Abstract
Introduction
Arts on Prescription is a social prescribing model in which health professionals refer adults to community-based artistic activities led by artists or musicians, with the goal of promoting wellbeing and mental health. This systematic review aimed to investigate whether Arts on Prescription improves wellbeing and reduces symptoms of depression and anxiety.
Methods
PubMed, PsycINFO and Cochrane CENTRAL were searched in November 2024, following a previously developed search strategy. The selection of studies, quality assessment and data extraction were carried out independently by two authors. Disagreements were resolved by consensus. We included randomised control trials, quasi-experimental studies, observational studies and mixed-methods studies assessing the impact of art programmes prescribed by a health professional on adults’ wellbeing and symptoms of depression and anxiety. A narrative synthesis of the results was carried out.
Results
Of the 3,561 unique citations obtained, six studies met the inclusion criteria. Analysis of the reference lists of the included studies revealed two additional pertinent studies. Five of the studies were quasi-experimental studies and three were observational studies. The sample size ranged from 12 to 1,297 participants, with an average age ranging from 43 to over 80 years old. In all studies, an improvement in wellbeing was reported following participation in Arts on Prescription programmes. Evidence regarding depression and anxiety was limited to one study, which reported statistically significant but clinically modest reductions in both outcomes.
Discussion
Arts on Prescription programmes were consistently associated with improvements in wellbeing across a range of populations and settings. Preliminary evidence suggests potential benefits for depression and anxiety that warrant investigation in more rigorous study designs. Further studies are needed to overcome the limitations of the analysed studies, such as the lack of control groups, small and non-representative samples, and short follow-up periods.
Systematic review registration
https://www.crd.york.ac.uk/PROSPERO/view/CRD42024572685, identifier CRD42024572685
Keywords: anxiety, art programmes, arts on prescription, depression, mental health, social prescribing, wellbeing
1. Introduction
Interest in the arts and their effects on individuals’ health and wellbeing has been a relevant topic in the scientific community in recent years, leading to a significant increase in studies in this area since the beginning of the century. The arts are a promising tool for improving mental and physical health, with potential benefits at different stages of life (1).
Social prescribing is the process by which individuals are referred by health and social care professionals for non-clinical interventions to support their wellbeing (2). This involves referral to a link worker, whose role is to develop an intervention plan based on the preferences and values of the person and the community in which they live. The link worker will also introduce them to the intervention to be implemented (3). In this way, a link has developed between primary health care and community support, with the common goal of benefiting the individual’s health and improving the network of social relationships (4). In addition to referrals via the link worker, referrals can also be made directly by a general practitioner (GP) or by the individuals themselves (3). Social prescribing provides a non-medical referral option that can be used to complement existing treatments (5).
Arts on Prescription (AoP) was established in the mid-1990s as one of the earliest models of social prescribing (6). AoP Stockport began in 1994 as the first programme aimed at improving the wellbeing of patients with mild to moderate depression (2). AoP programmes are led by artists or musicians and involve people from the community (2). Examples of these activities include painting, sculpting, and visiting museums. Patients are referred to participate in artistic activities, usually once or twice weekly, lasting around 2 h, for 6 to 12 weeks (7).
Simon Opher evaluated the cost–benefit impact of Artlift, a programme funded by the Gloucestershire National Health Service (NHS). The evaluation found that participation in arts interventions led to a reduction in GP visits and in associated NHS healthcare costs (8). This type of intervention could have benefits not only for the individuals referred, but also for the economy and healthcare spending. It is worth noting that art therapy and AoP are different. Art therapy aims to support psychotherapy or emotional expression, whereas AoP emphasises creative exploration and process rather than skill acquisition (7).
Central to the World Health Organization’s definition of health, wellbeing can be defined as a dynamic state in which individuals are able to realise their potential, work productively and creatively, form positive relationships, and contribute to their community. Wellbeing is strengthened when people are able to achieve their personal and social goals while feeling that they have a purpose in society (9). Additionally, in 2019, around 970 million people worldwide were living with a mental disorder, including 301 million with anxiety disorders and 280 million with depressive disorders. Given their high prevalence and the consequences these diseases can have on individuals’ lives, including their effects on interpersonal relationships, work capacity, quality of life, and economic costs (10, 11), there is a need to look for new strategies to provide better support to patients (11).
We therefore aimed to systematically review and synthesize the existing evidence on the efficacy of AoP (the intervention) in reducing levels of depression and anxiety and improving wellbeing (the outcome) in adults (the population) compared with any other strategy.
2. Methods
We conducted this systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (12). Our protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO; CRD42024572685).
2.1. Data sources and searches
A search strategy was developed to enable a systematic, efficient, and sensitive search for articles (Supplementary file 1). We used free text and database-specific keywords for the terms “Arts on Prescription,” “depression,” “anxiety,” and “wellbeing,” and adapted them to the different databases. Articles identified in the reference lists of the articles included in this review were analysed manually by two authors to find additional articles in accordance with the inclusion and exclusion criteria.
PubMed, PsycINFO, and Cochrane CENTRAL (Cochrane Central Register of Controlled Trials) were searched in November 2024 using the specified search strategy.
2.2. Study selection
All citations were imported into Rayyan (13). After importing the citations into the platform, the Rayyan software was used to identify duplicates and to support the independent screening process by the authors.
Articles were analysed and selected independently by two authors according to previously defined inclusion and exclusion criteria, based on the title and abstract and subsequently on the full text. Disagreements were resolved by consensus.
We included articles that met the following criteria: (i) randomised controlled trials (RCTs), quasi-experimental studies, observational studies and mixed-methods studies; (ii) an adult population (≥18 years); (iii) intervention involving arts programmes prescribed by any health professional; (iv) outcomes including changes in patients’ levels of depression, anxiety, and wellbeing measured by validated scales or questionnaires. We excluded articles that were: (i) published in 1990 or earlier; (ii) studies in which the intervention consisted of the prescription of art as an adjunct to other therapies. No language restrictions were applied. We accepted the standard of care or any other intervention as a comparator strategy. Only studies published from 1990 were included, since AoP programmes became established in the mid-1990s. Previous studies were excluded because it is thought that they may not adequately reflect the properly structured intervention that is intended to be studied.
