Mobile Health (mHealth) for Diabetes Care: Opportunities and Challenges

The number of adults with diabetes has quadrupled since 1980, affecting more than 400 million people worldwide, and is estimated to increase to 642 million by 2040. ^1,2 Treatment and management of this disease account for 12% of global health expenditure. ² The disabling and life-threatening complications of diabetes can be avoided or delayed by intensive glycemic control. ³ However, less than half of the people with diabetes receive standard health checks. ⁴ A comprehensive model of diabetes care constitutes a collaboration between healthcare providers (often a multidisciplinary team), the patient, and care givers or the persons responsible for the patient. With the increasing prevalence of diabetes and limited health resources, it is unlikely that the traditional model of care for diabetes management can be sustained. Therefore, innovative and more efficient models of care for diabetes are warranted.

During the past decade, mobile technology has introduced a range of innovations in our daily life. Mobile health or mHealth is the use of mobile phones and wireless digital devices to support the achievement of health objectives. With an estimated 5 billion mobile phone users worldwide, ⁵ mHealth has the potential for improving quality of diabetes care and reducing the cost. Mobile digital devices can improve communication between patients and healthcare providers, enhance self-management, and facilitate remote monitoring of patients. ⁶ Although smartphones are becoming more popular, basic mobile phone functions (e.g., voice call and text messaging) still play a remarkable role in improving clinical outcomes of diabetes care. ⁷ However, it remains unclear what constitutes the success of mHealth interventions and who can benefit the most from these interventions. A review study by El-Gayar et al. highlights the need for research on factors that impact the usability and perceived usefulness of mobile apps to improve the adoption of mHealth for diabetes self-management. ⁸

In this issue of Diabetes Technology & Therapeutics, Nelson et al. report the results of a study on the intervention elements and user characteristics that determine the utility of MEssaging for Diabetes (MED). ⁹ This mHealth system is designed to improve medication adherence of people with type 2 diabetes mellitus (T2DM). They used personalized text messages (SMS) and interactive voice response (IVR) to help the patients overcome their perceived barriers to medication taking. The MED intervention comprised (a) a daily tailored SMS, (b) a daily assessment SMS at the end of the day, and (c) a weekly IVR call delivering customized feedback to the user. Eighty adults with T2DM who were taking antidiabetic medications trialed the intervention for 3 months and were asked to provide feedback on the utility of MED.

Sixty of the 80 MED users agreed to be interviewed during or at the end of the trial. The majority of respondents said they read the daily messages and found them easy to understand and personally relevant. Daily assessment SMS played the role of an extra reminder for most of the respondents to take their medications. Nearly all of the respondents reported that they knew how to work with IVR and most of them believed IVR helped them better adhere to medication taking. However, the respondents generally liked text messaging more than IVR. This could be partly because of technical difficulties of using the IVR system. Younger patients and those who were more recently diagnosed with diabetes provided more favorable feedback toward MED.

Although this study elaborates on the consumers' perspective of the use of an mHealth intervention, the findings should be interpreted in the light of the limitations of this study, such as low number of interviewees (60 patients), single-site trial, and relatively short duration of intervention (3 months). The fact that the patients who declined to interview had higher HbA1c and lower health literacy may indicate that the people who are more in need of diabetes care are less likely to use mHealth interventions. Moreover, for a comprehensive mHealth intervention, in addition to medication adherence, other aspects of self-efficacy and self-management for diabetes care should also be explored from the patient's perspective. These are especially important for more complex interventions that can be delivered through mobile health applications (apps) running on smartphones.

