Presently, I would think fatality rates mostly measure testing penetration, not treatment effectiveness. A country that's testing everyone will show a much lower fatality rate than a country testing only the sickest people, because the latter countries have few to no confirmed minor cases.