Co‐design to deliver service improvement: What does this mean and how can we do it? A qualitative study with upper gastrointestinal cancer patients and professionals

Abstract Background There is strategic objective to incorporate the principles and practice of co‐design into routine service development and improvement. Aim The aim was to explore the concept and feasibility of service co‐design with patients and health professionals with regards to the upper gastrointestinal (UGI) cancer care pathway. Methods and Results Qualitative telephone interviews and face‐to‐face focus groups in one region of England. Twenty patients completed interviews. Nine patients and ten professionals formed two focus groups. Patients were referred through the urgent (two week) GP referral route and were within six months of receiving their first treatment for an UGI cancer. Professionals were working as service planners and providers of the UGI cancer care pathway. Thematic analysis was undertaken. Six themes emerged: Responsibilities and expectations, Knowledge and understanding, Valuing patient input, Building relationships, Environment for co‐design activities, Impact and effectiveness. Based on the themes a checklist has been created to provide practical suggestions for both professionals and patients on approaching co‐design for service improvement. Conclusion This study offers policy and practice partners a clearer understanding of co‐design and factors to consider when approaching co‐design in real life settings.


| INTRODUCTION
Tailoring care to patient need and improving patient experience are frequently cited goals of healthcare systems internationally. 1 Co-design of services with professionals, patients and the public is heralded as a mechanism through which person-centred care may be achieved. For example, co-design, that is co-creation of services with stakeholders, is recognised as one of the nine principles required for the successful implementation of the 2015 Cancer Strategy for England. 2 Existing systematic reviews provide a synthesis of evidence on co-design within healthcare. As a result of co-design developments within health care services have been implemented. 3 The implemented changes included: the production of new or improved sources of information for patients; simplifying appointment procedures; extending clinic opening times; improving transport to treatment units; improving access for people with disabilities; and a cultural change within organisations to involve patients. The strategic imperative to embed co-design within service development is supported by It has been recommended that support is given to allow co-design to become more widespread 10 and there is a recognition that codesign should become embedded into local routine practice. However, survival rates for upper GI cancers are low (5%-20%), 5 years after diagnosis. 20 Therefore, in order for this to happen in cancer services, policy and practice partners require a better understanding of how to undertake co-design in real life settings. With so many existing definitions co-design remains an evolving concept. 21 Therefore, research investigating how co-design, with both health care professionals and patients, could be effectively incorporated within the cancer care pathways is vital to enable evidence to be translated into practice.
To date, no studies have been conducted within the UK using codesign specifically with UGI cancer patients. In addition, whilst existing studies articulate some of the issues raised when undertaking co-design (e.g., time constraints, power imbalances, [22][23][24] the current literature fails to provide an adequate procedural understanding of implementing codesign in practice). This study aimed to address this by: exploring the concept and feasibility of service co-design of the UGI cancer care pathway with patients and professionals (commissioners and providers); identifying barriers and facilitators to undertaking co-design in practice; and producing a practical 'how to' checklist, outlining preliminary and preparatory steps with practical implementation recommendations. The study research question is what barriers and facilitators do UGI patients and professionals identify in relation to co-design in practice? 2 | METHOD 2.1 | Methodological approach Two phases of data collection were undertaken. Firstly, qualitative semi-structured telephone interviews were conducted, in 2017, with UGI patients. The majority of the topic guide for these interviews covered the patient experience of UGI cancer care pathway, the findings of which have been reported separately. 19 The guide also included questions covering the concept of co-design, asking interviewees how they felt about patients and professionals working together to improve services. Findings relevant to this line of enquiry are reported here.
Subsequently, focus groups were conducted to generate discussion and greater depth around issues relevant to co-design. Two focus groups, conducted in 2017, one with patients and one with professionals (providers and commissioners) sought to identify barriers and facilitators to the adoption of co-design. Each group lasted 90 min and was facilitated in a semi-structured way, guided by PowerPoint presentations adjusted for each audience, in order to ensure the concept of co-design was introduced and its feasibility fully explored. Focus groups encourage participants to play active roles and share their points of view, 25 adding depth to the insights gained and maximising the opportunity for this work to inform the application of 'co-design' in the realm of cancer (and indeed other services).
The study was carried out in accordance with the consolidated criteria for reporting qualitative research (COREQ) 26 (COREQ checklist included within Supplementary information 1) and guidance on standards for reporting qualitative research findings. 27 3 | PARTICIPANT RECRUITMENT

