According to the latest global statistics, the prevalence rate of stroke was approximately 25.7 million and was the second leading cause of death in 2013 (Benjamin et al., 2017). Although older people have a higher risk of stroke, the onset of stroke among people aged 20–64 years has increased by 25% from 1990 to 2010 worldwide (Feigin et al., 2014). In the United States, around 795,000 people experience a new or recurrent stroke every year. Stroke is also the fifth leading cause of death and the leading cause of severe long-term disabilities in the United States (Benjamin et al., 2017). In Hong Kong, there are ~2 million patients with chronic illnesses and 2.2% of them are stroke patients (Census Statistics Department, 2015). Every year, ~3000 people die of stroke (Hospital Authority, 2018). Cerebrovascular diseases, including stroke, are the fourth leading causes of death in Hong Kong (Centre for Health Protection, 2018). These statistics indicate that the world is now facing a grave challenge with an increasing number of stroke survivors, thus more resources are needed for rehabilitation and post-stroke support.

Current rehabilitation for stroke survivors depends heavily on physiotherapies, occupational therapies, and speech therapies. These rehabilitation programs are essential for stroke survivors as they aid in functional recovery. Physiotherapy, also recognized as physical therapy, focuses on the movement and physical function of patients and aims at the maintenance, development, and restoration of mobility (World Confederation for Physical Therapy, 2017). The main goal of occupational therapy is to facilitate patients to participate in their daily life activities and enhance their self-autonomy (World Federation of Occupational Therapy, 2012). Speech therapy aims at treating the patients' speech, language, and swallowing problems and at enhancing the patients' verbal communication and swallowing abilities (American Speech-Language-Hearing Association, 2018). Functional recovery is of utmost importance in stroke rehabilitation, especially during the first 6 months after the onset of stroke. The sudden loss of physical ability as well as the changes in various aspects of life caused by stroke also trigger psychological distress, social withdrawal and confusion in the meaning in life (Knapp et al., 2000; Yeung et al., 2011), which together need to be further addressed. These psychosocial and spiritual needs are also suggested to be crucial for the quality of life for stroke survivors (Katona et al., 2015). Enhancing the post-stroke quality of life has therefore been increasingly emphasized in recent years, in particular, in the United States (American Heart Association, 2013) and Australia (Australian Stroke Foundation, 2008). Due to their non-intrusive processes, therefore, non-pharmacological approaches, in particular, have been drawing considerable attention. Among them, creative arts-based therapies, which are mostly playful and without side-effects, have been recommended for stroke survivors. Different art modalities are also being suggested to be useful in stimulating different parts of the brain, and the stimulations from diverse art forms may further enhance the neuroplasticity in the brain and this may be helpful for facilitating the recovery process after stroke (Demarin, 2017). In addition, neurological evidence has shown that regularly listening to music after stroke may lead to structural changes in the brain among stroke survivors, and these structural changes may, in turn, relate to improvements in cognition (Särkämö et al., 2014). Moreover, engaging in a regular visual arts intervention has also been proven to facilitate the spatial improvement in functional connectivity in certain parts of the brain which may be associated with the psychological resilience in adults (Bolwerk et al., 2014). Apart from the above evidence, dance movement has also been suggested as an innovative approach for the rehabilitation for stroke survivors, due to its nature of engaging both physical and cognitive functions, dance may thus have the potential to tackle both the physical and cognitive impairments simultaneously (Dhami et al., 2015).

There are various types of arts-based therapeutic approaches. While creative arts-based therapies emphasize specific art forms such as music therapy, art therapy, dance movement therapy, drama therapy, and expressive writing, other approaches such as expressive arts therapy make use of the integrative approach and multiple art forms. Although each approach has its own uniqueness and characteristics, all arts-based interventions have a common goal of providing stimulation of different sensations, cultivate a safe environment for self-exploration, and encourage self-expression, creativity, and imagination through the use of arts. Over the past decades, an increasing number of studies have been conducted in examining the outcomes of creative arts-based therapies for stroke survivors. Several systematic reviews and literature reviews on the outcomes of specific types of creative arts-based therapies for this population have also been published. For instance, Strzemecka (2013) completed a review of music therapy in stroke rehabilitation, focusing on the role of different types of music therapy in stroke rehabilitation. Reynolds (2012) also published a review of art therapy for stroke survivors, which gathered both quantitative and qualitative findings from various studies. Nevertheless, these review articles only focused on a single art modality approach. In this qualitative systematic review, however, different forms of creative arts-based therapies were included, as there appears to be a paucity of transparent understanding of the beneficial outcomes of different creative arts-based therapies. With a better understanding of the characteristics and specialties of therapies using different art modalities, art therapists, service providers, and users, may afford a better appreciation on how creative arts-based therapies can provide the most favorable experience to stroke survivors.

The present qualitative systematic review serves the specific aim of examining both the positive and negative experiences of creative arts-based therapies for stroke survivors. As the stroke survivors' personal and unique experiences of participation in creative arts-based therapies can provide in-depth information on their perspectives, the present review has thus adopted the qualitative approach.

Creative arts-based therapies refer to interventions that usually apply one major art form as a medium to achieve a physiological, psychosocial, social, or any other therapeutic goals. Hence, the type of art form can include music, dance movement, visual arts, creative writing, and drama. Creative arts-based therapies designed for survivors of the two major strokes, namely, ischemic stroke, and hemorrhage stroke, were included in the review because the severity and effects of the TIAs are short-lived.

