Dance for Parkinson's: The Best Supported Corner of Dance Research, and What That Means

Dance for Parkinson’s is the best known arts and health programme in the world: weekly community dance classes, led by dance artists rather than clinicians, usually with live music, and usually open to a partner or carer as a participant rather than a spectator. It is also the best supported corner of dance research, and this page is largely about how modest a statement that is.

I do not run one of these classes and I have never taught one. I run singing sessions in two care homes and on a stroke rehabilitation unit, which is a different job in a different room. What I have done is sit at the back of a Dance for Parkinson’s class as a visitor, and the thing I came away thinking about was the car park. People had driven a long way. When I asked why, nobody mentioned an outcome measure. They mentioned the pianist, and a woman said that it was the only hour in her week when nobody was assessing her. I have carried that around ever since, because it is a real answer to a question the trials are not asking, and it is also not evidence of anything a trial would recognise.

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 the classes are

A community dance class, weekly, an hour to ninety minutes, led by a dance artist, very often with a live musician in the room.

The common shape is recognisable across a lot of provision: a seated warm up, then standing work with a barre or a chair back available for anybody who wants it, then something travelling across the floor, then a sequence learned over several weeks. Carers, partners and adult children are generally welcome to take part rather than to sit at the side, which participants mention more often than any element of the movement itself. Classes are often held in dance studios, theatres and arts venues rather than in health buildings, and that is a deliberate choice about what the hour is meant to feel like.

What it is not is a treatment. There is no assessment, no clinical goal, no notes, no discharge, and no protected title behind the person leading it. That places it firmly on the unregulated community side of this field, alongside the classes described on finding a dance for health class, and not alongside the clinical work described on dance movement psychotherapy, where an accredited register and a postgraduate training sit behind the practitioner¹. Neither is a substitute for physiotherapy or for any part of Parkinson’s treatment.

Why this corner has more research than the rest of dance

Four things make Parkinson’s the easiest place in dance to run a study, and none of them is about dance being more effective here.

The outcomes are physical and there are validated instruments for them. Motor scales, balance scales and timed walking tests exist, are widely used in Parkinson’s research, and produce numbers without asking a participant how they feel. Compare that with the outcomes claimed for community arts generally, which are mostly wellbeing scales.

There is an existing rehabilitation research community. Parkinson’s already has trialists, outcome measures, patient groups willing to be recruited and clinicians used to being asked. Most community arts activity has none of that infrastructure anywhere near it.

The intervention is unusually well specified. A weekly class of a stated length, over a stated number of weeks, taught to a recognisable structure, is something a protocol can describe. A great deal of arts in health cannot be pinned down that precisely, which is one reason it does not get trialled.

The mechanism is plausible in a way that funders accept. Rhythmic cueing, weight transfer, balance under changing conditions and dual tasking are all things Parkinson’s rehabilitation already works on. A funder does not need convincing that movement to a beat could plausibly do something for gait.

Put together, those explain why the dance literature clusters here. They do not tell you that the results are strong, and it is worth separating the reasons a field gets studied from the findings it produces.

What best supported actually means here

It means better than the rest of a weak literature. It does not mean well supported, and the difference is the whole of this section.

Three features run through this body of work, and each one shrinks the conclusion a positive result can carry.

The trials are small. Not uniformly, but typically. Small trials produce wide confidence intervals and are more easily swayed by who happened to drop out, and pooling a lot of them produces a precise looking average built from imprecise parts. The clearest demonstration of that anywhere on this site is not in dance at all: the 2021 Cochrane review of music interventions in cancer care pooled 81 trials and 5,576 participants and still rated its findings very low certainty², which is Tier 2. Volume is not strength.

Blinding is close to impossible. Nobody can be unaware that they have spent an hour dancing. Participants know, teachers know, and unless outcome assessment is done by somebody kept in the dark, the assessor knows too. That does not invalidate a trial and it does routinely cost it a certainty rating, particularly on any outcome the participant reports themselves.

The comparator is usually nothing. This is the one that matters most and gets discussed least. A trial that compares a weekly dance class against no treatment or a waiting list can only tell you that people who were given something did better than people who were given nothing. It cannot separate the dancing from the attention, the company, the structure of a fixed weekly commitment, the expectation of benefit, or the plain fact of leaving the house. Every one of those is a real thing that helps people, and none of them is dance. A trial with an active comparator, another supervised group activity of similar length and sociability, asks the useful question. There is much less of that.

