Abstract

To examine the effect that a concise, objective, and potentially computer-extractable history would have on diagnostic accuracy in the interpretation of chest radiographs, we designed and tested a computerized patient-history form that could be integrated realistically into the clinical environment. We performed a series of studies in which 247 posteroanterior normal (79) and abnormal (168) chest radiographs were interpreted by four board-certified radiologists, both with and without accompanying clinical histories. The radiologists recorded their confidence rating of the presence or absence of one or more of the following abnormalities: interstitial disease, nodule, and pneumothorax. Analysis of receiver operating characteristics showed that, with the exception of interpretation of one abnormality by one radiologist, there were no statistically significant differences (p less than .05) between cases interpreted with and without the history form for any of the radiologists. The results of our study suggest that knowledge of clinical history does not affect the accuracy of chest radiograph interpretations for the detection of interstitial disease, nodules, and pneumothoraces. These results may not be applicable to other clinical situations.