Consent and Safeguarding in Arts in Health: Capacity, Photographs and What Cannot Be Shared

A person who lacks capacity to consent to being photographed, filmed or recorded cannot consent to it, and a relative’s permission does not reliably extend to publication. Those two sentences settle most of the difficult cases in arts in health, and both of them get overridden constantly by people acting entirely in good faith, because the footage is genuinely lovely and everybody wants to share it.

I have the photograph still and I have never used it. A daughter took it during one of my care home sessions: her mother, mid song, eyes closed, hands raised. She sent it to me and asked me to put it on anything I liked, because she wanted people to see that her mother was still there. I understood exactly what she meant and I have thought about that picture for two years. What stopped me was working out that the person in it could not have told me whether she wanted a photograph of herself on the internet, that her daughter’s permission was her daughter’s rather than hers, and that she had agreed to nothing except singing. The daughter was not wrong to want it seen. She simply did not have the thing she was offering me.

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.

Capacity, and the two things people get wrong about it

Capacity is decision specific, and it can fluctuate.

In England and Wales the framework starts from an assumption of capacity, requires that capacity be assessed in relation to the particular decision in front of the person, and requires that people be helped to make their own decisions before anybody concludes they cannot¹. Two consequences follow that matter enormously in a room with instruments or paint in it.

A diagnosis is not an answer. Somebody with dementia, a learning disability or a brain injury may have capacity for many decisions and not for others. Deciding on their behalf because of the label is the error the framework exists to prevent.

The decision has to be named. Somebody may be entirely able to decide whether to join in this afternoon, and unable to weigh up what it means for a video of them to exist indefinitely, in front of an audience they will never meet, in a context they cannot picture. Those are different decisions of very different sizes. The first involves the next hour. The second involves the rest of their life and beyond it.

That distinction is the whole of this page. Consent to take part is never consent to publish.

Why a relative’s permission is not enough

Because publication is not the kind of decision another person can usually make for somebody.

Relatives are frequently the right people to consult, and consulting them is proper. They know what somebody would have thought, what they were private about, what they would have hated. A programme that ignores families is not being rigorous, it is being careless in a different direction.

What a relative generally cannot do is give consent to publication on somebody’s behalf. Three features of publication make it a poor fit for any process of deciding in another person’s interests. It is open ended: nobody can say who will see it or for how long. It is irreversible: an image once shared cannot be recalled from the people who have seen it or the systems that have copied it. And it benefits the publisher, which means the person making the decision has an interest in the answer, which is exactly the situation these safeguards exist to handle.

There is a legal layer too. Photographs, video and voice recordings of identifiable people are personal data with obligations attached under UK data protection law², and organisations working in regulated care settings have their own governance duties³. Those matter and they are a floor rather than a ceiling, because something can be lawful and still wrong. The question this page is about is whether a person who cannot picture the audience has agreed to meet it.

The separate decisions nobody separates

Each of these is a different ask and each needs asking about on its own.

• Taking part in a session.
• Being photographed during it.
• Being filmed or recorded, which captures more and lasts differently.
• A photograph being shown to family.
• A photograph appearing in a printed newsletter, which has a small and known audience.
• A photograph appearing on a website, which does not.
• A clip appearing on social media, which can travel anywhere.
• Being quoted, by name or identifiably.
• Work being exhibited, with or without a name attached.
• Anything appearing in a funding bid, where the person becomes evidence for somebody else’s argument.

A single consent form covering all of that is not consent, it is paperwork. The practical version is to ask for the specific use in front of you, in terms the person can actually weigh: this photograph, on our website, where anybody can see it, for as long as the site exists.

Artwork is identifying too

Photographing the work rather than the person solves less than it appears to.

A painting is recognisable to anybody who knows its maker. Work made on a ward or in a care home carries its setting with it, and a caption naming a project narrows the field to a small number of people. Handwriting identifies. A named exhibition of work from a specific unit tells a viewer where every contributor was and roughly when.

So the same questions apply to the object as to the person: may this be shown, where, with what caption, and does the person understand who will see it. And in a registered therapy the work has an additional status, because material made in a therapeutic relationship is part of that relationship rather than a public artefact. What happens to it should be agreed at the start of the work rather than at the end, and the person should be told the arrangement rather than asked to assume one. See what happens in an art therapy session and art therapy for children, where the child’s ownership of their own work raises the question again with a further layer.

Touch, and the assistance that looks like inclusion

Physically moving somebody’s limb is a decision about touch, not a teaching technique.

