Abstract The notion of risk questions our dominant and traditional medical approach to health phenomena: a binary normal-pathological model anchored in the pathophysiological framework. For risk factors which are continuous biological variables such as hypertension or hypercholesterolemia, the demarcation line between a level which would be ‘normal’, ‘at risk’, or ‘pathological’ is drawn in a conventional and non-fixed manner. Above all, these risk factors, which are the subject of preventive treatment, are increasingly considered as diseases in their own right. This shift would reinforce the pathologisation of human life and the associated problems of overdiagnosis and overtreatment. However, while it is easy to see the value of clarifying and specifying the distinction between ‘risk’ and ‘disease’ when dealing with those problems, finding the best way to do so remains challenging. In this chapter, I begin by describing the confusion between these two categories. I then show that the conceptual analysis, in particular, the naturalistic definition of disease, fails to provide a satisfactory distinction. Finally, I argue that this distinction can be better established if we understand the epidemiological risk approach as a different way of modelling health phenomena than the binary and categorical approach of pathophysiology.