Nontraumatic chest pain is the second most frequent cause of emergency department (ED) visits among adults, resulting in more than 8 million visits annually in the United States. 1 Although the predictive value of many signs, symptoms, and diagnostic tests have been defined for life-threatening diagnoses, to our knowledge, the frequency of these diagnoses has not been determined in a nationally representative sample. The pretest probability of a given condition is critical to drive Bayesian analysis and help determine posttest probability when known predictive values of findings from the history, physical examination, and any laboratory, electrocardiographic, or radiologic testing are applied.
Hsia et al. (Mon,) studied this question.