| Patients
For both the focus group and interviews, clinical nurse specialists (CNSs) identified eligible patients within their clinics and completed a referral form for the research team. Eligibility criteria required patients to be within 6 months of receiving their first treatment for an UGI cancer and be referred initially through the TWW pathway in a region of England. Only patients who had already taken part in an interview for the study were asked whether they would be interested in participating in a follow-up focus group and hence recruited directly (full description of recruitment and interview details provided in a previously published paper). 19

| Professionals
Professionals were emailed an invitation to participate in the professional focus group via one of the NHS Cancer Alliance communications networks. Cancer Alliances bring together the key organisations in their area to coordinate cancer care.

| Patient and public Involvement statement
A patient representative formed part of the project steering group (together with various other stakeholders), and provided feedback on all documentation, including suitability of the recruitment method for patients, the interview topic guide and the structure and content of the focus groups. All participants provided written consent before the interviews and focus groups began and were assured of confidentiality. Participants were made aware that direct quotes in published materials would be anonymised. There were no financial or other incentives offered to participate.

| Data collection
To ensure analytical rigour, the focus groups were facilitated by two female researchers (AH & SS) both with PhDs and previous experience in interviewing and facilitating focus groups and without any previous relationship with the research participants. The focus groups were audio-recorded and transcribed verbatim. Recordings along with the researchers' field notes enhanced trustworthiness of the data analysis by supporting the credibility of findings. The transcripts were anonymised with unique participant IDs linking the participants' contributions to a password-protected spreadsheet. Once transcription has occurred and been checked, recordings were deleted.
All interviews (except one, which was face-to-face at the interviewee's request) were conducted by one researcher (AH) via telephone, audio-recorded and transcribed verbatim. Interviews lasted approximately 1 h.

| Data analysis
The interviews and the two focus groups were initially analysed separately, with findings subsequently combined. Data analysis used inductive coding that allowed themes to emerge from the data without prior categorisation. Thematic analysis was used to identify themes and followed six phases: (1) Data Familiarisation; (2) Generating initial codes; (3) Searching for themes; (4) Reviewing themes; (5) Defining and naming themes; and (6) Producing the report. 28 NVivo 11 software was used to facilitate coding and analysis of transcribed data. One researcher (AH) coded all interview transcripts, with a selection of the interviews (20%) then coded by a second researcher (SS). One researcher (AH) coded the two focus group transcripts. The coding framework was then discussed and agreed within the full research team. Data saturation was defined using the principle out- The thematic findings were developed into a draft checklist providing tips and advice on how to implement co-design in healthcare in the future. This was refined through discussion with the wider project team and also the project steering group.
Where findings arose only in the interviews, or one particular focus group, this has been made clear in the results. Illustrative quotes are provided to support the findings.

