Fair point - so taking the proportion of road deaths related to drink as the measure instead (which should control for your points to a great recent), 31% of US deaths are drink related, compared to 15.2% in the UK. (according to https://www.gov.uk/government/uploads/system/uploads/attachm...).
Also you have to consider a bunch of other variables such as the difference in the traveling distances between work, home, and play.
And take into consideration things such an increase of X in volume of traffic might increase deaths by kX. With k being 2 or 5 or 10.
And then the gap narrows.