>Generally when talking about how safe a given sort of nuclear power is the most important thing to consider is whether it has a positive void coefficient or not.

The void coefficient is only appropriate, when you are dealing with a criticality accident like in Chernobyl. Criticality accident (fission chain reaction getting out of hand) has never happened in commercially used Western reactors. In fact, in Western countries you can not get a license for a reactor design, which has positive temperature coefficient for power i.e. the fission power must decrease, if temperature increases.