The manifestation of participation within a co‐design process involving patients, significant others and health‐care professionals

Abstract Background Despite intentions to increase user participation in the development of health services, the concept of participation and how it unfolds within studies with a participatory design has rarely been addressed. Objective The aim of this study was to describe how user participation manifests itself within a co‐design process involving patients, significant others and health‐care professionals, including potential enablers or barriers. Methods This study was conducted in the context of a co‐design process of a new person‐centred transition from a hospital to continued rehabilitation in the home involving three patients with stroke, one significant other and 11 professionals. Data were collected by observations during the workshops, semi‐structured interviews and questionnaires. Results Four categories: ‘Composition of individuals for an adaptive climate’; ‘The balancing of roles and power’; ‘Different perspectives as common ground for a shared understanding’; and ‘Facilitating an unpredictable and ever‐adaptive process’, with all together nine subcategories, resulted from the analysis. Participation varied between individuals, groups and steps within the process, and on the topic of discussions and the motivation to contribute. Discussion/Conclusion Participation is not something that is realized by only applying participatory design methodology. Participation manifests itself through the interaction of the participants and their skills to handle different perspectives, roles and assignments. Participation is enabled by individual, group and facilitating aspects. Co‐design processes should allow for varying levels of participation among the participants and throughout the process. Patient or public contribution Patients, significant others and health‐care professionals participated as co‐designers of a care transition model between hospital and home.


| INTRODUC TI ON
User participation aims at the higher rungs on Arnstein's ladder of citizen participation, that is the redistribution of power by involving those who otherwise might be the subject of research or recipients of care to be deliberately included. 1 This is recognized as a key component to enhance quality in health and social care. 2 User participation in health care has been described as a continuum spanning from consultation to partnership and shared leadership on the levels of direct care and organizational design, and in policymaking. 3,4 On the level of the direct care, user participation has been linked to person-centred health care with active, informed and empowered patients where health care is provided in a patient-professional partnership, as opposed to a paternalistic health care where patients are seen as passive recipients of care. [5][6][7] On the level of organizational design, there has been an increase in the use of participatory study designs, intending to involve relevant users, for example patients, their families and health-care professionals, to co-design improvements and develop health services. Such an approach has been suggested to result in health-care services that better meet the actual needs and expectations of the users and are more likely to be implemented in practice. [8][9][10] However, it has been shown that these joint processes are complex [11][12][13] and the empirical evidence is inconclusive. [14][15][16] Previous research on co-design has highlighted the risk of unequal power relationships, difficulties making voices heard, barriers to contribute in a meaningful way and a tokenistic approach to involvement. [17][18][19][20][21] In co-design processes involving patients and professionals, there may be difficulties to achieve equal power relationships due to inherent hierarchical structures. 22 Further, in processes involving patients with limited cognitive and communicative skills, 4 such as after a stroke, it may be even more challenging to achieve equal power relationships. Calls have been made to critically explore participation in health-care design 23 and how 'reconfigurations of power relations' can be achieved within co-design processes. 24 Consequently, research with a specific focus on user participation in the co-design process and the enablers and barriers to user participation is needed. In this study, we define user participation as participants' actions and behaviours in the development of a new health service during a co-design process at the level of organizational design.
Therefore, the aim of this study was to describe how user participation manifests itself within a co-design process involving patients, significant others and health-care professionals, including potential enablers or barriers.

| ME THODS
We conducted a co-design process, intending to develop a new care transition process between the hospital and the home with continued rehabilitation in the home environment for people with stroke. The co-design process was conducted with design thinking, a human-centred approach and user-driven development processes that co-create solutions to problems in collaboration between users. 25,26 Design thinking is both a process/methodology and an approach. The co-design process, conducted in the Stockholm region, consisted of five half-day workshops, held at Openlab, a designlab and challenge-driven innovation community, starting in November 2019 and ending in January 2020.
Ethical approval was obtained from the Swedish Ethical Review Authority.

