What is art if not therapy? It’s easy to understand how engaging with art might improve well-being, but that isn’t the same as tackling serious mental health problems.

We’ve heard mixed messages in relation to art therapy for schizophrenia here in the woodland (Tomlin, 2012; Judge, 2014), but what about for non-psychotic mental health disorders? After all, depression, anxiety and phobias make up the majority of mental health problems and are not associated with psychosis.

A new HTA study addresses this, considering clinical effectiveness and cost-effectiveness (Uttley, 2015). Like most HTA studies, the analysis was multifaceted. The authors hoped to identify:

1. The clinical effectiveness of art therapy for people with non-psychotic mental health disorders
2. User and service provider perspectives on acceptability and potential harms and benefits of art therapy
3. The cost-effectiveness of art therapy for people with non-psychotic mental health disorders

What is art therapy?

Simply doing art is not art therapy. It can involve things like painting, clay work and the use of digital media as a form of nonverbal expression. Art therapy is seen as an option for people who have problems putting distress into words, or other problems with engagement. It’s important that it takes place alongside other modes of communication within a therapeutic relationship and in a therapeutic setting.

Variation in definition and delivery is surely one of the greatest challenges in the evaluation of psychological therapies. Unfortunately, like many complex interventions, art therapy lacks a clear single definition. The authors provide one for the purpose of their review, requiring that it involves:

• Definition of appropriate therapeutic boundaries
• Presence of a therapist, with an appropriate therapeutic relationship
• Therapeutic use of art materials
• Appropriateness to the cultural and service context

NICE recommends art therapy to be considered for people with psychosis and schizophrenia (NICE, 2014) and arts therapies are included in the autistic spectrum disorder Strategic Plan for Wales (Welsh Assembly Government, 2009). However, there are currently no national guidelines in the UK for the use of art therapy for non-psychotic mental disorders.

The literature search

The foundation of all 3 parts of the study is one mega literature search. 13 databases were searched along with 10 additional sources. The authors should be commended for including full details of their search terms. The searches allowed for all study types and used a flexible population definition, including mental health clinical samples (with or without formal diagnoses) as well as people with physical conditions where art therapy was used to reduce mental health symptoms. All possible comparators were considered and studies could be conducted in any setting.

The search identified 10,270 unique records. Studies conducted in people with a psychotic disorder were manually excluded through sifting, and each of the 3 stages of the study used different inclusion and exclusion criteria.

The authors describe their work as an ‘evidence portfolio’ because the research questions centre on the intervention rather than a specific population. However, the population is defined as ‘non-psychotic’ and therefore conditions like schizophrenia are excluded.

Quantitative review

Non-randomised studies were excluded from the quantitative analysis, which sought to identify treatment effects associated with art therapy. Of the 10,270 records, 49 full-text articles were assessed for eligibility. 18 references were included in the analysis, which were based on 15 different studies.

Art therapy had a positive impact on mental health symptoms in 10 out of the 15 studies, but the quality of the RCTs was generally low. Study characteristics included:

• Mean sample size of 52
• 11 of 15 studies examined group art therapy
• 3 of 15 studies targeted people with non-psychotic mental disorders
• 4 of 15 studies were in cancer patients
• 4 of 15 studies were in children or adolescents
• 5 of 15 studies involved only women; 1 only men
• Number of sessions ranged from 1 to 40 (mean of 9)

These facts highlight that there was a lot of heterogeneity. Primary diagnoses included cancer, dementia/Alzheimer’s, PTSD, depression, HIV/AIDS, asthma and sickle cell disease. Comparators included treatment as usual, wait-list, attention placebo, CBT and verbal dynamic psychotherapy. Therapists were variously described as ‘trained’, ‘licensed’, ‘qualified’, ‘certified’ and ‘registered’, while some studies made no reference to accreditation. 7 studies did not describe randomisation. The majority of studies had a high risk of bias.

It wasn’t possible to perform a formal pooled analysis as there were many different outcome measures used. The table below shows the number of studies identified as beneficial or otherwise for a selection of target symptoms. We can clearly see that most studies were finding art therapy to be beneficial, but many showed no statistically significant benefit.

