Predictive values: PPV and NPV
Sensitivity · specificity · prevalence · likelihood ratios · Bayes
Why a 99%-accurate test can be wrong most of the time, and how to compute the answer. Strict definitions of sensitivity, specificity, prevalence and the predictive values; the 2×2 table; Bayes' theorem in four equivalent forms; where the inputs come from; typical values for real tests; and the classic traps, including the inverse fallacy and the case–control table.

