Modelling of glycaemic control in diabetic patients

The use of short-term prediction algorithms of the subcutaneous (s.c.) glucose concentration may contribute significantly to the prevention of hypoglycemic events and the daily management of insulin-treated diabetes. We have treated s.c. glucose prediction as a multivariate regression problem using support vector regression (SVR). The method is based on variables concerning: (i) the s.c. glucose profile, (ii) the plasma insulin concentration, (iii) the appearance of meal-derived glucose in the systemic circulation, and (iv) the energy expenditure during physical activities. We have also extended our SVR model to predict separately the nocturnal events during sleep and the non-nocturnal (i.e. diurnal) ones over 30-min and 60-min horizons for a hypoglycemic threshold of 70 mg/dl. Additional variables have been introduced accounting for recurrent nocturnal hypoglycemia due to antecedent hypoglycemia, exercise and sleep. The application of RF and RreliefF feature evaluation algorithms on real-life Type 1 diabetes data has also been proposed as a means to customize the input of glucose predictive models.