(75g) Artificial Intelligence to Improve Blood Glucose Control for People with Type 1 Diabetes

An individual with type 1 diabetes (T1D) must be vigilant in monitoring their blood glucose values while considering the impact of meals, exercise, and other critical events. More and more people with T1D are wearing continuous glucose monitors (CGM, providing blood glucose values at 5 min intervals) and insulin pumps that continuously deliver rapid-acting insulin. However, only a few commercial closed-loop systems, that automatically adjust insulin in response to glucose changes, are currently available. These systems require manual meal announcements to provide insulin boluses at mealtime (feedforward control) and automatically suspend insulin when glucose is low. Still, missed meal announcements and exercise-induced hypoglycemia are major causes for suboptimal glucose control in people with T1D. Another common problem is CGM pressure-induced sensor attenuation (PISAs) that typically occur overnight when a user sleeps on their CGM, resulting in inaccurate CGM signals. PISAs can cause the pump to shut off unnecessarily. Evaluating the likelihood of CGM readings being PISAs can allow a controller to ignore readings with a high PISA probability.

Automatic activity detections and forecasts can allow a controller to prompt for meals and anticipate exercise. To this end, we have developed an automatic activity detection system using a smartwatch and smartphone. Long Short-Term Memory (LSTM) networks, convolutional neural networks (CNN), probabilistic graphical models (PGM), and lookup tables are used to create minute-based probabilities of other, sleep, eating, exercise, and null using sensor data. A particle filter is used to detect activity events by aggregating minute-based probabilities. Population level time of day prior probabilities are used to forecast exercise and sleep. Support vector machines (SVM), random forests, and neural networks are compared for classification of PISAs.

The CNN achieves a class-weighted accuracy of 85% using smartwatch sensor readings. The particle filter detects 99.9% of sleep, 99.4% of eating, and 94.4% of exercise events. 4.3 false sleep events/day, 2.5 false eating events/day, and 0.38 false exercise events/day are detected. Time of day probabilities provide higher likelihoods of impending exercise and sleep when they truly occur compared to when they do not. For PISA classification, the random forest achieves a true positive rate of 90% and a false positive rate of 6.3%.

Reliable activity detections, and forecasts of exercise and sleep can be achieved with daily patient review. Using a random forest, reliable likelihoods of PISAs can be obtained. Signals from the CGM and insulin pump can therefore be supplemented with activity detections/forecasts and PISA detections to improve glucose control. Furthermore, these techniques can reduce patient burden and allow people with T1D to lead more normal lives.