Yeah the way a question is set up makes a big difference and often has unspoken assumptions. For example your preference might indicate an assumption that cars are the natural and good way to get around, that low rates of car ownership are a sign of poverty and something lacking. In this framing, more people dying in cars is just the cost of progress.

I'd be interested in "per million kilometers traveled", including cars, trains, busses and cycling. Even then I'm assuming that we should optimize for the least deaths per km when actually no one likes traveling and less travel is better. Suburban sprawl is not the only way to structure a city, after all.