A reduced sampling schedule for estimating the parameters of the glucose minimal model from a labeled IVGTT

In the minimal model of tracer glucose disappearance, 30 blood samples are usually drawn to measure the insulin and the labeled glucose concentration time courses, which provide the database for the model identification. The problem of reducing the number of blood samples withdrawn without, however, deteriorating parameter estimation is addressed. By using a stepwise approach, largely based on optimal experiment design theory, the authors have derived a reliable reduced sampling schedule, which comprises only 14 total samples. The practical application of this reduced schedule has been validated by applying it in a new set of human experiments, completely independent of the ones used in the design phase.