What other statistic would you use? Deaths per year? Injuries per year? Injuries per TWh? Without further information we can't have a conversation about what wouldn't be misleading (or whether this is misleading in the first place - I don't see why).
Total deaths over 1000 years after a catastrophic release of nuclear material, for example 10 fully-fueled power plants being hit by missiles during a conflict.
Also, probability of nuclear waste being unearthed in the next 50.000 years multiplied by the estimated death toll of such event.
I mean come on, deaths by pollution and accidents is all you look at to declare it as safe? What about the almost existential risk and the unsolved final repository problem? (the US at least still doesn't have one). Maybe the risk is low and until now nothing serious has happened in the US, but the potential for damage is so devastating that you should factor it in somehow. It's just not as simple as one number is probably what I'm trying to say ...
