Driving isn't hazardous to your health. Crashing is. That's like saying eating is unhealthy to your health because you might choke.
Jumping off a building is not hazardous to your health, hitting the ground is.
A vehicular accident as a driver is not a direct cause of driving any more than a pedestrian accident by a vehicle is the direct cause of walking.
I object to the conflation of quantitative risk of experiencing an incidental event with the qualitative risk posed by a voluntary event.
How do you draw the line?
I think this is what people don't get about the study. The study is not saying no one should drink. It's not saying that the risks of drinking outweigh the benefit. It's saying that for any amount of alcohol, ingesting it, on aggregate, decreases life expectancy. Nothing more. It's a measurement of the increase of risk, which can then be combined and contrasted with other measurements and help make informed decisions.
You draw the line however you want. You now have more information to make that decision.
Not that there isn't acceptable tradeoffs with what is attained by driving.