LAUSR

striking and unsettling when researchers first found out that physicians’ intuitive assessments were largely off the mark. As early as 1982, David Eddy (credited for having introduced the expression “evidence-based medicine”), reported that physicians’ estimates of breast cancer turned out to be mistaken by up to one order of magnitude: close to 80%, the correct answer being about 7.5% (see below) [1]. The cognitive mechanism leading to so high an estimate is a variant of so-called representativeness heuristic. Many clinicians derive their inflated judgment of the positive predictive value—P(cancer|test+)—from the consideration that a positive mammogram is a very representative (or typical) feature of a woman with cancer, as indicated by the high sensitivity of the test, or by the sizable likelihood ratio of a positive result (80%/10% = 8).