I agree with Desrtopa in that "I don't feel that lives-saved is necessarily a very good metric of utility." Death is binary (dead / not dead), but human pain and suffering is not. This should impact the analysis. Assuming the same cost to save the life, if forced to decide between saving someone from a fatal gunshot wound (perhaps in a war or encampment somewhere) versus saving someone from pancreatic cancer (according to Livestrong, one of the most painful terminal diseases), the outcome (life-saving) may be the same in either case, but there is more utility in saving the latter because overall pain would be reduced.

Thanks for this article; it's a fantastic read.