A qualitative exploration of post-acute stroke participants’ experiences of a multimodal intervention incorporating horseback riding

Background Multimodal rehabilitation interventions delivered in late phase of stroke recovery involve physical (motor and sensory), social, and cognitively challenging activities. Horseback riding can be incorporated within such interventions, leading to meaningful long-term improvements when applied to individuals with moderate levels of disability. There is a lack of research illuminating stroke survivors’ experiences and perceptions of horseback riding in the context of multimodal interventions. Aim To explore stroke survivors’ experiences of participation in a multimodal group-based intervention that included horseback riding. Methods An explorative interview study was conducted with individual face-to-face interviews performed on a single occasion, utilising a semi-structured interview guide. Eighteen participants were purposively selected from a larger trial (mean age 62, 12 men, 6 women) within four weeks after treatment completion. The interview duration was between 17 and 50 minutes. The data was analysed using a qualitative content analysis method. Findings Four broad themes were identified from the analysis. These themes were: transformative experiences; human–horse interaction; togetherness and belonging; and the all-in-one solution. Interacting with the horse and peers had a profound emotional impact on the participants. The participants also reported having learned new skills, increased self-efficacy and self-esteem, and improvements in balance and gait, all of which could be transferred to everyday life. The horse itself played a central role, but other components, such as the other group members, the instructors, and the challenging tasks on the horseback, were also important. Conclusion A multimodal rehabilitation intervention that includes horseback riding may provide stroke survivors in a late phase of recovery with rich pleasurable experiences that may have life-changing and profound impacts on their emotional and physical state.


Introduction
that equine assisted therapy (EAT) did alleviate severe back pain [24]. It also identified EAT as a transitional process incorporating four dimensions, i.e. body awareness, competence, emotion, and environment. The authors discussed the importance of positive feelings and meaningful activities in the enriched environment that the horse and stable offer [24]. Close relationship with a horse, including a properly aligned position on the horse's back, may also contribute to the development of non-verbal communication skills [24], suggesting that horsebased activities may be beneficial for people with aphasia.
Due to the complexity of multimodal stimulating interventions and the multitude of experiences that may be involved, it is recommended that the subjective in-depth perspectives of such interventions be explored, in addition to obtaining objective measurements [25,26]. In this regard, a qualitative approach may be appropriate for generating complimentary personcentred knowledge not revealed by quantitative data [27][28][29]. Within the field of medicine, only few qualitative studies have been identified exploring the riders' perspectives. Apart from the study concerning people with back pain [24], one qualitative study demonstrated psychosocial benefits for five people with psychiatric disabilities [17], and one study including 15 people with various disabilities (two with stroke) revealed that horseback riding may have a positive influence on the participants' identity construction [30]. There is, however, a lack of research illuminating stroke survivors' experiences and perceptions of horseback riding in the context of multimodal interventions. A qualitative study in this context may better capture experiences that reflect the complexity of the intervention design. The aim of the present study was thus to explore stroke survivors' experiences concerning participation in a multimodal group-based horseback riding intervention.

Study design
This qualitative study consisted of semi-structured interviews with participants selected from a 12-week multimodal intervention to allow in-depth exploration of participants' perspectives. The study followed recommended guidelines for qualitative inquiry within the field of medicine and public health [28,31]. The reporting of the study has been done according to the COREQ criteria [32] (S1 Table).

Theoretical framework
In order to obtain a deeper understanding of the dynamic interactive experiences of taking part in a multimodal rehabilitation programme incorporating horseback riding, the philosophical position underlying this study relied on an interpretivist orientation to qualitative inquiry. According to Guba & Lincoln (1985), the epistemological position of interpretivism is subjectivism, and the ontological position is relativism, where reality is seen as subjective and varying from person to person [33]. In addition, an interactionist perspective was adopted as methodological approach where people are seen as actively participating in the creation of their own development through interaction with the social world [34,35].

