The DIAMONDS program is underpinned by a previous work conducted as part of a National Institute for Health and Care Research Programme Development Grant (RP-DG-1214-10002). During this 12-month development period, we conducted a scoping review of the existing diabetes self-management and education programs for people with type 2 diabetes combined with a series of multidisciplinary workshops to explore how to adapt diabetes self-management for people with SMI and conducted 5 patient and public involvement (PPI) panels to build PPI capacity.

The scoping review identified >150 unique interventions for supporting the self-management of diabetes. Although differing in format, duration, and delivery, the interventions were broadly similar in content, and there was no leading candidate approach for adaptation for people with SMI. Many of the interventions from the scoping review were included in a previous systematic review of diabetes self-management interventions by Chrvala et al [37]. This review identified 118 interventions that included goal setting in at least 1 area of self-management and concluded that interventions with a combination of group and individual sessions and those with a mean contact time of >10 hours were more effective than those without these features. We used 2 intervention planning workshops alongside consultation with our PPI panel to obtain feedback on the suitability of different intervention types (eg, group or individual or combined) and to identify how self-management interventions may need to be adapted to meet the needs of people with SMI. The workshops were attended by the DIAMONDS research team and clinical professionals from primary care, diabetes services, and psychiatric services. A total of 19 people attended the first workshop, and 21 people attended the second workshop.

In these workshops, we explored the specific barriers to and facilitators of self-management in people with SMI with experts in the field. Discussions around the educational content of programs concluded that any intervention should cover the 7 main areas of self-management included in the American Association of Diabetes Educators 7 Self-Care Behaviors (healthy coping, healthy eating, being active, taking medication, monitoring, problem-solving, and reducing risks) [17]. The workshops also specified a need for additional one-to-one support, which was not often provided in the existing diabetes self-management and education programs. There were mixed opinions about the extent to which we should use digital technologies and the potential acceptability and feasibility of this approach for people with SMI. The attendees of the workshops also discussed opportunities to train people working in mental health services to deliver a tailored intervention. Diabetes and mental health clinicians, including physicians, specialist nurses, and dietitians, agreed that this approach was feasible. However, they pointed out that although people working in mental health services have the skills to support people with SMI, they may know little about diabetes and would need specific training for this.

PPI panel meetings confirmed that there were mixed opinions regarding the use of digital technologies to support health. The panel agreed that digital applications could form part of an intervention but should not be used to the exclusion of other platforms and intervention content. PPI meetings also identified that in-person support would need to be an essential aspect of any intervention to help reduce anxiety and address low motivation. There was consensus that both group and individual intervention sessions would be needed to support people with SMI and diabetes. There was agreement that group sessions need to be optional and stand-alone and not compulsory so that people could attend if able to.

Our preparatory research informed our decision to develop a bespoke self-management intervention for people with type 2 diabetes and SMI. The intervention would need to include both group and one-to-one support to ensure accessibility and engagement for people with SMI. The intervention should also incorporate a digital application, but there should be options available for participants who do not have access to digital technology or who struggle with using it, such as a participant workbook. In addition, the barriers to self-management experienced by people with SMI meant that the intervention would need to be adapted to individual needs. These design principles were taken forward and served as the foundation for a phased approach to intervention development using a co-design methodology.

Stage 1 identified MoAs that were mapped against the determinants of self-management. This stage combined a secondary analysis of published qualitative and quantitative evidence about the enablers of, barriers to, and determinants (mapped against MoAs) of self-management for adults living in the community with SMI and LTCs [33,34] with an analysis of in-depth qualitative interviews with staff, informal caregivers, and service users in England about the lived experience of self-management of LTCs and SMI [23]. Evidential links between all MoAs and the determinants of self-management behavior were cross tabulated. An example of this cross-tabulation mapping the MoA “attitude towards the behaviour” with the determinants identified from the primary qualitative research [23] is shown in Table 1.