Mixed-methods studies were eligible for inclusion according to the same criteria. For the analyses, we focused exclusively on the quantitative component of such studies, corresponding to the study designs defined in the inclusion criteria (RCTs, quasi-experimental studies, and observational studies). This approach is consistent with established practice in systematic reviews that include mixed-methods studies (14, 15).
Articles in which the intervention was art therapy—a type of psychotherapy that involves an art therapist working with individuals or small groups, using visual and tactile media as a way to express themselves (2, 16)—were excluded.
We contacted the authors of articles for which full texts were unavailable, or in which only a protocol was identified, to determine whether the respective study had been completed and published.
2.3. Data extraction, quality assessment, and data synthesis
Data extraction was performed independently by three authors, with two authors assigned to each article. The data to be extracted were pre-defined and included study author, publication year, study design, country (and city), total sample size, sample size by group, loss to follow-up, mean age and standard deviation (SD), gender (number and percentage of females), referral process, intervention type, intervention duration, control group or comparator, outcome measurement methods (including reporting format), follow-up time, changes in patients’ levels of depression, anxiety, and wellbeing, other outcomes, missing outcome data, and any other relevant data. Any disagreements arising during this process were resolved by consensus.
Two authors analysed the articles and determined the study design according to their characteristics, reaching consensus. We used the Joanna Briggs Institute (JBI) critical appraisal tools to evaluate the quality of the studies according to their design: RCTs (17), quasi-experimental (18), or observational (19) studies, as appropriate. These JBI checklists are specific to the different study designs and comprise multiple assessment parameters recorded as “yes,” “no,” “unclear,” or “not applicable,” with space for comments. This assessment was performed independently by three authors, with two authors assigned to each article. Discrepancies were resolved by consensus.
A narrative synthesis was then performed. Data were presented in a table organised according to the different topics extracted. The results were also presented in different sections according to the effect of AoP: wellbeing, anxiety, depression, and other relevant effects.
3. Results
3.1. Study selection
The PRISMA flow diagram (Figure 1) illustrates the search process. Following the search of the databases, 4,041 articles were identified, of which 480 were duplicates and were deleted. Thus, 3,561 articles were analysed. Of these, 3,402 were excluded based on the title and abstract, according to the inclusion and exclusion criteria. In addition, 7 protocols that corresponded to previously selected studies, 2 protocols from unpublished studies, and 9 protocols/studies for which the author did not respond were excluded. Thus, 141 articles were analysed in the second phase, based on the full text. Of these, 134 were excluded, and 7 were included.
Figure 1.
PRISMA flow diagram.
Upon analysing the reference lists of the included articles, we identified two additional articles that met the inclusion and exclusion criteria. Three of the articles included are part of the same programme (20–22), and one of them (20) reports preliminary results from the cohort analysed in another study (21), which followed the cohort over a longer period. Therefore, both articles (20, 21) were analysed together. In total, nine articles, corresponding to eight studies, were included in our systematic review.
3.2. Assessment of risk of bias
The quality of the articles included in this review was assessed using the JBI critical appraisal tools. The results are presented in Table 1.
Table 1.
Quality assessment.
| Author(s) (reference) | JBI critical appraisal checklist for cohort studies | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Question 1 | Question 2 | Question 3 | Question 4 | Question 5 | Question 6 | Question 7 | Question 8 | Question 9 | Question 10 | Question 11 | Total | |
| Crone et al. (20, 21) | No | Yes | Yes | Yes | No | Not applicable | Yes | Yes | No | No | Yes | 6/11 |
| Sumner et al. (22) | Not applicable | Not applicable | Yes | Yes | Yes | Not applicable | Yes | Yes | Yes | No | Yes | 7/11 |
| Holt et al. (27) | Not applicable | Not applicable | Yes | No | No | Not applicable | Yes | Yes | No | No | Yes | 4/11 |
| JBI critical appraisal checklist for quasi-experimental studies | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Question 1 | Question 2 | Question 3 | Question 4 | Question 5 | Question 6 | Question 7 | Question 8 | Question 9 | Total | |
| Thomson et al. (23) | Yes | No | Yes | Not applicable | Wellbeing - No | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - Yes | 6/9 |
| Poulos et al. (28) | Yes | No | Yes | Not applicable | Wellbeing - No; Levels and frequency of creativity - No; Frailty - Yes | Wellbeing-Yes; Levels and frequency of creativity - Yes; Frailty - Yes | Wellbeing - Yes; Levels and frequency of creativity - Yes; Frailty - Yes | Wellbeing - No; Levels and frequency of creativity - Yes; Frailty - Yes | Wellbeing - Yes; Levels and frequency of creativity - Yes; Frailty - Yes | 5/9 |
| Van de Venter et al. (24) | Yes | No | Yes | Not applicable | Wellbeing - No | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - Yes | 6/9 |
| Thomson et al. (25) | Yes | No | Yes | Not applicable | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - Yes | 7/9 |
| Vogelpoel et al. (26) | Yes | No | Yes | Not applicable | Wellbeing - No | Wellbeing - Yes | Wellbeing - Yes | Wellbeing - No | Wellbeing - Yes | 5/9 |
The mean quality score for the three cohort studies was 5.7 out of 11 (JBI cohort tool), with the main sources of bias being: incomplete follow-up (JBI item 9), absence of strategies to address missing data (item 10), and limited identification and control of confounding factors (items 4–5).
The mean quality score for the five quasi-experimental studies was 5.8 out of 9 (JBI quasi-experimental tool), and the main sources of bias were: the absence of a control group (item 2), lack of multiple outcome measurements (item 5), and absence of strategies to address missing data (item 8).
