Researchers used machine learning techniques to add virtual CGM data to the famous DCCT study in order to assess the predictive value of CGM data for the development of diabetes complications.
Free access research article:
https://www.liebertpub.com/doi/10.1089/dia.2024.0404
@Boerenkool Thanks for posting this link! I hadn’t seen the study and found it interesting. Seems a good use of ML and the DCCT data. Would be interesting to see similar done with the Tidepool collection data (represents the CGM data of an equivalent number of people to the DCCT study), though for a bit less time (I think) than the DCCT. The Tidepool data might be used as confirmation of the accuracy of the ML inference data.