A framework for involving coproduction partners in research about young people with type 1 diabetes

Abstract Background Involvement of end‐users in research can enhance its quality, relevance, credibility and legitimacy; however, the processes through which these changes occur are unclear. Our aim was to explore a coproduction research team's experiences of their involvement in research about young people with type 1 diabetes mellitus (T1DM). Methods Semi‐structured interviews conducted with two young people with T1DM, two parents, one diabetes educator, one endocrinologist‐scientist and one research‐engineer explored experiences of coproduction research and its impact on both the research and the participants. Drawing on grounded theory, we undertook inductive analysis and storyline mapping to develop a theorized framework of mechanisms supporting the process of coproduction in T1DM research with young people. Findings The framework involving coproduction partners in research about young people with type 1 diabetes centres on the unique expertize that different team members bring to the research and describes conditions that enable expert contributions through the enactment of a variety of expert roles. The framework also describes outcomes—the impact of the expert contributions on both the research and the team members involved. Conclusion The findings of this small exploratory study provide a sound foundation to develop further understanding about structures and processes that are integral for the success of coproduction research teams. The framework may provide a guide for researchers planning to incorporate coproduction, on elements that are important for this model of research to succeed. It may also inform coproduction impact assessment research and be used for hypothesis testing and expansion in future studies.


| INTRODUCTION
While there are different perspectives on what involvement in research should look like, the principle of active involvement in research has been adopted broadly by policymakers, researchers and research funding bodies. 1,2 This includes patient and public involvement (PPI), coproduction, codesign, cocreation and participatory action research, which share philosophical underpinnings and fall under the umbrella of collaborative research. 1,3 The active involvement of end-users in the research process, or PPI, has been shown to enhance the quality and relevance of research. [4][5][6] Studies examining PPI impact have been conducted in the United Kingdom 7,8-10 and Canada. [11][12][13] In Australia, where the term 'consumer involvement' is often used in place of 'patient and public involvement', 14 there is a well-established body of consumer involvement research. [15][16][17] In addition to enhancing the scientific and ethical aspects of research, consumer involvement can increase a study's credibility and legitimacy. 18 Respectful and effective engagement requires the inclusion of numerous and diverse consumer perspectives. 19 People who draw on their lived experience of a health condition to inform research decision-making as part of the research team are often referred to as PPI contributors. 12 Researchers' perspectives on the benefits and challenges of including PPI contributors range from scepticism to enthusiasm and support. 10,20 Patients and public, and researchers have reported elements that drive PPI, naming shared goals, a supportive environment and provision of feedback as elements for success. 20 Results of a survey of PPI contributors with type 1 and 2 diabetes found that a critical part of the role was to provide constructive criticism; it was also important to them to receive feedback from researchers. 21 The involvement of young adults with type 1 diabetes mellitus (T1DM) in research improved understanding of what was needed to enhance health service delivery. 22 Incorporating teenagers' perspectives and values when designing T1DM interventions is acknowledged as important. 23,24 Other important end-users of research are clinicians and policymakers. The inclusion of these and other stakeholders in research, in addition to PPI contributors, is usually referred to as a coproduction model of research rather than PPI alone. 25,26 Coproduction extends on PPI to ensure that all relevant aspects of a concept are included in research, enabling the construction of complete knowledge, 25 with the potential to add value to and improve PPI. 27 The focus is on researchers, PPI contributors and other stakeholders-referred to as 'co-production partners' 28 working together, and sharing decisions and power within a project from beginning to end. 27,28 Evidence of what does and does not work in the coproduction of research knowledge is scarce. 29 To our knowledge, no research has examined the experience of including young people with T1DM and their carers, and clinicians in research about young people with T1DM.
Our aim was to explore a coproduction research team's experi- The HEDT comprises five people living with T1DM (three young people and two parents), two diabetes educators, one endocrinologistscientist and three health services researchers with backgrounds in nursing, sociology and medical anthropology. At the time of writing, for over 2 years, the HEDT has met monthly, sometimes more, to discuss the project, including study design, data collection and analysis. All members contribute to writing papers and grant applications, and are named investigators on these. Some examples of the research and activities conducted by the team are presented in Boxes 1-4.

