An interaction of public policy and technology has reduced (and continues to reduce) deaths per million vehicle miles traveled.

Traffic accidents are hard because risk homeostasis works to negate improvements, and a lot of research has already been put into improving automobile safety, so the outside view is that we should expect continued diminishing returns. On the other hand, autonomous/self-driving cars are a new innovation which could potentially make a big difference. So that seems like a good area to target.

Next you have a cluster of poisoning and suicide (which I guess is classified as unintentional?).

Good luck trying to reduce suicide! Poisoning... I dunno. Many of those might be suicide, and the obvious tactics of child-safety locks and warning labels have been implemented for a long time. Is there any low-hanging fruit there?

I've noticed that "risk of falling" is often tracked in studies of the elderly, although I don't know anything more about this.

Falling is a huge problem for elderly, and also one that seems to me like it could be easily tackled. As Joshua says, there's architectural improvements to housing that would reduce risk of falling which are currently uncommon, and there are other tactics: exercises and vibrating platforms may be able to improve the balance of elderly, and there's a biological aspect in weakened bones (vitamin D clinical trials sometimes show reductions in all-cause mortality, which seems to be largely due to better bones leading to fewer damaging falls, and this may be why the bisphonates also reduce all-cause mortality).