Validation of mathematical models of complex endocrine-metabolic systems. A case study on a model of glucose regulation.

The validation process is an essential component of the modelling ofin vivo endocrine and metabolic systems. In the paper a validation study of a comprehensive model of the glucose regulation system, previously developed for intravenous testing, is performed on a new data set based on oral glucose tolerance studies. A novel approach based on a ‘partition and input/output inversion’ technique which allows the full use of this new data set for validating a submodel of the previous model, instead of having to do an extra modelling effort, is presented. The necessity of a new description for a number of glucose unit processes emerges from this study. The new submodel has been successfully tested and incorporated into the comprehensive model. Although the approach presented has been developed and illustrated with reference to a specific case study, it may have wider applicability in developing appropriate validation strategies for complex physiological control systems.