The article pins part of its argument on the claim that advances in diagnostics have not changed the death rate for (e.g.) breast cancer during the past quarter century, but a superficial search immediately suggests this to be grossly false.
"Death rates from female breast cancer dropped 39% from 1989 to 2015." - all other things being equal, it ought to have increased: age is a risk factor for cancers, and we have an older population. Instead, it has greatly decreased.
So I dispute the veracity of several of the article's claims, but also: what is considered unnecessary changes with the benefit of hindsight.
I would argue that it's expected that a majority of diagnostic tests deliver a negative result (i.e. turn out to be "unnecessary"), and this is actually a mark of a healthy healthcare system.