Type 2 diabetes mellitus (T2DM) is a serious public health concern globally. The prevalence of T2DM is increasing at a rapid pace in developed regions, such as Western Europe including the Netherlands, and causes a substantial economic burden [1-3]. Various lifestyle factors are important for the development of T2DM [4,5]. To manage diabetes sufficiently, adherence to regular physical activity and a healthy diet are important. Several studies have reported the positive effects of lifestyle interventions, including regular physical activity and healthy food intake, on the stabilization of blood glucose levels and health status of patients with T2DM [5-10].

T2DM requires self-management and support. Self-monitoring of blood glucose levels, food intake, and physical activity can support the self-management of patients with T2DM. In recent years, there has been an increase in the development and availability of technologies for diabetes self-monitoring, especially mobile health apps [11], for example, an app to integrate and keep track of blood glucose levels in combination with data regarding physical activity and food intake. These apps have considerable potential to support diabetes self-management and have a positive effect on a person’s lifestyle [12-14]. However, studies have shown that the uptake of apps for managing diabetes is rather low [15-18]. Insight is needed into how apps can be integrated into diabetes self-management care [19]. Research regarding the acceptance of apps for managing diabetes among patients is important for their successful implementation. Several quantitative and qualitative studies have been performed to gain further insight into the acceptance of apps for managing diabetes among patients with diabetes [20-24]. Zhang et al [22] investigated predictors of the intention to use apps for managing diabetes using a web-based questionnaire. They found that performance expectancy (ie, perceived usefulness) and social influence were the most important determinants of the intention to use apps for managing diabetes. Torbjørnsen et al [23] conducted interviews to obtain an in-depth understanding of users’ acceptance of a mobile app for diabetes. They found that users’ acceptance of mobile apps for diabetes self-management differed. Regular use of an app could be useful (supportive and educational) but could also become a burden, requiring too much time and not contributing enough to the effort needed to change lifestyles. Furthermore, Torbjørnsen et al [23] concluded that both practical (ie, usability and utility) and social aspects (ie, attitude and shared understanding) are important for the acceptability of mobile apps for diabetes. Jeffrey et al [24] conducted semistructured phone interviews among patients with T2DM and found that a lack of knowledge and awareness of apps as health care tools was one of the barriers.

The unique aspect of this study is the use of an explanatory sequential design (mixed methods) among both users and nonusers. To explore the actual use of apps for T2DM and gain greater insight into the main barriers and drivers, a web-based questionnaire using the Unified Theory of Acceptance and Use of Technology (UTAUT) was deployed along with in-depth interviews. The UTAUT is a unified model that was developed by Venkatesh et al [25] and is commonly used globally in studies regarding health technology acceptance. The UTAUT is based on the Social Cognitive Theory with a combination of 8 prominent information technology acceptance research models (Theory of Reasoned Action, Theory of Planned Behavior, Technology Acceptance Model, Motivation Model, a model combining the Technology Acceptance Model and the Theory of Planned Behavior, Model of Personal Computer Use, Diffusion of Innovation theory, and Social Cognitive Theory). Validation by Venkatesh et al [25] showed that the UTAUT accounts for 70% of the variance in behavioral intention to use and about 50% in actual use. On the basis of the results of this study, recommendations are described for future research and the integration of apps with diabetes self-management care.

The aim of this study is to investigate the actual use and the barriers and drivers among users and nonusers of mobile health apps for T2DM. The use of a mixed methods approach is different compared with most previous studies using the UTAUT and investigating the use of eHealth among patients with diabetes. Most studies have applied quantitative research methods, and some have conducted interviews. The addition of in-depth interviews based on the quantitative findings is therefore different from previous studies and could show more explicitly where patients with T2DM need support and how the barriers and drivers influence their use of apps to self-manage diabetes. Furthermore, the focus on patients with T2DM is rather novel, as is the inclusion of nonusers of apps. These nonusers were included to understand their view toward apps, and the hypothesis of this study is that some of them are not unwilling to use apps but rather are unaware of the availability.