| Design
We adopted a constructivist grounded theory approach for its underlying assumption that reality is influenced by context, within which each individual socially constructs his or her own reality. 30 Our aim was not to

BOX 1: OHIOH codesign project
In 2019, the HEDT collaborated with the OHIOH Devices team in codesigning a new breath sensor device aimed to enhance and improve diabetes management. The HEDT met with engineering students and their supervisor on several occasions to provide their perspectives on the characteristics required for a device to be both useable and desirable. Based on their experience, they described preferred physical attributes of a device, such as size and shape, durability and visibility of data. The engineering team used this advice to inform their design-making the device small enough to fit into a pocket, and the data output easily accessible. The engineers returned to subsequent meetings with updated iterations of the device until it was complete. DESBOROUGH ET AL. | 431 develop theory, but to elucidate the experiences and impact of involving PPI contributors and clinicians in the HEDT, and to explore relationships between these, hence our use of the term 'grounded theory approach' 31 also described as grounded theory-lite. 32 Grounded theory positions the researcher as one who reconstructs participants' experiences and their meaning through the use of constant comparative analysis, firmly grounding the research in the data (participants' experiences) and findings. 33 In view of the risk of prior experience and knowledge influencing the outcomes of research, interviews and initial coding were conducted by F. L. who had a background in biology and was a junior medical student with no prior knowledge of the team or of PPI or coproduction. However, she was encouraged to become familiar with these concepts before conducting interviews to stimulate her thinking about the data. 34 She was trained to conduct interviews by an experienced researcher (M. B.); they conducted the first interview together. Following a reflective exercise to debrief about the experience, F. L. conducted subsequent interviews alone. Interview protocols were designed by two health service researchers (J. D. and M. B.). Other than the student, all researchers had doctoral degrees and qualitative research expertize.

| Sample
The sample, approached via email and drawn from the OHIOH HEDT, consisted of two young people living with T1DM and their mothers (PPI contributors), one female diabetes educator from the Canberra Hospital Paediatric Diabetes Clinics (clinician), one male endocrinologist-scientist from the Canberra Hospital Diabetes Service and a male OHIOH research-engineer, who was working with the team to design a new device to support diabetes management.
While the endocrinologist-scientist was also a clinician, his primary role in this team was as a researcher. The two teenaged participants (one male and one female) had been diagnosed with T1DM for 2 and 9 years, respectively. Both were managing their diabetes independently with support from their parents if required. One young person was unavailable when interviews were conducted. Two health services researchers from the HEDT (J. D. and A. P.) supervized the student (F. L.) in the conduct of interviews and study analysis, deliberately adopting an arms-length approach to optimize participants' sense of being able to speak freely during interviews.

| Ethical approval
Ethical approval for the study was granted by the Australian National University Human Research Ethics Committee (HREC 2019-568).
Participants provided signed consent to be interviewed and for these interviews to be recorded and transcribed.

| Data generation and analysis
Semi-structured interviews were conducted on the university campus, in two parts. The first explored coproduction partners' (PPI contributors and the diabetes educator) experiences of their involvement in OHIOH and researchers' experiences of involving coproduction partners, and how they felt this impacted both the research and themselves. Crocker et al. 7 have outlined a typology of impactful roles for PPI contributors, presented in Table 1. We were also interested in exploring whether these roles resonated with the HEDT and if they might identify other roles. In the second part of each interview, participants were provided descriptions of these six roles and were invited to share their perspectives on whether these were reflective of their roles, and if there were additional roles they might identify from their involvement in the OHIOH project. BOX 2: OHIOH T1DM cohort research A core component of the OHIOH project is a longitudinal cohort study of young people with T1DM. The HEDT worked closely with cohort study researchers in the development of the study protocol for the cohort, including survey design and input into aspects such as the timing, regularity and location of data collection.

