Physical activity in patients with rheumatoid arthritis - an agile lifelong behaviour: a qualitative meta-synthesis

Background Physical activity (PA) in rheumatoid arthritis (RA) is considered a cornerstone in the treatment. To highlight aspects involved in supporting a positive PA behaviour, it is important to understand the patients’ perceptions of the phenomenon. Objective The aim of this qualitative meta-synthesis was to explore and synthesise patient perceptions of PA in RA. Methods A purposeful search was conducted across three online databases (PubMed, CINAHL and Web of Science). The methodological quality of the included studies was appraised, and data were extracted and analysed using an interpretive inductive thematic synthesis. Results Fifteen studies met the inclusion criteria and were included. PA was identified as an agile lifelong behaviour, with one main theme: The disease as a persistent catalyst for or against PA illustrating how the constant presence of the disease itself underlies the entire process of a life with or without regular PA. Seven subthemes: ‘considering aggravated symptoms’, ‘acknowledging the impact on health’, ‘becoming empowered and taking action’, ‘keeping informed to increase awareness’, ‘creating body awareness’, ‘dealing with social support’ and ‘feeling satisfied with circumstances and achievements’ were interpreted as facilitators and/or challenges. Conclusion This synthesis has identified PA as an agile lifelong behaviour in which the disease pervades all aspects of an individuals’ perception of PA. Placed in a theoretical context, our findings outline a model for tailoring PA support to the drivers and determinants of a certain individual, which will improve clinical practice for the benefit of both health professionals and patients with RA.

INTRODUCTION Good-quality evidence has accumulated over the past two decades on the effectiveness of aerobic and muscle-strengthening physical activity (PA) to reduce disease-related symptoms and comorbidity risk in people with rheumatoid arthritis (RA). [1][2][3] Since PA is effective and safe, it is recommended to be included in the standard care of patients with RA. 4 Nevertheless, PA levels remain low in people with RA [5][6][7] and a number of factors challenge the participation in and maintenance of regular PA. Such factors include, for example, time and cost, pain, fatigue and activity limitation, poor selfregulation skills and low autonomous motivation, as well as a lack of knowledge and reluctance on the part of health professionals (HPs) to implement evidence-based PA guidelines. [8][9][10][11] PA is a complex behaviour. Moving from a sedentary state to a more physically active state can be challenging for an individual, who may benefit from facilitating factors but also needs to manage several potential barriers. Facilitating factors and barriers for behaviour change include, but are not limited to, current and past behaviour, beliefs, emotions and context, social aspects and physical abilities. 12 13 Thus, HPs who aim to support improved PA behaviour among their patients should have a good understanding of these aspects, as well as the complex interplay between them. Findings from studies with qualitative design may be useful for understanding the nature of facilitators in

Key messages
What is already known about this subject? ► Physical activity (PA) in rheumatoid arthritis (RA) is considered a cornerstone in the treatment of RA and patients' perceptions of phenomena related to PA have been studied over two decades in order to gain understanding of this complex behaviour.

What does this study add?
► This qualitative synthesis of 15 studies identifies PA as an agile lifelong behaviour with a main theme illustrating how the constant presence of the disease itself underlies the entire process of balancing facilitators and challenges to regular PA.

RMD Open RMD Open RMD Open
reinforcing PA promotion and the barriers that need addressing.
In recent decades, a growing body of knowledge on the experience and perception of PA in people with RA has emerged from studies with qualitative designs. [14][15][16][17] A synthesis of their findings would identify the knowledge gaps that need to be addressed in future research and would also help HPs to better grasp their essence and make use of them. A previous meta-synthesis focused solely on the experience of exercise participation in people with RA. 18 However, there is currently a need for a synthesis that has a broader scope and includes a wide variety of phenomena related to PA in people with RA. Thus, the aim of our study was to review studies with a qualitative design and synthesise their findings in order to achieve a more comprehensive understanding of phenomena related to perceptions of PA in RA.

METHOD
This study used an interpretivist approach 19 for conducting a qualitative meta-synthesis 20 on phenomena related to PA. A meta-synthesis connects and synthesises findings from several individual qualitative studies at a higher abstraction level aiming at understanding the perceptions of different individuals from different contexts and make the findings more accessible for clinical practice. 21 The review followed the Enhancing Transparency of Reporting the Synthesis of Qualitative research statement 22 in order to report the different stages of synthesis and methods for qualitative findings (online supplemental file 1).

