Interesting thread, I started reading for the KPI aspect.
For federal departments, hospitals and insurance company bean counters, they place a $ value on a human's life. In the manufacturing plant, the accident goal is always zero.
Don't be that guy..... That guy that makes his powerpoint presentation that shows five deaths last year, so the goal this year is to reduce deaths to two. He plans on only having two deaths this year. I laugh with sadness at the idea, who will be the two operators we pick to die this year?
The thoughts of targeting time between accidents and root causes to make it safer work place is the right path.
Welcome to the Cove.
As I've said before, and maybe in this thread,
the term "acceptable" screws everything up. Of course no number of workplace injuries is "acceptable." I have also never seen a Powerpoint presentation that suggests that instances of fatal injuries should be reduced from
x to
x-y, especially with the idea that some number of fatalities is "acceptable."
There is no way to formulate targets for injuries that makes sense in the same way that it makes sense for product defect rates. To think that it's sensible is to commit a major category error.
The "target" should be to understand the variables that contribute to accident risk and control them to the extent possible. If some yahoo overrides controls and someone is injured as a result, it doesn't necessarily mean that the controls aren't effective, and it also doesn't mean that such an injury is "acceptable." If something is
inevitable, "acceptable" is moot.
excellent way to say it! 
