fbpx

What AI Can/Cannot (and Could!) Do for Type 1 Diabetes

Living with type 1 diabetes (T1D) involves constant efforts to keep blood sugar levels within the recommended range to avoid potential complications. This is a significant burden, both physically and mentally, for people with T1D and their loved ones.

Artificial intelligence (AI) enables machines to perform tasks that normally require human intelligence (e.g., decision making). AI has represented an exciting revolution in the management of T1D these past few years. 

An ongoing revolution

Although we may not always realize it, AI is all around us, including in T1D management technologies (e.g., mobile applications, CGMs, and artificial pancreases). Data is collected in real time and sent to AI algorithms to help predict upcoming blood sugar levels and suggest insulin doses accordingly (e.g., artificial pancreases, smart pens).

AI is also used to analyze health data (e.g., lab test results, medical history, lifestyle habits) and to predict and assess the risk of long-term diabetes-related complications. With the help of AI, programs can “learn” (e.g., from mistakes that were made) and improve their accuracy with each diagnosis. 

A few years ago, Canadian researchers developed an AI algorithm that can detect diabetic retinopathy (eye damage) early and very accurately (94%) from a simple picture of the retina. By identifying risks, interventions to prevent them can be implemented more easily and quickly.

What we can hope for in the next few years

The next few years should see significant developments in AI for the prevention and detection of T1D and the personalization of treatment. These improvements should help ease the burden of managing this disease. Here are a few examples of current research at various stages of development: 

  • CGMs that measure blood sugar more accurately and more frequently to improve predictions.
  • Mobile apps to accurately scan and count carbs (e.g., from a picture of a plate of food).
  • Insulin pumps with algorithms that detect eaten meals or physical activity to adjust the administration of insulin in real time more accurately. Just imagine not having to count carbs ever again!

AI can also help accelerate scientific research by collecting, reconciling and cross-referencing information from studies around the world to target candidates for research or to analyze results more accurately. 

What AI cannot do

Although AI shows great potential to facilitate the management of T1D, there are still significant limitations to this technology, particularly in terms of cost and the amount of data required to function optimally. 

Also, because AI algorithms are based on mathematical models, they may not consider certain factors that can affect blood sugar, such as stress or hormonal changes. So, this technology is no match to the judgment and knowledge of people with T1D and their healthcare team.

References:

  • Nomura, Akihiro et al. “Artificial Intelligence in Current Diabetes Management and Prediction.” Current diabetes reports vol. 21,12 61. 13 Dec. 2021, doi:10.1007/s11892-021-01423-2
  • Ellahham, Samer. “Artificial Intelligence: The Future for Diabetes Care.” The American journal of medicine vol. 133,8 (2020): 895-900. doi:10.1016/j.amjmed.2020.03.033
  • Grzybowski, Andrzej et al. “Artificial intelligence for diabetic retinopathy screening: a review.” Eye (London, England) vol. 34,3 (2020): 451-460. doi:10.1038/s41433-019-0566-0
  • Rollo ME, Aguiar EJ, Williams RL, et al. eHealth technologies to support nutrition and physical activity behaviors in diabetes self-management. Diabetes Metab Syndr Obes 2016;9:381–90. https://doi.org/10.2147/ DMSO.S95247.
  • Zhang W, Yu Q, Siddiquie B, Divakaran A, Sawhney H. “Snap-n-Eat”: Food Recognition and Nutrition Estimation on a Smartphone. Journal of Diabetes Science and Technology. 2015;9(3):525-533. doi:10.1177/1932296815582222
  • Dankwa-Mullan, Irene et al. “Transforming Diabetes Care Through Artificial Intelligence: The Future Is Here.” Population health management vol. 22,3 (2019): 229-242. doi:10.1089/pop.2018.0129

Written by:

  • Sarah Haag RN. BSc.

Reviewed by:

  • Amélie Roy-Fleming Dt.P., EAD, M.Sc.
  • Ana Teresa de Luna Pallone, MD. 
  • Andréanne Vanasse, Nathalie Kinnard, Jacques Pelletier, Michel Dostie, Marie-Christine Payette, Claude Laforest, patient partners of the BETTER project

Linguistic revision by: Marie-Christine Payette