Smartphones are becoming much more popular than basic phones in almost all nations. In 2016, 80% of the mobile phones sold worldwide were smartphones. This disruptive technology has transformed the way we conduct many of our daily activities such as shopping, banking, and socializing. Proliferation of mobile health apps is rapidly increasing. Despite tightening the conditions for publishing health apps, ¹⁰ the number of available health apps in major app stores increased by 57% in 2016 and reached 250,000 apps. The number of downloads of health apps in 2016 is estimated at 3.2 billion, 7% growth in the current year. This trend indicates an obvious oversupply of mHealth apps. Nevertheless, there seems to be no quality control over these increasing number of mobile health apps. ¹¹

Diabetes is expected to remain the top target of mobile apps (73%), followed by obesity (40%) and hypertension (27%). ¹¹ Moreover, emerging digital health technologies such as smartwatches, wireless physiological sensors, and wearable devices could transform the healthcare in the near future. ¹² The high penetration rate of smartphones and production of health apps provide an unprecedented opportunity for enhancing self-management of diabetes and provision of a range of patient-centered and personalized diabetes care. A recent systematic review of the randomized trials showed that using mobile apps in people with T2DM was associated with a reduction of HbA1c of 0.49%. It also revealed that younger age and receiving feedback from the healthcare professionals enhance the effect of mHealth for diabetes care. ¹³

Despite the availability of a high number of health apps (with a large proportion of them available for free), the adoption of mobile health remains low. The majority of freely available diabetes apps for self-management of diabetes do not meet the minimum requirements of an effective app. ¹⁴ Given the massive number of diabetes apps, it is not an easy task for the patients and doctors to choose a “good” app for diabetes managment. ¹⁵ A 2016 survey shows that after the app stores, health insurers and physicians are the main distributing channels of mobile health apps. Physicians would not recommend a mobile app unless shown to be safe, clinically effective, and preferably approved by a regulatory body (e.g., Food and Drug Administration [FDA]). Similarly, health insurers would less likely advocate a mobile health app unless shown to be capable of reducing the cost or improving the quality of the care. A holistic approach involving all the stakeholders should be taken for the design and implementation of mHealth services. ¹⁶

Although poorly designed medical apps have potential to harm the users, very few of them have been assessed and approved by regulatory bodies such as the FDA. ¹⁷ Similarly, a very low proportion of available medical apps have been scientifically evaluated and the majority of those apps, which have been scrutinized in a research study, are not developed beyond proof of concept or efficacy evaluations. Review studies highlight the dearth of evidence on the effectiveness of mHealth and call for robust evaluation of long-term effectiveness and socioeconomic impact of mHealth. Clearly, rapid advancement of mobile health technology outpaces research studies in this field and traditional designs of research are not necessarily capable of filling this gap.

Research studies involving mHealth face several challenges. The recruitment rate of these studies is often low, and it is difficult to retain the participants in the clinical trials. Most reported trials are no longer than a few weeks. Research should focus on how to develop mobile health apps to engage more patients and retain users over time. Strategies to engage patients who do not use health apps are required. Often this group of patients has the worst diabetes control and self-management and thus would theoretically gain the most from such interventions. Uptake and retention might improve if the health apps involve entertainment characteristics. Gamification of health apps may be a useful strategy.

Development of medical apps should be based on established theories or guidelines to be clinically effective. However, wide adoption of mHealth will not be possible without the support of appropriate policy and legislation. Lack of legislation for reimbursement of mHealth services is a major barrier for sustainability of interventions beyond the research phase. For policy makers to advocate mHealth, evidence of the effectiveness and efficiency of new interventions is required. Future studies should move on from efficacy studies to clinical and cost-effectiveness studies and ultimately assess the impact of mHealth at a social level. ¹⁸

The MED intervention of Nelson et al. is one step to further understanding the characteristics of patients who utilize mHealth. Mobile apps should be safe and clinically effective for physicians to recommend them, should reduce costs of care for insurers to support them, and should be acceptable and engaging for users to adopt and retain them. With the current trend of publishing new health apps, the production rate of health apps clearly exceeds the potential evaluation and assessment capacity of researchers and regulatory bodies. Flexible and more responsive research approaches are required to ensure that the full potential of mobile health apps is examined and understood.