Practicing and perfecting the art of medicine demands recognition that uncertainty permeates all clinical decisions. When delivering clinical care, clinicians face a multiplicity of potential diagnoses, limitations in diagnostic capacity, and "sub-clinical" disease identified by tests rather than by clinical manifestations. In addition, clinicians must recognize the rapid changes in scientific knowledge needed to guide decisions. Cushing's syndrome is one of several disorders in which there may be considerable difficulty and delay in diagnosis. This article describes a current model of clinical reasoning, some of its challenges, and the application of the principles of clinical epidemiology to meet some of those challenges.