During this study, we would like to assess the efficacy in keeping blood glucose in a safe range (70-180 mg/dL) of the second generation of a Glucose Predictive and Therapy Advisory system in diabetic patients treated by basal-bolus insulin regimens using pumps or multiple daily injections. We want to compare the time spent in safe range while using the predictor and advisor outputs and when the patient is not using them. This is a randomized controlled crossover trial. Patient will come for two admissions of 3 days, one with the DIAdvisor system fully activated, and one with the prediction and advice features disabled.
