The future of diabetes technology: Enhancing patient experience with predictive continuous glucose monitoring

• Diabetes is a complex condition requiring a lifetime of careful management through blood glucose monitoring
• Solutions for blood glucose monitoring have greatly improved in the last 40 years, but challenges to optimal glucose control remain
• Artificial intelligence-enabled continuous glucose monitoring is the future of diabetes technology, helping people with diabetes to proactively manage their condition

The latest figures from the International Diabetes Federation estimate that 590 million people worldwide have diabetes, with 1 in 9 adults (aged 20 – 79 years) living with the condition. By 2050, these already stark figures are projected to rise dramatically, reaching 853 million people—a 46% increase.¹ As one of the most expensive health conditions to manage, the growing rate of diabetes globally presents a pressing challenge for healthcare systems worldwide and a demand for better diabetes technology.²

Most of the cost of diabetes care is not a result of day-to-day management, but comes from complications associated with the condition. In the United Kingdom, for example, out of the £10.7 billion spent on diabetes care, approximately 40% is spent on routine diabetes care, and the remaining 60% is spent on diabetes-related complications such as heart failure, strokes and amputations.³ However, effective diabetes management, essential for avoiding complications, remains a complex task for anyone living with the disease.

To stay within target glucose ranges, people with diabetes must make dozens of health-related decisions on a daily basis. Decisions regarding diet, insulin administration, doctor visits, and blood glucose monitoring must be constantly weighed against a complex matrix of personal calculations. Simple activities like dining out, walking the dog, commuting, and sleeping require careful planning around blood sugar. People with diabetes make most of these choices alone, without medical support, but even with meticulous preparation, blood sugar responses can be unpredictable. The potential “what-ifs” in the day ahead, which are difficult for everyone in this busy, modern age, are doubly hard for people with diabetes, which adds an additional layer of stress and mental health burden.

While solutions for at-home monitoring have significantly improved over time, helping people with diabetes to better plan around their disease, most glucose monitors still offer a reactive approach to disease management. That is, they are limited in their effectiveness because they provide only past and present glucose values. Recent advances in artificial-intelligence-enabled monitoring devices, however, are shifting diabetes management from reactive to proactive self-management. These technologies predict where future glucose levels are headed so that people with diabetes can take action before an incident occurs.⁴

People with diabetes have insufficient production or utilization of insulin, necessitating intervention to regulate blood glucose levels. Blood glucose monitoring is an integral part of diabetes management since fluctuations in blood glucose levels can result in serious consequences.⁵ Over time, raised glucose levels, known as hyperglycemia, can lead to cardiovascular disease, nerve damage, and vision loss. On the other hand, low blood glucose levels, known as hypoglycemia, can cause seizures, loss of consciousness, and in some cases death.⁵ Monitoring glucose levels informs decisions on treatment, nutrition, and physical activity in order to avoid dangerous fluctuations.

Since the 1980s, blood sugar meter technology, self-monitoring of blood glucose (SMBG), has been available for people living with diabetes.⁶ SMBG allows people with diabetes to check blood values at their convenience with a quick prick of the finger and, thanks to its accuracy and ease of use, it remains an invaluable diabetes technology. However, this technology’s usefulness is limited to point-in-time information, and repeated finger pricks throughout the day can be uncomfortable and inconvenient.⁶

In the late 1990s, diabetes management had a breakthrough with the approval of a wearable technology, the first continuous glucose monitoring (CGM) device. As the name suggests, CGM tracks glucose levels continuously, providing people with near real-time information about their glucose levels and removing the need for finger pricking. Glucose levels are shared to a receiver, such as a smartphone, allowing users to easily and conveniently check glucose levels.⁷ By allowing a fuller picture of what’s happening inside the body, people can make more informed diabetes management decisions.⁸ For example, studies have shown that people with type 1 and type 2 diabetes using CGM have fewer hypoglycaemic episodes and an overall lower long-term blood glucose average.^9,10

Despite CGM providing a more advanced option for people with diabetes to manage their condition, there are still areas for improvement in care. As CGMs provide near-constant information on glucose levels, people using them can feel overwhelmed by large amounts of data and may be confused about how to interpret or act on results.¹¹ This can lead to a feeling of powerlessness or overstimulation unfortunately compounded by a key component of CGM devices: an alarm announcing either high or low glucose levels. Although these alarms can be effective, their intrusive nature can cause stress, disturb sleep, and create alarm fatigue.¹² Current CGM devices can sometimes lead to lower levels of treatment adherence, and people still experience hypoglycaemia.¹³

Whether using SMBG or CGM to manage diabetes, the management style is reactive. But what if people with diabetes could predict what their glucose levels will be an hour from now, two hours from now, or overnight while sleeping? Such predictions could encourage an active role in diabetes management, with the aim of empowering people to take proactive measures to reduce or prevent diabetic complications.

It is here where artificial intelligence is playing a new and powerful role. By enhancing the recognition of glucose patterns and effectively forecasting glucose dynamics, AI can translate vast amounts of data into useful, actionable insights for people with diabetes and their care providers.^4,14

One key area that adds to the burden of diabetes management is nocturnal hypoglycaemia – episodes of very low blood glucose that occur during nighttime. A single severe incident of nocturnal hypoglycaemia can have dire, even fatal, consequences. Understandably, the fear of nocturnal hypoglycaemia impacts the mental, physical, and social well-being of people with diabetes, causing anxiety, tension, lack of sufficient sleep, or sleeplessness.⁴

The assistance of an AI-enabled CGM solution offers the ability to predict the likelihood of hypoglycaemia throughout the night, and in silico studies have shown that time spent in nocturnal hypoglycaemia can be reduced by 37% with this technology.^{15, 16}  In addition, by acting in advance to a predicted nighttime hypoglycaemia, threshold alarms and associated alarm fatigue can be reduced, subsequently decreasing distress, improving quality of life, and leading to overall better glycaemic control.¹²

Effective glycaemic control is not only crucial for people managing diabetes, it also translates into long-term economic benefits by reducing healthcare spending. An algorithm-driven CGM that predicts short-term blood glucose levels as well as imminent and nocturnal hypoglycaemia, can prevent the costs associated with days spent in the hospital, changes to medications, visits to healthcare providers, and treatment for associated mental health/distress.¹⁷

With the help of AI, people living with diabetes can anticipate upcoming complications and take corrective action, empowering them to manage their own care with less worry. AI-enabled medical devices create the conditions for deeply personalized care, tailored to the unique metabolism of each individual. Acting not just on past and present glucose values, but on an individual’s predicted glucose values represents a paradigm shift transforming diabetes care. 

Importantly, this advancement also has the potential to reduce clinical and economic burden on already taxed healthcare systems, by preventing hospital admissions from diabetes-related complications. Over the next decade, predictive CGM technologies are expected to become the gold standard for at-home diabetes management, bringing peace of mind to the hundreds of millions of people likely to be impacted by this challenging condition.¹⁸