The mean quality score for the study analysing anxiety and depression was 7 out of 11 (JBI Cohort tool), with the main source of bias was the absence of strategies to address incomplete follow-up (item 10).
3.3. Study characteristics
The characteristics of the studies are presented in Table 2 and Supplementary file 2. The studies were published between 2014 and 2021. Seven of the studies (88%) were conducted in the United Kingdom (20–27) [four of these were specifically conducted in England (20–22, 25, 27)]. One of the studies (13%) was conducted in Australia (28).
Table 2.
Study characteristics.
| Author(s) (reference) | Study year | Study design | Country (city) | Sample size (total and by group) (N) | Follow-up losses (N) | Age (years), mean (SD) | Gender (female), N (%) |
|---|---|---|---|---|---|---|---|
| Crone et al. (20, 21) | 2018 (DC 2009–2016) | Observational | England (Southwest) | 1,297. Attenders (referred and attended): 818 (completed: 651). Non-attenders (referred but did not attend): 440. | 39 | 51.1 (15.87) | 980 (77.0) |
| Thomson et al. (23) | 2020 (DC February–April 2018) | Quasi-experimental, sequential mixed methods | United Kingdom | 20. NA | 0 | 53 (44–70) | N* |
| Poulos et al. (28) | 2019 | Quasi-experimental, based on mixed methods | Australia (Sydney) | 127. NA | 20 | 78.1 (7.99) | 94 (74.0) |
| Van de Venter et al. (24) | 2014 | Quasi-experimental, based on mixed methods | United Kingdom | 44. NA | 0 | 43 (27–73) | 36 (82) |
| Sumner et al. (22) | 2021 (DC 2017–2019) | Observational | England (Southwest, Gloucestershire) | 245. NA | 0 | 50.5 (15.71) | 196 (80) |
| Thomson et al. (25) | 2018 (DC 2015–2017) | Quasi-experimental, based on mixed methods | England (Central London and Kent) | 115. NA | 0 | 65–94 | 72 (63) |
| Vogelpoel et al. (26) | 2014 | Quasi-experimental, based on mixed methods | United Kingdom | 12. NA | 4 | Mean age of over 80 (61–95) | 9 (75) |
| Holt NJ (27) | 2020 (DC September 2017–July 2019) | Observational | England (Bristol) | 66. NA | 0 | 47 (25–75) | 58 (88) |
| Author(s) (reference) | Intervention type | Intervention duration/follow-up time (in weeks) | Control group or comparator | Changes in patients’ levels of depression |
|---|---|---|---|---|
| Crone et al. (20, 21) | Led artist, activities: poetry, ceramics, drawing, mosaic and painting; Groups 3–10. | 8 or 10 | B/A | NA |
| Thomson et al. (23) | Led by a horticultural specialist, an art tutor and a museum volunteer. Activities: Combination of outdoor horticultural activities and indoor nature-based creative activities. Weekly. | 10 | B/A | NA |
| Poulos et al. (28) | Led by professional artists with the support of a community care worker or volunteer. Courses: visual arts, photography, dance and movement, drama, singing and music. Groups: 6–8. | 8–10 | B/A | NA |
| Van de Venter et al. (24) | Led artist. Activities: painting, textiles, music, photography and film. | 20 | B/A | NA |
| Sumner et al. (22) | Activities: visual arts (painting, ceramics, mosaics, photography) or performing arts (playwrighting, creative writing, singing). Weekly. | 8 | B/A (both initial referral and re-referral) | Method: Self-reported; Instrument: PHQ-8; Results: ↓ (initial referral: 13.4 ± 6.46 vs. 11.5 ± 6.45, p < 0. 001; re-referral: 13.2 ± 6.23 vs. 10.7 ± 6.21, p < 0. 001; across both cycles: 14.0 ± 6.37 vs. 11.28 ± 6.07, p < 0.001). In the multimorbid subgroup, depression also ↓ (initial referral: 14.3 ± 6.00 vs. 11.7 ± 6.10, p < 0.001; re-referral: 13.9 ± 6.12 vs. 10.8 ± 5.57, p = 0.002; across both cycles: 14.5 ± 6.29 vs. 11.2 ± 5.36, p = 0.001). |
| Thomson et al. (25) | Led by museum staff and volunteers. Activities: curator talks, behind-the scenes tours, object handling and discussion, and art activities inspired by the exhibits. | 10 | B/A | NA |
| Vogelpoel et al. (26) | Led by visual and tactile arts facilitators, supported by sense support staff and communicator guides. | 12 | B/A | NA |
| Holt NJ (27) | Led by skilled arts and health practitioners. Activities: art workshops. Weekly. | Thirty participants attended a 12-week programme, and thirty-six completed two programmes. | B/A | NA |
| Author(s) (reference) | Changes in patients’ levels of anxiety | Changes in patients’ levels of well-being | Outcome missing data, N (%) |
|---|---|---|---|
| Crone et al. (20, 21) | NA | Method: Self-reported; Instrument: WEMWBS; Results: ↑ (who attended: 38.1 ± 9.59 vs. 44.6 ± 9.84, p < 0.001; who were engaged: 38.0 ± 9.61 vs. 44.6 ± 9.79, p < 0.001; multimorbidity subsample: 36.7 ± 9.94 vs. 42.8 ± 9.32, p < 0.001; all categories of participation: 37.8 ± 9.63 vs. 44.4 ± 9.98, p < 0.001). | 0 (0). |
| Thomson et al. (23) | NA | Method: Self-reported; Instrument: UCL Museum Wellbeing Measure, specifically the positive generic wellbeing measure; Results: ↑ (16.70 vs. 25.30, p < 0.001). Each of the individual mood items also ↑. | 0 (0). |
| Poulos et al. (28) | NA | Method: Self-reported; Instrument: WEMWBS; Results: ↑ (mean increase: 6.86, 95% CI: 5.33–8.38, p < 0.001). | 17 in wellbeing data (percentage not available). |
| Van de Venter et al. (24) | NA | Method: Self-reported; Instrument: WEMWBS; Results: ↑ (B/A: 38.2 vs. 46.2; after a mean of 14 sessions: mean increase 8.0, 95%CI: 4.8–11.2, p < 0.0001). | 0 (0). |