| FINDINGS
Six themes related to co-design emerged from the analysis: responsibilities and expectations of involvement in co-design, knowledge and understanding, valuing patient input, building relationships, environment for co-design activities, and impact and effectiveness. One aspect of co-design that emerged as important was the need for expectations to be outlined clearly for each person involved in the process; this related to each individual's responsibilities but also as the co-design group as a collective. This was deemed necessary to allow people to be involved to levels that were suitable for them, for example, not adding more pressure to professionals' working schedule or overburdening patients undergoing or recovering from treatment.
'It depends which aspect you're looking at…you can only comment on the treatment if you're being treated…hospitals follow you up for five years…ten, which I think is incredible. But it's ten years, I don't know how often they'll want to see me or any others but I suppose you can give feedback as to how your years has progressed or your last six months have progressed but I don't know how relevant that would be to somebody who's currently in treatment.' (Patient focus group).
As illustrated by the quote above, expectations were also discussed in terms of the timeframes that would be necessary and appropriate in a process such as co-design. It was generally acknowledged that patients should be involved in the process of improving services as well as professionals.
Participants suggested that neither the professional nor the patient view was more important, but rather that each was important in different ways and for different areas of potential improvement Both patients and professionals believed patients would be assets to a co-design process. This finding was largely supported in the patient interviews. Patients described how working together could enable everyone to gain a better perspective of each other's experiences, that is, patients and health care professionals.  Training and education were suggested as a way of preparing people to be involved in a process such as co-design.   Building relationships within a group can also create supportive environments in which participants may be more productive and honest. To aid this, feedback throughout the process was considered important, so that participants felt that their involvement and views were valued. This was seen as important to both explain why and what service changes had taken place but also as a method of explaining if input from a co-design group had not been implemented. It was believed that providing this level of transparency to the process would embed trust and worth in the participants, whilst reducing the chance of participants becoming disheartened.
Theme 5. Environment for co-design activities.
The environment in which co-design occurs was discussed predominantly within the professional focus group. Participants spoke about how people tend to naturally stay within their traditional environments and require others to adapt (e.g., researchers conducting focus groups in a university building) but that this may not be conducive to successful co-design. 'We give precious little attention to the actual environment in which we do these things. Even if it's about going to the environment within which those individuals are working and living, we might be more successful than the reverse of bringing them to our environment'.
The format of 'a meeting' may be alien to some patients and may discourage them from participating in a co-design progress.  Constraints, such as time and resources, were considered essential to consider prior to the co-design process. Participants spoke about identifying clearly, from the outset, the expectations of participants and what is required for the process to be successful. Theme 6. Impact and effectiveness.
The issue of how effective co-design really is was identified predominantly within the patient focus group. Patients described providing feedback for healthcare they had previously received and pondered how much impact this had had.  Have you considered… what the ground rules of the meetings are? (e.g., terms of references) who will chair meetings? (e.g., using patient representatives to chair meetings) the environment that will be used for meetings? (e.g. public, NHS or university settings) allowing time in the process for members to build relationships and communicate effectively? (e.g., may need to focus on smaller or personal issues initially in order to progress to larger, significant pathway specific problems) using different communication styles? (e.g., technological advances-video calls) the need for preparation time as well as time for meetings?

Feedback
Have you considered… how will feedback be provided on members' considerations that have been put forward within the group?
how feedback will be provided on end decisions and the rationale behind them?
how long does knowledge stay relevant?
how members will know they/the group have made a difference?
using events as an opportunity for patient feedback? (e.g., annual events) another hospital or somewhere else that was responsible for it' (Patient focus group).
Participants discussed how impact and effectiveness relied on when patients were approached in terms of disease course. This emerged as an important consideration, and echoes findings from theme one in relation to expectations within the co-design process, as to how you involve participants, when and for how long. Table 1 shows the checklist of practical suggestions for both professionals and patients on how to approach co-design for service improvement. The checklist includes a series of initial prompts and reflective questions that any group looking to undertake healthcare service redesign or development through a co-design process would be advised to consider in terms of (a) preparing the ground work for the venture, (b) considering membership for the working group, (c) thinking through how any co-design meetings/activity would be structured and (d) how feedback will be handled.

