Space-time variation of cancer incidence data and healthcare resources allocation

In this work we analyze cancer incidence data collected between 1992 and 2005 in a specific area of France. Our primary aim is to monitor changes of cancer incidence in space and time taking into account of age group categories. We assume that the distribution of the different age categories is not the same in all of the geographical units of the region and change in time. Such analysis can be useful for healthcare resources allocation, in order to replan the allocation of hospitals according to patient age. To do so we use Generalized Additive Mixed Models (GAMMs) with a Poisson response, using the methodology presented in Wood (2006).