It seems to me that China measured exactly opposite the way Italy and Spain (must be) measuring.
However, it is important to measure into the general population of "suspected" maybe/maybe not infected people and get some recovery data. Not all, but some. It is as if Italy and Spain are measuring only those more or less that are "serious/critical" and that does not do any justice to the general population.
-- if you try to test everybody, you have the problem of (How many times?), not infected today but maybe tomorrow or in a few hours on your way home in the bus...
I would hope China reported real death numbers, even if only 1/2, still "fine". I'm talking about the vast difference in total recovered...