BOX 3: Support for coproduction partners
The HEDT coproduction partners assisted with data extraction and analysis for a systematic review. To support their capacity to contribute to this, they took part in a 2-hour session about research methodologies, including systematic reviews. Our project attracted some media attention in 2019 and in anticipation of further contact, the team attended media training arranged by the university.

BOX 4: Coproduction partners' research and other outputs
• Coproduction partners were involved in two online conferences in 2020, for which they prepared prerecorded slide presentations.
• Coproduction partners were named coauthors on a published systematic review and a second paper currently under review.
• Coproduction partners gave prerecorded presentations about their work with the HEDT at the 2020 OHIOH symposium.
We used constant comparative analysis, amending interview protocols after each interview to iteratively explore concepts as they were identified. This was facilitated through post-interview discussions between the interviewer (F. L.), who wrote memos after each interview, and supervisors (J. D. and A. P.), whereby concepts were clarified and relevant amendments to subsequent interview protocols were made. Separate protocols were used for PPI contributors and the clinician (Table S1) and researchers (Table S2). Purposeful ordering of interviews facilitated alternative comparison of concepts from different perspectives, enacting Charmaz's assertion that grounded theory can function like a camera-changing one's lens to view scenes from different angles. 33 For example, the first interview was conducted with a parent, the second with a young person and the third with a researcher, and repeated.
Interviews were of 20-40 min duration and were recorded, professionally transcribed and identifying details were removed.
NVivo software 35 was used to support data storage and analysis.
Line-by-line coding was conducted by F. L., and used to identify initial concepts, explicit statements and implicit meanings. Initial codes were analysed iteratively for categories and recurrent concepts, resulting in focussed codes. 33 After identifying and clarifying categories and focussed codes as a team, we used Strauss and Corbin's 34 coding paradigm based on conditions; actions and interactions; and consequences to support axial coding. On three occasions, the team met and used a whiteboard to identify conditions and mapped out relationships between conditions, the roles and the perceived outcomes. Initially, only roles identified by participants in the first part of interviews were included. Later, these roles and data from the second part of the interviews, where participants described differences from and similarities to those identified by Crocker, were carefully teased out. A theorized, three-part framework was identified through this process. This inductive approach enabled deep engagement with the data, facilitating interpretation of how specific conditions supported actions and interactions, and participants' perceived outcomes of this process. While analysis of data obtained from the first and main part of the interviews was inductive, analysis of data of participants' perceptions of the six impactful roles was more deductive in that data were examined to elucidate whether participants' experiences in the HEDT described the same roles, synergies with these, or new roles.
We then used the storyline technique 36 to describe and make further sense of the findings. This iterative process of writing and revisiting the coded interviews enabled refinement of the framework.
The student presented the draft framework to study participants at a team meeting and asked whether they believed this was an accurate representation of the process and outcomes of their involvement in the HEDT. Suggested amendments focussed on reducing confusion in the diagram rather than changing content. The team (J. D., A. P. and F. L.) met at a later date and discussed the suggestions, and amended the framework accordingly. This revised framework was again presented to participants, who confirmed that it was an accurate representation of their experiences in the HEDT.

| FINDINGS
Our aim was to explore a coproduction research team's experiences of working on research about T1DM in young people, including the impact on the research and on the team members themselves.
Researchers were asked to relate their experiences and the impact of working with coproduction partners, and coproduction partners were asked to relate their experiences of working in the team and their perceived outcomes of this. Our findings provide insight into the breadth of perspectives and dialogues that are brought to this study team through involving a range of stakeholders, 37 in this case young people with T1DM, their parents, a diabetes educator, an endocrinologist-scientist and a research-engineer. Participants also reflected on the role of other researchers in the team, who were not interviewed for this study.
While our aim was not to develop theory, we did theorize a framework, involving coproduction partners in research about young people with type 1 diabetes, that participants agreed provides an accurate representation of their experiences in the research about T1DM in young people. The framework centres on the unique expertize that different team members (including coproduction partners T A B L E 1 Six impactful roles for PPI 7

