I think I'm well versed in statistics, but I don't understand the problem. Why not test everybody, and then adjust the threshold of intervention to reduce overa

I think I'm well versed in statistics, but I don't understand the problem. Why not test everybody, and then adjust the threshold of intervention to reduce overall mortality? (Or maximize quality of life * years, or some other metric taking patient well-being into account.)