| Participants
Three people with stroke, one significant other and ten health-care professionals took part as seen in Tables 1 and 2. To enable participation, the employers of the participants (patients, significant others and professionals) were offered reimbursement for the time the participants were absent from work. The workshops were moderated

| Co-design process
The overall focus of the co-design process was to develop solutions (prototype) to improve the experience of care transitions for patients and significant others from the hospital to continued rehabilitation in the home environment. At the start of the co-design process, the facilitator emphasized that the experiences of the care transition process of the patients and significant others were in focus. Codesign activities such as patient narratives, patient journeys, interviews with patients and the significant other were conducted to give precedence to the needs of patients and significant other. The codesign process comprised five half-day workshops, held in Swedish, with different elements and modules and breaks in between for food and drinks. 26,28 The participants were divided into three groups, the same across all five workshops, in order to facilitate discussion and creative collaboration. The groups consisted of one patient per group, and one group also included the significant other. All groups included three professionals of different professions, from both hospital and multidisciplinary rehabilitation teams. The researchers composed the groups based on both personal acquaintances with TA B L E 2 Characteristics of health-care professionals (H1-H10) and facilitator (F1) the participants and the experiences of the care transitions as a patient, significant other or professional.
The co-design process followed the foundation of the doublediamond model 28 and used varying design methods to meet the aim of the specific workshop as seen in Figure 1. Before and in between the workshops, the researchers collaborated with the facilitator in an iterative process of planning and structuring the workshop process. verbatim. As one of the interviewees did not want to be audiorecorded, notes were taken, and the interview lasted for 50 minutes.

| Questionnaires
Those health-care professionals who were not interviewed were

| Data analysis
Since the aim was to describe how user participation manifests itself within a co-design process including potential enablers or barriers, qualitative content analysis 29 was chosen as the data analysis method. The transcribed data from field notes, reflections, interviews and questionnaires were analysed with inductive qualitative content analysis. Each source of data was initially analysed separately with manifest and latent content analyses according to the steps presented in Table 3, in accordance with Graneheim and Lundman. 29

| RE SULTS
The results showed how user participation manifested itself including potential enablers or barriers within a co-design process involving patients, significant others and health-care professionals.  A relaxed and easy-going atmosphere where the participants spoke their minds enabled the forming of relationships with the

| Composition of individuals for an
Step 1 The transcribed unit of analysis was read through to get an overview of the content.
Step 2 The unit of analysis was divided into meaning units comprising words, sentences or paragraphs related to each other.
Step 3 The meaning units were condensed without interpretation through the shortening of the text without removing the core meaning of the unit.
Step 4 Each meaning unit was labelled with a code through the interpretation of the underlying meaning.
Step 5 Codes of similar meaning were clustered into subcategories. The initial subcategories were reduced through a back and forward comparison of codes and groups to capture the underlying meaning.
Step 6 Subcategories from the different sources of data were analysed by their latent content and compared in order to form categories based on the underlying meaning of the subcategories.  There may be a general problem that emerges quite clearly in the group at times. There is a kind of accusation that the patients make due to painful, bad experiences they had of healthcare. Obviously, this clashes a great deal with your professional identity.
It demands a great deal, to understand and not de-

| Balance of power
There were examples when the view of participation was hierarchical and an uneven balance of power became evident. Some patients viewed the professionals as experts and felt that they as patients had less to contribute than the professionals who had everyday health-care experience and skills.
I felt inferior in that way; of course, they were experts and I wasn't.
Health-care professionals, in their turn, could steer the discourse by contrasting the experience-based perspective of the patients/significant other from one unique occasion with their long-term experience from working with these issues every day. This appeared to risk cementing the traditional expert-novice roles. The use of role-play, where participants were required to clarify and change roles, was experienced as neutralizing the roles between the patients/significant other and the professionals. Interview, significant other.

| Different perspectives as common ground for a shared understanding
A recurrent use of cumbersome language, that is sector-specific words from the design area including a mix of Swedish and English, hindered the understanding of new tasks and concepts used in the process. This, in turn, hindered the participants' ability to comprehend the tasks that the groups would perform, limiting the possibility to participate as an equal group member.
What I found hard to understand was the process itself that she had drawn up and discussed. In some way it was above, or outside what I think is important.
It was more fancy words than genuine. It was a little bit too flashy for my taste.
Interview, significant other.

