Contents
- The 30-second summary
- What you will find here
- What is demonstrated: wellbeing
- Where the evidence breaks
- Art, exercise or company: the meta-analysis of 26 trials
- The finding from 23,706 people
- Why nobody knows whether it pays for itself
- The active ingredients, and why they matter
- Practical steps for application
- Frequently asked questions
- What we do not know
- References
A GP prescribing a painting class instead of an anxiolytic is a powerful image, and arts-on-prescription programmes have spread across half of Europe on the back of it. The most complete systematic review finds a significant improvement in wellbeing. And, at the same time, that the evidence on anxiety and depression comes down to practically a single study.
There are three further findings that ought to change how these programmes are designed and rarely do. Among older adults, arts-based activities outperform both physical and social ones for reducing loneliness. In a British household panel of 23,706 people, the benefit appears only when the activity takes place in a group, and solitary arts participation shows no significant association. And the largest randomised trial conducted in Spain, with 320 participants and observer-blinded assessment, was published in 2026: until very recently, almost this entire field rested on studies without a control group.
This piece categorizes the metrics that are tracked, the ones that are poorly quantified, and those that remain completely unassessed.
The 30-second summary
It is worth distinguishing «wellbeing» from «symptoms» in every row: the evidence is not equally solid for the two, and the field frequently treats them as though it were.
| Finding | Result | Study and year |
|---|---|---|
| Effect on wellbeing | Statistically significant improvement in the meta-analysis | Jensen and others, FPH (2024) |
| Basis for that conclusion | 25 studies selected from 7,805 records | Jensen and others (2024) |
| Evidence on anxiety and depression | Limited to one single study, with a clinically modest effect | Sá and others, FPH (2026) |
| Which modality works best in older adults | The arts-based one, ahead of physical and social | Wang and others, Geriatric Nursing (2026) |
| Size of that meta-analysis | 26 randomised trials, 3,892 participants | Wang and others (2026) |
| What happens if the activity is solitary | No significant association with mental health | Kaya and Mathieu, SJPH (2025) |
| First large randomised trial in Spain | 320 participants, blinded assessment, nature-based | Coll-Planas and others, Lancet RH Europe (2026) |
| Australian controlled trial on loneliness | 114 participants, eight-week outcomes | Dingle, Sharman and others, FP (2024) |
| Broadest synthesis of randomised trials | 40 studies, still a preprint without peer review | Feng, Kanukula and others (2026) |
| State of economic evaluation | Standard health economics methods are barely applied | Lynch and others, FPH (2026) |
What you will find here
- What the meta-analysis reveals regarding wellbeing, alongside the cautionary note issued by its creators.
• Why the evidence on anxiety and depression is so fragile, with numbers.
- The meta-analysis of 26 trials that compared art, exercise and social activity among older adults.
- The finding from 23,706 people: what matters is not the art, it is the group.
- The randomised trials that have arrived in the last two years and what they change.
- Why nobody has yet calculated whether these programmes pay for themselves.
What is demonstrated: wellbeing
Anita Jensen, Nicola Holt, Saeko Honda and Hilary Bungay published in Frontiers in Public Health in 2024 the first systematic review with meta-analysis of the impact of arts on prescription on individual health and wellbeing. They started from 7,805 records and retained 25, from Australia, Denmark, Sweden, the United Kingdom and the United States.
The evaluated programmes display a greater similarity than anticipated, thereby simplifying the comparison process: the majority of these initiatives spanned eight to ten weeks, took place within community venues, creative hubs, general practices, and primary healthcare clinics, and featured a diverse spectrum of activities.
The quantitative result is clear: the meta-analysis reveals a statistically significant enhancement in wellbeing. Furthermore, three distinct benefit categories emerge from the qualitative assessment: social advantages, psychological gains, and progression pathways.
