The cost-effectiveness of various diagnostic strategies in patients with clinically suspected pulmonary embolism (PE) was analysed using a modelling approach. In 451 consecutive patients with clinically suspected PE data on the performance of diagnostic tests were collected prospectively in two large teaching hospitals in Amsterdam, The Netherlands. The ventilation-perfusion lung scan was used as the primary diagnostic test in all patients. In patients with a non-diagnostic lung scan result the performance of a clinical decision rule, a D-dimer test, and ultrasonography of the leg veins was evaluated with pulmonary angiography as the gold standard. It was estimated that the strategy recommended by a 1992 Dutch consensus meeting costs about Dfl 4400 per patient and that 97.02% of the patients can be expected to survive the first 6 months after the primary PE. The nation-wide annual costs for the diagnosis and treatment of patients by this strategy were estimated at 163 million Dutch Guilders. Subsequently, the costs and effects of alternative strategies were evaluated in a modelling approach, and compared with those of the consensus strategy. One strategy was selected that produces the best results in terms of survival and leads to considerable savings as compared with the consensus strategy. In this strategy subsequently a ventilation-perfusion scan, a clinical decision rule, a D-dimer test, a pulmonary angiography and leg ultrasonography are performed. Patients with a high probability ventilation-perfusion scan, an abnormal angiography or leg ultrasound test are treated, whereas treatment is withheld in patients with a normal ventilation-perfusion scan, a normal clinical decision rule, a negative D-dimer test, a normal angiography, or a normal leg ultrasound test. This strategy will have to prove its value and usefulness in clinical practice in a subsequent prospective validation phase.