An explanatory sequential design [26], using a mixed methods approach, was applied to assess the actual use, barriers, and drivers among users and nonusers of apps for T2DM. This study started with the collection and analysis of a web-based questionnaire using Qualtrics software (Qualtrics International Inc). The web-based questionnaire was designed and reported based on the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) [27]. The web-based questionnaire was pretested for usability and technical functionality before fielding. The qualitative data were used to design the interview guide (Multimedia Appendix 1) so that it follows from the quantitative phase and provides the opportunity to obtain in-depth information. The following in-depth semistructured phone interviews lasted between 25 and 50 minutes, audiotapes were made, and field notes were taken. Ethical approval was obtained from the University of Twente Ethical Committee (reference number 201213). Participants were informed about the purpose of the study and length and time of the questionnaire and which data were stored and where before the start of the web-based questionnaire. They provided web-based consent and were informed about their right to withdraw at any time. Data were anonymized, confidentiality was maintained, and the data will be retained for 10 years.

The study population included patients with T2DM aged ≥16 years spread across the Netherlands. Recruitment for the web-based questionnaire was performed using convenience sampling by publishing the link on web-based platforms for patients with T2DM, in the newsletter of the Dutch Diabetes Association, and social media. A total of 183 respondents completed the questionnaire. In total, 80 respondents did not complete the questionnaire sufficiently (missing data on actual use or barriers and drivers) and were excluded. Questionnaire data were analyzed from 103 respondents. At the end of the questionnaire, participants were asked to leave their phone number if they were willing to participate in a follow-up phone interview. A total of 16 respondents participated in the phone interviews; 10 were users and 6 were nonusers of apps to manage T2DM. Data were collected between July and December 2020.

The questionnaire data were analyzed using IBM SPSS Statistics (version 25.0). Descriptive statistics were performed for the actual use of apps among patients with T2DM and the individual items of the UTAUT. Differences between users and nonusers were tested using chi-square tests for the individual items. Independent 2-tailed t tests were performed to test for differences in mean scores per UTAUT construct. Therefore, 3 items were reverse-coded: Using apps would cost me a lot of time and energy (effort expectancy), Other people would think bad of me if I used apps (social influence), and When I think about using apps, I fear that confidential information could end up in the wrong hands (trust in data security). To analyze the qualitative data, interviews were transcribed verbatim. Content analysis with a deductive approach was performed on all transcripts guided by the UTAUT [32]. The UTAUT was used in this analysis to align the integration of quantitative and qualitative data. Data management was performed using the NVivo (version 11) software package. During the organizing phase of the analysis, a matrix was developed comprising the components of performed expectancy, effort expectancy, social influence, facilitating conditions, anxiety, and trust in data security and knowledge. Categories were created within each component of the analysis matrix (Multimedia Appendix 2).

Credibility was established using several procedures [33]. Method triangulation was performed using multiple data collection methods, namely questionnaires and individual interviews with audio recording. Researcher triangulation was achieved by 3 researchers (MB, CMvL, and TJJO) who performed the interviews and read and compared the findings. Peer debriefing took place during weekly meetings with the project team, where scientific and organizational aspects were discussed. When data collection was complete, a member check was performed by sharing an infographic with the respondents. A thick description was developed for transferability [33], which included the sampling method, recruitment, data collection, questionnaire and interviewing method, and analysis process.

Of the respondents who completed the questionnaire, 55.3% (57/103) were men and 41.7% (43/103) were women (Table 1). The mean age was 69 years (SD 10.8; range 27-90 years). Most respondents (97/103, 94.2%) were of Dutch origin, and 28.2% (29/103) had a lower education, 17.5% (18/103) had an intermediate education, and 50.5% (52/103) had a higher education. Most respondents (84/103, 81.6%) had been diagnosed with T2DM ≥10 years ago. Demographic characteristics of the respondents from the interviews (n=16) were comparable with the demographic characteristics of the overall respondents from the questionnaire (N=103; Table 1).

