The PRAESIIDIUM consortium, in which CheckHealth is a partner, is developing solutions aimed at providing real-time prediction of the risk of prediabetes in individuals. The project is now launching its clinical studies to collect more data for the prediction modelling.
To predict the risk of prediabetes, the project uses physics-informed machine learning based on a rich dataset of real-life data, combining registry-based research with behaviour tracking through wearables and nutritional data collection using AI.
To acquire additional bio-clinical data for the modelling and testing of wearable sensors, the project is launching clinical studies in Austria, Lativa and Italy with the following focus points:
- Patients with low risk of prediabetes
- Patients with metabolic risk factors and
- Plasma bio-banking and clinical data collection
CheckHealth is supplying LinkWatch as the backend platform for organising and preparing all collected patient data for analysis together with VoiceRPM for individual nutritional data collection, and solutions for wearable device connectivity.
The PRAESIIDIUM approach to risk prediction goes beyond traditional black-box based approaches by combining mathematical models and explainable AI techniques while improving the prediction performances and reducing the computational time of the risk calculation based on the simulations of the mathematical models.