The Patient Path: a first approach

International audienceEuropean healthcare systems are faced with multiple challenges, including anaging population, an increase in chronic diseases and patients withmulti-morbidity, and limited financial and human resources. The response tothese challenges is based in particular on the organisation of care into carepathways. Namely, Once the data necessary for the construction of a carepathway are acquired and processed, one has to model the patient pathwaymathematically in a generic way. After that, using clustering algorithms, one canidentify patients subgroups, then, mine for common treatments, predict thefuture of patient pathways and answer clinicians’ questions. All these steps wouldlead to an automated processes which has to be evaluated by medical experts.Available statistical methods remain limited and inefficient to construct carepathways. Indeed, data obtained from health care providers and insurancecompanies are all time-dependent, highly heterogeneous, qualitative in part, withseveral thousand possible modalities and mainly made up of missing data. Wepropose an approach based on functional data analysis combined withlongitudinal data analysis in order to construct care pathways