| Sumner et al. (22) | Method: Self-reported; Instrument: GAD-7; Results: ↓ (initial referral: 11.9 ± 6.00 vs. 9.6 ± 5.80, p < 0.001; re-referral: 11.7 ± 5.87 vs. 9.5 ± 6.02, p < 0.001; across both cycles: 11.8 ± 6.05 vs. 9.9 ± 6.02, p = 0.008). In the multimorbid subgroup, anxiety also ↓ (initial referral: 12.0 ± 5.99 vs. 9.8 ± 5.33, p = 0.001; re-referral: 11.7 ± 5.77 vs. 9.5 ± 5.81, p = 0.010; across both cycles: 12.1 ± 6.44 vs. 9.8 ± 5.74, p = 0.036). | Method: Self-reported; Instrument: WEMWBS; Results: ↑ (initial referral: 37.1 ± 9.71 vs. 41.9 ± 10.40, p < 0.001; re-referral: 38.6 ± 9.92 vs. 42.4 ± 10.16, p < 0.001; across both cycles: 37.4 ± 9.77 vs. 42.3 ± 10.17, p < 0.001). In the multimorbid subgroup, wellbeing also ↑ (initial referral: 37.3 ± 8.84 vs. 43.0 ± 8.89, p < 0.001; re-referral: 38.4 ± 9.94 vs. 42.6 ± 10.12, p = 0.001; across both cycles: 36.6 ± 8.6 vs. 42.4 ± 10.11, p < 0.001). | 0 (0). |
| Thomson et al. (25) | NA | Method: Self-reported; Instrument: MwM-OA; Results: All emotions ↑ from pre- to post-session (start-programme: F(1,88) = 72.228, p < 0.001, partial eta squared = 0.451; mid-programme: F(1,83) = 67.651, p < 0.001, partial eta squared = 0.449; end-programme: F(1,76) = 54.689, p < 0.001, partial eta squared = 0.418) | 0 (0). |
| Vogelpoel et al. (26) | NA | Method: Self-reported; Instrument: WEMWBS; Results: mean score ↑ (from 41 to 47; positive gains were made in 11 of the 14 areas). | 0 (0). |
| Holt NJ (27) | NA | Method: Self-reported; Instrument: WEMWBS; Results: ↑ (first programme: mean of 5.01 units; second programme: mean of 4.73) | Not specified. |
B/A: before versus after design. DC: data collection. GAD-7: generalised anxiety disorder scale. MwM-OA: museum wellbeing measure for older adults. NA: not applicable. PHQ-8: patient health questionnaire eight-item version. ↑: increase. ↓: decrease. * similar. In mixed-method studies, we focused our analysis on the quantitative phase. Referral process for all studies: Referral person.
Although our analysis focused on the quantitative phase of the studies, only three of them employed exclusively quantitative methods (20–22, 27) and were observational studies. Five studies adopted a mixed-methods approach (23–26, 28), and their quantitative components were designed as single-arm quasi-experimental studies. In all studies, the sample underwent AoP and the outcomes were compared before and after the intervention.
The sample size ranged from 12 to 1,297 participants. The mean age of participants ranged from 43 to over 80 years old. Seven studies had a predominance of female participants (20–22, 24–28), while one study (23) had a similar number of participants of each sex. In three studies (20, 21, 26, 28), loss to follow-up was reported (between four and 39 participants).
In all studies, patients were referred by a healthcare professional who considered that the patients could benefit from the intervention and had the necessary conditions to participate. The interventions included a range of activities such as poetry, ceramics, drawing, mosaics, painting, textiles, photography, film, dance, movement, drama, singing, music, playwriting, and creative writing. Two of the studies involved a museum-based intervention (23, 25), one of which combined outdoor horticultural activities with indoor, nature-based creative activities (23).
Programmes were led by artists (20, 21, 24, 28), horticultural specialists, art tutors, and museum volunteers (23), museum staff and volunteers (25), visual and tactile arts facilitators (26), and skilled arts and health practitioners (27). One study (22) did not provide information about who led the intervention. Only two studies (20, 21, 28) provided information on the number of participants that could participate in each programme [3–10 participants (20, 21) and 6–8 participants (28)].
The intervention duration and follow-up times varied across studies: 8 weeks (22), 10 weeks (23, 25), 8 or 10 weeks (20, 21, 28), 20 weeks (24), and 12 weeks (26, 27). One of these studies allowed participants to take part in two programmes (27).
Wellbeing was measured in all studies. Six studies (20–22, 24, 26–28) used the Warwick–Edinburgh Mental Wellbeing Scale (WEMWBS) (29, 30), one study (23) used the UCL Museum Wellbeing Measure, specifically the Positive Generic Wellbeing Measure (31, 32), and one study (25) used the Museum Wellbeing Measure for Older Adults (MwM-OA) (31, 32). One study (22) measured anxiety symptoms using the Generalised Anxiety Disorder Scale (GAD-7) (33). One study (22) measured depression symptoms using the eight-item Patient Health Questionnaire (PHQ-8) (34).
Levels and frequency of creativity, and measures of frailty criteria, were also assessed in one study (28). Individual mood was self-reported using the Short Mood Scale (SMS) (35, 36) in one study (27). All of the studies measured outcomes before and after the intervention, and only one reported missing data (28). Two studies (21, 22) analysed these measures in a multimorbid subgroup.
3.4. The effect of arts on prescription on wellbeing
Wellbeing increased in all studies (20–28), as reported in the “Changes in patients’ levels of wellbeing” section of Table 2. The effect of AoP on wellbeing was evaluated, using one of three tools: the WEMWBS (20–22, 24, 26–28), the UCL Museum Wellbeing Measure, specifically the Positive Generic Wellbeing Measure (23), and the MwM-OA (25).