Qualitative review

The qualitative review sought to investigate issues relating to acceptability for both patients and practitioners and how these might affect clinical effectiveness. 42 full-text articles were assessed for eligibility and 13 references (12 studies) were included in the analysis. 11 papers assessed patients’ attitudes and 2 assessed practitioner attitudes. In total the studies contained qualitative data from 188 patients and 16 health practitioners.

Benefits of art therapy identified included:

• Relationships
• Understanding
• Perspective
• Distraction
• Personal achievement
• Expression
• Relaxation
• Empowerment

Some respondents indicated that they did not think art therapy would be beneficial. In patients with cancer, a number of potential harms were identified including the potential for causing anxiety, increasing pain and leading to the activation of emotions that were not resolved.

Overall, art therapy was viewed as acceptable. In particular, a good relationship between the patient and art therapist was viewed as essential for effective treatment.

Cost-effectiveness analysis

Essentially, the economic evaluation was a bit of a flop. The authors did their best but the data just weren’t there; 26 full-text articles were assessed but only 1 study was identified as useful. Not only that, but this was a case study of a single patient.

Nevertheless, the authors developed a simple model to evaluate cost-effectiveness in terms of cost-per-QALY, with outcomes in the primary analysis based on a single study of mindfulness-based art therapy on 111 women with cancer in the US. The analysis assumed – based on data from the British Association of Art Therapists – that treatment would be 43 sessions in a year, at a cost of £750.

This analysis suggested that art therapy, compared with wait-list control, was very likely to be cost-effective.

Conclusion

The authors concluded that:

Art therapy appears to have statistically positive effects compared with controls in a number of studies in patients with different clinical profiles.

Furthermore, the report concludes that art therapy is reported to be an acceptable treatment associated with a number of benefits, and that art therapy appears to be cost-effective versus wait-list control.

Discussion

The authors acknowledge that their approach is non-standard in that they are considering an intervention rather than a population; they essentially review art therapy being used for any health problem. This is not adequately addressed by the methodology.

My conclusions would be more circumspect than the authors’. In regard to the review of effectiveness, the authors identify various potential sources of bias. I think publication bias could be playing an important role here, but this study does not reveal much in this regard. I also do not think a claim can be made that we would expect art therapy to be cost-effective. The analysis is based on a very specific population, and outcomes are based on a mapping approach that could be unreliable. Costs might also be far from realistic; it’s worth noting that the cost-effectiveness analysis assumed that 43 sessions would be carried out over 52 weeks, yet no study identified included this many and the mean number of sessions was 9.

The authors acknowledge that “it is difficult to define or exhaustively list comparators for art therapy”. This is a key requirement of decision analysis; we cannot make the best decisions without knowing about the full range of alternatives.

Art therapy has been heavily studied, but there remains a dearth of randomised trials. Art therapists are clearly engaged with the research, but there is a need to develop better designed and more standardised randomised studies to identify whether art therapy might be worth the investment.

Links

Primary paper

Uttley L, Scope A, Stevenson M, Rawdin A, Taylor Buck E, Sutton A, et al. Systematic review and economic modelling of the clinical effectiveness and cost-effectiveness of art therapy among people with non-psychotic mental health disorders. Health Technol Assess 2015;19(18) [PubMed]

Other references

Judge J. Art therapy for schizophrenia: an effective add-on treatment? The Mental Elf. 27 Jun 2014.

Tomlin A. Group art therapy does no better than standard care at improving symptoms for people with schizophrenia. The Mental Elf. 5 Mar 2012.

NICE (2014) Psychosis and schizophrenia in adults: treatment and management. NICE clinical guideline, CG178, 2014. (See recommendations section 1.4.4.3: “Consider offering arts therapies to all people with psychosis or schizophrenia, particularly for the alleviation of negative symptoms. This can be started either during the acute phase or later, including in inpatient settings”).

Welsh Assembly Government (2009) The Autistic Spectrum Disorder Strategic Action Plan for Wales (PDF). A report produced by the Adult Task and Finish Group, Oct 2009.