Winter 2016: bSMARTbSMART predicts future blood sugar levels from past history.
Abraham Pabbathi is a Chicago Booth Alum and is also the father of a 4 year old Type 1 diabetic daughter. Ever since his daughter was diagnosed Abraham has been working on building an app which can help him manage his daughter’s diabetes. He started working on this solution after he did extensive research on existing solutions and found none that was simple and useful. He has been working on this solution for six months and is now ready to package it in the form of a smart phone app.
What’s bSMART trying to solve?
My app will help predict blood glucose level 2 hours into the future and make suggestions on insulin dose adjustments. The app requires at least two weeks of insulin, carbs intake and activity data to start making the predictions. The app will have an interface to log blood glucose numbers, carbs intake, insulin dose and activity levels. The app will run a proprietary data analytics algorithm and predict how the blood glucose levels would be in the next two hours given certain amount of carbs, insulin and activity levels. The predictions are usually within 20% of the range 75% of the time. Even the best CGMs out there could be off by 20% so prediction within 20% range is actually quite good. The app also allows you to test out what-if scenarios to help you administer the right amount of carbs, activity levels and insulin dose. But the human body is quite unpredictable and especially growth hormones and epinephrine (adrenalin) can affect the blood glucose levels which are very difficult to predict..
What are we going to build?
We’re going to build a basic MVP of the app.