The recipe for a healthy lifestyle is to eat well, stay active, reduce stress and take medication as prescribed. For someone with type 2 diabetes, small changes in any of those categories can have big impacts – positive or negative.
Living with diabetes requires ongoing access to diabetes care to help manage the condition, so patients meet with their health care team about once every three months. But what if they’re having trouble between appointments?
Hamilton Health Sciences (HHS) researcher Diana Sherifali set out to discover if it’s possible to prevent a small issue from becoming a bigger one by using artificial intelligence. She’s taking the first steps to develop a health coaching algorithm to help people with diabetes. This algorithm could be added to existing fitness or wellness apps that already track diet and exercise.
Diabetes health coaching
As a clinical nurse specialist at HHS, an associate professor in McMaster University’s School of Nursing and associate scientist at the Population Health Research Institute (PHRI), Sherifali has been exploring if a type of artificial intelligence called machine learning could be the solution. PHRI is a joint institute of HHS and McMaster University.
A computer can eventually learn winning strategies for nearly any situation.
“Individuals living with diabetes are managing the condition every day. This means at least 95 per cnet of diabetes management occurs outside of the health-care system,” says Sherifali, who is also a certified diabetes educator who understands the benefits of health coaching. “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.”
Type 2 diabetes occurs when your pancreas doesn’t make enough insulin, which regulates sugar within the body, or your body doesn’t respond well to insulin. There is no cure, so it requires managing blood sugar levels. This is done through dietary modifications, exercise, maintaining a healthy body weight, monitoring blood sugar levels and possible medication or insulin injections.
Digital health care experts
If the health coaching algorithm is applied to a wellness tracker it can use the existing diet and exercise data. Then, once individuals add their weight, blood sugar levels and medications, the algorithm can determine if changes need to be made and provide recommendations on what to do.
To explore and develop this idea further, Sherifali partnered with experts in digital health and data science at HHS’ CentRE for dAta Science and digiTal hEalth (CREATE).
“With my basic knowledge of artificial intelligence I knew I needed to work with CREATE, so I approached them with the idea right from the start,” says Sherifali. “I was excited when I was told it was worth exploring.”
Teaching the computer
When you play chess on your phone do you ever wonder how the computer knows how to play? By automatically playing thousands of games and being rewarded for victories and penalized for losses, a computer can eventually learn winning strategies for nearly any situation. This approach to machine learning is called reinforcement learning and is leading to cutting-edge advancements in many applications of artificial intelligence, including self-driving cars.
The same approach has been used by CREATE for computerized diabetes health coaching, says Jeremy Petch, CREATE’s founding director.
“We provide the algorithm with health data, a medical professional’s recommendations and the outcomes – good and bad,” says Petch. “This allows it to learn the best strategy under all different circumstances, just like the computer opponents you play against on your phone.”
In this case, the data includes blood sugar levels, medications, nutrition, physical activity, weight and stress.
After developing and testing the algorithm, the team determined that it does provide accurate initial recommendations.
More data makes a stronger algorithm
“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,” says Petch.
The next stage of the study will teach the algorithm how to provide accurate recommendations with more complex data. Then it will eventually be tested with clinicians and finally, patients.
“The challenge is that there’s always more data,” says Sherifali. “The algorithm isn’t meant to replace in-person appointments so we’ll need to determine at what stage there is enough data for the algorithm to be effective with coaching individuals through many different issues, while leaving complex scenarios for a medical team to address.”
Since fitness and wellness apps are already well established as a vehicle to house and track the kind of data that people with type 2 diabetes are already monitoring, the implementation of the algorithm into these apps won’t be the most challenging part of the project. The first steps – determining if the algorithm will work, are actually the hardest steps.
So Sherifali says she’s excited that this first stage is already showing success. It means that heath coaching for those with type 2 diabetes could be at their fingertips in the near future.
