Ageing is nowadays a major public health problem. Its description remains complex, both individual and collective conceptualization being interlaced with a strong subjective dimension. Health professionals are increasingly required to integrate ageing and prevention into their thought and to create adapted protocol and new tools. Ageing characterizes unavoidable changes in the body. It is usually measured by the age dependent on time and called “chronological age”. However, the criterion « chronological age » reflects imperfectly the actual ageing of the body depending on many individual factors. Also, this criterion has for a long time been replaced by another composite criterion called « biological age » supposed to better reflect the ageing process. In order to build a score of ageing adapted to general practice, a new methodology is proposed suitable for general practitioners. First of all, a first phase of this work consisted in a qualitative and quantitative survey conducted among general practitioners in France. This survey was done to obtain data on the use of predictive scores by general practitioners in their daily practice and their appropriateness, as well as to know the reasons of their non-utilization. Results showed that predictive scores are useful tools in daily practice to target a complex systemic approach insofar as they are simple to use (few items, items suitable for general practice) and their scientific validity is easily understood. In addition, patient’s age has been cited as a major criterion influencing general practitioners use of a predictive score. Results of this first phase have been used to propose a model of biological ageing, with reflexion on mathematical model as well as on component variables of this model. A selection of variables as markers of ageing was carried out from a review of the literature, taking into account their capacity of integration in general practitioners’ daily practice. This selection was completed by a mathematical approach based on an ascending process on a regression model. A control sample, assumed to be "normal ageing" on the basis of current knowledge in general medicine, was then used. This sample was first carried out from a review of the literature and then from a K-means method that classified this sample into several groups. The statistical dependence of measured variables was modeled by a Gaussian copula (taking into account only linear correlations of pairs). A standardized biological age was defined explicitly from these correlation coefficients. The tails of marginal distribution (method of excess) were estimated to enhance the discriminating power of the model. Results suggest interesting possibilities for a biological ageing calculation, and the predictive score they provide, suitable for general practitioners’ daily practice. Its validation by a morbidity and mortality survey will constitute the final phase of this work
