Hampson correctly points out that Heijboer et al1 demonstrated a very low incidence of objectively proved venous thromboembolism (2.2%) in patients with suspected deep vein thrombosis who had normal findings on 3 serial impedance plethysmographs. The tests were performed on days 1, 2, and 8. It is therefore possible that fewer serial impedance plethysmographic evaluations than were used in our study2 may be adequate. However, in our study, of 16 cases of proximal vein thrombosis detected on serial testing, 7 were detected between days 8 and 14 and 4 were detected between days 10 and 14. In performing our cost analysis, the more conservative approach using 5 serial tests was chosen. The use of 3 as opposed to 5 serial tests would achieve further cost savings.
Our more recent report2 extended the previous study3 to 1564 consecutive patients with suspected pulmonary embolism. Of the 627 patients