precision vs recall sensitivity vs specificity type I vs type II positive likelihood ratio = (sensitivity/(1-specificity)) Anchor examples from hemoptysis lecture by Jason Hughes (pulmonary embolism vs infectious causes (50-60% non-TB + 8% TB). EM Lecture, 17:30 "Differential dx for fever is huge." Without localizing signs, acute fever will probably be influenza (industrialized nations), malaria (non-industrialized), TB, Lyme, autoimmune complications, etc