Infectious agents have to cope with the immune system of their hosts. In particular, it is the immune system that records past exposures as

well as vaccination history. As a consequence, the interaction between various strains of a pathogen is mediated by the immune status of the individual hosts. We therefore need nested models, i.e., models for transmission at the population level that have models for within host parasite-immune system interaction as a building block. The aim of this talk is to sketch (in wishful thinking spirit) a possible top-down approach based on the delay equation formulation of physiologically structure population models. (By 'top-down' I mean that specification of a submodel for within-host processes is postponed.)