Study context
For this interview study, a sub-population of 18 participants was selected among 123 individuals consecutively recruited to participate in an intervention trial in a late phase of stroke recovery. The inclusion process for this trial has been thoroughly described in the study protocol [12]. After inclusion, the participants were randomised to either the horseback riding group, to a rhythm-and-music based group, or to a control group. The randomisation was stratified with respect to gender and hemispheric location of the stroke [12]. The contents of the horseback riding intervention are shown in the supplementary file 'S2 Table'. In all, there were eight separate groups with between four and seven participants in each group. The rhythm-andmusic intervention is described elsewhere [12].
All participants in the active interventions were asked if they would agree to be interviewed, with the motivation that physical assessments and questionnaires would not sufficiently capture their personal experiences and perceptions of taking part in the multimodal interventions. All participants in the horseback riding therapy group agreed to be contacted, except three. The reasons for refusal are outlined in the supplementary file 'S1 Appendix'. Prior to the large trial, ethical approval was obtained from the Regional Ethical Review Board in Gothenburg (ref. no. 698-09). The study was conducted in accordance with relevant ethical guidelines including written informed consent from the participants [11]. Informed consent was obtained by experienced health professionals according to specified inclusion and exclusion criteria that assured capacity to consent. The study was registered on ClinicalTrials.gov, identification number NCT01372059.

Participant selection
Sample size. As the research question was considered to be narrow in scope, it was estimated a priori that between 15 and 20 participants would be sufficient to capture a variety of experiences, instead of using data saturation to determine sample size. This is seen as a pragmatic approach to sample size [31].
Sampling and method of approach. A list of all participants within the horseback riding group containing basic demographics was distributed to GC, an occupational therapist with several years of experience from conducting qualitative research. Based on the need for a rich diversity in age, sex, stroke location and stroke severity, GC purposively selected participants to contact. Representatives from all eight groups were selected. Efforts were made to also include people with aphasia, despite their lack of ability to fluently express themselves verbally. Eighteen participants were identified who represented the required diversity and were contacted by telephone by GC within four weeks after treatment completion. All participants (twelve men; six women; mean age 60.3 years, range 51 to 74 years) agreed to be interviewed.
Setting. All interviews were conducted at a rehabilitation facility at Sahlgrenska University Hospital in Gothenburg, Sweden. This out-patient setting was chosen to ensure confidentiality in a relaxed atmosphere. During one of the interviews, a personal assistant, employed by the participant for verbal support, was also present. This interview was also visually recorded to later facilitate the verbatim transcription. Table 1 presents demographic details of participants.
Data collection. Data was collected through individual face-to-face interviews on a single occasion between December, 2010 and August, 2013 using a semi-structured question guide ( Table 2). All interviews were audio recorded. Sixteen interviews were conducted by GC and seven with the assistance of a speech therapist specialised in aphasia who facilitated interviews with supportive conversational techniques (e.g., pictures of the horses and the riding centre) [36]. Two interviews were conducted by the speech therapist alone. One pilot interview was carried out by GC, but no adjustments were considered necessary. The pilot study was thus included in the analysis. In all interviews, the initiating prompt was "Can you, in your own words, describe how you experienced taking part in the horseback riding therapy?", followed by frequent promptings such as: "If so, in what way?" to encourage the participants to elaborate their thoughts. This approach suited our objective of obtaining in-depth data. No repeat interviews were conducted. No field notes were taken during the interviews. The duration of the interviews was between 17 and 50 minutes. Each interview was transcribed verbatim in Swedish by a person external to the research group and thus without influence over the research process. Transcripts were anonymised by a unique number prior to the analysis (P1 to P18). The transcripts were verified for accuracy by PP (a physiotherapist and trained medical secretary) by cross-checking the transcripts with audiotapes. No transcripts were returned to participants for comment or correction.

Data analysis
The method used for data analysis was qualitative content analysis [37,38], which is appropriate when the aim is to focus on the subjects as well as on the contextual meanings [37]. The emphasis lies on describing variations, e.g. similarities and differences within parts of the text [37,39]. Qualitative content analysis may comprise both manifest contents (close to the text) and latent meanings (close to the lived experiences, but more distant from the text) [39]. In this study we have mainly described the manifest contents, which still also requires some degree of interpretation [39]. An inductive (data driven) approach was used in order to identify patterns of the content [39]. With an inductive approach, little or no predetermined theory, structure, or framework is used to analyse data [38]. This approach is recommended when there is little former knowledge about the phenomenon of interest [38]. Thus, the themes were derived from the data and not on the basis of previous knowledge or theories. The process of deriving themes was done using a step-wise procedure. The 18 transcribed interviews were considered as the unit of analysis. All interviews were initially read repeatedly and independently by PP and GC to start the process of identifying meaning units corresponding to the objective and to acquire a good grasp of the whole. Next, the same two authors compared their meaning units, and when differences appeared, they were discussed until consensus was reached. All text units (i.e., the interviews) were then transferred to a computer software system (Open Code 4.0, freely available from Umea University at http://www.phmed.umu.se/enheter/epidemiologi/forskning/opencode), and the meaning units were subsequently condensed into labelling codes by both authors together. A code is close to the text (manifest level) and describes the content of the meaning unit with a few words [37]. The codes were then grouped into themes with similar contents and labelled with a description of their contents. The themes, sub-themes, and examples of meaning units and codes are shown in the supplementary file 'S3 Table'. The process of grouping codes into themes was conducted through several meetings between GC and PP. The final set of themes was presented along with the codes to the other authors (LBK, MN and CB) and determined by consensus in the research team during two study meetings.

