e-Health

PEPPER: Patient Empowerment through Predictive Personalised Decision Support

The PEPPER project will empower people with diabetes by providing intelligent decision support for self-management through a predictive, knowledge-based system that is adaptive, personalised and patient-centred. The work will model the effects of a variety of diverse factors (e.g. clinical, biological, therapeutic, and behavioural data) on glycaemic outcomes. The resulting decision support system will make recommendations to optimise care using a novel approach that will build on recent research in artificial intelligence for insulin dosing. The system will connect a case-based reasoning engine with a glucose forecasting algorithm to maximise safety. As such, the system will offer advice that is highly adaptive to the needs of individuals who need to solve uncertain and variable problems on a day-to-day basis. The rationale for using a knowledge-based approach rather than a model-based technique is that existing models for glucose forecasting are not able to make reliable estimations sufficiently far into the future to be beneficial for open-loop therapy (i.e. where the patient administers the insulin).