Role
Mechanism of impact The expert in lived experience PPI contributors are able to consider the acceptability and feasibility of research proposals through the lens of their lived experience The creative outsider PPI contributors bring a fresh perspective from outside the research system, and can help to solve problems by thinking 'outside the box' The free challenger PPI contributors are able to challenge researchers without fear of consequences The bridger PPI contributors bridge the communication gap between researchers and patients or the public, making research more relevant and accessible The motivator PPI contributors increase researchers' motivation/enthusiasm, for example, by emphasizing how the research will benefit people The passive presence PPI contributors can change the way that professionals think just by being present at meetings Abbreviation: PPI, patient and public involvement. Hence, the process was dynamic-experts (coproduction partners and researchers) could take on new roles and continue learning through their contribution to the team.

| Conditions
Participants described six conditions that enabled them to con- Who was present at a meeting was a condition that supported participants' contribution, influencing how comfortable they were with others in the group, also linked to a sense of respect for their expertize. One health professional expressed an occasional reluctance to speak up due to concern about upsetting perceived hierarchies amongst the various health professionals in the team (e.g., medical vs.  impactful roles directly corresponded to the roles that they already saw present in the team. These roles and participants' reflections are presented in Table 2. Participants believed that one expert role could be enacted by multiple individuals or one individual could enact multiple expert roles. All participants recognized the value of lived experience. Sharing these experiences was important to ensure that researchers understood the reality of life with the disease, which could generate novel ideas due to their different way of thinking compared with the researchers. The 'professional expert' was identified by the diabetes educator and confirmed by others. A professional's knowledge comes from his or her experiences with many others, which provides a different lens. The 'research implementer' was identified as an individual who acts on information from meetings and implements it into the research. Both parents said that just being present at meetings has impact through making the researchers mindful of the young people throughout the research process.  [What] they do for us is they actually help us define our projects a bit more. After talking to them I've got a bit more [of an] idea… (P4, Researcher)

| Outcomes
Preventing wasteful research was recognized as a key benefit of engaging with coproduction partners to ensure that the research being conducted would be useful to people living with diabetes.
You might be on the wrong track and it might not be important to, or it might be, you might be dreaming up something that they would never ever dream that they wanted to use or be part of and then you sort of waste a lot of research effort. (P3, Researcher) Researchers also came to understand the importance of ensuring that their research was not burdensome for the OHIOH cohort.
[I]f someone wants to come along and interview all the patients in the cohort like you're interviewing me, you know, just to take into consideration the time commit-

| DISCUSSION
The coproduction model of research emphasizes equality and the importance of power-sharing among stakeholders and researchers to foster inclusive research practices and build relationships. 26,38 At the core is valuing and respecting knowledge from different sources, 38 which underpins our framework. Our framework describes conditions that optimize contributors' capacity to be involved, and captures value in terms of impact-both identified as gaps in the current literature. 26 Our findings examine how conditions and roles interact, and how they influence the research and the individuals involved.