Search strategy and inclusion criteria
A purposeful sampling strategy 23 was applied and performed through a search of electronic databases (PubMed, CINAHL and Web of Science) using combinations of key search terms: 'qualitative methods, qualitative study, interviews, focus groups, PA, exercise and RA' (online supplemental file 2). To ensure maximum retrieval, no date limitation was used on the search which was performed up to 31 December 2020. A modified Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart 24 was used to illustrate the purposeful search and selection process (figure 1). Studies using a qualitative research design were included if they explored perceptions of different phenomena related to PA in individuals with RA, were empirical, published in English in peer-reviewed journals and included only informants with RA who were at least 18 years of age. Studies were excluded if they were published as short reports or letters, used questionnaires to collect data, explicitly explored body-mind exercises such as yoga, mindfulness or tai chi or PA interventions supported by e-Health or m-Health technology. Two independent reviewers (ES and NB) screened the titles and abstracts, excluded studies that did not meet the inclusion criteria and evaluated full-text versions for eligibility. Further, backtracking of the reference lists of the retrieved articles was performed.
Quality appraisal/quality assessment method The quality of the studies included in the synthesis was screened and appraised using the 10-item Critical Appraisal Skills Programme (CASP) checklist for qualitative research. 25 Two researchers (NB and ES) independently appraised all of the included studies and resolved any discrepancies through discussion (table 1). An independent researcher (Adrienne Levy-Berg), familiar with qualitative methods, appraised the quality of the included studies published by the authors of the present study.

Data extraction
All of the included studies were read independently by two researchers (ES and NB) and all text related to the findings and quotes from each study were extracted for analysis, as well as the findings in abstracts and conclusions. The following data were also extracted in order to provide a summary of each of the included studies: author, year, country, research aim, sampling methods, participant characteristics (number, gender, age, disease duration) and method of data collection and analysis.

Data synthesis and analysis
An interpretive inductive thematic synthesis, in accordance with the techniques of Thomas and Harden,26 was conducted by two of the authors (ES and NB) and discussed with the third author who was a peer expert (CO). Thematic synthesis involves identifying key concepts across studies based on the findings from each study and creating themes that go beyond the content of the original studies. The analysis comprised three stages: (1) line-by-line coding of the findings of the studies, (2) developing descriptive themes and 3) generating a higher level of analytical themes. 26 The studies were first read several times to familiarise with all the text to be analysed from each study and codes were identified by one of the authors (ES) and grouped together based on their similarities and differences. The codes were organised to form descriptive themes by two of the authors (ES and NB), who subsequently developed analytical themes inductively (theme and subthemes) by interpreting the descriptive themes and answer the aim of the present meta-synthesis. Data were repeatedly explored, compared and contrasted across the studies and the main theme and subthemes were discussed and reviewed with CO in order to generate new conceptual interrelationships and understandings. A thematic map that interpreted the interrelationship between the main theme and the subthemes was developed as a final step.

FINDINGS Literature search results
Fifteen primary studies with qualitative designs including a total of 233 individuals with RA (177 women, 56 men, aged 21-87 years) were included in the review and synthesis 14-17 27-37 (table 2). The studies were conducted in Denmark, Ireland, New Zealand, Sweden and the Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis UK. All studies captured patient perspectives on PA but varied in terminology, with eight studies using the term 'PA' 14-16 33-37 and seven using 'exercise'. 17 27-32 The focus of the studies varied between perceptions of specific PA programmes, future PA programme designs, barriers and facilitators, maintenance, attitudes, effects, intensity, as well as PA in everyday life. Most studies included participants with low to high levels of PA, although PA levels were unclear 31 or not specified in six studies. 15 16 28 29 36 Most studies used purposive sampling but two used consecutive sampling 15 30 and another used a snowball sampling method, 29 while the sampling strategies were unclear in two of the studies. 16 28 Sample sizes varied from eight to 19 participants. Most studies recruited their informants from rheumatology clinics, physiotherapy clinics or arthritis associations. Twelve studies used individual interviews, 14-17 27 28 30 32-35 37 and three employed focus group methodology. 29 31 36 Six studies used thematic analysis, 16 27-29 35 36 four used phenomenographic analysis, 14 15 17 37 three used content analysis 30 31 33 and individual studies used systematic text condensation, 34 and interpretative phenomenological analysis. 32

Quality appraisal
The methodological strength of the included studies varied (table 1). Thus, 13 studies adequately justified their research design, while two did not. 27 28 Five studies did not adequately describe their recruitment strategies. 15 16 28 29 36 Fourteen studies met the quality criteria for data collection but one study did not discuss data saturation or provide a justification for the sample size. 34 Eleven studies discussed the relationship between the researcher and the participants; the other four studies 15 29 35 37 failed to discuss their own role and potential bias and how this may have influenced the research process and findings of the study. One study did not confirm whether ethical     15 Thirteen studies provided adequate in-depth descriptions of the analysis process, the other two did not. 27 29 Bias was addressed in all studies using one or more independent researchers in parts of the analysis. One study did not provide thick interpretative descriptions to support its findings. 35 All studies  used quotes to support their findings. All findings were discussed in relation to the original research aims and the credibility of the findings was discussed in all of the studies. All studies were appraised as valuable. Thus, none of the studies were excluded because of methodological flaws.

