DIABET: Dual sensing AI for non-invasive diabetes management
Background
Diabetes is a global epidemic. Every year, around 5 million people die globally from diabetes, or related conditions, according to World Health Organisation (WHO). However, many of these complications are predictable and preventable by better self-management (SM) and control of the disease, which must be a core feature of any effective care plan for a person with diabetes.
A major challenge in the successful management of diabetes is the accurate self-monitoring of blood-glucose (SMBG). Unfortunately, the invasive nature (finger prick test) and cost of best-in-class current glucose monitoring solutions means SMBG remains low. Semi invasive solutions are available, but these are also costly and still require combination with daily blood testing using finger prick method.
Summary
DIABET addresses the need for a non-invasive continuous monitoring solution and enables better self-management of diabetes. The DIABET project has developed an innovative digital device for high accuracy non-invasive blood glucose monitoring.
This project has allowed the improvement of accuracy using dual sensing technique (Radiofrequency and Infrared). Machine learning and data fusion help overcome imperfections arising from each technique acting individually.
Benefits
The combination of advanced sensors, data processing, and artificial intelligence developed by the consortium partners enables the first truly non-invasive continuous monitoring solution for people with diabetes
Through the continuous monitoring of blood-glucose levels, the DIABET smart monitor can also be able to act as a decision support system (DSS) e.g. offering dietary and lifestyle advice. With patient consent, DIABET can also enable "clinical decision-making support" through sharing of this patient-specific data between health care professionals (e.g. General-Practitioners, endocrinologists and psychiatrists), social-care professionals and diabetes nurses.
Outcome
The research led to the development of a table top device, to be used at home, or in Point of Care settings, coupled to a bracelet for overnight blood glucose trend monitoring. The developed prototype system combines advanced sensors for multiwavelength electromagnetic transmission (Infrared and Radiofrequency), data processing, and Artificial Intelligence algorithms developed by the consortium partners to enable a truly non-invasive continuous monitoring solution for people with diabetes.
Project Partners