Support  >  Mobile App

Predictive Insights | Automated Decision Support

One Drop uses proprietary machine learning technology and over 3 billion data points, collected from One Drop users, to power 8-hour blood glucose forecasts—delivered through the One Drop mobile app.

These blood glucose predictions are coupled with actionable, behavioral advice known as Automated Decision Support (ADS). 


To use this feature you must:

  • Use One Drop mobile on iOS or Android
  • Select type 2 diabetes, gestational diabetes or prediabetes during app on-boarding, or via "My Account" profile section of the app


When will I receive a blood glucose forecast?


After entering or uploading your very first blood glucose value to the One Drop app! Moving forward, you will receive a pop-up notification after every blood glucose reading and as you continue to manually and/or passively collect other health data points (e.g., food, activity, steps).

One Drop will notify you of your expected blood glucose direction (e.g., BG to rise or fall) for the next 8-hours and couples this prediction with advice—what you can do right now to maintain or achieve healthy blood sugars (e.g., take a 15 minute walk).


How can One Drop make these predictions?

One Drop's machine-learning model gathers data from you, as well as from other users with similar profiles and behaviors. One Drop has over 3 billion data points to pull from, making our predictions extremely accurate.

 
How accurate are my blood glucose predictions?

One Drop ADS accurately predicts future blood glucose values with 91% of predictions within +/- 50 mg/dL of glucose meter readings and 75% within +/- 27 mg/dL. Overall accuracy improves as more data is fed to the machine learning models, and the accuracy for any given individual improves as he/she enters more personal health data into the app.


I don't want predictions!

Continue tapping “Not Useful” when prompted by the pop-up notifications.

Updated: January 22, 2019

Was this article helpful?
100% of people found this helpful.