Trustworthiness and reflexivity
Qualitative research has been gaining acceptability in the medical literature [28]. In order for the reader to judge the scientific rigour of the findings, it is therefore important to establish trustworthiness [29]. As suggested by Lincoln & Guba (1985) [33], we have considered the following aspects of trustworthiness in relation to our study: credibility, transferability, dependability, and confirmability. To achieve credibility, efforts were made to include individuals with a wide variety of disabilities following the stroke, for example also including aphasia. This contributed to a richer variation of the phenomena under study. To further enhance credibility, quotations from the interviews are provided to illustrate and support the findings in order to highlight the relevance of the themes and categories. The use of direct quotes also increased the confirmability. To further enhance the trustworthiness, all quotes were translated by a professional English language editor. To facilitate the transferability, we have tried to give a clear and distinct description of the context, selection, and characteristics of the participants, data collection, and process of analysis. Dependability relates to the ability of the researchers to be flexible and change perspective in accordance with the emerging process [29]. This was done by including a full audit trail of all steps and decisions taken during the process (S1 Appendix), a technique proposed by Lincoln & Guba [33].
Depending on positions and perspectives, different researchers might access different representations of the situation under study [40]. Members of the current research team represented different professions, allowing for triangulation of the analytical process. Methodological triangulation could have been further enhanced by including other data sources, for example by interviewing the instructors. Reflexivity is further discussed in the audit trail (S1 Appendix).

Findings
Eighteen participants aged 51 to 74 years took part in individual face-to-face interviews. The time elapsed since the stroke insult was on average 3.3 years, ranging from 19 months to 9.9 years. Participating in the multimodal intervention affected individuals in many dimensions, all of which revolved around the experiences with the horse (Fig 1). Analysis generated four broad themes: transformative experience; human-horse interaction; togetherness and belonging; and the all-in-one solution.

Transformative experience
This theme described the deep experiences associated with the intervention and the changes that it brought about with regard to how participants viewed themselves and life after stroke, as well as actual functional improvements. All but seven of the participants had no previous  The participants initially felt that horseback riding was a dangerous activity, and that horses were big smelly animals that one should be wary of. At the same time, they were curious to try something new, and the appearance of a big horse was experienced with excitement, challenge, and respect. Mounting the horse resulted in an explosion of emotions. The participants were filled with positive feelings, gratitude, and humbleness. They felt that horseback riding did them good, that it was a warm and enlightening experience, and one of the best things that had ever happened to them. The same participant also experienced memory related improvements: And my memory was improved by the frequent mental repetitions that I did, when I was preparing myself for the next lesson: "Today I will trot, and how is this done again?" So, I had to think about it a bit in advance.
Grooming the horse and holding the reins were also perceived as improving fine motor skills. This led to an enhanced motivation to continue with the horseback riding intervention.
Moreover, being able to manage the huge challenge of horseback riding also led to gradually increasing self-esteem and self-pride. Individuals who were initially afraid of performing the exercises on horseback, and in some cases felt challenged by trotting, experienced a growing sense of capability and success throughout the intervention, for example: It's very much about confirming that "I can!". Obviously this has been a very good experience! [---]

