GP’s concern is that the quantisation scale is not representative of linear severity. It’s more like classification of disjoint characteristics tagged with colour

I won't offer an opinion of my own but I don't disagree with that take.