Drs. Kaul and Abbott [1] provide an excellent overview in their editorial. Unfortunately, although they advocate common-sense clinical approaches to identifying anginal pain, they overlook the importance of identifying causes of nonanginal chest pain, which are by far the most common cause of thoracic pain, even in the elderly. Moreover, the authors reject probability-based computer models (based on clinical findings) because they imply that diagnostic certainty is achievable in the real world. Further, they are confused by the difference between pre- and post-test probability when they state that "average experiences" are unreliable in predicting an event for a single person. Yet, it seems ingenuous to reject a probabilistic approach simply because it can be embedded in an expert system when sound clinical judgment is itself based on probabilities derived from experience [2].

Experienced clinicians have always known that the pursuit of diagnostic certainty is an idiot's chimera and the cause of untold medical mischief. They know that most clinical presentations, especially chest and abdominal pain, cannot even be named, much less given an ICD-9 code [3]. It is a platitude that failure to listen to or expertly examine the patient is the cause of most needless hospital admissions for chest pain due to the hyperventilation syndrome, costochodritis, and dyspepsia. In the rare case of undecidable myocardial infarction the single best test remains cardiac imaging with radionuclides, preferably using single-photon emission computed tomography with Tc-99m sestamibi. Yet, even if this technology can be provided in the emergency department, a small subset of patients will either be subjected to unnecessary hospitalization or sent home with an acute myocardial infarction. No road leads to diagnostic perfection, the legal system notwithstanding.