The USA list is an interesting conflation of what I think are two distinct classes of error: ones which can be made impossible with the right processes followed correctly, and ones which can't.<p>"Wrong site surgery", "retained instrument post-operation", "infant discharged to wrong person", sure, absolutely agree processes can and should be such that this can <i>literally</i> never happen.<p>On the other hand, "Intraoperative or immediately postoperative death in an ASA Class I patient" seems like one which is ultimately in the lap of the gods: we certainly can and should make those odds really good (maybe much better than they currently are) by improving processes and ranking surgeons by their error statistics and so on. But sometimes people do just die, and the best surgeon could make a one-in-one-thousand slip of the hand (because human bodies just aren't built for such perfect fine motor control). And "serious injury or death associated with a fall": again, I'm not sure there are processes even in principle which could prevent a determined patient from injuring themselves by taking a fall, short of tying them to the bed or otherwise disabling them. There's a solid chance I, a basically-healthy patient in a ward, could be socially engineered into helping the miserable elderly patient next door to get out of bed for a short walk; so now your processes need to be robust to having young healthy people actively trying to break them! This problem seems not like the others.