Diagnosis of pulmonary embolism: a cost-effectiveness analysis.

OBJECTIVE Pulmonary embolism is a major cause of maternal death. The work up for suspected pulmonary embolism is complex, with many potential diagnostic options. We performed a cost analysis to evaluate which of several diagnostic strategies was the most cost-effective with the least number of deaths from pulmonary embolism. STUDY DESIGN We created a decision tree to evaluate the following strategies: (1) Compression ultrasonography followed by anticoagulation (if there is a positive result) or secondary tests, ventilation perfusion scans or spiral computed tomography (if there is a negative result); high probability ventilation perfusion scans (a positive test result) resulted in anticoagulation; low probability ventilation perfusion scans (a negative test) resulted in no treatment; intermediate tests that resulted in a second test (computed tomography or pulmonary angiography). (2) Ventilation perfusion scans as a primary test followed by anticoagulation. (3) Computed tomography followed by anticoagulation (if there is a positive result). The following assumptions were made: The incidence of pulmonary embolism in pregnant women with suspected pulmonary embolism is 5%; 40% of documented pulmonary embolisms have a positive compression ultrasound result; 10% of ventilation perfusion scans for suspected pulmonary embolism are high probability, 60% are indeterminate, and 30% are low probability for pulmonary embolism; the sensitivity of computed tomography is 95%; the sensitivity of angiography is 98%. The assumed mortality rate of treated pulmonary embolism is 0.7% and of untreated pulmonary embolism in pregnancy is 15% (range, 10%-50%). The angiography-associated mortality rate is 0.5%, and the anticoagulation associated mortality rate is 0.2%. The following costs were used for the model: compression ultrasonography, 200.00 dollars; ventilation perfusion scans, 400.00 dollars; angiography, 1000.00 dollars; computed tomography, 500.00 dollars; and anticoagulation, 5982.00 dollars. With baseline assumptions, spiral computed tomography as the initial diagnostic regimen was found to be the most cost-effective at 17,208 dollars per life saved. Sensitivity analyses were performed over a wide range of assumptions that included alteration of the probability of pulmonary embolism, the sensitivity of computed tomography, ventilation perfusion scans, and compression ultrasonography, the cost of computed tomography, and the mortality rate of untreated pulmonary embolism. Our findings remained robust over a wide range of assumptions. CONCLUSION Suspected pulmonary embolism remains a diagnostic quandary. Our analysis indicated that spiral computed tomography offers the most cost-effective method for diagnosing this potentially fatal condition.