Is The Age-Period-Cohort Model Well Suited to an Epidemic Context? The Case of the French BSE Epidemic