This answer includes a good breakdown on the difference between correlation and causation, the article claims that the mortality ratio was adjusted to account for drug use, alcohol use and body mass, but a quick look at the actual results of the paper show that sleep chronotype has no effect on mortality rate, but habits do. While sleep deprivation is a totally different story when it comes to mortality, going to bed late doesn't seem to be much different from going to bed early, as long as you are sleeping enough.