That's an excellent description of trading reduced expected return for reduced variance actually means. What you seem to be saying is that your utility functio

That's an excellent description of trading reduced expected return for reduced variance actually means. What you seem to be saying is that your utility function isn't actually linear in lives saved, but is instead slightly concave or perhaps has a step change where the utility goes from 0 for lives < n to some positive number for lives > n (for some relatively small n).