Significant increases in wellbeing were reported over time (β = 0.85; 95% confidence interval (CI) 0.45–1.25; p < 0.001), with scores improving at first programme (mean increase: 5.01), at second programme (mean increase: 4.73) and across both programmes (p < 0.001) (27). Significant increases was also reported in the other study (22) across both referral cycles (p < 0.001). However, no difference in the increase in mean WEMWBS scores was found according to the number of courses attended (28).
Two studies (21, 22) analysed a subsample of participants with multimorbidity and found a significant increase in WEMWBS scores [(21): 36.7 ± 9.94 vs. 42.8 ± 9.32, p < 0.001; (22) at initial referral: 37.3 ± 8.84 vs. 43.0 ± 8.89, p < 0.001; (22) at re-referral: 38.4 ± 9.94 vs. 42.6 ± 10.12, p = 0.001; (22) across both referrals: 36.6 ± 8.6 vs. 42.4 ± 10.11, p < 0.001], though this increase was more modest than that observed in the total sample (21).
The results obtained were influenced by the characteristics of the programme. Crone et al. (20, 21) reported that participants in the eight-week programme experienced a greater increase in WEMWBS scores (8-week programme: 37.8 ± 9.18 vs. 43.9 ± 9.65, p < 0.001; 10-week programme: 38.6 ± 10.19 vs. 45.7 ± 10.62, p < 0.001), were more likely to complete the programme, and demonstrated greater engagement than those in the 10-week programme. The impact of patients having a broader choice of intervention was also analysed. Those referred under these conditions were more engaged, however no statistically significant differences in the rate of change in wellbeing were observed (after the change: 38.3 ± 8.95 vs. 44.6 ± 9.65, p < 0.001; before the change: 37.5 ± 10.34 vs. 44.4 ± 10.56, p < 0.001).
The impact of the intervention is also influenced by individual factors. Those who completed the intervention or were classified as engaged reported higher baseline WEMWBS scores and differences between those who attended and others (by occupation, mean number of referral reasons, length of the referral course, and reason for referral) and between those who engaged and others (by occupation and mean number of referral reasons) were observed (20, 21).
The increase was slightly greater for women (women: 8.1, 95% CI 4.3–12.0; men: 5.7, 95% CI 0.4–10.8) and for Black and Minority Ethnic participants (Black and Minority Ethnic participants: 9.8, 95% CI 0.3–19.5; White British participants: 6.4, 95% CI 2.4–10.5), and slower among participants with lower baseline scores (24).
33% of the variation in WEMWBS scores could be explained by gender, ethnicity, baseline scores, and the number of sessions attended (24). In fact, the expected increase in wellbeing scores was not experienced by all individuals (with varying increases, maintenance, and decreases in WEMWBS scores over time), but the variation in slopes was not statistically significant (β = 0.35; 95% CI 0.05–2.28; p = 0.29) (27). However, a relatively low level of wellbeing was observed in one study (26), with a mean score of 47 compared to 49.9 in the general population in Scotland.
Thomson et al. (23) analysed global changes in the UCL Museum Wellbeing Measure scores and changes in each individual mood item assessed by participants using this tool. The results showed a significant increase for each mood item, with no significant differences between them. In another study (25), all emotions also increased significantly from pre- to post-session at the start, mid- and end-programme. Significant increases at items “Excited” (23), “enlightened” and “absorbed” (25) were observed. However, Vogelpoel et al. (26) observed that aspects such as “feeling optimistic,” “dealing well with problems”, and “feeling good about oneself” were assessed as slightly lower at the post-intervention evaluation than at the pre-intervention evaluation. In addition, the authors reported that the largest positive change for the group was “feeling more relaxed”, as this increased from “rarely” to “some of the time”.
In one study (25), multivariate analysis of variance (MANOVA) showed a highly significant main effect of programme, session, and emotion. A significant interaction between session and emotion was observed, but not between the programme and session, programme and emotion, or programme, session, and emotion when evaluated simultaneously. Comparing the start- and mid-programme measures, a highly significant difference between pre-session wellbeing scores (p < 0.002) and between post-session wellbeing scores (p < 0.036) was observed, but no significant differences were observed between mid- and end-programme.
3.5. The effect of arts on prescription on anxiety
The effect of AoP on anxiety was evaluated in one study (22). The results are presented in Table 2 (“Changes in patients’ levels of anxiety” section). The Generalised Anxiety Disorder Scale (GAD-7) was used to measure anxiety levels at the beginning and end of each referral cycle through self-reporting. These assessments were then compared between pre- and post-intervention at initial referral, re-referral, and across both referrals.
A decrease in anxiety symptoms was observed in all comparisons (at initial referral: 11.9 ± 6.00 vs. 9.6 ± 5.80, p < 0.001; at re-referral: 11.7 ± 5.87 vs. 9.5 ± 6.02, p < 0.001; across both referrals: 11.8 ± 6.05 vs. 9.9 ± 6.02, p = 0.008). The minimal clinically important difference (MCID) for the GAD-7 (four points) was not met, and significant differences were observed in the clinical categorisation of the scales (p < 0.001; scores ≥5, ≥10, and ≥15 indicating mild, moderate, and severe anxiety symptoms, respectively).
This study analysed a multimorbid subgroup and observed a decrease in anxiety symptoms across all comparisons (at initial referral: 12.0 ± 5.99 vs. 9.8 ± 5.33, p = 0.001; at re-referral: 11.7 ± 5.77 vs. 9.5 ± 5.81, p = 0.010; across both referrals: 12.1 ± 6.44 vs. 9.8 ± 5.74, p = 0.036).
3.6. The effect of arts on prescription on depression
The same study (22) evaluated the effect of AoP on depression. The results are presented in Table 2 (“Changes in patients’ levels of depression”). Depression symptoms were self-reported using the PHQ-8 before and after the intervention for each referral cycle.
