While it looks like a reasonable attempt, it’s only looking at directly related effects. For instance, it notes that pedestrian fatalities have decreased since 1960 due to, among other things, discouraging people from walking. The effect of cars on population health would be much harder to quantify but might end up dwarfing many of the other factors.