[Presentation and automatic treatment of evolutive data in medicine. Application to clinical pancreatic pathology]

In this paper we propose a method allowing the representation of time-dependent changes in symptomatology. In algorithm, which automatically generates the sequence of stable states in patient's symptomatology evolution, describes slow, as well as fast, variations of diseases. This sequential representation of data can then be analysed by classical statistical procedures in a population of patients suffering from pancreatic diseases, to estimate the rate of global modification of clinical symptomatology in time.