The authors point out, nevertheless, an important qualification that usually gets lost in press summaries: the narrative review identified promising outcomes requiring further evidential support, such as reductions in anxiety and depression. In other words, wellbeing is being assessed, whereas clinical symptoms are not yet being evaluated.
| Study | What it analysed | What it found | What it cannot establish |
|---|---|---|---|
| Jensen and others (2024) | 25 studies across five countries | Significant improvement in psychosocial wellbeing | Effects on anxiety and depression remain unconfirmed |
| Sá and others (2026) | Six studies on wellbeing, depression and anxiety | Wellbeing in all; symptoms in only one | No control groups and no long follow-up |
| Wang and others (2026) | 26 randomised trials with 3,892 older adults | Arts-based outperforms physical and social | Population aged over 60; not extrapolable |
| Kaya and Mathieu (2025) | British panel of 23,706 people, fixed effects | The benefit appears only in groups | Longitudinal association, not an experiment |
| Coll-Planas and others (2026) | Randomised trial with 320 vulnerable adults | Nature-based group intervention, blinded assessment | One city and one specific modality |
| Lynch and others (2026) | 18 studies from health economics | Standard methods are barely applied | Does not conclude it is unprofitable, only unmeasured |
Where the evidence breaks
Érica Frade Sá, Matilde Monteiro-Soares and six co-authors examined specifically the clinical question in Frontiers in Public Health in 2026: whether arts on prescription improves wellbeing and reduces symptoms of depression and anxiety.
The selection funnel gives the measure of the problem. From 3,561 unique citations, six studies met the criteria, and two more appeared on checking the references. Five were quasi-experimental and three observational. Sample sizes ranged from 12 to 1,297 participants, with mean ages between 43 and over 80.
The wellbeing result agrees with the earlier meta-analysis: enhancement in psychological wellness following participation was documented across every study.
The symptom result is the one not to overlook: the evidence on depression and anxiety was limited to one study, which reported statistically significant but clinically modest reductions in both. One study, and with an effect its own authors describe as clinically modest.
Why the field has this problem
It is neither coincidence nor carelessness: it is the history of the field. The systematic reviews that have audited it converge on the diagnosis. Gina Reinhardt and Dragana Vidović assessed in Perspectives in Public Health in 2021 the impact of social prescribing initiatives on loneliness and started from nine studies; the same team, widening to individual and community wellbeing in IJERPH, reached 51. Matthew Cooper, Leah Avery and five co-authors reviewed in BMJ Open in 2022 22 UK interventions targeting mental health and wellbeing, with the aim of establishing not only effectiveness but the «active ingredients».
The issue shares a consistent layout across the board: numerous modest, isolated-group trials featuring prior and subsequent assessments, lacking any control group, and tracked over brief durations. Such frameworks prove helpful for outlining an occurrence yet fall woefully short in establishing causation, since any progress might stem simply from the elapsed time, placebo anticipation, or the inherent tendency of workshop participants to be recovering already anyway.
Two examples illustrate that dominant format well. Linda Thomson, Bridget Lockyer and two co-authors evaluated in Perspectives in Public Health in 2018 the British Museums on Prescription programme, with 115 older participants, measuring change over time in six self-rated emotions: absorbed, active, cheerful, encouraged, enlightened and inspired. Rachel Sumner, Diane Crone and two co-authors analysed in Public Health in 2021 the observed changes in anxiety, depression and wellbeing among 355 British participants across eight-week referral cycles.
Both are serious pieces of work and both describe real changes in the people who took part. Neither has a control group, and that is why they cannot settle the clinical question.
Art, exercise or company: the meta-analysis of 26 trials
The work that resolves part of that methodological deficit is by Yi-Ke Wang, Yanzhongcheng Xu and two co-authors in Geriatric Nursing in 2026, and its aim was precisely comparative: to systematically evaluate the differential effects of arts-, physical- and social-based interventions on quality of life and loneliness reduction in older adults.
Adhering strictly to the PRISMA 2020 guidelines, nine domestic and international databases were thoroughly searched for randomised trials appearing between 2016 and 2024, setting inclusion criteria that focused on community-dwelling adults aged 60 or older who received structured social prescribing alongside conventional care, while methodological rigor was carefully guaranteed through the application of the Cochrane risk-of-bias 2.0 tool.
The outcome: 26 randomised trials encompassing 3,892 participants. Social prescribing yielded substantial enhancements in quality of life alongside moderate decreases in loneliness.
And the finding that should most influence programme design: subgroup analyses indicated that arts-based interventions yielded the largest effect sizes. Above physical and purely social ones, in a population where exercise is habitually what gets prescribed.
The authors anticipated substantial clinical heterogeneity due to cross-cultural variations in how the intervention is delivered, so they used a random-effects model and subgroup analyses by intervention type, duration — short-term up to three months, long-term beyond — and cultural dimensions. That the heterogeneity was anticipated in advance rather than discovered afterwards is one of the reasons to give this work weight.