| The need for a facilitator
The facilitator was an enabler for participation and the workflow of But it was good that there were different methods.
There were lectures, we could discuss, we could paint, we could write; I thought that all of that kind of thing was great. Then perhaps one might start to think in a different way as well. It's also good in as much as we are different people in the room and there is always something that suits someone.
On the other hand, parts of the co-design process also seemed to negatively affect the participants' possibilities to participate on equal terms. Assignments that were forced, unclear and difficult to understand hindered participation. The confusion that arose when the participants did not understand how to take on an assignment made participants cautious about speaking their minds and participating in the task. This insecurity, together with stress due to the time-limited process, made the participants focus less on interactions between each other and more on the fulfilling of the given task. Time was mentioned as both an enabler to avoid getting stuck in endless discussions and a barrier to participation.
It was slightly stressful at times, bang, bang. So we had to finish some things when we had barely started.
I could lose a little focus at times.
As dissatisfaction, minor conflicts and co-operation problems occurred along the process, the possibility to adapt and revise the process was necessary to enable participation. Customized support, such as individual meetings and simplified information between workshops, was provided. One patient desired to explain their own experiences in greater detail, and arrangements to meet these needs were provided. An ever-adaptive process was dependent on a reflective approach from the researchers and facilitator to perceive the needs of the participants and pay attention to changes needed to be made. The establishment of a positive group climate and inclusive atmosphere was described as important aspects to enable trust, facilitate speaking out and participate in the co-design process. These results are in line with previous studies 21, [30][31][32] and also align with the first step of our co-design process, which was to establish a mutual understanding, empathy and trust among the participants. 33 Building relationships and creating social connections were experienced as important means to reach an inclusive climate within the group.

