Diagnostic variability in suspected pulmonary embolism.

Over a 12-month period, we observed adult patients with suspected pulmonary embolism referred for lung scanning to determine variability in the diagnostic process. Among 269 studies, 157 lung scans were judged necessary by predetermined criteria. Ninety-three of these 157 patients had inconclusive results (low probability, intermediate probability, or indeterminate). Of these 93 patients, 42 had pulmonary angiograms, ten of which were positive. Of the 51 patients with necessary but inconclusive scans, five were poor candidates for angiography, 15 had other indications for anticoagulation, seven refused the study, and 24 had physicians who considered further studies unwarranted. Patients with and without pulmonary angiography were demographically and clinically similar. Although confirmatory testing such as pulmonary angiography was used frequently (45%) after an inconclusive lung scan, the question of pulmonary embolism was often left unanswered (55%). Methods for linking clinical judgment to lung scan results are necessary to select proper patients for invasive confirmatory testing.