A recent study has revealed that more than just carbohydrates influence blood sugar levels, indicating that current automated insulin delivery systems might be oversimplifying the management of Type 1 Diabetes (T1D). Researchers from the University of Bristol have found that the varying insulin needs of people with T1D are not adequately addressed by a ‘one size fits all’ approach.
The study, published in JMIRx Med, delved into the complex patterns of insulin needs in T1D patients using OpenAPS, an advanced automated insulin delivery system. It was found that these needs are frequently influenced by factors beyond just carbohydrate intake.
Lead researcher Isabella Degen, from Bristol’s Faculty of Science and Engineering, stated: “Our findings reinforce the idea that factors other than carbohydrates significantly impact blood glucose levels. However, without a way to measure these influences, automated insulin delivery systems struggle to maintain optimal blood glucose levels, potentially resulting in levels that are too low or too high.”
T1D is a chronic disease where the body doesn’t produce enough insulin, a hormone needed to control blood glucose. The main treatment involves administering insulin, which must be carefully balanced with carbohydrate intake to prevent spikes in blood glucose levels. Other factors, like exercise, hormones, and stress, also affect insulin needs. The new study shows that these factors often cause unexpected, significant impacts on blood glucose levels, making insulin dosage a complex and often imprecise task.
The research underscores the intricate nature of glucose regulation in T1D and the diverse insulin needs among T1D patients, emphasizing the need for a personalized approach to treatment.
In order to incorporate these non-carbohydrate factors more systematically into clinical practice, scientists need to find ways to measure and quantify their effects on insulin needs. This could also enhance the accuracy of blood glucose predictions, which the study found can’t be reliably made based on insulin and carbohydrate information alone.
Degen added: “Managing Type 1 Diabetes is much more complex than simply counting carbs. The invaluable insights we’ve gained from studying automated insulin delivery data prove it’s worth the effort to analyze this type of real-world data. What surprised us was the wide variety of patterns we observed, even among a relatively uniform group of participants. Clearly, there’s no ‘one size fits all’ solution to managing diabetes.”
The research team is now working on refining methods to analyze the diverse and complex real-world medical data, including addressing issues of irregular sampling and missing data. They are actively seeking long-term, open-access automated insulin delivery datasets and collaborations with experts in time series and machine learning to aid their future research. The goal is to uncover more nuanced patterns and complexities in the data to drive innovations in personalized diabetes care.