The broadest available synthesis goes one step further, but with an important caveat. Xiaoqi Feng, Raju Kanukula and six co-authors have quantified the effects of social prescribing on mental, physical and social health outcomes in adults across 40 studies, searching Medline, Embase, Cochrane Central, AMED, CINAHL, PsycInfo, Web of Science and several trial registries. Their justification describes the field’s situation exactly: expansion has outpaced evidence from randomised trials. The work is a preprint, still without peer review, and should be treated as such until it is published.
The finding from 23,706 people
If only one result from this whole field were to be retained, it should probably be this one. Gokhan Kaya and Chloé Mathieu analysed in Scandinavian Journal of Public Health the association between arts participation and mental health using the UK household longitudinal study, with 23,706 people and fixed-effects models, separating solitary participation from group-based participation.
The result: the advantage emerges solely when the activity involves a group. Individual artistic engagement reveals no meaningful connection to psychological well-being.
That discovery reshapes how everything prior is understood. If the collective rather than the creative medium drives the outcome, then these initiatives are actually addressing isolation, making the painting workshop merely the delivery mechanism instead of the core therapeutic agent. This perspective also aligns with the fact that loneliness is the metric where impacts appear most reliably.
A clarification about what the British panel adds relative to the trials, because its value is easy to misread. Fixed effects compare each person against themselves over time, which rules out the result being explained by stable differences between those who make art and those who do not: education, income, personality. That is a powerful control. What it does not rule out is reverse causation, that is, people participating more in groups during the stretches when they already feel better.
Two recent randomised trials have begun to test that hypothesis with controlled designs. Genevieve Dingle, Leah Sharman and nine co-authors published in Frontiers in Psychology in 2024 the eight-week outcomes of a trial with 114 adults in Queensland experiencing loneliness, comparing GP treatment-as-usual against that same treatment combined with social prescribing, in which a link worker supports the person in connecting with community-based social activities. Their own justification flags the gap: there have been few controlled evaluations of social prescribing.
Furthermore, the methodologically most robust investigation within this domain in Spain came out in The Lancet Regional Health – Europe throughout 2026, courtesy of Laura Coll-Planas, Marta Masó-Aguado alongside twenty-eight additional co-writers. Constituting an observer-blinded randomised controlled trial with 320 participants, this study evaluates «Friends in Nature», which operates as a facilitator-led group and nature-based social prescribing intervention targeting loneliness and health-related quality of life among vulnerable urban adults residing in Barcelona.
Its framing sums up why this field matters: loneliness is a major public health problem disproportionately affecting vulnerable populations and has well-recognised health consequences.
| Design | What it allows you to claim | How much of this exists in the field |
|---|---|---|
| One group, before and after | That participants improved | Most of the studies |
| Quasi-experimental | That the improvement might be due to the programme | Five of the six in the clinical review |
| Longitudinal panel with fixed effects | That the association persists within the same person | One, with 23,706 people |
| Randomised trial | That the programme caused the improvement | 26 in older adults; two recent ones in adults |
| Randomised trial with blinded assessment | That the result was not biased by the assessor | One, from 2026, in Barcelona |
| Standard economic evaluation | Whether the programme pays for itself | Practically none |
Why nobody knows whether it pays for itself
Social prescribing is frequently defended as a saving: fewer appointments, fewer referrals, fewer prescriptions. Mary Lynch, Aoife Keating, Elizabeth Morrow and Lesley Spencer reviewed that claim in Frontiers in Public Health in 2026 from a health economics standpoint, across 18 international studies.
Their starting point is neutral: social prescribing is an approach to improving health and wellbeing that links individuals to community-based supports — arts programmes, physical activity initiatives, financial or housing advice, volunteering and social groups — and the evidence on its effects on physical, functional and mental health is expanding.
And their conclusion highlights a void rather than a finding: no systematic review of the global data had previously evaluated this particular dimension. Their analysis reveals that conventional health economics methodologies are scarcely utilized for these initiatives.
That needs to be analyzed carefully, as it frequently gets misconstrued either way. It does not state that arts on prescription lacks cost-efficiency. Rather, it indicates that its value has not been assessed through the analytical frameworks required to determine it. Any financial savings figures currently being shared do not stem from a comprehensive economic assessment.
| Data point | Figure |
|---|---|
| Records screened in the wellbeing meta-analysis | 7,805, of which 25 remained |
| Typical duration of the interventions | Between eight and ten weeks |
| Studies with evidence on anxiety and depression | One |
| Trials and participants in the older-adult meta-analysis | 26 trials, 3,892 people |
| Modality with the largest effect size in older adults | The arts-based one |
| Sample of the British panel | 23,706 people |
| Participants in the Barcelona randomised trial | 320 |
| Studies reviewed from health economics | 18 |
The active ingredients, and why they matter
There is a question this field has started asking late and which decides whether a programme can be replicated: which specific part of the intervention produces the effect. Cooper and colleagues put it explicitly in 2022 when setting out to establish not just the effectiveness of British interventions but their «active ingredients».