| DISCUSSION
This study has examined the concept and feasibility of cancer service relationships. Adding to these previous findings, the current study participants discussed the need to identify and embed vital stakeholders that could influence change in a service and identified the importance of high authority representation for buy-in and to implement changes. In addition, providing feedback throughout a co-design process was also deemed to be an essential requirement to promote clarity and transparency as to what could and could not be adopted and the rationale behind end decisions. These findings, together with our finding regarding the importance of concrete practical considerations, expand on the current literature base and illustrate how preparatory work and planning is essential to try and ensure the most effective process and end outcomes.
A remaining challenge of co-design is the imbalance of power between patients and healthcare professionals; this has been highlighted as a potential barrier to the involvement of patients within health care improvement. [22][23][24] These findings relate to the theme 'environment for co-design activities' within the current study which identified the need to consider a variety of meeting settings and discussion styles to ensure patients feel welcomed and valued in the pro- be represented and-critically-truly valued, within the co-design process. This is an important aspect as it enables a variety of opinions on how the overall pathway works not just a particular viewpoint.
As far as we are aware, this is the first study to explore the conceptualisation and feasibility of undertaking co-design of service improvement for UGI cancer in the UK. The professional focus group included a range of professionals, with representatives from primary and secondary/tertiary care. These professionals were based across various NHS organisations within the North East and Cumbria. We were able to get not only professional but also patients' viewpoints in relation to co-design. For patients, we were able to triangulate the findings from two different data collection methods (individual indepth interviews and a patient focus group). The consistency of thematic content uncovered using these two data collection processes lends weight to the findings identified from the data. A potential limitation is the inclusion of patients and professionals solely from the UGI cancer pathway, which may limit the generalisability of the findings. However, findings tended to relate to the use of co-design in healthcare improvement generally rather than specifically within the UGI pathway. Clinical Nurse Specialists were used in the recruitment process; this may have resulted in selection of particular patients to be approached. However, participants expressed a range of views.
Another limitation was the limited number of female participants, although UGI cancer is more common in males the proportion is still lower than expected and may have impacted on findings. The study lacked ethnic diversity among participants (all participants were white British). In part this reflects the ethnic diversity within the study area.
However, previous research has identified ethnic minority cancer patients as reporting lower satisfaction, less positive experiences of care overall and less understanding of healthcare professionals. 32 This has implications for ethnic inequalities in healthcare and the study misses the opportunity to uncover these. The professional and patient focus group were run separately, which is not entirely coherent with the concept of co-design. However, our intention was to explore attitudes towards co-design rather than undertaking it. From experience of working with professional and patient groups, and the relative power dynamics between the two, running separate focus groups for patients and professionals was deemed most appropriate to allow each group space to openly express their views and opinions around co-design.
While there is strategic imperative to implement co-design within healthcare for service improvement, co-design is also increasingly being used in the development of behaviour change interventions for cancer survivors or to support survivors' self-management of cancer and its consequences (see, e.g., [33][34][35][36]. Therefore these study findings also have wider applicability for intervention development research It has been recommended that understanding of the concepts of codesign is an important aspect to avoid not only the implementation of ineffective or inefficient services or service solutions but also mistrust in health and social care services. 21 Transferrable insights have been identified which are likely to be relevant to co-design within other complex patient pathways (cancer and beyond), and these have been developed into a practical checklist. The checklist requires further research and dissemination to examine whether it is beneficial in aiding the preparatory stages of the co-design process. Sarah Sowden: Conceptualization (lead); formal analysis (equal); methodology (equal); project administration (equal); supervision (equal); writingreview and editing (equal).

ACKNOWLEDGMENTS
Thank you to our patient representative within the steering group.
We would like to acknowledge the Northern Cancer Alliance who collaborated on the study and helped to identify health professionals for the focus group. We would like to thank the Clinical Nurse Specialists who helped to identify eligible patients. Finally, we would like to thank all the health professionals and patients who took part in the study.

CONFLICT OF INTEREST
The authors have stated explicitly that there are no conflicts of interest in connection with this article.

DATA AVAILABILITY STATEMENT
The data collected and analysed during the current study are available from the corresponding author on reasonable request.

ETHICS STATEMENT
The study was approved by the Newcastle University Faculty of Med-