So when this page says best supported, it is making a comparative statement about the dance literature and not a claim about strength. On this site the label is Tier 3, and specifically the nobody has looked properly kind rather than the null controlled evidence kind. Those are different situations and conflating them is a mistake in both directions.

Why no effect size appears on this page

Because no dance effect size is locked in this site’s source document, and importing one from a secondary source is precisely the practice this site exists to stop.

That deserves a plain explanation rather than a shrug, because a reader arriving here wants a number and is being told there is not one. The rule this site runs on is that every efficacy figure has to be traced back to the study or review that produced it, quoted in the units that document used, with its design, its interval, its certainty rating, its sample size and its year printed alongside. A figure that has been forwarded through a policy report, a press release and a funding bid has not been checked, it has been repeated, and the only honest thing to do with it is go back to the source or leave it out.

For dance and Parkinson’s, that source work has not been completed here, so the number is absent and the gap is named. I would rather write that sentence than publish a percentage I could not defend, and I am aware that it is a less satisfying page as a result. The one dance review that has been traced to source is the Cochrane review of dance movement therapy for depression: 3 studies, 147 participants, whole sample SMD -0.67 (95% CI -1.40 to 0.05), very low quality, with its authors stating that no firm conclusions can be drawn³. That is what a traced dance figure looks like when you find one, and it is set out in full on dance for depression.

What would change this page, stated so it can be held against: a current systematic review reporting a pooled effect on a named outcome, with a confidence interval, a certainty rating, a sample size and a date, all in a document a reader can open. Better still if the pooled trials include active comparators rather than waiting lists, and if the outcome measures are consistent enough across trials to be pooled without straining. Until then the figure stays off.

What a Tier 1 claim looks like, for calibration

The only unambiguous Tier 1 finding in this field concerns walking, and it comes from music rather than dance.

Rhythmic auditory stimulation after stroke improved gait velocity by 11.34 m/min (95% CI 8.40 to 14.28), from 9 trials and 268 participants, at moderate quality⁴. Five things sit next to that claim: the effect size, the interval, the certainty rating, the sample size and the year, all from one review you can open. That is the bar. It is also worth noticing how narrow the claim is: gait velocity, and nothing about mood, confidence or quality of life. See rhythmic auditory stimulation, and dance and health for why that finding sits on the dance pillar deliberately.

The related case study in how a small striking result becomes a global claim is tango and Parkinson’s research, which is a better lesson in method than in dance. And the corner where dance evidence is most often said to be strongest, balance and falls, is covered on dance, balance and falls, where the same comparator problem recurs in a different form.

How large the mapped literature is, and why that is not the same thing

A great deal has been written about the arts and health, and the size of that literature is regularly used to imply a strength it does not have.

The WHO Europe scoping review covers over 900 publications, made up of 200 plus reviews and 700 plus individual studies, with those reviews between them covering over 3,000 studies⁵. Those counts nest inside one another rather than adding up, and presenting them side by side inflates the apparent size of the field several times over. More importantly, a scoping review maps a literature rather than pooling it, so it carries no tier and supports no claim about effect at all. It is evidence that a field has been studied. The sector’s own reports and case studies, gathered in one place⁶, are worth reading the same way: as descriptions of what is happening, not as evidence that it works. That distinction is worked through on creative health policy and reports.

If you are deciding whether to go

Nothing above is an argument against going, and it would be a poor reading of this page to take it as one.

A Tier 3 label is a statement about the research literature and not about the world. Exercise, company, live music and a weekly reason to be somewhere are worth having in their own right, and none of them requires a trial. What the label rules out is a promise. Nobody can honestly tell you that a class will improve your gait, your balance or your mood, and nobody can tell you it will not, and you are entitled to weigh that up knowing which situation you are in.

Practical questions worth asking a class: what training the leader has and whether they have worked with Parkinson’s before, what happens if somebody becomes unsteady, whether there is a seated option for every part of the class, how long the class is funded for, and whether a partner can join in. Funding is the one people forget, and provision in this sector runs on short grants, so a class recommended online may already have closed.

If you have fallen in the past year, are unsteady, or your medication timing affects your movement through the day, speak to the clinician who knows all three before you start. Nothing on this page is a reason to change, delay or decline any part of anybody’s treatment, and no class substitutes for one.