The prevalence of type 2 diabetes is increasing globally, and health services in many countries are struggling with the morbidity, mortality, and costs associated with the complications of this long-term condition [1]. Diabetes self-management education (DSME) is important as it can reduce the risks of developing diabetes-related complications and improve glycemic control, at least in the short term [2-4]. However, uptake of DSME and behavioral support is low: in England, less than 10% of newly diagnosed people with type 2 diabetes have been recorded as attending structured education [5], and in the United States, less than half of the patients have been found to receive DSME [6]. Barriers to attending DSME include inconvenience, fear of stigma, and a lack of knowledge about the potential benefits [7]. Education and behavioral support requirements by patients extend beyond initial DSME, and there is often a need for ongoing diabetes self-management support beyond educational courses in the light of patients’ evolving needs [8]. Digital health interventions (DHI) can provide sustained support and may overcome challenges associated with attending DSME. They have the potential for delivery at multiple locations at convenient times, anonymity, and presentation of content in attractive and tailored formats [9].

Examples of features used in existing interventions to improve the user experience include making fonts consistent, using logos and pictures, using bold to emphasize key points and improve design, repeatedly emphasizing the basic structure of the site and the program, including a screencast video to demonstrate the site, removing unhelpful jargon and terminology, managing expectations about the site in general, and personalizing the “source” by providing details about the development team [10]. User-expressed needs about the detail and degree of tailoring of information provision varies considerably, and it may not be possible to have a consensus on a format that satisfies all users [11].

Challenges with DHIs include low levels of uptake in the target population with high levels of attrition and reduced user engagement over time [12]. The literature suggests there are two main approaches to improving uptake and maintaining ongoing engagement: (1) maximizing acceptability and usability of the intervention itself so patients want to use it [13,14] and (2) providing a degree of human support or facilitation so that patients are able to use it [15]. User engagement with interventions can be influenced by user perceptions [14] and intervention design [13]. User perceptions have both cognitive and affective elements [16]. Proposed cognitive drivers for user engagement include efficiency (ease of finding what users are looking for), effectiveness (impact of use), and trustworthiness. Important affective elements have been identified as enjoyment and interest [17]. Enjoyment can be defined as a general positive disposition and liking of media content [18], whereas interest motivates learning about something new and complex [19]. Users need to be interested to initially visit the website and then need to enjoy using the intervention to stay engaged [20]. With regard to understanding attrition, revisiting a website is associated with higher levels of education, being older, and a positive affective user experience [21]. Effective interventions therefore need to be interesting, enjoyable, and useful for patients. Qualitative research that explores the patient perspective can be a key tool to inform the development of such interventions.

This study describes qualitative research that was conducted with patients to identify user requirements before the development of a new Web-based self-management intervention for adults with type 2 diabetes called HeLP-Diabetes [22]. HeLP-Diabetes was developed as part of a 5-year National Institute for Health Research (NIHR) Programme Grant for Applied Research and provides comprehensive self-management support that includes health information, behavior change support, emotional support, self-monitoring tools, and access to online peer support. It can be used by adults with type 2 diabetes who are able to understand written English at any stage of their diabetes journey, and has been shown to be effective in improving diabetes control [23].

User requirements were conceptualized as “needs” and “wants,” where “needs” were features that were expected to achieve therapeutic benefit [24,25], whereas “wants” were features that users (patients) desired in an intervention and that were likely to make them want to return to the program.

The theoretical basis for understanding the needs of people living with type 2 diabetes was based on Corbin and Strauss’s model of the work of living with chronic illness [26]. Their study has informed much of the subsequent literature on self-management [27]. Corbin and Strauss described three types of work when living with a chronic illness: illness work, everyday life work, and biographical work. These tasks have also been described as medical management, emotional management, and role management [28]. Illness-related work consists of the tasks of managing treatment regimens, preventing and managing crises, symptom management, and diagnosis-related work. Everyday work describes the mundane work of everyday living and includes the sentimental work of managing emotions and relationships. Biographical work describes the work done in extracting meaning from life experiences and creating personal socially constituted identities as patients, parents, spouses, partners, professionals, or friends. Being diagnosed with a chronic illness can have a significantly disruptive impact on a person’s biographical narrative [29,30].

As mentioned previously, support or facilitation in using DHI can help increase use and potential impact of such interventions. Most of the data on the effects of providing human support or facilitation for DHI come from mental health, where studies on Internet cognitive behavioral therapy demonstrate that including therapist support significantly improves engagement with Internet cognitive behavioral therapy programs and their effectiveness [15,31]. In these studies, therapist input has been limited to encouraging engagement with the program, rather than providing therapy. However, human support, and in particular, health professional support, is an expensive and scarce resource, so it is important to determine how much human input is required and who should provide it.

This study was undertaken to determine patient needs and wants for a DHI to support self-management in patients living with type 2 diabetes.

The study was reviewed by the North West London Research and Ethics Committee (REC reference 10/H0722/86).