| Conditions
We identified six key conditions: 'having relevant background knowledge', 'having a sense of duty to use experience to help others', with previous research indicating the supportive role of such strategies. 11,39 'Having a sense of duty to use experience to help others' influenced participants' readiness to contribute, supporting previous evidence that committed and engaged participants are a key element to PPI success. 20 Understanding people's motivation for being involved is important to optimize their effectiveness and potential outcomes. 40,41 Participants highlighted the benefits of 'feeling respected and listened to' and 'feeling comfortable with others in the group' as enablers of contributing. A positive, open and trustworthy atmosphere 11,20 and consumers being respected for their expertize and opinions 18 are key to successful and sustainable collaboration.
The OHIOH team provides remuneration for meeting attendance, which also contributes to the 'support' theme identified in previous studies. 11,39 We included interdisciplinary researchers and clinicians in our research team to ensure that our research was applicable and useable, although on reflection, we realized that in meetings, we tended to primarily focus on the young people with T1DM and their parentsensuring that their voices were heard and that they were comfortable.
PPI and coproduction involve complex social processes 42,43 that may underpin the challenges identified in relation to 'performing a professional role in a new context'. We were surprised by this finding and that one participant was occasionally reluctant to contribute due to perceived professional hierarchies. This highlighted for us the greater consideration that is needed when implementing a coproduction model, in particular, the need to manage perceptions of hierarchy and power. 28,37,43 Acknowledgement of this in a large and diverse group is essential, as perceptions of inequality can reduce individuals' motivation to engage. 3 We subsequently worked to address these concerns through raised awareness and varied meeting configurations. Practical strategies for group discussions such as demonstrating value and respect for diverse views, creating opportunities for quieter voices to be heard and encouraging equality (e.g., using first names instead of official titles) may help address these issues, but require active and constant attention.
Participants referred to the need to manage conflicting priorities, which, for young people, was related to sport and school activities and for professionals, integrating time for their role with work activities. Managing hierarchies and conflicting priorities are acknowledged costs 3 as teams expand to include more stakeholders. 39,44 Overall, conditions identified in our study that facilitate PPI mirror those elucidated in previous research, indicating that conditions that enable PPI contribution also enable coproduction.

| Roles
We included nine roles in our framework, seven of which were identified independently by participants (lived experience expert, professional expert, research implementer, teacher/translator, free speaker, supporting another and being present), and two previously identified by Crocker et al. 7 (creative outsider renamed as 'outside-the-box thinker' and 'motivator'). Participants acknowledged and confirmed Crocker's six impactful roles -renaming some ( Table 2). Several roles were described as pertinent to a variety of team members; for example, the research implementer mostly referred to the research-engineer, the professional expert referred to the diabetes educator and the outsidethe-box thinker referred to both coproduction partners and the research-engineer. Similarly, the teacher/translator applied to all team members at different times. The roles described are representative of a more holistic understanding of both the context of T1DM and a potential technological solution, reflective of the substantive value of coproduction. 27,37,41 Our study is unique in its examination of young people's experiences of PPI and, as a consequence, the inclusion of their parents.
Most PPI impact studies involve adult patients, with relatively little focus on carers of people with diabetes. 7 The role 'supporting another' was identified by a parent and is unique to our study due to carer involvement in the OHIOH HEDT. One parent reduced their participation to provide more space for their child to contribute.
However, carers' lived experience is acknowledged as separate and valuable, 45 and this parental action to prioritize the child's contribution might have consequentially muffled the parent's voice. Finding a way to balance the needs of both is something that we are considering with our team.
Participants described the roles as fluid; they may alter with changing conditions. The cyclical nature of coproduction in our framework reflects this; roles change over time as participants gain knowledge through personal and research learning, which feeds back into conditions.

| Outcomes
Echoing recent research about PPI, 11,20 participants described distinct benefits of coproduction in terms of both personal and research learning, including the output of research tailored for, and relevant to those involved, to improve life for people living with diabetes. Colearning and research benefits for researchers resulting from PPI have been described previously, 11 plus therapeutic and education benefits 46 and enjoyable aspects, such as working towards a common goal. 20 The outcomes described in our study provide new evidence of the impact of coproduction, advancing knowledge to fill an identified gap. 37,47 Seeing evidence of their involvement in the research output, including device codesign, was a valued outcome for coproduction partners, feeding back into the condition of 'feeling respected and listened to', improving readiness to contribute. Responsive design contributes to a sense of meaningfulness and research relevance for PPI contributors. 20 A lack of involvement of young people and their caregivers in the design of diabetes technologies has been highlighted, 24 and the need to improve the depth and breadth of PPI in health technology assessment more broadly. 48 Our research is addressing these gaps.
Researchers identified preventing wasteful research, clinical usefulness and ensuring that projects have value for people living with T1DM as key outcomes in this study. The latter two may be considered to be aspects of preventing wasteful research, the cost of which has been highlighted and substantial efforts have been made to reduce this in recent years. 49 Our findings elucidate how coproduction can address issues of waste, 49 in particular by optimizing the quality of the design and conduct of research, enhancing its impact on research, clinical practice and policy.
Our findings also report impact on the people involved in the research, some of whom will be end-users. The true impact of our research on policy and practice will be tested when we work to implement the outcomes into clinical practice, including the device being codesigned by our team members. We aim to facilitate this through continued involvement of clinicians and the local health service, requiring consideration of our needs in terms of advocacy skills and tools. 37

