Since most epi data comes in the form of incidence, would be very useful to have this as a model output. Maybe as a derived parameter.