Transcript – Personal Continuous Glucose Monitoring System for Diabetes Self-Management

Hello, my name is Bobak Mortazavi, and I’m here to present our T-3 project personalized continuous glucose monitoring system for diabetes self-management, along with my colleagues, Dr. Sherecce Fields and Dr. Hu. The purpose of this study is to aid those suffering from pre-diabetes and Type two diabetes in managing their diabetic care. This really means minimizing excess glucose levels and preventing potential hypoglycemic events through Just-In-Time alerts and interventions. However, this automated Just-In-Time monitoring requires strict control of diet and exercise. Now, while exercise tracking exists, the diet monitoring portion of this is really quite challenging. A number of solutions exist that involve manual logging through things like my fitness pal, but with continuous glucose monitors, we have an opportunity to automate this through wearable sensing techniques. And in particular, we have a potential to automate diet monitoring by knowing the physiological changes to glucose response based on meals and the macro nutrient composition of those meals. So we aim to develop a tool that will track activity in context of what a user is doing, so when are they active? How much energy are they spending? When are they eating? When are they sleeping? When are they working? When are they driving to and from work and using all of this as context to track energy expenditure and when it will lower glucose versus when eating will raise glucose.

We have a preliminary study that looked at identifying contexts and users and identify moments of eating that will help label when they’re having breakfast or when they’re having lunch without the user having to label, without having this excess logging burden. And it will allow us to signal segments, the glucose response to specific meals and for their two three study to be two zero one nine zero seven nine three. We intend to take this platform of activity recognition, which has seen a 15 percent increase in eating detection rates with very minimal user labeling, then standard algorithms for activity recognition. And we’ll take this with our glucose monitoring and provide a solution that is able to track diet as you’re taking food in and let you know when you’re going to be hypoglycemic and potentially need to exercise or hypoglycemic and potentially. And we look forward to continuing this use subject studies to validate that this kind of intervention tool can work.