It is the same question pharmacology resolves before authorising anything, and here it remains open. Without an answer, a council wanting to copy a programme that worked does not know what it is copying: the workshop, the group, the link worker, the ten weeks, the referral from a doctor, or the simple fact of having a weekly appointment to turn up to.
| Candidate component | What points in its favour | What is missing to confirm it |
|---|---|---|
| The artistic activity itself | Largest effect sizes in older adults against physical and social | Direct comparison in other age groups |
| The group | Alone, no significant association | A trial isolating group from content |
| The link worker | Present in the trials with positive results | An arm without a link worker and the same activity |
| Referral from a doctor | Gives legitimacy and reaches people who would not attend | Comparing self-referral with clinical referral |
| The eight-to-ten-week duration | It is the regime evaluated in most studies | Comparison with shorter and longer schedules |
| Regularity of the appointment | Explains part of the effect in any group intervention | An active comparator with no cultural content |
The last row is the most uncomfortable and the most important. An active comparator — a group meeting at the same frequency to do something with no artistic content — is the design that would settle the question, and it practically does not exist in the literature. Until it does, the claim that culture improves health is compatible with the claim that meeting people once a week improves health, and the available data does not distinguish between them.
That takes nothing away from the programmes: if they reduce loneliness and improve wellbeing, they serve their health purpose regardless of the mechanism. What it takes away from is the rhetoric. «Prescribe a painting instead of a pill» describes something different from «get people together once a week», and only the second claim is unambiguously supported by the current data.
One practical consequence follows, and it is worth stating because it cuts against how these schemes are usually sold to funders. If the mechanism is the group, then the scarce resource is not the artist: it is the link worker and the room. A programme that economises on the person who gets people through the door and keeps them coming back is economising on the part the evidence actually supports, while paying for the part it does not. That is a budgeting error with a research basis, not a matter of taste.
It also changes who should be referred. The outcome with the most consistent evidence behind it is loneliness, not mood, which means the referral criterion that fits the evidence is isolation rather than a symptom score. Those two populations overlap but are not the same, and a scheme designed around the second will be evaluated against outcomes it was never shown to move. The Barcelona trial is instructive here precisely because it chose loneliness as its primary outcome and recruited vulnerable urban adults rather than patients with a diagnosis: it is the design that matches what the rest of the literature had already pointed at, which is probably why it is also the first one in this field robust enough to be published in a general medical journal. Whether its result replicates elsewhere is now the question that matters most, because a single blinded trial in one city is a foundation rather than a conclusion.
Practical steps for application
- Promise wellbeing, not symptom remission. The meta-analysis backs the former; findings concerning anxiety and depression are restricted to a single trial demonstrating a clinically modest impact.
• Design the activity as a group, always. In the panel of 23,706 people, solitary arts participation shows no significant association with mental health.
- Prioritize artistic pursuits over physical workouts for seniors. Within the meta-analysis encompassing 26 trials and 3,892 subjects, interventions grounded in the arts generated the most substantial effect sizes.
• Keep to eight to ten weeks. That is the duration of most evaluated interventions, and therefore the regime on which evidence exists.
• Include a control group in the evaluation. It is what most of the field lacks and what distinguishes the two recent trials from everything before them.
• Do not use saving figures. The 2026 review finds standard health economics methods have barely been applied to these programmes.
Frequently asked questions
Does arts on prescription work?
For wellbeing, yes: the meta-analysis of 25 studies across five countries finds a statistically significant improvement. For symptoms of anxiety and depression, the evidence comes down to one study with significant but clinically modest reductions.
Does it replace a treatment?
There is no basis for claiming so. The existing controlled trials compare usual treatment against usual treatment plus social prescribing, not one instead of the other, and the reviews expressly warn that effects on symptoms require further evidential support.
Which activity should be prescribed?
In adults over 60, the meta-analysis of 26 randomised trials revealed that arts-based interventions produced the largest effect sizes, outperforming physical and social ones.
