In the era of digital health, mobile apps have increasingly been used to improve diabetes care. These platforms are vital for individuals living in geographically marginalized settings with limited access to conventional, in-person diabetes support [1]. Particularly, adults with type 1 diabetes (T1D) living in rural and remote communities in British Columbia have reported that peers with T1D share first-hand lived experience [1,2], from which they may benefit. As such, digitally delivered mental health support may be a promising solution for underserved populations.

Several systematic reviews and meta-analyses involving children or adults with T1D have examined mobile apps [3-7] designed to improve self-management [8-10], monitor treatment through data visualization [11], and deliver psychosocial interventions [12-15]. Notably, research in this area has mainly focused on mobile app evaluation, while descriptions of app development lag behind.

Digital platforms that engage end users early in the development process are more likely to produce favorable results [16]. Specifically, co-design is a research methodology that, in principle, involves shared decision-making through the collaborative engagement of end users, researchers, and health care professionals to create and develop the technology to be used [17,18]. Thus, research exploring the mobile app preferences and perspectives of individuals with T1D has grown steadily [1,19,20]. Not surprisingly, 2 of the 4 best practices in app development involve “prioritizing user preferences” and “engaging stakeholders” [21].

While there is increased interest in collecting user feedback, how these data are integrated during the app development process has not been clearly described. Although several other T1D-specific apps provide detailed descriptions of their development, their goals are to improve glycated hemoglobin and blood glucose management by promoting self-care practices (carbohydrate counting, injection site rotation, insulin dosing, physical activity, etc) [22-25]. To our knowledge, no studies describe the comprehensive development process for a mobile app focused on reducing diabetes distress by offering peer-led mental health support to adults with T1D.

T1D REACHOUT (REACHOUT) is a mobile app designed to reduce diabetes distress by offering peer-led mental health support to adults with T1D living in rural and remote areas of Interior British Columbia [2]. Based on self-determination theory, which highlights what motivates patients to make support-related decisions [26], REACHOUT is a 6-month intervention that offers a range of support delivery modalities to meet users’ individual needs.

The use of the app is not intended to replace standard health care provision but rather to enable increased access to peer-led support and fill a gap in the health care system. While interdisciplinary diabetes care typically involves a team of endocrinologists, diabetes nurses, and dietitians, what is often overlooked and undervalued are the peers who understand first-hand the day-to-day experience of living with T1D.

This paper describes the iterative co-design and development process of the REACHOUT mobile app, including its use in a pilot trial and subsequent revisions made prior to its implementation in a randomized controlled trial (RCT). While empirical outcomes from the REACHOUT pilot intervention have been reported elsewhere [2], we focus specifically on the technical development and implementation of the app used in the intervention and aim to share generalizable lessons with other mobile health (mHealth) research teams.

Specifically, this paper reflects on the (1) collaboration process between researchers and end users (ie, adults with T1D) to identify information needs and support requirements that informed app functionality, (2) iterative design of the app with end users to enhance usability and meet key user needs, (3) deployment of the app in a pilot study to understand feasibility, and (4) incorporation of feedback from the pilot study to further refine the app before launching the RCT.

This paper is organized in the following sections: Ethical Considerations; Recruitment; Platform Design and Implementation, which includes all stages of the co-design process; Platform Features, which outlines the functionality of the designed app; Initial Platform Utilization Patterns, which includes a network analysis of app usage during the pilot study; Insights and Observations, which specifically addresses challenges to co-design in mHealth; and Discussion, which includes an overview of our findings, limitations, and future work.

Our development process occurred in 3 stages: predevelopment, co-design and development, and revision (Figure 1). As people with T1D are the end users of the REACHOUT app, their involvement was imperative; hence, their feedback and guidance were sought throughout the app’s development process to ensure it met end user needs. Through focus groups, informal think tanks, and feedback sessions, the co-design approach defined requirements and critical features for the peer support–based mobile app (Figure 2).

The iteration of the REACHOUT app that was released for the RCT contains the following 6 key features identified in the initial stage 1 focus groups: a 24/7 community chat room (group text messaging), topic-specific discussion boards, a peer supporter library, peer supporter profiles, small group virtual sessions (ie, virtual happy hours), and direct (one-to-one) messaging. Each of these features is described below, along with an accompanying figure showing the progression of development based on user feedback, from earliest design (conception) to the final iteration used for the RCT.

The 24/7 chat room is a feature built upon Rocket.Chat’s existing microservices architecture (Figure 5). Users can access a general-purpose discussion forum and send text, images, and voice messages. A banner reminds users of community rules, such as not providing peers with medical advice. In fact, group exchanges are monitored by research assistants and health care professionals (registered psychologist, dietitian, diabetes nurse, and pharmacist).

