Diagnostic decision-process in suspected pulmonary embolism. Report of the Herlev hospital study group.

An analysis was made of how a diagnosis was arrived at in 60 consecutive patients with suspected pulmonary embolism. Patients underwent an initial clinical assessment, a chest X-ray examination, pulmonary scintiphotography, and final clinical assessment. Later, chest films and scintiphotos from these patients and from 120 controls were mixed and reread "blind". The results suggest that clinical physiologists were biased by the clinical information when interpreting the scintiphotos and that clinicians in turn believed implicitly in the scintigraphic report when they made the final diagnosis. The degree of agreement between the first and the second readings of the scintiphotos was unsatisfactory. The radiological findings did not correlate with the clinical and scintigraphic assessments. The study reveals some of the problems which arise in routine clinical practice, when the true diagnosis cannot be established by independent means.