English-speaking adults in England with type 2 diabetes were recruited for this study. Printed leaflets and posters were distributed to general practitioner (GP) surgeries and local diabetes support groups across London. An advertisement was placed in the Diabetes Balance magazine of Diabetes UK. Online recruitment included an advert on the Diabetes UK website, a local council website, ethnic minority forums, and other diabetes forums. Respondents were sent an information sheet and consent form and were invited to complete a questionnaire that was used to recruit a maximum variation sample. Factors that the literature suggested were likely to influence wants and needs included demographic factors (eg, age, gender, ethnicity, and first language), clinical factors (eg, duration of diabetes, current treatment, presence or absence of diabetes-related complications, and previous experience of self-management programs), and factors related to health and computer literacy (such as educational attainment, previous experience with computers, and access to the Internet) [32-35]. Participants were purposively sampled to vary across these characteristics.

Four focus groups with 3 to 6 participants were held in a community center in London. Focus groups were chosen as they encourage interactions between participants and allow ideas to be generated, reflected on, and debated by participants [36,37]. They were facilitated by two or three researchers from the team (CD, KP, EM, and FS), audio-recorded, and transcribed verbatim. Previous reviews of patient information materials and websites provided the basis for the structure of the focus groups [38,39]. At the start of the focus group, participants were shown 3 examples of existing websites that support diabetes self-management and asked to explore them. These had been selected by the research team to demonstrate the range of interventions and component parts available. They varied in terms of content, complexity, tone, navigation, and presence of interactive features such as forums, Ask the Expert, or self-monitoring tools. The websites were selected based on a number of criteria: up-to-date, accredited by national diabetes organizations (eg, American Diabetes Association or Diabetes UK), and written in English. The topic guide was piloted in an interview by one researcher (CD) to ensure it stimulated discussion across the target areas.

The 4 focus groups were held in 2 community centers. The facilities had computer access and rooms suitable for small group discussions. A venue outside health care settings was chosen to put participants at ease and to minimize the impact of the power differentials that might be created in such settings from participants taking on the role of a patient.

Each focus group was run by 2 to 3 researchers and lasted up to 4 hours in total. The first 15 min was allocated to welcoming participants and introductions, followed by up to 90 min exploring the 3 websites. This was followed by a 30-min break and then 90 min of group discussion. Having more than one researcher was helpful as they took a more observational role to monitor the interactions between facilitator and participant and could pick up on undue prompting or dominance from the main facilitator [40]. Having nonmedical cofacilitators helped monitor for dynamics that would limit the data generated from the focus groups if the facilitator took on the role of the “expert” and inhibited participant discussions [40]. Discussions were semistructured with a list of topics to be covered (not necessarily in a defined order) during the session.

Data collection continued until no new data emerged regarding content and design or patient-defined wants and needs for self-management.

Analysis was conducted by a multidisciplinary team. Transcripts were independently read by five authors (CD, KP, JR, EM, and FS), and emerging themes were discussed at a multidisciplinary meeting. As described previously, the underlying sociological theory guiding the analysis of patient needs was Corbin and Strauss’s study on living with a chronic illness. The main constructs of the model were used to sensitize and encourage a holistic perspective that explored the impact of type 2 diabetes mellitus on people’s day-to-day activities, relationships, and emotions (everyday life work); the burden of having to take medicines or make lifestyle changes for the medical management of the condition (illness work); and the disruption or changes to the roles that patients played within their families and at work (biographical work). This model was not used to define a priori codes or categories, but was used as a sensitizing tool to organize codes generated using inductive thematic analysis. Two authors (EM and KP) coded the transcripts and mapped themes onto the Corbin and Strauss’s model. The mapping was then discussed and agreed upon in a meeting between the 5 authors (CD, JR, FS, KP, and EM). The resulting structure was used to determine patient needs and wants for a DHI. None of the themes developed inductively fell out with the Corbin and Strauss’s model.

Illustrative extracts of the data are presented in the Results, with identification by focus group number and participant number together with age, gender, ethnicity, duration of diabetes, and computer experience.

Atlas Ti (version 6.2, Scientific Software Development, Berlin, Germany) was used to manage the transcripts and coding and to facilitate the final data analysis.

The demographics of the 20 participants who took part in the pilot interview and 4 focus groups are summarized in Table 1.

Just over half the participants were male with a mean age of nearly 57 years. Almost half were retired and over half had degree-level qualifications. Moreover, 70% (14/20) of participants were white; the remaining participants described their ethnicity as black, Asian, or other. Time since diagnosis ranged from 3 months to 36 years. An overwhelming majority of participants had home Internet access, and most had used the Internet to look up information about diabetes. In addition, 60% (12/20) of participants had been on a diabetes self-management course, but most had never used a computer-based self-management program.

The data mapped easily onto the Corbin and Strauss’s model, with the resultant themes and subthemes summarized in Table 2.

