INSULIN TITRATION ALGORITHM BASED ON PATIENT PROFILE

摘要

A device, system and method is provided for predicting risk of hypoglycemia and adjusting an insulin medicament dose size for a subject being on pen based multi-daily injections. A first algorithm is provided for computing a first insulin dose size based on a plurality of timestamped blood glucose (BG) measurements and insulin injected dose sizes. A risk prediction module is adapted to, based on machine learning algorithms, predict risk of hypoglycemia based on the timestamped BG measurements and calculate a second insulin dose size resulting in a non-hypoglycemic state. In case the second dose size is lower than the first dose size, the first dose size is automatically adjusted down to the second dose size and communicated as the recommended dose size to the subject.