DIABETES is a global problem, with the disease’s incidence and prevalence increasing annually. Globally, approximately 537 million adults had diabetes in 2021, although in Australia it is the fastest-growing chronic condition, with around 1.3 million Australians living with it in 2020. Diabetes is estimated to cost almost $2.5 billion every year to the Australian health care system , and if combined with its complications, a significant $6.86 billion annually.
Diabetes is associated with considerable morbidity and mortality; thus, it is paramount not only to detect diabetes in a timely fashion but also to prevent and delay its complications. Management options for diabetes include insulin therapy, oral hypoglycemics, regular glucose monitoring, and behavioral modification such as healthy eating and increased physical activity.
Diabetes also requires ongoing self-management to prevent complications and reduce the risk of long-term disability. Some self-management techniques include monitoring diet and blood glucose trends, complying with medication or insulin schedules, and increasing physical activity levels.
Self-management applications, or in other words, mobile health applications, are rapidly evolving and have the potential to improve diabetic care as they are convenient to use, easily accessible, portable, have a broad reach, and are economical. They allow users to perform many self-management activities through medication reminders, dose calculators, food and activity tracker, community support, glucose management diaries and many more in one place rather than maintaining many paper-based records.
Mobile application use has been associated with positive self-management behavior and increased physical activity and has been able to significantly reduce glycated hemoglobin (HbA1c) and improve quality of life while reducing health care costs and preventing complications (here, here, here).
A supportive technology for diabetes prevention and management that has gathered much attention in the past few decades is artificial intelligence (AI). AI has been predominantly used in four critical areas of diabetic care: clinical decision support, predictive population risk stratification, automated retinal screening, and patient self-management applications.
This article will mainly focus on the self-management approach using mobile health applications for diabetes management, its mechanism of action and its effectiveness.
Mobile applications with or without AI capabilities are typically developed keeping the concept of behavioral modification in mind. There is a plethora of clinical evidence and research studies emphasizing that behavior is the driving force and key to improved health outcomes in diabetes. The data from multiple studies cutting across diabetes types and age span point to the fact that 30–50% of overall control of diabetes is due to behavioral change. Interventions modifying behavior alter diabetes management behavior, which has an impact on HbA1c level and other outcomes such as ketoacidosis, hospitalizations, and hypoglycaemic events.
Mobile applications are a powerful way by which people with diabetes can take control of their own diabetes management. Coupled with AI, these applications provide people with diabetes an opportunity to make decisions about activity and diet while supporting the calculation of the caloric and nutritional value of different meals. Another benefit AI adds to these applications is that its algorithms can provide meaningful insights on the progress charts and figures, which allows one to make an informed decision based on real insights.
It is impressive how AI adds so many capabilities to mobile health applications; however, a critical aspect of diabetic care that developers often overlook is the bond between the patients and their health care providers. It is not uncommon for people with diabetes to have multiple professionals providing multidisciplinary care. And, for diabetes, regular monitoring and easy access to care providers are a must. Unfortunately, people with chronic diseases such as diabetes often find it difficult to access their providers promptly, leaving them with poor management of the condition.
Some applications also offer the functionality of virtual behavioral coaching where a virtual coach, either human or AI, reads entered data and gives personalized recommendations in real time. Adding virtual coach functionality has shown more promising results over a simple application use for psychosocial support, glycemic control, and behavior change due to data-driven personalization facilitating better engagement of users and higher satisfaction.
Personalization is the key to modern medicine, including diabetes management, whether therapeutics or self-management. Self-management applications should not be based on a one-size-fits-all approach; instead, they should be designed with a one-size-does-not-fits-all approach. The result of a systematic review of reviews and meta-analyses indicates that when the applications provide personalized feedback, there is a significant improvement in HbA1C level, clearly underlining the importance of personalization in diabetes self-management applications. To be effective, interventions should incorporate the needs of the end users and be personalized.
The critical benefit of using a virtual coach is that their human-like appearance allows users to interact with them with familiarity and realistically through facial expressions, body movements and speech, giving users a personalized experience. This enables the development of a working and social relationship between the virtual coach and the user, enabling sustained engagement of the user with the application, an essential element needed in the context of diabetes management and application use.
To summarize, mobile health applications using AI capabilities are a promising intervention in helping people with diabetes to self-manage their diabetes better. Developed with relevant features, these applications are bringing a personalized touch to self-management of diabetes. Their uptake and use, however, can be improved when the personalized needs of the end users are taken into account.
Dr. Jeetendra Mathur is the Co-founder and Project Manager at Healea, a technology-oriented solution for diabetes self-management. His extensive experience in the health care industry involves working in various medical specialties, including emergency medicine and burns/plastic surgery at large public and corporate hospitals in both India and the Maldives. He is passionate about making health care better and safer with a focus on digital health and artificial intelligence.
Dr. Sandeep Reddy is the Director of the MBA (Healthcare Management) program at Deakin University. In addition to a medical degree, he has qualifications in medical informatics, management, and public health. He has managed various health service projects and formulated high level policies in Australia, New Zealand, and Europe.
The statements or opinions expressed in this article reflect the views of the authors and do not necessarily represent the official policy of the AMA, the MJA or InSight+ unless so stated.
Subscribe to the free InSight+ weekly newsletter here. It is available to all readers, not just registered medical practitioners.
If you would like to submit an article for consideration, send a Word version to firstname.lastname@example.org.