The fatality rate is based off of having required facilities and equipment. With those run out the fatality rate will jump from .5% to 5%.
The WHO published one number for case fatality rate. According to you, they should have published two (or more) numbers.
The reality of the situation is that epidemiologists use certain fixed constants to understand the virus as part of a very crude model. The WHO is and was providing a number for this model. In these models, the distinction you made is not captured whatsoever (thus "crude model").
The WHO got to its number based on averaging unreliable data. This has two problems. The first is garbage in, garbage out. The second is that this averaging is an unscientific method.
I get what you're trying to say though, and broadly agree. Epidemiology models, insisting on a single fatality number, are so crude they are nearly not useful for making predictions.