A significant decrease was observed when the assessments were compared between pre- and post-intervention at initial referral (13.4 ± 6.46 vs. 11.5 ± 6.45, p < 0.001), at re-referral (13.2 ± 6.23 vs. 10.7 ± 6.21, p < 0.001), and between the pre-intervention (initial cycle) and post-intervention (re-referral; 14.0 ± 6.37 vs. 11.28 ± 6.07, p < 0.001).
The MCID for PHQ-9 (five points) was not met, and significant differences were observed in the clinical categorisation of the scales (p < 0.001; scores 0–4 indicating no significant depressive symptoms, 5–9 indicating mild symptoms, 10–14 indicating moderate symptoms, 15–19 indicating moderately severe symptoms, and 20–24 indicating severe symptoms).
As with anxiety, a decrease in depression symptoms was observed in a multimorbid subgroup, at initial referral (14.3 ± 6.00 vs. 11.7 ± 6.10, p < 0.001), at re-referral (13.9 ± 6.12 vs. 10.8 ± 5.57, p = 0.002), and when comparing the pre-intervention (initial cycle) and post-intervention (re-referral; 14.5 ± 6.29 vs. 11.2 ± 5.36, p = 0.001).
3.7. Other relevant effects of arts on prescription
In one study (27), the interruption of the programme was associated with a decrease in wellbeing. The same was observed when participants were re-referred, with increases in anxiety and depression levels and a decrease in wellbeing (22). Sumner et al. (22) observed that the baseline measures of anxiety, depression, and wellbeing were significantly associated with the outcome.
Even though the evaluation focused on wellbeing, Poulos et al. (28) also analysed levels and frequency of creativity, and measures of frailty criteria. No statistically significant differences were observed in the proportion of participants scoring on each of the frailty criteria between pre- and post-intervention. However, a statistically significant increase was observed in both the level of creativity (mean difference = 1.56, 95% CI 1.44–1.98, p < 0.001) and the frequency of creativity (mean difference = 1.60, 95% CI 1.06–2.14, p < 0.001).
In one study (27), individual mood was self-reported using the SMS, and the participants reported feeling significantly calmer, more relaxed, alert, energetic, content, and well, with mood significantly predicted by time. A larger reduction in tense arousal after art making was associated with increases in global wellbeing over time (γ = 0.41, 95% CI 0.07–0.76, p = 0.019). Changes in energetic arousal and hedonic tone were not significant predictors.
4. Discussion
4.1. Summary of main findings
This systematic review synthesised evidence from eight studies on the impact of AoP on wellbeing, depression, and anxiety in adults. Wellbeing improved consistently across all eight studies. Evidence for the reduction of depression and anxiety symptoms derives from a single study (22), in which statistically significant reductions were observed but the minimal clinically important difference was not met for either outcome.
Multimorbid individuals, who have multiple medical conditions and take multiple drugs, have also benefited from AoP programmes (21, 22). Therefore, this intervention could complement existing treatments and provide a more holistic approach to caring for people in these complex situations.
In one study (26), the level of wellbeing after the intervention was relatively low compared with the general population score in Scotland. This could be explained by the fact that the sample had lower levels of wellbeing than the general population, and that the intervention may have been insufficient as a stand-alone treatment for the patients’ condition. The fact that some aspects remained problematic for the participants (26) also highlights the idea that AoP programmes may not be sufficient as a stand-alone intervention.
Participants showed greater engagement when they had more freedom to choose the characteristics of the programme, including art type and location (21). This should be considered when designing and commissioning AoP programmes.
Characteristics such as gender, ethnicity, and baseline scores influenced the outcomes (22, 24). These factors should be considered when developing practical prescribing guidelines. Future studies should investigate additional patient characteristics that may influence outcomes.
4.2. Strengths and limitations
No randomised controlled trials (RCTs) were included, and none of the studies had a control group. Therefore, the observed results could be attributable to chance or to other factors rather than the intervention alone.
A further critical limitation is the substantial heterogeneity of both interventions and outcome measures across studies. The AoP programmes included in this review varied considerably in their arts activities (visual arts, music, creative writing, museum-based activities), session frequency, group size, duration (8 to 20 weeks), and facilitator background. This heterogeneity is a recognised challenge in arts and health research: Jensen, Bungay and Holt (37) emphasise that more specific data is necessary to enable the comparison of similar studies. Similarly, the use of three different wellbeing instruments across the included studies (WEMWBS, UCL Museum Wellbeing Measure, MwM-OA), each with distinct conceptual frameworks and psychometric properties, limits the comparability of wellbeing outcomes across studies. The small number of studies analysed, most of which were carried out in the United Kingdom (20–27), could affect the generalisability of the results. Small sample sizes (23, 24, 26) and interventions of short duration are also limitations.
The methodological quality of the included studies was limited by the absence of control groups, limited identification and control of confounding factors, incomplete follow-up, a lack of strategies to address missing data, and a lack of multiple outcome measurements. These limitations should be considered when interpreting the present findings.
4.3. Literature gap and further research
Although the results of this systematic review are promising, there is still a significant lack of research in this field. Randomised controlled trials and other rigorous study designs should be developed with the aim of including control groups. This would enable a clearer assessment of causal relationships between the intervention and the outcomes. Additionally, larger and more representative samples should be used to enable the results to be more widely generalised. Future research should also include longer follow-up periods to determine whether the improvements observed at the end of the intervention are sustained over time.
In addition, future studies should consider the identification and control of confounding factors, strategies to minimise loss to follow-up, avoidance of baseline differences between groups, and the inclusion of multiple outcome measurements.
As only one study has investigated the effects of AoP on anxiety and depression, these outcomes should be prioritised in future research.
5. Conclusion and future directions
We conclude that AoP programmes are consistently associated with improvements in wellbeing across a range of populations and primary care settings, and that preliminary evidence suggests potential benefits for symptoms of depression and anxiety that warrant investigation in more rigorous study designs. Further studies with more robust designs, larger samples, and longer follow-up periods are needed to better understand the conditions under which AoP is most effective and to inform its implementation in clinical practice and health systems. Referral guidelines should also be developed to help clinicians identify patients who may benefit most from this intervention.
