The earlier statistic of deaths per total population was worthless and a complete non-starter for conversation. It would have no relevance. To take it to an extreme, put the billion plus Chinese in the denominator, most of who don't have cars and have never driven. It would make it look like the Chinese have a great road system and great drivers. Anyone who has been there knows this is not the case.
I believe the statistic comparing deaths per billion kilometers per year would be much more valid. However, I would concede Hakan's point that it could be distorted by professional drivers putting on thousands of miles per week. A statistic based on deaths per number of vehicles would also be closer than deaths per total population but can be distorted by numbers of vehicles owned per driver and by the drivers who hardly ever drive. For example, my brother- in-law in Europe has a car that sits in the garage except for about one use per month. He walks and uses city buses for almost all his needs. Most intercity transportation is by bus or train. Most inter-country transport is by train or plane.