| D ISCUSS I ON
However, even though diversity among participants is an important element for a creative co-design and design thinking, 26 this entailed difficulties and challenges in collaboration in practice. Factors such as personality, creativity and communication skills, and ability to act within a group affected how the participants interacted, collaborated and contributed to the process. The composition of groups with a majority of professionals and one single patient may also have affected the interaction. Putting together, participants in co-design processes require careful consideration and the knowledge that a diversity of skills and capacities can require extensive efforts in facilitating collaboration within the groups. 34 Our results point to the need for intense and adaptive facilitation to enable participation of all group members. Further, the necessity of teamwork in design education, to prepare designers/facilitators as they will be confronted with this matter in practice, has also been described. 35 Co-design includes a team approach, 36 which stresses the importance of creating well-working groups. Groups act and form collaboration in different ways depending on the composition of individuals. 37 Groups who coordinate their work and manage group conflicts more easily create cohesiveness and an environment within the group where all members can contribute with their unique knowledge and skills. 37,38 This study highlights that to enable the participation of all participants, considerations must be taken concerning group composition and to support an inclusive group climate.
Ambiguity around the roles of the participants was described to hinder participation. Patients and significant others struggled to combine their binary roles as both users with lived experience and being an accountable partner in the design group. They experienced difficulties with contributing more than their own experience and felt they had insufficient knowledge (ie about the health-care system) to participate and contribute to the discussions. Patients' and the significant other's insecurity about their own ability to contribute 19,21 and difficulties in contributing more than their own experience have previously been reported. 31 One of the main purposes of co-design is to empower patients to become legitimate and acknowledged members of the design group. 37 However, not all people, whether a patient, a significant other or a professional, are capable of assuming or want to assume the role of a designer in co-producing health care. 3 It has been recognized that the capability of individuals to participate in the co-design varies 17 and that users cannot always take on binary roles, as it depends on the level of expertise, passion and creativity of the user. 36 In the present study, participation varied between participants and between different groups due to both capacities, will and contextual circumstances. Fischer points to the fact that people in some situations want to be designers and in other situations want to be 'consumers', and therefore advocates for the emergence of an 'adaptive design'. 39 This participation continuum has also been highlighted by Carman, showing that participation may vary depending on topic and context. 4 This, together with the results of our study, highlights a need to allow for flexible levels of participation. Education of professionals and patients, to gain knowledge and skills about co-design and the aim of the co-design process, was requested and has been described to enhance the co-design of new health-care services 3 and to strengthen their position within co-design processes. 3,16 Our findings therefore call for specification and clarification of the different roles needed before recruiting to a co-design process. This should also be clearly communicated to eligible participants. Further, it must be ensured that each of the participants understands their importance and unique significant value to the process.
The ambiguity of roles was also related to the power relationships between the participants which hindered participation. Some of the traditional hierarchical roles from health-care practice were present in the co-design workshops. The patients viewed the professionals as experts and downgraded their own experience as an equivalent contributor to the process, and some professionals considered patients' experiences as anecdotal. This strengthened the boundaries between participants and hindered participation. This type of discreditation has been described by Fricker as epistemic injustice. 40 Despite the fact that the co-design process itself should facilitate a user-centred approach with an emphasis on the experience of all users, it is of critical importance to acknowledge the possible presence of epistemic injustice. Hence, such injustice increases power asymmetries and separates rather than unites individuals. In the co-design groups, an epistemic injustice could therefore be seen in how professionals valued their own and their peers' long-term experience more than the patients'/significant other's experience.
However, there may also have been an epistemic injustice in that the workshops were based on the patients'/significant other's needs and their experiences were prioritized rather than the professionals'  16,44,45 We suggest that the ever-adaptive facilitation must consider the differences in the individuals, how used they are to work in co-design processes and how much collaboration with each other the co-design process entails.
The data collection with four different data sources generated rich data and enabled us to validate data through triangulation of sources, which strengthens the credibility of our study. Further, the use of both questionnaires and interviews enabled that participants in questionnaires to more freely express potential negative experiences of the process, while also ensuring rich data from the interviews. At the same time, conducting interviews with all the participants might have generated even richer data. The observers and interviewers were the same individuals who conducted the analysis, facilitating participation for all involved stakeholders is difficult. We divided the participants into three groups to balance the representation and designed the co-design process to facilitate participation.
Despite this, patients and significant others were underrepresented in number in all groups. We do further acknowledge the lack of participation from users in the analysis and manufacturing of this manuscript.
Our study shows that participation varied between individuals, groups and steps within the process, and on the topic of discussions and motivation to contribute. It has been argued that a co-design process inherently aims at collective ownership, 24 that is the higher rungs on Arnstein's ladder of citizen participation. 1 However, based on the result of our study, participation is not something that is realized by only applying participatory design methodology. Several authors have engaged in further elaboration on Arnstein's view of participation, [46][47][48] Tritter and McCallum claim that Arnstein does not acknowledge the complexity, and dynamic and evolutionary nature of user participation. 46 This is in line with how participation manifested itself in this study. Andersen et al. reason that participation should be a 'matter of concern' rather than a 'matter of fact'. 48 We acknowledge that the aim of co-design processes might not be to strive towards the highest rung on Arnstein's participation ladder in terms of all participants involved, but that a coherent use of co-design in the development of services might contribute to citizen power. Participation must be recognized and assessed on an individual basis and be voluntary in the sense that participants must have the possibility to refrain from participation and disclaim their right to have a say in all specific topics.
In conclusion, the individual's participation in a co-design process manifests itself through the interaction of the group members and their skills to handle different perspectives, roles and assignments.
The individuals' participation is enabled by an ever-adaptive facilitation. Furthermore, co-design processes should allow for varying levels of participation among the group members and throughout the process. Future research should explore how different levels of individuals' participation impact the outcome of the co-design process.

ACK N OWLED G EM ENTS
The authors would like to thank the people with stroke, significant others, health-care professionals and facilitator who participated in the study. The authors would also like to thank the funders of the study: Kamprad Family Foundation for Entrepreneurship, Research and Charity; Swedish Stroke Association; Neuro Sweden; and the Doctoral School in Health Care Sciences, Karolinska Institutet.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.