| Other considerations
The nature of working with young people is that they grow up and seek different paths. Similar to recent research involving young people with T1DM, 44 it will be important for us to continue our relationships with current members as they grow, and to also engage with and nurture younger members. Participants reported subjective experiences and short-term outcomes. We acted on their suggestion to broaden the team expertize, now including a younger male with T1DM-a medium-term outcome. Participants described personal learning of transferable skills such as data analysis, media presentation skills, grant writing and research dissemination, also described in recent research, 44 which may be described as long-term impacts.
From a research perspective, the long-term impacts of coproduction in this study about young people with T1DM will also describe the impact on the research agenda and culture for the broader project and within the university, including when, where and how it can be most effective. 26 These represent valid ways of understanding the impact of coproduction in this study. 50

| Future directions
Our experience affirms that coproduction is hard but rewarding work. 3,28,37 Involving people living with T1DM in research requires an organisational commitment to fund this activity and create conditions conducive to PPI contribution, which we have worked to embed into our project. Our experience of involving clinicians and interdisciplinary researchers in addition to PPI contributors has been that it requires even greater active and continual outreach and engagement. A dedicated support position is required to support this ongoing engagement in large projects.
A variety of frameworks for supporting PPI in research exist and include similar components to our framework; however, the literature is not clear on how such elements can be combined to understand how these processes work together. We believe that our findings and our framework make a valuable contribution to understanding this and can inform future research embedding either PPI alone or coproduction. At the same time, we acknowledge that there is no 'onesize fits all approach' and challenges and enablers to coproduction will vary with context. 43 It takes time and commitment to build relationships and capacity, and to create a team that can work together in a safe space where different perspectives and opinions can be aired. There is a need for studies that focus on the effect of coproduction and tests of when, where and how it can be used most effectively. We hope our study will contribute to this new knowledge. 37

| VALIDITY AND RELIABILITY
Regular, iterative analysis at weekly team meetings ensured the methodological rigour of this study, in particular, the determination of codes and categories as they were identified and how to explore these. This enhanced the transferability of our findings by incorporating contributions from debriefing following interviews into the subsequent interviews and ultimately into the framework. 51 Incorporation of feedback from the OHIOH HEDT into the final framework established credibility. 51

| LIMITATIONS
Our study reports the experiences of a small team of researchers focusing on T1DM. Due to the finite size of our sample, we were unable to establish if we had reached data saturation, although the analysis yielded enough insight to inform a theoretical model. Confirmation of our model with another team inclusive of research contributors will increase the transferability of our framework, confirming its applicability to different groups and settings. 51 The majority of people involved in this study were closely linked to the project, which may have added an element of bias.

| CONCLUSION
The outcomes or impact of coproduction research are shaped by many contextual factors, including the skill and experiences of the people involved, and the subject matter. 42 Understanding the mechanisms that operate in specific contexts to produce particular outcomes can inform research planning. Our findings contribute new evidence to the understanding of conditions that enable the contribution of unique and broad expertize to research and outcomes arising from this. While this was a small exploratory study, it provides a sound foundation to further develop understanding, and also tools to support coproduction research. The framework may provide a guide for researchers planning to incorporate coproduction, on DESBOROUGH ET AL.
| 439 elements that are important for this model of research to succeed. It may also inform coproduction impact assessment research and be used for hypothesis testing and expansion in future studies.