Study objectives: To compare physician judgment in the use of chest radiographs for diagnosing pneumonia with decision rules developed by Diehr, Singal, Heckerling, and Gennis.
Design: Propsective observational investigation with preradiograph survey of physicians' intent to order chest radiographs for patients presenting with respiratory complaints. All patients had uniform clinical data collected, including chest radiographs and sufficient information to retrospectively apply the four clinical prediction rules.
Setting: The emergency department and medical outpatient clinic of a major urban teaching hospital.
Participants: Adult patients presenting with recent history of acute cough or exacerbation of chronic cough plus either fever, sputum production, or hemoptysis.
Results: Of 290 patients, 21 (7%) had pneumonia. The sensitivity of physician judgment (0.86) exceeded that of all four decision rules. The specificity of the Diehr (0.67), Heckerling (0.67), and Gennis (0.76) rules exceeded that of physician judgment (0.58). The accuracy of the Gennis (0.76) and Heckerling (0.68) rules also exceeded that of the physicians (0.60).
Discussion: Physicians' diagnostic and therapeutic decisions were characterized by high sensitivity but lower specificity for ordering chest radiographs to diagnose pneumonia. The higher specificity and accuracy of two of the decision rules suggest that they may have a role in patient evaluation.