Stage 2 involved a 2-phase survey and consultation with expert stakeholders (who participated in stage 1) to prioritize and select candidate BCTs based on established theoretical and empirical evidence about behavior change interventions. Expert stakeholders were first tasked with ranking (on a 4-point Likert scale) the MoAs identified in stage 1 in terms of the perceived strength of their association with the self-management of diabetes and modifiability. A total of 21 people responded to the electronic survey; the stakeholders included health care professionals working in primary care (n=1, 5%), diabetes services (n=5, 24%), chronic obstructive pulmonary disease care (n=4, 19%), and mental health services (n=2, 10%); clinical academics (n=3, 14%); carers (n=1, 5%); service users (n=4, 19%); and health care managers (n=1, 5%).

The DIAMONDS research team then mapped the average expert-rated rankings for MoA-BCT associations with the cross-tabulation mapping from stage 1 to generate a list of candidate MoAs and BCT links. The second phase of the survey was administered again to the expert stakeholders approximately 2 months after the completion of the first survey to judge each MoA and BCT link according to whether it was acceptable and practical using the APEASE (Affordability, Practicability, Effectiveness, Acceptability, Safety and Equity) criteria. The BCTs and MoAs that the stakeholders reached a consensus on were carried forward to inform the potential content in the co-design process.

For each MoA, between 4 and 11 BCT links were identified. Following the prioritization process, the DIAMONDS research team selected the BCTs that offered the most promise in terms of potential efficacy, acceptability, and deliverability using the results of both surveys, preparatory work, expert consultation, and PPI. The 22 BCTs that were included in the intervention are listed in Table 2.

Stage 3 happened in parallel with the first 2 stages and involved the creation of animated trigger films. Trigger films, also known as catalyst films, are films designed to generate discussion on a specific topic [32]. The films were based on key themes and illustrative points from our qualitative study [23] and qualitative evidence synthesis [34] about the experiences of people with coexisting SMI and LTCs, their informal caregivers, and health care professionals. These themes included the impact of SMI on self-management behaviors, interaction between SMI and LTCs, and barriers to and facilitators of self-management. During stage 3, we worked with an external collaborator, the digital design agency HMA (HMA Digital Marketing Ltd) [38]. We used animations within the trigger films to integrate and illustrate various themes and points rapidly. Subsequently, we used these trigger films and the personas we developed in stage 4 as part of the co-design process. We developed 2 trigger films, one consisting of an overview of 3 characters who had SMI and LTCs and another providing different self-management scenarios for each of the 3 characters to highlight the difficulties people with SMI experience when trying to manage their health.

Stage 4 included an initial discovery day workshop that lasted 5.5 hours, followed by 5 co-design workshops with staff and service users over 8 months. The 5 co-design workshops took place face-to-face and lasted approximately 15 hours in total. The co-design workshops were facilitated by mHabitat (currently Thrive by Design Ltd [39]) and were attended by our digital design partner HMA [37]. During the workshops, the DIAMONDS research team, digital design partner, service users, and health care professionals worked closely together to inform the development of the intervention.

Workshop participants were recruited from three groups: (1) service users (ie, individuals with SMI and LTCs); (2) informal caregivers; and (3) health care professionals involved in the care of people with SMI, LTCs, or both.

We recruited participants from the existing DIAMONDS research cohorts (ie, people who had taken part in the program development work or related research and had given consent to be contacted again), DIAMONDS Voice, and 3 National Health Service (NHS) mental health trusts in Yorkshire, United Kingdom. The eligibility criteria for workshop participants are listed in Textbox 1.

All potential participants received an invitation pack about the study via post, including an invitation letter, a participant information sheet, and an expression of interest form that included consent to contact. The DIAMONDS research team contacted individuals who expressed an interest and explained the study to them. All participants signed a consent form at the beginning of the first discovery day for the duration of the co-design workshops. Overall, 24 people were recruited for stage 4 of the study. Of them, 8 (33%) people were recruited and provided consent but were unable to attend any of the workshops, 7 (88%) of whom were health care professionals, and 1 (12%) was a service user. One of the health care professionals attended the first discovery day workshop and then withdrew from the study. The number of attendees by group for each workshop is listed in Table 3.

