Those sorts of things affect disease incidence rates but not "death rates after diagnosis of major disease."

And Japan, for example, has an aging population of heavy smokers.