How objective is the observed mortality following critical care?