The participants in this study described a range of unmet needs for supporting their self-management efforts and identified a number of potential ways that DHI could help them. The features of DHIs desired by participants included specific content relating to diabetes (eg, hypoglycemia, medication side-effects, weight loss, and physical activity) and emphasized the strong desire of participants for reliable and accessible dietary advice. Patients in the United Kingdom have good access to primary care doctor and nursing support with structured templates to help standardize diabetes care. However, where patient self-management needs extend beyond the remit of these structures that often focus on using medication to optimize glycemic control, it can be difficult to get support.

A DHI could help overcome some of the barriers to self-management currently posed by limitations of existing health care systems. An evidence-based, well-written, up-to-date DHI could be available 24/7, and could improve access to high-quality information, DSME, and behavioral support where patients are not able to access face-to-face services that provide tailored information and help with behavior changes such as increasing physical activity, dietary change, and weight loss.

Access for patients to their electronic medical record via the DHI was seen as a potential benefit that could engage patients, especially if access to these systems allowed administrative functions such as booking medical appointments. Technology-based prompts have been shown to have positive effects on engagement with other DHIs [10], and participants in this study reported that regular emails would be an acceptable strategy for increasing intervention use.

The DHI developed as a result of these data, HeLP-Diabetes, was effective in improving diabetes control for users. We believe a key reason for this was the careful attention paid during the development phase to users’ wants and needs, and we commend this approach to others. The data generated during the focus groups were underpinned by a strong sense of the burden that the diagnosis of diabetes placed on participants, which had negative impacts on their emotional well-being, work, social life, and physical health. This finding fitted with our overall theoretical framework, based on the Corbin and Strauss’s model, and underlined the importance of ensuring that the self-management program addressed the three key tasks of medical, emotional, and role management. Although participants’ experiences of the health care services varied, participants reported difficulties getting the information and patient-centered care they needed to support self-management when consultations with health professionals were too time-pressured or protocol-driven to accommodate individual patient needs.

The strengths of this study included the use of focus groups and semistructured topic guides, which allowed participants to raise their own concerns and determine the direction and content of the discussions. The use of a sociological model [17] encouraged a holistic approach to the data that provided a broader perspective on self-management than usual bio-medical definitions of DSME.

The main limitation was that, although we were successful in recruiting a sample that was diverse in terms of ethnicity, duration and treatment of diabetes, and gender, the participants who volunteered for the study were relatively well-educated and computer-literate and the sample may have over-represented patients who were motivated and actively engaged in self-management. People with lower health literacy and from other cultures may have different needs that were not explored in this study. The majority of participants had also not had any previous experience of using a computer-based self-management program. The risks of using a naïve population are that interventions might not be optimized for the needs of potential users who are most comfortable using Web-based tools and suggestions for intervention development would not be grounded in a detailed understanding of existing interventions. However, the advantage of using a “naïve” population is potentially better generalizability with the needs and preferences representing the general population rather than early adopters who might be more technically savvy. The exponential growth of the Internet and mobile phone use illustrates how useful and user-friendly technology can have a mass-market impact; thus, exploring the needs of a study population that represents a more generalizable population beyond early adopters would be important in maximizing the potential benefits of such an intervention. Well-designed DHIs for diabetes have been shown to engage and benefit users with a wide range of health literacy, and findings can be generalizable to a wide range of users [41].

There were many areas where a DHI could help address the unmet needs described above. There have been a number of systematic reviews of DHIs in type 2 diabetes, including narrative syntheses, meta-analyses, and meta-ethnographies [9,42-44]. The interventions described in these reviews have tended to focus on improving the medical management of type 2 diabetes through information provision and behavior change support [45-49]. However, for participants in this study, the “work” of emotional and role management was more important for them than most aspects of medical management other than food and diet. Previous studies have highlighted inconvenience and poor motivation as reasons for poor engagement with DSME and DHI for diabetes [7,49]. This study suggests that another important aspect might be a potential mismatch between the content and emphasis of existing DSME interventions and patient-defined needs and priorities for diabetes self-management. Increasing the focus on emotional management for such interventions also has clinical importance as cross-sectional and prospective evidence suggests that changes in diabetes-related distress correlate with changes in glycemic control, possibly through changes in adherence with medication, whereas depressive symptoms are correlated with self-management behavior [50,51].

By focusing on medical management and information provision, existing health care services and education programs may not be adequately meeting the needs of patients with type 2 diabetes. DHIs have the potential to help improve access to DSME and extend the range of content offered by health services to meet a wider range of patient needs. The features of a DHI that could address the unmet needs described by participants in this study included an emphasis on emotional and role management, being available 24/7, having up-to-date evidence-based guidance for patients, and providing access to peer-generated and professional advice. The findings of this study have been used in the development of an effective DHI for adults with type 2 diabetes, called HeLP-Diabetes, which has been evaluated in a randomized controlled trial and implementation study and has been shown to be acceptable to a wide range of users [23].