The history of medicine is a path of progress and discoveries, characterized by a constant search for knowledge for human care and well -being. Nowadays medical practice is based on three principles: knowledge (model), experience and statistics (evidence-based medicine, EBM). Mathematics is increasingly entering medicine with two distinct approaches: namely, data-driven and model-driven. However, this must not limit professional freedom or distort the traditional pathophysiological modelling approach that is the only one accepted by doctors, who must remain the protagonists of the therapeutic decisions. This paper wants to show how model-driven approaches offer more guarantees, compared to data-driven ones, to maintain a pathophysiological model (accepted by clinicians) and therefore bring an important contribution to the knowledge for human care. On the other side, only data-driven approaches can discover variables involved in a complex system. So, the integration of these two approaches is the future. In addition, only the collaboration between different skills and the construction of a common language can bring innovative solutions for the management of health care and patient treatment.