I think it's purely a media thing. When the public health folks I know have been modeling, they've been modeling with a primary eye toward number of ICU beds needed and number of ventilators needed. For instance if you look at https://mn.gov/covid19/data/modeling/index.jsp which has a pretty decent model (on Github at https://github.com/MN-COVID19-Model) you can see all their pdf slides predict peak infection, number of ICUs/vents needed, and then mortality. But that's not what gets headlines, as you observe.
It is certainly true that survival rates have improved as treatments have changed based on experience. I think folks know better how to manage things at home, to some extent. Earlier testing also allows earlier intervention, which allows treatment that may avoid ventilator usage. Vent usage corresponds with pretty poor outcomes, although that too is improving.
It is certainly true that survival rates have improved as treatments have changed based on experience. I think folks know better how to manage things at home, to some extent. Earlier testing also allows earlier intervention, which allows treatment that may avoid ventilator usage. Vent usage corresponds with pretty poor outcomes, although that too is improving.