Pulmonary embolism is being overdiagnosed and overtreated, especially in previously normal women using oral contraceptives, with undesirable consequences in heparin treatment. This is in part a consequences of high estimates of its occurrence based on postmortem data and in part of technologic developments in diagnostic methods. Blood gas data are not diagnostically very helpful. Perfusion scans should be used, largely to exclude the diagnosis. Ventilation scans are only occasionally helpful and are expensive. Pulmonary angiography is the most accurate diagnostic means currently available. Further studies are needed to elucidate the prevalence and natural life history of pulmonary embolism in the previously healthy person.