Acknowledgments
We would like to express our deepest gratitude to Professor Dr. John Yaphe for his valuable contribution to the final review of the systematic review.
Funding Statement
The author(s) declared that financial support was received for this work and/or its publication. This article was supported by National Funds through FCT – Fundação para a Ciência e a Tecnologia, I.P., within the project RISE-Health - UID/06397/2025.
Edited by: Claudia Chiavarino, Istituto Universitario Salesiano Torino Rebaudengo (IUSTO), Italy
Reviewed by: Leonhard Thun-Hohenstein, Paracelsus Medical University, Austria
Shannastaniar Aisya Adif, State University of Malang, Indonesia
Abbreviations: AoP, arts on prescription; CI, confidence interval; GAD-7, generalised anxiety disorder scale; GP, general practitioner; JBI, Joanna Briggs Institute; MANOVA, multivariate analysis of variance; MCID, minimal clinical important difference; MwM-OA, museum wellbeing measure for older adults; NHS, National Health Service; PHQ-8, patient health questionnaire eight-item version; PRISMA, preferred reporting items for systematic reviews and meta-analyses; RCT, randomised control trial; SD, standard deviation; SMS, short mood scale; WEMWBS, Warwick–Edinburgh mental wellbeing scale.
Data availability statement
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.
Author contributions
ÉS: Resources, Formal analysis, Project administration, Writing – review & editing, Visualization, Methodology, Writing – original draft, Investigation, Validation, Supervision, Conceptualization, Data curation. MM-S: Writing – review & editing, Validation, Conceptualization, Methodology, Data curation. MR: Resources, Investigation, Validation, Writing – review & editing, Conceptualization, Methodology, Data curation. AB: Validation, Conceptualization, Methodology, Writing – review & editing. JH: Conceptualization, Validation, Writing – review & editing, Methodology. DL: Methodology, Validation, Conceptualization, Writing – review & editing. LM: Writing – review & editing, Data curation, Methodology, Supervision, Investigation, Validation, Conceptualization, Project administration. SB: Conceptualization, Supervision, Writing – review & editing, Investigation, Data curation, Validation, Project administration, Methodology.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that Generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpubh.2026.1833798/full#supplementary-material
References
- 1.Fancourt D, Finn S. What is the Evidence on the Role of the Arts in Improving Health and Well-Being? A Scoping Review. Copenhagen: WHO Regional Office for Europe; (2019). p. 9–10. [PubMed] [Google Scholar]
- 2.Bungay H, Clift S. Arts on prescription: a review of practice in the UK. Perspect Public Health. (2010) 130:277–81. doi: 10.1177/1757913910384050 [DOI] [PubMed] [Google Scholar]
- 3.Fancourt D, Warran K, Aughterson H. "Arts in social prescribing and health outcomes". In: Evidence Summary for Policy: The Role of Arts in Improving Health and Wellbeing: Report to the Department for Digital, Culture, Media & Sport. London: University College London; (2020). p. 13–4. [Google Scholar]
- 4.South J, Higgins TJ, Woodall J, White SM. Can social prescribing provide the missing link? Prim Health Care Res Dev. (2008) 9:310–8. doi: 10.1017/S146342360800087X [DOI] [Google Scholar]
- 5.Bickerdike L, Booth A, Wilson PM, Farley K, Wright K. Social prescribing: less rhetoric and more reality. A systematic review of the evidence. BMJ Open. (2017) 7:e013384. doi: 10.1136/bmjopen-2016-013384, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Chatterjee H, Polley M, Clayton G. Social prescribing: community-based referral in public health. Perspectives. Public Health. (2018) 138:18–9. doi: 10.1177/1757913917736661, [DOI] [PubMed] [Google Scholar]
- 7.Bungay H, Jensen A, Holt N. Critical perspectives on arts on prescription. Perspectives. Public Health. (2024) 144:363–8. doi: 10.1177/17579139231170776, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Opher S. Cost-Benefit Evaluation of Artlift 2009–2012: Summary. Gloucester: Gloucestershire County Council and NHS Gloucestershire; (2011). [Google Scholar]
- 9.The Government Office for Science . Foresight Mental Capital and Wellbeing Project: Final Project Report. London: The Government Office for Science; (2008). p. 10. [Google Scholar]
- 10.World Health Organization. (n.d.). Mental health. Available online at: https://www.who.int/health-topics/mental-health#tab=tab_2 (Accessed July 20, 2024)
- 11.World Mental Health Report: Transforming Mental Health for All. Geneva: World Health Organization; (2022). p. 5–40 Licence: CC BY-NC-SA 3.0 IGO. [Google Scholar]
- 12.Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. (2021) 372:n71. doi: 10.1136/bmj.n71 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan—a web and mobile app for systematic reviews. Syst Rev. (2016) 5:210. doi: 10.1186/s13643-016-0384-4, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, et al., editors. Cochrane Handbook for Systematic Reviews of Interventions version 6.5. London: Cochrane; (2024) Chapter 5. [Google Scholar]
- 15.Lizarondo L, Stern C, Carrier J, Godfrey C, Rieger K, Salmond S, et al. "Chapter 8.3: the JBI approach to mixed method systematic reviews". In: Aromataris E, Lockwood C, Porritt K, et al., editors. JBI Manual for Evidence Synthesis. Adelaide: JBI; (2024) [Google Scholar]
- 16.British Association of Art Therapists. Art therapy. Available online at: https://baat.org/art-therapy/ (Accessed January 30, 2025).
