This is truly the problem with evidence based medicine. You can say how a given treatment will impact a population (i.e. 1 out of 100 patients with stage 4 colon cancer will survive past 6 months) and say it's not worth it, but when you're talking to a single patient, you don't know what the outcome will be. Maybe the treatment will be in vain or maybe he or she will be the one patient that goes into complete remission.

What do you do then?