From the knowledge engineering point of view, the observation of patients subjected to clinical protocols of therapy constitutes a domain characterized by the existence of strongly structured knowledge. We have approached the problem from the perspective of a homogeneous and modular knowledge representation theory, based on the concept of Generalized Magnitude. This concept arises from identifying and collecting all possible facts of a domain established a priori, and being inspired by the concept of physical magnitudes. The Generalized Magnitudes scheme includes temporal extensions necessary to solve a medical problem for which exists a therapy and a follow-up plan with temporal specifications, and also facilitates the creation of advisory expert systems.