Human-horse interaction
This theme describes the physical and emotional closeness to the horse and the significant impact that the strength of the bond and interaction with the horse had on the participants. A sense of emotional bonding was formed between the participants and the horse during their time together, and the participants felt that a mutual understanding between the horse and the rider evolved. One of the participants even spoke of "falling in love" with this sensitive and beautiful animal: I. . . fell in love. . . I think so. . . (trying to find the words) . . . no, oh, no [---]. Well, it was very educational that a horse could mean so much to me, I didn't think so! It's their eyes. What eyes they have! [P16, female with aphasia, 67 years] Touching the soft coat and inhaling the smell of the horse reinforced the sensation. Participants felt a strong relationship with the horse and were astonished that such an animal could mean so much and be experienced as a good friend. This new friendship stirred them very deeply and evoked strong emotions. One participant said: I have never been so close to a horse before, it felt really great. It was almost as if the horse was, well not a person perhaps, but I wanted to hug the horse (

laughs). It felt like communing with an individual! [---] The riding itself was also fun, but the most rewarding aspect was the horse! [P8, male, 60 years]
Another participant described that the time in the stable was just as important as the riding itself: In order to ride, you can't just sit on top of a horse, you also have to try to get close to the horse and get acquainted with it. When you groom it, you get this close contact and you get this feeling "it's you and me working together". [---] That contact is just as important as getting up on the horse. [P7, female, 66 years] In addition, sitting only on a thin saddle pad gave an intense sensation and strong feedback from the horse´s moving body. The own body, that in many ways felt different after the stroke, was now perceived in new ways through the movements of the horse. The tight body-to-body connection was restful, and it became natural to follow the horse´s rhythmic gait, mimicking movements during riding, forcing the person to indulge in and completely rely on the horse, A feeling of mastery also developed as the ability to control the horse improved. The participants told of initially clinging to the horse, while at the end being able to manage the horse with ease and joy: The first times were a bit hard. I had no real control, I thought, when I was sitting there on the horse. . . happy as long as I didn't fall off. [P13, male with aphasia, 51 years]

I was helped to get up on the horse, I needed help to get the left leg over the withers. But then you could steer the horse by yourself, and that was really fun! [P3, male, 58 years]
Learning was enhanced by observing the others, by listening to the instructor, and by trying themselves. The unpredictability of the horse was an additional challenge, as it could sometimes startle when exposed to a sudden noise or movement. This was experienced as pushing the limits and even led to a couple of fall accidents, but without injuries.
The horse itself was seen as a tutor, a friend, and a co-worker. Working together in a partnership with the horse was deeply rewarding. To be able to succeed with horseback riding, it was essential to have an interaction and collaboration between the horse and the rider. As one participant [P4, male, 55 years] described this: It' s me and the horse who are doing this together. That felt really nice.
The close contact with the horse also reinforced communication skills, in that the participants needed to be able to make the horse understand the participants' intentions: In the beginning I was just clinging to the horse to make sure I didn't fall off. But

Togetherness and belonging
A strong relationship between the group members developed during the intervention due to the amount of time spent together and the experience they shared. The opportunity to meet and interact with other stroke survivors was fruitful, both during the stable activities and the shared lunch. Even the time spent on the bench while waiting for one's turn and observing the other group members ride was appreciated by most, although a few participants expressed that this wait was quite long and occasionally boring.
Several participants described how wonderful it was to follow the sometimes surprising progress of the others. This was seen as very inspiring. One woman told about a person with severe hemiplegia and aphasia who suddenly became more confident on horseback: X [mentions a person's name] who was in my group, he was terrified of horses! He had no previous riding experience, he was so frightened, and he was so good. That day when his fear let go, that was a fantastic sight! [P7, female, 66 years].
The participants were also generous with positive affirmations and telling each other about the progress they had observed, for example: We continuously told her that we noticed the improvements, and I think that spurred her to take on more challenges and not dwell upon "No, I can't". [P7, female 66].
To meet other persons with the same condition also gave opportunities to discuss common issues and gave a feeling of being able to be oneself: The social part was really important! You met some fellow sufferers, and we had a really good time! I think all of us enjoyed it. And eating lunch together afterwards was also very rewarding. [P11, male, 56 years] The rich opportunity to meet afterwards during lunch gave them time to reflect upon and process their experiences. It also gave a sense of landing after the intense activity. During lunch, when the staff and personal assistants were present, the participants were able to help each other analyse the riding experience and to provide generous feedback to each and affirmation of progress. The presence of the staff was highly valued in this process. One participant told that she had continued with horseback riding after the intervention, but had missed the social part very much and did not enjoy the riding as much.

