High yield criteria for obtaining non-trauma chest radiography in the adult emergency department population.

To develop a clinical decision rule for predicting significant chest radiography abnormalities in adult Emergency Department (ED) patients, a prospective, observational study was conducted of consecutive adults (>or=18 years old) who underwent chest radiography for nontraumatic complaints at an urban ED with an annual census of 85,000. The official radiologist interpretation of the film was used as the gold standard for defining radiographic abnormalities. Using predefined criteria and author consensus, patients were divided into two groups: those with clinically significant abnormalities (CSA) and those with either normal or nonclinically significant abnormalities. Chi square recursive partitioning was used to derive a decision rule. Odds ratios and kappa statistics were calculated for derived criteria. The results showed 284 (17%) of 1650 patients had clinically significant abnormal radiographs. The presence of any of 10 criteria (age >or= 60 years, temperature >or= 38 degrees C, oxygen saturation < 90%, respiratory rate > 24 breaths/min, hemoptysis, rales, diminished breath sounds, a history of alcohol abuse, tuberculosis, or thromboembolic disease) was 95% sensitive (95% CI: 92-98%) and 40% specific (95% CI: 37-43%) in detecting CSA radiographs. Positive and negative predictive values were 25% (95% CI: 23-27%) and 98% (95% CI: 96-99%), respectively. A highly sensitive decision rule for detecting clinically significant abnormalities on chest radiographs in nontraumatized adults has been developed. If prospectively validated, these criteria may permit clinicians to confidently reduce the number of radiographs in this population.