Digital health apps for diabetes have made strides in addressing key challenges from access to engagement and cost of diabetes care. In the virtual Scientific Sessions of Endocrine Society 2021 held this past week, a symposium entitled “Data Overload - How to Integrate Diabetes Technology Into Your Research and Practice” highlighted the current status of diabetes management mobile apps. They leveraged this new telehealth medium to reach patients outside of the healthcare setting and leveraged AI to increase patient monitoring and help medical guidance in the real world. These are valuable improvements, but much more lies ahead.
These apps have increased the frequency of encounters between providers and patients from a few times a year to multiple touchpoints within a week or even a day. However, for chronic conditions like diabetes, it’s all the small decisions made throughout daily life by the patients themselves that drive successful management. To date, the digital health landscape is still focused on developing products to bridge gaps between healthcare providers and diabetes users.
While digital health apps have lowered the barriers to provider encounters, scalability is still limited, and cost remains an issue. For the most part, these apps use artificial intelligence to collect data remotely but still rely on high-cost healthcare providers for clinical analysis and, more importantly, coaching. To reach true scalability in diverse socioeconomic demography, the AI must be able to operate independently of human providers in support of its users (patients) throughout their lives by delivering personalized, meaningful and actionable lifestyle guidance at the right place and the right time.
Digital health apps have primarily focused on improving the delivery of care, but beyond the delivery, the therapeutic approach also needs to change. The outcome for lifestyle therapy, as with all therapies, is driven by the degree to which benefits outweigh side effects. It’s well known that in diabetes lifestyle therapy, the side effects — unpleasantness of changes — significantly hinder compliance and suppress outcomes. AI has the potential to achieve precision in lifestyle change such that its users are only asked to make changes if it will work for that user. This amplifies outcomes and lowers burden, which in turn drives compliance. It is a powerful evolution from the overly broad mantra of “diet and exercise” to a personalized and precise understanding of each patient as an individual.
To take this further, AI has the opportunity to seek out enjoyable foods, activities, and behaviors that its users can indulge in without negative impacts on their health. This enables a departure in tone away from avoiding the forbidden toward pursuing the enjoyable that is largely missing from the current approaches to lifestyle coaching.
Whether it’s trying to prevent root causes for onset and progression instead of treating symptoms, asking patients to make major lifestyle changes now to avoid complications ten years later, or trying to get a few hours ahead on predicting insulin needs, timescale remains one of the biggest challenges to effective diabetes care. For the most part, current therapies and guidances are reactive and corrective instead of preventative. Even with the most advanced technology on the market, glucose prediction still relies largely on trending short-term data. The ultimate goal of AI is to be able to predict glucose far enough in advance to enable smooth insulin delivery and to be able to bridge the feedback gap between big lifestyle changes now with the benefits that manifest decades later.
The COVID-19 pandemic accelerated the evolution of digital medicine to an unprecedented pace. We are likely at the dawn of a new era that will eventually allow us to live longer, happier, and healthier through inspiring digital tools at lower costs than traditional healthcare providers.