- 17.Barker TH, Stone JC, Sears K, Klugar M, Tufanaru C, Leonardi-Bee J, et al. The revised JBI critical appraisal tool for the assessment of risk of bias for randomized controlled trials. JBI Evid Synth. (2023) 21:494–506. doi: 10.11124/JBIES-22-00430, [DOI] [PubMed] [Google Scholar]
- 18.Barker TH, Habibi N, Aromataris E, Stone JC, Leonardi-Bee J, Sears K, et al. The revised JBI critical appraisal tool for the assessment of risk of bias for quasi-experimental studies. JBI Evid Synth. (2024) 22:378–88. doi: 10.11124/JBIES-23-00268, [DOI] [PubMed] [Google Scholar]
- 19.Moola S, Munn Z, Tufanaru C, Aromataris E, Sears K, Sfetcu R, et al. "Systematic reviews of etiology and risk (2020)". In: Aromataris E, Lockwood C, Porritt K, Pilla B, Jordan Z, editors. JBI Manual for Evidence Synthesis. Adelaide: JBI; (2024) [Google Scholar]
- 20.Crone DM, O'Connell EE, Tyson PJ, Clark-Stone F, Opher S, James DVB. 'Art lift' intervention to improve mental well-being: an observational study from U.K. general practice. Int J Ment Health Nurs. (2013) 22:279–86. doi: 10.1111/j.1447-0349.2012.00862.x, [DOI] [PubMed] [Google Scholar]
- 21.Crone DM, Sumner RC, Baker CM, Loughren EA, Hughes S, James DVB. 'Artlift' arts-on-referral intervention in UK primary care: updated findings from an ongoing observational study. Eur J Pub Health. (2018) 28:404–9. doi: 10.1093/eurpub/cky021 [DOI] [PubMed] [Google Scholar]
- 22.Sumner RC, Crone DM, Hughes S, James DVB. Arts on prescription: observed changes in anxiety, depression, and well-being across referral cycles. Public Health. (2021) 192:49–55. doi: 10.1016/j.puhe.2020.12.008 [DOI] [PubMed] [Google Scholar]
- 23.Thomson LJ, Morse N, Elsden E, Chatterjee HJ. Art, nature and mental health: assessing the biopsychosocial effects of a ‘creative green prescription’ museum programme involving horticulture, artmaking and collections. Perspect Public Health. (2020) 140:277–85. doi: 10.1177/1757913920910443, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Van de Venter E, Buller AM. Arts on referral interventions: a mixed-methods study investigating factors associated with differential changes in mental well-being. J Public Health. (2015) 37:143–50. doi: 10.1093/pubmed/fdu028 [DOI] [PubMed] [Google Scholar]
- 25.Thomson LJ, Lockyer B, Camic PM, Chatterjee HJ. Effects of a museum-based social prescription intervention on quantitative measures of psychological wellbeing in older adults. Perspect Public Health. (2018) 138:28–38. doi: 10.1177/1757913917737563, [DOI] [PubMed] [Google Scholar]
- 26.Vogelpoel N, Jarrold K. Social prescription and the role of participatory arts programmes for older people with sensory impairments. J Integr Care. (2014) 22:39–50. doi: 10.1108/JICA-01-2014-0002 [DOI] [Google Scholar]
- 27.Holt NJ. Tracking momentary experience in the evaluation of arts-on-prescription services: using mood changes during art workshops to predict global wellbeing change. Perspect Public Health. (2020) 140:270–6. doi: 10.1177/1757913920913060, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Poulos RG, Marwood S, Harkin D, Opher S, Clift S, Cole AMD, et al. Arts on prescription for community-dwelling older people with a range of health and wellness needs. Health Soc Care Community. (2019) 27:483–92. doi: 10.1111/hsc.12669, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Tennant R, Hiller L, Fishwick R, Platt S, Joseph S, Weich S, et al. The Warwick-Edinburgh mental well-being scale (WEMWBS): development and UK validation. Health Qual Life Outcomes. (2007) 5:63. doi: 10.1186/1477-7525-5-63, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Stewart-Brown S, Tennant A, Tennant R, Platt S, Parkinson J, Weich S. Internal construct validity of the Warwick-Edinburgh mental well-being scale (WEMWBS): a Rasch analysis using data from the Scottish health education population survey. Health Qual Life Outcomes. (2009) 7:15. doi: 10.1186/1477-7525-7-15, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Thomson L, Chatterjee H. Assessing well-being outcomes for arts and heritage activities: development of a museum well-being measures toolkit. J. Appl. Arts Health. (2014) 5:29–50. doi: 10.1386/jaah.5.1.29_1 [DOI] [Google Scholar]
- 32.Thomson LJ, Chatterjee HJ. Measuring the impact of museum activities on well-being: developing the museum well-being measures toolkit. Mus Manag Curat. (2015) 30:44–62. doi: 10.1080/09647775.2015.1008390 [DOI] [Google Scholar]
- 33.Spitzer RL, Kroenke K, Williams JB, Löwe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. (2006) 166:1092–7. doi: 10.1001/archinte.166.10.1092, [DOI] [PubMed] [Google Scholar]
- 34.Kroenke K, Strine TW, Spitzer RL, Williams JB, Berry JT, Mokdad AH. The PHQ-8 as a measure of current depression in the general population. J Affect Disord. (2009) 114:163–73. doi: 10.1016/j.jad.2008.06.026, [DOI] [PubMed] [Google Scholar]
- 35.Wilhelm P, Schoebi D. Assessing mood in daily life: structural validity, sensitivity to change, and reliability of a short-scale to measure three basic dimensions of mood. Eur J Psychol Assess. (2007) 23:258–67. doi: 10.1027/1015-5759.23.4.258 [DOI] [Google Scholar]
- 36.Kanning M, Ebner-Priemer U, Schlicht W. Using activity triggered e-diaries to reveal the associations between physical activity and affective states in older adult's daily living. Int J Behav Nutr Phys Act. (2015) 12:111. doi: 10.1186/s12966-015-0272-7, [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Jensen A, Bungay H, Holt NJ. Apples and oranges? The importance of recognizing heterogeneity in arts and health research publications. Front Public Health. (2025) 13:1670166. doi: 10.3389/fpubh.2025.1670166, [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.