They appear to have proven that by walking more and driving less, you have a lower chance of dying prematurely in an accident, but for some reason which they have yet to uncover, that effect is not due to the extra fitness you get from walking.
You also do have a higher chance when walking of getting fatally hit by a car (because right turn on red is legal in most of the US). So it probably isn’t quite that simple. I think that was one of the original Freakonomics book studies?
The study is from Sweden. Most of the world doesn't run people over nearly as much.
'accidents' includes non-vehicle related accidental deaths, such as "Hold my beer and watch me do this."