Synthesis
Across the studies and based on the thematic synthesis, we identified PA as an agile lifelong behaviour, with one main theme: The disease as a persistent catalyst for or against PA illustrating how the constant presence of the disease itself underlies the entire process of a life with or without regular PA ( figure 2). The disease can act as a driving force to motivate PA by not letting the disease take over, but can also be perceived as a burden responsible for a person's inability to be physically active. The disease also affects PA both positively and negatively by prompting different facilitators and/or challenges. Seven subthemes: 'considering aggravated symptoms', 'acknowledging the impact on health', 'becoming empowered and taking action', 'keeping informed to increase awareness', 'creating body awareness', 'dealing with social support' and 'feeling satisfied with circumstances and achievements' were interpreted as these facilitators and/or challenges for lifelong PA behaviour.

Considering aggravated symptoms
This theme refers to individuals' fears and concerns about aggravating the disease or its symptoms. Thus, concerns about the safety of PA may represent a challenge to performance and speculation as to whether PA helps or damages joints and/or worsens disease activity is common. Such concerns are present at the time of disease  Acknowledging the impact on health This theme illustrates the need to experience and perceive the physical and mental benefits of PA. Fear of losing independence or exacerbation of symptoms due to inactivity increases motivation for PA because of its well-known general benefits. Habitual and regular PA can act as a way of coping with the disease, keeping a strong body, decreasing symptoms and retaining independence. Insights, either self-experienced or gained through conversations with others, into the potential health benefits of regular PA can help individuals achieve a more active and conscious involvement in PA.
The joints are lubricated when I move, it doesn't make me stiff but simply more agile, it gives me a better general condition and all that (Kamwendo et al 15 ) it becomes easier to get up from a chair, to use the toilet, so many things in your everyday life becomes easier (Loppenthin et al 34  Introducing it into the routine and being able to manage time was probably the key difficulty…this is an activity that is now inbred within your general activity of a day and it becoming a habit is then just something that now evolves through naturally, and I think that's a key difference for me ( I got to exercise based on what I could manage, you could say. But anyhow, I got… when I had gotten started, the opportunity to try different things, so to say. It wasn't just for me to do certain things, but I could add or try what worked for me, and that felt good for me. I thought that was good (Lange et al 30 ) Dealing with social support This theme includes positive and negative aspects of social support related to PA. PA in a clinical environment with support from a physical therapist can be highly valued, since it provides a sense of guidance and supervision, shared responsibility and expertise regarding safety. HPs must also make external demands regarding the adoption and maintenance of PA behaviour in case the internal motivational drives are insufficient. In contrast, support from HPs can be negative in the case of individuals with a high level of autonomy and independence who want to make their own decisions and be responsible for their own regular PA. A sense of not wanting to join regular gyms or exercise groups with healthy people confers a desire to receive social support and be part of exercise groups for people with rheumatic diseases. Such groups also motivate and serve as a discussion forum for the disease itself. However, peer support is not always appreciated due to varying social needs. PA is encouraged through positive attitudes and support from family members, although support will be sought elsewhere in the event that family members are worried and limit the opportunities for PA.
It was so tough as I wanted to do what I know I could do. But I just couldn't. I would get into fights a lot with (my husband) because of that. I was stressed and sad and the he would be stressed and worried that I had done too much (Baxter et al 28 ) Just this help, that you've got someone who says, "now you've done this," and you move on…you need encouragement (Swärdh et al 17 ) I do not want to go somewhere where it is just about disease and arthritis-for me it is also important to be with ordinary people ( 38 In order to go beyond only summarising and verifying the results, using theory to understand the findings from meta-synthesis investigations could facilitate their implementation into clinical practice. 39 Our findings clearly illustrate how the three main drivers of behaviour described in the Behaviour Change Wheel (BCW) 40 are perceived and articulated in the context of PA in RA (figure 3). All our subthemes, interpreted as facilitators and/or challenges for lifelong PA behaviour, describe, resemble and can be explained by the BCW drivers 'social and physical opportunities', 'automatic and reflective motivation' and 'physical and psychological capability' for engaging in PA. Thus, our subthemes can interact in such a way that, in order to engage in regular PA, a patient would have to know more about why it is important (capability), have more access and support to engage in it (opportunity), and/or believe that it would be beneficial and not harmful (motivation). Using the BCW as an explanatory model might thus enhance decisions on the specific drivers to address, based on a certain individuals' perceptions, for the adoption and maintenance of healthy PA in RA. Furthermore, our subthemes illustrate perceptions of the potential determinants of behaviour, as explained by the Theoretical Domains Framework (TDF) 41 42 which underpin the BCW drivers (figure 3). These determinants (TDF) can provide guidance on strategies that can be used to modify the behaviour and include individual-level factors.