The all-in-one solution
This theme was a reflection upon the organisation of the intervention programme itself in relation to other training programmes, such as physiotherapy, as well as other aspects regarding the programme's environment.
The arrangement of the programme that included stable activities, riding lessons led by skilled and devoted instructors, and lunch or refreshments with peers and staff was much appreciated by the participants. The social interplay and the practical activities were considered as equally important: Well, the all-in-one solution was great fun when we had lunch together and drank coffee and so on . . . [P3, male, 58 years] Some participants felt that it might be valuable to start with such a programme earlier in the rehabilitation process, but others confirmed that it was actually possible to make progress many years after stroke. It was stressed that others with disabilities, not only stroke induced, should be given the opportunity to participate in such a rich programme.
The instructors were present throughout the entire session. All participants spoke highly of the instructors. They were seen as devoted, professional, trustworthy, pedagogical, and competent, with great knowledge about both stroke symptoms and horseback riding. They gave the participants a feeling of equality-a feeling of being treated as adults. One participant said: I think it was good that it wasn't-excuse the expression-childish! We are in fact grown-ups! I think they managed to keep that distance, I really appreciate that. Instead we were able to meet like adults, and not looked down upon or treated as tiny little kids. . . [P11, male, 56 years] Another participant expressed that the instructors knew exactly what they were doing: It was obvious that they knew what sort of people they were dealing with, they were very familiar with the problems one might have. That it can be so different from person to person. And they were also horse lovers, so they knew about those things too. They were able to "kill two birds with one stone", so it was really good. [P13, male with aphasia, 51 years] It was experienced that the instructors tailored the exercises individually with a clear aim of progress, always challenging the limits of the ability of the participants. The participants were continuously given challenging tasks to perform, such as memorising a certain pathway, and this was especially appreciated. As one participant put it: They increased the challenges little by little I would say, gradually we were expected to ride by ourselves. [---] They also gave us different riding paths, or we could invent our own paths, and we had to memorise them and tell them how we planned to go, and then we had to actually do it. That was fun, and always a challenge.
[P3, male, 58 years] Another participant put it like this: And then of course we got better at it, so they raised the bar gradually, both with things that were difficult and not so difficult. All in all, it was terrific! [P13, male with aphasia, 51 years] Some of the participants expressed that these tasks were experienced as unvaried and occasionally boring, for example: We were often asked to follow a certain pathway, and some of these became a bit too familiar. [P15, male with aphasia, 66 years] The instructors were also extraordinarily good at picking the right therapy horse for each participant's specific needs. One participant who had observed the instructors test riding a couple of horses said: Carefully selected, because of safety reasons. [---]

