Clinical evaluation of a novel adaptive bolus calculator and safety system in Type 1 diabetes

Bolus calculators are considered state-of-the-art for insulin dosing decision support for people with Type 1 diabetes (T1D). However, they all lack the ability to automatically adapt in real-time to respond to an individual’s needs or changes in insulin sensitivity. A novel insulin recommender system based on artificial intelligence has been developed to provide personalised bolus advice, namely the Patient Empowerment through Predictive Personalised Decision Support (PEPPER) system. Besides adaptive bolus advice, the decision support system is coupled with a safety system which includes alarms, predictive glucose alerts, predictive low glucose suspend for insulin pump users, personalised carbohydrate recommendations and dynamic bolus insulin constraint. This thesis outlines the clinical evaluation of the PEPPER system in adults with T1D on multiple daily injections (MDI) and insulin pump therapy. The hypothesis was that the PEPPER system is safe, feasible and effective for use in people with TID using MDI or pump therapy. Safety and feasibility of the safety system was initially evaluated in the first phase, with the second phase evaluating feasibility of the complete system (safety system and adaptive bolus advisor). Finally, the whole system was clinically evaluated in a randomised crossover trial with 58 participants. No significant differences were observed for percentage times in range between the PEPPER and Control groups. For quality of life, participants reported higher perceived hypoglycaemia with the PEPPER system despite no objective difference in time spent in hypoglycaemia. Overall, the studies demonstrated that the PEPPER system is safe and feasible for use when compared to conventional therapy (continuous glucose monitoring and standard bolus calculator). Further studies are required to confirm overall effectiveness.Open Acces