Diabits - Glucose Monitoring Made Easy
See your future glucose values and manage your fluctuations with ease. Diabits Calculates future glucose using a Machine Learning Algorithm designed to learn your unique physiology.
-
Most popular related searches
Software Overview
60-minute Predictions
See your predicted glucose values 60-mins ahead*
Nocturnal Hypoglycemia
See Hypoglycemia predicted 3 to 8 hours ahead.
24/7
Freedom
Know before dangerous fluctuations happen.
No more surprises!
Industry Leading Accuracy
How Diabits Helped Jenn
Jenn always had confidence in her ability to manage her type 1 diabetes - she had done so for most of her life. But one day, she experienced a critical low that changed everything.
While studying at the University of British Columbia, she was resting in her dorm room and unintentionally put pressure on her continuous glucose monitor (CGM). The CGM stopped reading her blood sugar correctly, and failed to alert her of a low.
Jenn felt something was off, but by the time she got up to have a snack, she was already lightheaded. She fell, struck her head on the coffee table and was unconscious in her dorm for two hours. If she hadn’t been discovered in time, she could have died.
After two days in hospital, Jenn recovered, but her confidence didn’t. The incident had dealt a heavy blow to her sense of independence.
When we met Jenn two weeks later, she was still shaken by her critical low. To prevent a repeat, she had kept her blood sugar high by delivering less insulin, but that takes a damaging toll on the body over the long term. She needed a solution that could help her identify potential lows before they occurred, and give her the confidence to keep her blood sugar in range again.
Diabits was that solution. Jenn is now able to view her estimated future blood sugar values and take action to prevent lows. And in the event of a critical situation, her parents receive ample warning with Diabits’s follower feature.
Before using Diabits, Jenn did not have the confidence to keep her blood sugar in a safe range, and it was high all the time.
After using Diabits, Jenn could see where her blood sugar was headed, and was confident in knowing that she would not go low. This helped Jenn increase her time in range and reduce her blood sugar variability and HbA1c.
Diabits Accuracy
The Diabits algorithm performed with 93.6% accuracy for OhioT1DM data set.
We use Deep Learning methods to analyze and model glucose metabolism. Our approach takes advantage of Neural Networks ability to 'learn' and 'think' like a pancreas.
Dataset
The OhioT1DM was developed for blood glucose level prediction research. The dataset consists of 8 weeks of continuous glucose monitoring via Medtronic sensors and self-reported life-event data for 12 people with type 1 diabetes.
CGM 8 Weeks
This dataset is chosen because in addition to blood glucose values, carbohydrate and insulin events are also recorded.
Results
We recorded the sensitivity (true positive rate), specificity (true negative rate), and accuracy of our predictions for previously unseen patients.
The model predictions and the actual blood glucose values were given a label, post prediction. The labels are based on ADA’s recognized blood glucose ranges.
- 93.6% Accuracy
- 96.1% Specificity
- 83.6% Sensitivity
The labels for actual and predicted values were used to calculate the accuracy, specificity, and sensitivity.
Predicted vs. Actual
Below is the partial overlay of actual vs predicted blood glucose values.
Use this link to join the Beta version on TestFlight with experimental features.
Customer reviews
No reviews were found for Diabits - Glucose Monitoring Made Easy. Be the first to review!