(2019-2020) Reference: 2018 PROD 00041. Funded by the European Regional Development Fund (ERDF) of the European Union under the FEDER Operational Program of Catalonia 2014-2020. 99.530,10 €. P.I. Josep Vehí. Other researchers: Iván Contreras.
The HYPOGLYCEMIA MINIMIZER is a software package that can be embedded in any diabetes management app or system. It combines four machine learning approaches to assess the risk of having a hypoglycemic event:
- grammatical evolution for the mid-term continuous prediction of blood glucose levels,
- support vector machines to predict hypoglycemia during postprandial periods,
- artificial neural networks to predict hypoglycemia overnight, and
- data mining to profile diabetes management scenarios.
The HYPOGLYCEMIA MINIMIZER combines these four methods to help the patient to take decisions before an insulin injection. The system provides advanced calculations that minimizes the risks, alarms and warnings, reducing to less than half the number of hypoglycemic events. It can be adapted for any pump and for MDI and it is also compatible with any CGM currently in the market, or even could also work without a CGM.
At the end of the project we will have a technologically and clinically validated complete prototype. The clinical trial will include patients in real-life everyday situations for sufficient time to prove their efficiency.