A literature review of publications on creative arts-based therapies for stroke survivors was conducted. This qualitative review was based on the materials retrieved from a systematic literature search using six computerized databases, namely, PsycINFO, PubMed, MEDLINE, CINAHL, Cochrane, and Embase. The search was completed on February 20, 2018 and there were no restrictions on the publication period. The search was based on whether the combination of the following keywords appeared in any abstract: “stroke” and “music OR art OR dance OR dance movement OR drama OR writing” and “qualitative OR interview OR focus group OR observation OR thematic analysis OR grounded theory OR content analysis OR framework approach OR phenomenographic.” The key words for selecting the materials related to any kind of creative arts-based therapies were inspired by Beard (2012), who conducted a systematic review in dementia care that also required different kinds of creative arts-based therapies. To enhance the coverage and accuracy of the studies, reference lists of the included studies were also screened to extract relevant articles that were overlooked in the databases. Manual searches on the internet and the Google Scholar website were also performed.

Studies included in this review were based on the inclusion criteria listed in Table 1.

The process of data extraction began with comprehensive searches on different databases and from other resources. All searches were run independently in each database. Each result from each database was extracted into EndNote. For the first screening, duplicated studies were identified and removed. For the second screening, titles, and abstracts of the studies were examined. Studies, however, were excluded if the titles were not relevant. For the third screening, full texts of the remaining studies were examined by the first two authors, LTLT and LJLC, based on the eligible criteria. Appropriate reasons were also provided for any excluded studies. The last screening was the quality assessment and was conducted by LTLT in cooperation with LJLC. Data extractions were conducted on the final remaining studies. In order to enhance accuracy and to avoid any bias, LTLT and LJLC also jointly conducted the data extraction process. Data extractions were based on a tailor-made Excel file that included the study design, data collection methods, setting, informants, interventionists, use of art forms, samples, and sampling methods.

The data synthesis process adopted the thematic synthesis approach developed by Thomas and Harden (2008) which involves three steps: the first step is to conduct line-by-line coding, the second, to develop descriptive themes, and the last, to generate analytical themes. The result sections from the included studies were then extracted for free line-by-line coding. The two authors, LTLT and LJLC, conducted the coding independently. The two authors then joined together to discuss the consistencies and discrepancies of the codes and the interpretation of the findings. Subsequently, the two authors decided on the descriptive themes based on the similarities and differences of the preliminary codes. Finally, after several rounds of discussion, the analytic themes were developed. A summary of the results based on the decided descriptive and analytical themes was then drafted by LTLT and reviewed by LJLC and HRTH together. All discrepancies throughout the data synthesis process were resolved by consensus. Quotations from the studies were also extracted to support the descriptive and analytical themes.

After searching for relevant materials, 367 studies were identified. After eliminating duplicated studies, 215 studies were selected. After screening through title and abstract, 38 studies were selected, and their full text were examined. After that, 26 studies were further excluded with reasons provided. The remaining 12 studies underwent quality assessment, after which one study was excluded (Demers and McKinley, 2015). Finally, 11 studies were identified and included in the qualitative systematic review. The process of identification and selection of inclusion materials is illustrated in Figure 1. The flow diagram is based on the PRISMA guideline.

To evaluate the quality of qualitative research suggested by Murphy et al. (1998), the quality of the retrieved materials was based on two main aspects, namely, credibility and relevance. The credibility of the studies was assessed by the following five criteria: appropriate sampling and data collection method, auditability, reflexivity, handling negative cases, and fair dealing (Murphy et al., 1998). The relevance of the studies was assessed by the following two criteria: transferability and analytic generalization (Murphy et al., 1998). This quality assessment method was adopted from a qualitative meta-synthesis study on the life experiences of stroke survivors conducted by Salter et al. (2008). Thus, studies which did not meet the standards of the two main aspects, namely, credibility or relevance, were excluded. Quality assessments were also conducted by LTLT and LJLC independently. Results of the quality assessments were later discussed among themselves and their disagreements resolved by consensus. Supplementary Material (Table S1) presents the results of the quality assessments.

Eleven studies published between 2005 and 2017 were included in the systematic review. Table 2 lists the general characteristics of those studies. Eight studies were qualitative studies and three were mixed-method studies. The studies were conducted in different countries and cities which included Australia, Brazil, Hong Kong, New Zealand, Sweden, the United Kingdom, and the United States. Nine studies were conducted in a community setting and two in a hospital setting. Six studies applied the convenience sampling method, three adopted the purposive sampling method, and two used the voluntary sampling method for sample recruitment. Nine studies involved semi-structured interviews for data collection, three used focus group interviews, two involved observations, and one applied structured interviews.

As to the details of the interventions, six studies applied music-based interventions, three applied visual arts-based intervention, one applied dance-based intervention, and one study applied literature-based intervention. Among the six music-based intervention studies, five music-based interventions were led by qualified music therapists, and one music-based intervention was led by music facilitators. Other art modalities interventions were led by artists, fine arts students, nurses, actors or actresses, and dance instructors. Other details of the interventions are listed in Table 3.

Overall, five analytical themes were identified, namely, “functional restoration,” “psychological support,” “social engagement,” “spiritual experience,” and “short-comings and barriers.” Each analytical theme was supported by its corresponding descriptive themes which are listed in Tables 4–8. Supporting quotations from stroke survivors, significant others, and interventionists were also retrieved to support each descriptive and analytical theme.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.