Does painting at home count?
According to the only study that separates them, no. In a panel of 23,706 people with fixed effects, the benefit appears only when the activity is group-based, and solitary practice shows no significant association.
So is it the group that heals, not the art?
That is the hypothesis best fitting the data: that these programmes treat isolation and the artistic activity is the vehicle. That the most consistent effects appear on loneliness points that way, although no study has yet decomposed the effect into its components.
How long should the programme last?
The available evidence corresponds to programmes of eight to ten weeks, which is the duration of most evaluated interventions. The older-adult meta-analysis distinguishes short-term — up to three months — from long-term.
Does it save the health system money?
Nobody knows. The review of 18 studies from health economics concludes that standard economic evaluation methods have barely been applied to these programmes. That is not a negative conclusion: it is a gap.
What we do not know
• Whether it reduces symptoms of anxiety and depression. The evidence is limited to one study with a clinically modest effect, and the reviews themselves ask for more.
• Which part of the effect is the art and which the group. The contrast between solitary and group practice suggests the latter weighs more, but nobody has decomposed the effect experimentally.
• Whether the results in older adults transfer to other ages. The meta-analysis of 26 trials includes only community-dwelling adults aged 60 and over.
• How long the benefit lasts after the programme. Follow-ups are short and most studies end with the intervention.
- Whether it pays for itself. Until standard health economics methods are applied, any saving figure is an estimate with no basis in the literature.
References
1. Jensen, A., Holt, N., Honda, S. and Bungay, H. «The impact of arts on prescription on individual health and wellbeing: a systematic review with meta-analysis». Frontiers in Public Health, 2024. 10.3389/fpubh.2024.1412306
2. Sá, É. F., Monteiro-Soares, M. and others. «Arts on prescription for wellbeing in adults: systematic review». Frontiers in Public Health, 2026. 10.3389/fpubh.2026.1833798
3. Wang, Y.-K., Xu, Y. and others. «Impact of social prescribing on quality of life and loneliness in geriatric populations: A comprehensive meta-analysis of RCTs». Geriatric Nursing, 2026. 10.1016/j.gerinurse.2026.104098
4. Kaya, G. and Mathieu, C. «The association of solitary versus group-based arts participation with mental health: evidence from the UK Household Longitudinal Study». Scandinavian Journal of Public Health, 2025. 10.1177/14034948251370234
5. Coll-Planas, L., Masó-Aguado, M. and others. «Effectiveness of a nature-based social prescribing intervention on loneliness and health-related quality of life among vulnerable urban adults (RECETAS-BCN): an observer-blinded randomised controlled trial». The Lancet Regional Health – Europe, 2026. 10.1016/j.lanepe.2026.101821
6. Dingle, G., Sharman, L. et al. «A controlled evaluation of social prescribing on loneliness for adults in Queensland: 8-week outcomes». Frontiers in Psychology, 2024. 10.3389/fpsyg.2024.1359855
7. Feng, X., Kanukula, R. and others. «Effects of Social Prescribing on Mental, Physical, and Social Health Outcomes: A Systematic Review and Meta-Analysis of Randomised Trials». 2026. Preprint, not peer reviewed. 10.64898/2026.08.02.26359484
8. Reinhardt, G. and Vidović, D. «Understanding loneliness: a systematic review of the impact of social prescribing initiatives on loneliness». Perspectives in Public Health, 2021. 10.1177/1757913920967040
9. Cooper, M., Avery, L. and others. «Effectiveness and active ingredients of social prescribing interventions targeting mental health: a systematic review». BMJ Open, 2022. 10.1136/bmjopen-2021-060214
10. Thomson, L., Lockyer, B. and others. «Effects of a museum-based social prescription intervention on quantitative measures of psychological wellbeing in older adults». Perspectives in Public Health, 2018. 10.1177/1757913917737563
11. Sumner, R., Crone, D. and others. «Arts on prescription: observed changes in anxiety, depression, and well-being across referral cycles». Public Health, 2021. 10.1016/j.puhe.2020.12.008
12. Lynch, M., Keating, A., Morrow, E. and Spencer, L. «The health economics of social prescribing: systematic review of the international evidence». Frontiers in Public Health, 2026. 10.3389/fpubh.2026.1753435
This article summarises peer-reviewed scientific literature on social and arts prescribing, flagging in the text the preprint included. It is not medical advice and does not replace assessment by a health professional. Nobody should modify a treatment on the basis of this text.