Critically ill-patients often experience stress-induced hyperglycemia. This research demonstrates the effectiveness of a simple automated insulin infusion for controlling the rise and duration of blood glucose excursion in critically ill-patients. Heavy derivative controllers derived from a simple, two-compartment model reduced blood glucose excursion 79–89% after a glucose input in proof-of-concept clinical trials. Modelled performance is very similar to clinical results, including a strong correlation between modelled and actual insulin consumed, validating the fundamental models and methods. However, the need for additional dynamics in the model employed is clearly illustrated despite capturing the essential dynamics for this problem.