Living with type 2 diabetes is quite straightforward – eat well, exercise and meet with a health care team every three months!
However, it’s possible to have problems in between appointments, of course, which is why Hamilton Health Sciences (HHS) researcher Diana Sherifali is taking the very first steps to try and use artificial intelligence in order to address all of these potential small issues.
More precisely, the plan is to develop a coaching algorithm specifically created to provide people with diabetes with the help they need.
To make things easier, the algorithm could be simply added to an already existing wellness or fitness app many use to track their activity and diet already.
This allows the algorithm to track all the data it needs from the app or device of choice and then provide accurate advice.
Sherifali explained that “Individuals living with diabetes are managing the condition every day. This means at least 95% of diabetes management occurs outside of the health-care system. It’s unrealistic for their care teams to be helping them on a daily basis. But, if a machine learning algorithm is developed and applied to existing technology that these individuals already use, it could fill in some of the support needed between appointments.”
She decided to partner with HHS’ CentRE for dAta Science and digiTal hEalth, also known as CREATE, in order to further explore and develop her idea.
“I approached them with the idea right from the start. I was excited when I was told it was worth exploring.”
As for how it works, CREATE’s founding director, Jeremy Petch, says that “We provide the algorithm with health data, a medical professional’s recommendations and the outcomes – good and bad. This allows it to learn the best strategy under all different circumstances, just like the computer opponents you play against on your phone.”
The algorithm has been created and tested and sure enough, it is able to provide correct recommendations for basic needs but more data will make the algorithm even more accurate and useful.
Of course, the team is not done, Petch explaining that “Now that we’ve determined the algorithm can learn the appropriate recommendations for common issues encountered by those with type 2 diabetes, we need more detailed data to continue to refine it.”
The next step involves teaching the algorithm how to provide patients with the appropriate advice after analyzing much more complex data.
Finally, it will be tested by clinicians before finally going out to the public.