There are a lot of ways to slice the data, and each of them tells a different part of the story. Suppose two countries have the exact same number of fatalities per km, and the same population, but one country drives twice as much. It is still a relevant fact that the country that drives twice as much has twice as many deaths. Policies that discourage driving will help reduce death rates in that country.

Assuming death rates from driving are the only death rates affected. There are unintended consequences worth studying.

It certainly can get very complex. Diesel emissions vs electric cars. Emissions in a city where everyone is inhaling the pollution vs in the countryside where it disperses quickly...

Infrastructure that forces you to drive more is more dangerous. Do we measure "deaths per bullet purchased"?

Shorter drives means fewer deaths, generally speaking.

Yeah, that ought to skew the results a bit. Looks like the top countries remain the same though https://read.oecd-ilibrary.org/transport/road-safety-annual-...