People drinking in urban centers are more likely to walk home drunk (since they can). Urban centers are more dangerous for pedestrians.
People drinking in remote areas are more likely to drive home drunk (since there is no other choice). These drunk drivers will come in potential contact with fewer cars, but generate a lot more miles for the statistic.
While the quoted statistic is true at a literal level ("you are 8 times more likely to die per mile walked drunk than mile driven drunk"), the intuitive implication that you'd be safer using your car is not. And that, of course, is not even considering the safety of others.
People drinking in remote areas are more likely to drive home drunk (since there is no other choice). These drunk drivers will come in potential contact with fewer cars, but generate a lot more miles for the statistic.
While the quoted statistic is true at a literal level ("you are 8 times more likely to die per mile walked drunk than mile driven drunk"), the intuitive implication that you'd be safer using your car is not. And that, of course, is not even considering the safety of others.