The diagnostic usefulness of the medical history may depend on the type of problem confronted. It has been suggested that dyspnea is an example of a condition the causes of which cannot be easily distinguished based on identification in the history of stereotypical disease patterns presented in standard texts. To evaluate this assertion, faculty members independently interviewed 146 consecutively admitted patients with dyspnea, and following the history of the present illness, made a diagnosis. After discharge of the patients, another faculty member, using preselected criteria, independently reviewed each record to make a final diagnosis. History-based diagnoses predicted final diagnoses 74% of the time. Therefore, the history appeared to be useful in identifying the primary diagnosis for most dyspneic patients admitted to the hospital. However, it is not known whether this identification provides sufficient rationale for therapy or leads to more efficient use of laboratory tests.