On the basis of the questionnaire, 55.3% (57/103) were nonusers and 44.7% (46/103) were users of apps for managing T2DM. Of the 46 respondents who reported using apps for T2DM, 35 (78%) had used them for ≥12 months.

Table 2 shows the differences in mean scores between users and nonusers of apps for T2DM per the UTAUT construct. Mean scores were significantly higher among users of apps for managing T2DM for the constructs performance expectancy, effort expectancy, facilitating conditions, and knowledge compared with nonusers (P<.001). In addition, users scored significantly higher regarding social influence (P=.007) compared with nonusers, whereas mean scores regarding anxiety were significantly lower among users compared with nonusers (P=.003). Finally, the mean scores for trust in data security were significantly higher among users compared with nonusers (P=.02).

The aim of this study was to investigate the actual use and the barriers and drivers among users and nonusers of mobile health apps for T2DM. This study showed that the main drivers were the belief that using apps for managing diabetes will result in better personal health and well-being and that, by sharing their data with health care professionals, users will receive improved support. The barriers included time and energy to keep track of data and understand the app, lack of support by health care professionals, and no financial support or reimbursement by insurance companies. Nonusers had more doubts regarding the improved support provided by apps and showed more anxiety and distrust toward the use of apps. However, most nonusers were willing to try apps to help manage their diabetes.

One strength of this study is the triangulation method of applying an explanatory sequential design. Hence, rich data were collected, which provided in-depth insight into the actual use of mobile health apps among patients with T2DM and the main barriers and drivers for use. Second, the well-established UTAUT model was used as the base for the questionnaire [25]. The interviews provided in-depth insight and a more explicit description of each UTAUT item. Finally, the study was performed, and data were analyzed by 3 researchers, which contributed to researcher triangulation. There were some limitations to this study. A possible bias considers digital literacy and respondents’ interest in technology. Most respondents were of Dutch origin and had a higher educational level. Barriers to and motivators for the use of mobile health apps for managing diabetes may differ among specific subgroups (ie, people of non-Dutch ethnicity or people with a lower level of education). Approximately half of the respondents were app users for T2DM, which is a rather high percentage compared with other studies [15-18]. It might be the case that most respondents who filled out the questionnaire had a personal interest in this topic. Furthermore, the response and completion rates (103/180, 57.2%) of respondents who completed the web-based questionnaire were relatively small.

The literature, as well as this study, has shown that apps may increase diabetes self-management. It is important to integrate apps for managing diabetes into the daily practice of diabetes self-management care, with health care professionals playing an important role. On the basis of the findings of this study, we recommend that health care professionals get more involved and acquire relevant knowledge about mobile health apps specifically for patients with diabetes. The technology involved assists patients with T2DM to self-manage diabetes and assists professionals in supporting clients in their self-management. Currently, multiple apps are available for the management of diabetes. Therefore, it is difficult to know which apps would be most beneficial to whom and why. A study to investigate patient experiences among specific subgroups (ie, people of non-Dutch ethnicity or people with a lower level of education) in the use of different apps for managing diabetes is the next step for research. In addition, training patients with T2DM and professionals regarding the availability and use of apps is recommended. Important topics for such training include digital knowledge and competencies, learning about apps, how to track data, and how to read and use the collected data. The implementation of apps for managing diabetes in daily practice is complex. This study provides recommendations that focus on the main drivers and barriers. However, other factors (such as the type of organization, availability, and type of patients) also play a role in the implementation process.

One of the main drivers for use was the belief that using apps for managing diabetes would result in better personal health and well-being. The time and energy required to keep track of the data and understand the app were mentioned as barriers. Patients with T2DM stated that health care professionals’ engagement is of the utmost importance in supporting them in app use. In addition, patients stated that insurance companies can play a role in facilitating the use of diabetes apps, for example, by assuring reimbursement. Further research should focus on the evaluation of patients’ experiences with different apps for managing diabetes, how to integrate apps into diabetes self-management care, and investigating barriers and motivators in the use of mobile health apps for the management of diabetes among specific subgroups (ie, people of non-Dutch ethnicity or people with a lower level of education).