According to pilot study feedback, exchanges on the 24/7 chat room were perceived as overwhelming to some users, and message threads could easily be lost [31]. This formed a new requirement: creating topic-specific discussion boards for focused messaging channels limited to a specific T1D topic. These discussion boards were developed as separate public messaging channels using Rocket.Chat’s channel creation feature. The app was modified to create a designated menu page for topic-specific discussion boards. User-identified topics on current discussion boards include exercise, multiple daily injections, insulin pumps, insurance coverage, continuous glucose monitors, and traveling (Figure 6). The research team can add additional discussion boards, and frequently initiated topics will be retained for future platform instances.

The peer supporter library invites participants to self-select a peer supporter [2]. The implementation of this feature was initially explored in two ways: (1) a participant (one who receives support) browses profiles from the library and then chooses their own peer supporter (active matching) or (2) a participant conveys their preferences to the research team, which assists with selection (passive matching). The active approach was adopted so that users could identify a peer supporter who best fits their unique support needs. For example, participants considering having children might prefer a peer supporter who is already a parent. Rocket.Chat’s mobile app was customized to add this peer supporter library view (Figure 7). The library previews the first name and last initial, location, and age displayed as a card, with access to review the complete peer supporter profile enabled by tapping each profile card.

To initiate an active matching approach, participants must determine which peer supporter best fits their support needs. To aid this, each peer supporter completes a profile that can be accessed through the peer supporter library [2]. The characteristics included in each profile (Figure 8) are first name and initial of last name, age, age at the time of diagnosis, and devices used (eg, type of insulin pump and continuous glucose monitor). Each profile also contains a written biography and a 1-minute video biography accessible via a private YouTube link [2].

“Happy hours” are virtual synchronous small group meetings led by peer supporters. They highlight distress-related themes, such as workplace bias and discrimination. Currently, information on virtual happy hours is shared via announcements posted to a discussion board, and sessions are held on Zoom (Zoom Video Communications Inc.). For the RCT, virtual happy hours replaced virtual huddles, featured in the previous pilot study [2].

Adolescent focus groups identified a new requirement [33]: an enhanced user interface and user experience, achieved by making the home page less intimidating and more visually appealing. One example was a landing page shown when the app opened (Figure 9). The home page also functions as a directory where users can navigate to other app areas.

In collaboration with adults from the T1D community in British Columbia, we designed and implemented REACHOUT, a mobile app for adults with T1D in British Columbia, which seeks to reduce diabetes distress (a core facet of diabetes-specific mental health). This project builds upon previous work establishing virtual peer support as a viable strategy for reducing diabetes distress [1,13]. Six key user-defined features for the app were identified through focus groups and refined through co-design and a pilot deployment: peer supporter profiles, a peer supporter library, one-to-one messaging, group messaging, topic-specific discussion boards, and small group virtual sessions.

The REACHOUT app uniquely features a user-driven “match-making” process to access one-on-one peer support [2]. Specifically, each user chooses their peer supporter based on the factors that matter most to them (eg, shared hobbies, lifestyle, insulin pump and/or continuous glucose monitor brand, location of residence) rather than the selection determined by the research team or those administering the app. An iterative co-design approach enabled early and comprehensive user engagement, identified app requirements, enabled prompt identification of pain points (eg, usability issues and bugs), and provided us with ongoing formative feedback from end users.

Our app development process had its limitations. First, no formal usability testing was undertaken because (1) the importance of such testing was not initially identified and (2) the tight timelines and constrained resources of our behavioral research study were challenging. Formal usability testing using standardized questionnaires, such as the System Usability Scale [36], Questionnaire for User Interaction Satisfaction [37], or Post-Study System Usability Questionnaire [38], could have ensured that users were sufficiently able to use the app and contribute to our overall understanding of feasibility. While the lack of formal usability testing may have affected the app’s quality, user feedback from focus groups was incorporated before the app was used in the RCT. Additionally, focus group participants were predominantly female and self-selected, which may not reflect the larger T1D community in British Columbia. Future work should ensure diverse voices are represented when establishing app requirements, for example, by increasing the number of male participants.

The app has been deployed and is being evaluated in the REACHOUT RCT (NCT05668507). A formal usability evaluation should be performed to examine the app’s functionality and identify opportunities for additional features or device optimization. Features currently in development, based on further user feedback, include allowing peer supporters to limit the number of participants they connect with. When evaluating usability, participants’ usage practices, motivations, and intents should also be examined to understand ideal usage patterns. We are also codeveloping an adolescent-focused version of the app [33], which considers adolescents’ unique support needs and will be evaluated in a future trial.

The REACHOUT app was developed to reduce diabetes distress by providing peer-led mental health support to adults living with T1D in British Columbia. The app development process was described to guide others on similar projects. An iterative co-design process with end-to-end user engagement and continual user feedback was beneficial, enabling a thorough understanding of end user needs, preferences, and context, and ensuring the final app was impactful and relevant to end users. The REACHOUT app offers choice-based, just-in-time, and customizable support, making it the optimal peer support model for individuals of all socioeconomic backgrounds, geographic residences, and life circumstances [2]. Its potential to reduce diabetes distress and improve diabetes-specific mental health outcomes is being evaluated in a large-scale province-wide RCT.