Likelihood Ratios & ROC Curves

This lesson is part of Premium

LR+ = sens/(1−spec); LR− = (1−sens)/spec. LR+ >10 and LR− <0.1 = strong. Independent of prevalence; combine with pre-test odds (Fagan nomogram concept). ROC curve: sensitivity vs 1−specificity across cut-offs; AUC (0.5 useless → 1.0 perfect). Choosing a cut-off. Diagram: ROC curve. Exam tip: LRs let you revise probability without redoing the 2×2.

Premium unlocks all 40 lessons, every calculator and the timed mock.