The discovery day workshop was split between separate morning sessions for staff and service users and a joint afternoon session. To address the barriers to using a digital application to support self-management, the participants were asked to write down their frustrations surrounding digital technology on Post-it notes and place these on a “wall of frustration,” which included a poster illustrating participant quotes about experiences with digital technology from our qualitative research [23]. Following this initial exercise, the trigger films were shown separately to each group.

The service user group was asked to complete a journey mapping activity, mapping how a typical day in their life compared with and related to the characters in the trigger films. Following journey mapping, the service users were invited to identify problematic moments they experienced during a typical day when trying to manage their health. During this activity, the service user participants expressed that their ability to engage in self-management behaviors was heavily compromised by their mental health symptoms. Furthermore, as people with SMI, the service users described how they often felt intimidated by going to public places, such as gyms, and that this acted as a barrier to physical activity. The service users also identified that deciding where to do food shopping and their ability to cook were important considerations that underpinned their ability to manage their physical health.

The staff group was asked to draw up a care pathway for service users with SMI and type 2 diabetes and to determine whether there were any points along this pathway where digital technology could be incorporated. Following these separate discovery sessions, service users and health care professionals were brought together for a joint discovery exercise. The health care professional group presented the pathway they had designed, whereas the service user group presented an outline of “a day in the life of” to highlight the self-management activities they engaged in each day. The service users highlighted the challenges they often experienced, such as access to health care services and the acceptability of the services provided. Health care professionals recommended that optimized care pathways for people with SMI should consider the inclusion of structured education about type 2 diabetes. Figure 2 captures the perspectives of the workshop participants about the challenges of living with SMI and LTCs.

In the joint final session on the discovery day, the 2 groups reviewed the outputs from the earlier sessions. Variable sleep patterns, taking medication, and eating were highlighted as issues that contributed to a person’s ability to engage in self-management. Mental health was identified as having a much greater impact on quality of life and health and, therefore, was typically prioritized over the self-management of type 2 diabetes.

Following the discovery day, an in-person synthesis session was held to develop paper prototype components for a self-management intervention that could be delivered within an app or a workbook with prioritized functions and placeholder content. This prototype was based on a synthesis of the information gained during the discovery day and the MoAs and BCTs identified in stages 1 and 2. The prototype was used as a framework for the following co-design workshops.

During the synthesis session, the research team, DIAMONDS Voice, and staff from HMA and mHabitat categorized the notes from the discovery day as follows: (1) content (eg, smoking cessation), (2) functionality (eg, any interaction with the intervention), (3) insights (into the behavior of people with SMI), and (4) risk (eg, clinical or data risk).

Next, the group collaboratively prioritized content in the following order:

For each of the prioritized functions, the workshop participants explored how that function could be delivered. For example, the development of a daily routine could be supported by self-monitoring, consisting of prompts within an app.

A total of 5 co-design workshops were conducted after the discovery day and discovery day synthesis session. At least 1 member of the DIAMONDS research team attended each workshop to observe, take notes, and provide information about the existing evidence. Workshop content was developed iteratively based on a priori objectives, preceding synthesis sessions, and both verbal (in-person) and written inputs from the wider DIAMONDS team. Photographs of the workshop outputs were used as a basis for the synthesis sessions, during which the DIAMONDS team and DIAMONDS Voice iteratively tested and refined intervention prototypes, which included both the digital app and the manualized paper-based content.

The main outputs from the co-design process were the Change One Thing app and workbook and the DIAMONDS coach manual and training handbook. We hosted several collaborative meetings with Nifty Fox Creative (Nifty Fox Creative Ltd) [40] and DIAMONDS Voice to ensure that the layout, language, design, and organization of the workbook were accessible, engaging, fit for purpose, and matched the content of the app. We worked with the Leicester Diabetes Centre [41] to design the coach manual and training handbook. The Change One Thing app was designed and developed during the co-design process by our digital design partner HMA [38] and finalized by the app developer Pipe and Piper (Pipe & Piper Ltd) [42].