Discussion
This qualitative study showed that a group-based multimodal intervention that includes horseback riding and social activities may provide stroke survivors in late phase of recovery with rich pleasurable experiences having profound emotional and physical impacts. Participating in the intervention provided the participants with a plethora of intense sensual stimuli; touch, proprioception, vision, hearing, smell, and the vestibular system were all involved. The intimate closeness to the horse had a profound impact on the participants by affording a multitude of new sensations, such as those arising from stroking, grooming, and sitting on the large animal. The physical closeness helped to establish an emotional attachment to the therapeutic animal. Moreover, the close physical contact, in addition to the friendly and warm social environment within the horseback riding intervention, is essential for the release of oxytocin, a hormone that is associated with the "calm and connection" system [41]. This is something that is difficult to accomplish with a mechanical horse, although simulators have been found to be beneficial for stroke patients in other ways [22,23].
Horses also stimulate communication skills, as the relationship between man and horse builds upon other values and demands than interpersonal connections [24]. Horses understand and respond to communication primarily through body language, so the participants must learn to become more aware of their bodies, their body language, and expression of emotion through their bodies [42]. The intervention also provided the opportunity to experience authentic relationships with peers and staff members, adding to emotional awareness, communication skills, and a sense of group cohesion. Experiencing the intervention in a social context with a small group of peers provided numerous opportunities for the participants to relearn essential skills for living and to employ activities such as observation, reflectivity, problem solving, and socialising.
Apart from emotional boosts, the participants also described improved body functions. Horseback riding is generally thought to affect many body systems, achieving improvements in endurance, flexibility, symmetry, and body awareness, and development of trunk and postural control. The mechanism through which the action of riding confers an effect is thought to be the equine walk, which provides a multi-directional input resulting in movement responses that closely mimic the movement of the pelvis during the normal human gait, a movement that is both rhythmic and repetitive [43]. The movements are continuously repeated in that the horse usually takes many steps during a session. Such repetition is considered to be an important component for driving neuroplastic changes in the brain [44]. In order to drive activity-dependent neuroplasticity further, it is also crucial that the activity is perceived by the individual as meaningful and significant. If the training is experienced as rewarding and stimulating in a positive way, it should increase the likelihood of neural changes [44]. Experiences of success may also help to generate motivation for further improvements and add to a feeling of hope [45].
Furthermore, the participants learned the difficult task to master a horse. This complex skill was, for most participants, completely new. The learning experience, as described by the participants, is in line with social cognitive theory [46]. According to this theory, people do not learn new skills solely by trying them and either succeeding or failing, but also by observing others. Learning is also affected by positive reinforcement. Through the horseback riding intervention, the participants experienced qualities such as self-pride, courage, and self-efficacy, which ultimately had an impact on their ability to constructively cope with new situations and everyday life problems. Self-efficacy, defined as an individual's belief in his or her capability to organise and execute the courses of action that are required to achieve his or her desired outcomes, is an important aspect of emotional well-being [47], and consequently, building self-efficacy is often a goal of therapeutic interventions. According to Bandura (1997), self-efficacy beliefs influence an individual's actions, effort, and resilience when faced with obstacles and failure, as well as the amount of stress and depression experienced during times of difficulty [47]. Outcomes of self-efficacy have been found to be very important for people with long-term neurological conditions; therefore, activities facilitating self-efficacy are valuable additions to community rehabilitation [48]. The value of horseback riding in building self-efficacy is well recognised and has been studied in various settings [17,49,50]. Learning how to relate to a horse and master its movements may very likely have the same effect on people in a late phase of stroke recovery.
The findings of this study are important for informing future rehabilitation development and delivery. A paradigm shift has occurred within stroke rehabilitation, and it is now believed that combining activities with information from different sensory modalities facilitates poststroke recovery process in the late phase [4,7]. The real challenge after stroke, however, is the lack of opportunity to participate in such activities. It is important to point out that a stroke need not affect a person's response to pleasurable physical and social activities. It may therefore not be surprising that the intervention experiences had such a profound effect on the participants. It should be noted that the participants did not solely consist of persons who knew or believed that they would enjoy horseback riding, and the preconceived notions about horses were quite diverse among the participants. In addition, participants were randomised to either horseback riding or rhythm-and-music based intervention and were thus not able to choose which activity they would take part in.
The study has some limitations. The interviews were conducted over a long time span. This was due to the fact that it was difficult to recruit participants to the randomised controlled trial and because there was a pre-determined maximum number of persons who could participate in each horseback riding group. We believe that this prolonged process did not have a major negative impact on our findings. The small sample size limits the transferability, and therefore the findings are not transferable to the broader population of stroke survivors. In addition, we included stroke survivors with slight or moderate disability (modified Rankin Scale 2 and 3), and therefore, the findings are not generalisable to stroke survivors without apparent disabilities or those with severe disabilities. Inclusion of severely disabled survivors would, however, have increased the risk of adverse events and dropouts. Another limitation was that the inclusion of willing participants resulted in a 2-1 ratio of men to women, which may have resulted in a gender-biased view. However, a male overweight is seen for stroke in a middle-aged population. A strength of the study was the addition of a full audit trail, showing how the results were discussed until consensus between all researchers was reached, further enhancing the dependability. This audit trail also increased the confirmability, i.e., showing that efforts were made to keep our position as neutral as possible during the analytical process.

Conclusion and relevance for practice
This qualitative study showed that rehabilitative activities based on principles of multimodal sensory stimulation that includes horseback riding may provide stroke survivors in late phase after stroke with rich pleasurable experiences that may have life-changing and profound emotional and physical impacts. This work adds to the body of knowledge about how such a program is perceived by stroke survivors and how multimodal therapies can be used in the community to facilitate the recovery process and build self-efficacy after stroke. Further research is warranted in order to explore long-term experiences, as well as assessing cost-effectiveness for providing this complex intervention in a late phase after stroke. The findings should also be confirmed in other cultural contexts and within other health care systems.
Supporting information S1