This comes up constantly in seated work. Somebody is not joining in, everybody else is moving, and a member of staff takes their hands and moves them in time. It looks like inclusion and reads to the room as kindness. What it does is remove the one thing that person had control over in that half hour.

Somebody who has not responded may be listening. They may be tired. They may be declining in the only way available to them. Waiting long enough to receive an answer matters especially for people who process slowly, because those are precisely the people most likely to be moved before they have replied.

The defensible practice is short. Offer a hand rather than taking one. A hand held out and not taken is an answer. Where somebody has repeatedly shown that they like hand to hand contact, that is different, and it is still offered each time rather than assumed. Where a person cannot consent to being touched, the default is not to do it. Seated and chair based dance covers the settings where this arises most.

Safeguarding, beyond consent

Four things that belong to any programme working with people who may be vulnerable.

Checks appropriate to the role. Criminal record checks at the level the work requires, for volunteers as much as for paid practitioners, and for anybody working unsupervised. Volunteers are where this is most often missed.

Induction on escalation. Every visiting practitioner should know who to tell if somebody becomes unwell, if somebody discloses something, or if they see something that worries them. Most have never been told this anywhere they have worked.

A route to complain that bypasses the practitioner. Participants, relatives and staff all need one, and it must not run through the person leading the session.

Clarity about what is not the practitioner’s job. An artist or musician is not there to assess risk, to counsel, to give advice about treatment, or to hold information that belongs in a clinical record. Where somebody discloses something, the job is to pass it on through the agreed route, not to manage it.

A registered arts therapist works within a different framework again: assessment, clinical notes, supervision and a statutory register, which held 6,103 arts therapists across four protected titles in July 2026⁴. That is a small workforce, so most sessions in care settings are delivered by people without any of that structure behind them, which is precisely why the arrangements above have to be made explicitly. Starting an arts programme in a care setting is the practical checklist.

Why the moving clip is usually the one not to share

Because the material that makes the strongest case is generally the material gathered with the weakest consent.

The clip of somebody with advanced dementia singing a whole verse is the most persuasive thing any programme will ever produce. It is also, almost always, footage of a person who could not have agreed to it, taken in a moment nobody planned, in a place they did not choose to be, and wanted by an organisation that benefits from having it. Every factor points the same way.

There is a further reason to be careful, and it belongs on a site that grades evidence. The clip is not evidence. The 2025 Cochrane review of music based interventions in dementia covers 30 studies and 1,720 participants randomised, and reports depressive symptoms at SMD -0.23 (95% CI -0.42 to -0.04) at moderate certainty, which is Tier 1 and small, alongside agitation and aggression at SMD -0.05 (95% CI -0.27 to 0.17), also moderate certainty, which is Tier 1, null result, with no evidence of anything persisting four weeks after treatment ends and no separate quality of life estimate at all⁵. A moving clip is a moment. It cannot tell you what happened that afternoon, or the next week, or to anybody else. Publishing somebody who could not consent, in order to illustrate a claim the evidence does not support, is the worst version of this and it happens regularly. See creative activity in dementia care.

What a programme should have in place

Six arrangements, all cheap.

1. A named decision maker for images, who is not the practitioner and not the person who wants the photograph.
2. A standing rule of no recording unless arranged in advance for a stated purpose.
3. A record of what each person agreed to and when, specific to the use.
4. A review date. Permission given three years ago for a leaflet is not permission for a website today.
5. A stop route for participants, relatives and staff that does not go through the session leader.
6. A default of not publishing where there is any doubt, written down so that it does not have to be argued each time under pressure from a deadline.

The sector’s own reports and case studies are gathered in one place⁶, and they are worth reading with this page in mind, because the images in them were gathered somehow and the arrangements behind them are rarely described.

What this site does

The same rules, applied to itself, which is the only way this page is worth anything.

No image of an identifiable participant from any session I run appears here. No description of an identifiable individual appears without their agreement, and where somebody could not agree, they are not described in a way that could identify them. No named individual is assessed on this site, practitioner or participant. No cited body is presented as a partner, endorser or affiliate, and no organisation named anywhere on this site is connected to it. Where a figure is quoted, it is traced to the document that produced it.

None of that is difficult, and the reason to say it out loud is that it is checkable. The wider picture of programmes inside health buildings, and who runs them, is on arts in hospitals. A page about consent published by a site that had helped itself to a photograph would be worth nothing at all.

Nothing on this page is legal advice, and anybody making decisions about a specific person should be working with the framework their organisation operates under and with the people responsible for it. Nothing here is a reason to change, delay or decline any part of anybody’s care or treatment, and no arts activity substitutes for either.