The idea is that nicotine may lower your chances of infection, but once established I imagine that smoking will definitely reduce your chances of survival.
Presuming this is true (lots of evidence but still much too early to be sure), there's two possibilities: nicotine makes a person less likely to be infected, or nicotine makes it less likely that the infected will develop any symptoms. No symptoms, no test, that's been the rule until quite recently.
If it's the latter, it could explain what's going on here. I doubt that's the explanation, but it's possible.