Aren't we comparing drivers vs drivers here?

If we aren't then you're just measuring the fraction of the population that drives (which is largely a reflection of wealth and demographics) which isn't really a measurement of the safeness of the roads and driving.

We’re talking about people killed trying to move around the place they live. Butt-in-seat hours is a fine way to measure automobile vs automobile stats, but when we’re talking about deaths from transportation, counting only automobile hours is silly. There are various ways to get around a city, and the way you do that hardly matters if it suits the purposes of that trip.