Clinical implications
Previous studies on phenomena related to PA in RA identified a range of perceptions on PA, but our synthesis suggests that PA can be understood as an overall agile behaviour with the disease affecting PA by prompting different facilitators and/or challenges. In clinical practice it is important to address them by acquiring structured information on the patients' previous successes and failures in adopting and/or maintaining a healthy PA behaviour. Linking the patient perceptions of PA to theory ( figure 3) and, in a subsequent step, to specific behaviour change techniques, such as education, training, enablement, persuasion, environmental restructuring, modelling, incentivisation or coercion/ demands 43 would provide structured support for both HPs and patients if implemented into clinical practice. It would hence enhance HPs' understanding of who Rheumatoid arthritis Rheumatoid arthritis Rheumatoid arthritis needs to do what, when, where and how often, to develop facilitators or to overcome perceived challenges related to PA. Thus, some patients may need education while others will benefit from skills training, persuasion or role modelling, for example. Due to the constant presence of the disease, the perceptions described as drivers and determinants of PA behaviour may not always act in the same way or in the same time frame for all patients with RA. However, they will still serve as key elements in conjunction with the disease in the development of this agile lifelong behaviour and will therefore be important to identify when tailoring PA support to a specific patient by selecting appropriate behaviour change techniques.

Methodological considerations
In order to allow for a wide variety of perceptions of PA, no limits were used in our literature search regarding time or country of origin of the studies. The risk of including older studies from a time when PA was not as accepted by patients with RA as it is today, is probably negligible since the oldest study included was published in 1999. Studies on mind-body PA such as yoga, tai-chi and chi-gong were not included in our synthesis since they are expected to be perceived differently than regular muscle strengthening or cardiorespiratory PA. Studies using e-Health or m-Health technology were excluded since they tend to focus on the content, features, user friendliness and the outcome of the service per se rather than on PA. Thus, our findings are not transferable to these types of PA or PA support. One strength of our synthesis is that we did not exclude any studies because of low CASP grading. However, in order to optimise the quality of the results, we did not include unreviewed grey literature, which might be a limitation. A challenge in the present study was the different epistemological views applied in the original studies that might have generated different kinds of knowledge. However, despite different underlying epistemology, the research methodologies used, for example, data collection and data analysis had more similarities than differences which implies that findings across the studies could be interpreted and synthesised meaningfully. 20 We also found some heterogeneity in the sampling methods and the samples included in the original studies, which we consider positive since it adds to a broad illumination of the phenomenon under study. Since the participants' levels of PA were not presented as a variation sampling criterion, a bias towards recruiting participants with similar experiences and/or levels of PA may have occurred in some of the included studies, which might reduce transferability.

RMD Open RMD Open RMD Open
Future research Many studies that have explored aspects of PA and that span over a period of 19 years were identified in our literature search. Similar themes found across these studies indicate saturation in the area and raise doubts about the need for further replicating research. 44 Rather, an exploration of the perceptions of innovative interventions based on our synthesis of patient perspectives, linked to theory, could create a basis for reaching beyond current practice. This should also be developed according to global changes that require sustainability by using digital solutions for PA support. The findings of the present synthesis might also be transferable and useful for HPs working with individuals with other inflammatory rheumatic diseases in which PA is also a cornerstone of the treatment. However, in such diseases, other aspects of both drivers and determinants might be present, which would need to be explored in future studies.

CONCLUSION
This synthesis has identified several phenomena related to PA in RA and describes PA as an agile lifelong behaviour in which the disease pervades all aspects of an individuals' perception of PA. Placed in a theoretical context, our findings outline a model for tailoring PA support to the drivers and determinants of a certain individual, which will improve clinical practice for the benefit of both HPs and patients with RA.