than have inferred. Tier 3 is not a verdict that dance does not help after a stroke. It means the research needed to answer the question has not been done, and there are reasons for that which have nothing to do with whether it helps. Nothing on this page is a reason to change, delay or decline any part of anybody’s rehabilitation, and no class substitutes for it.