Methods
We used a decision analysis model to determine the clinical and economic effects of the following interventions: (1) stopping prophylactic antithrombotic therapy after the in-hospital perioperative period (7-14 days), (2) extending prophylactic antithrombotic therapy with oral anticoagulants up to 6 weeks or 3 months, and (3) screening for DVT at the time of hospital discharge by compression ultrasonography or venography. We measured for a hypothetical cohort of 10 000 patients the risk of developing recurrent DVT and symptomatic pulmonary embolism, the risk of major bleeding complications, and the costs associated with each strategy. Data were drawn from the published literature (MEDLINE search) and from our hospital cost manager.

Results
Compared with stopping prophylaxis at the time of hospital discharge, a 6-week course of prophylactic oral anticoagulant therapy would reduce the number of cases of recurrent DVT from 1172 to 493 and the number of symptomatic pulmonary embolisms from 234 to 99, and would be less costly. This gain would be achieved at the cost of 29 major bleeding complications if the bleeding risk is low (0.2 per month) and 71 if the bleeding risk is moderate (0.5 per month). Compared with stopping prophylaxis at discharge, ultrasound screening would reduce symptomatic pulmonary embolisms from 234 to 142, induce only 13 major bleeding episodes, and be cost-effective, with marginal costs per additional pulmonary embolism averted ranging from $3000 to $7000, depending on ultrasound sensitivity. Venography screening would be the most effective strategy; however, compared with ultrasound, its marginal costs per additional pulmonary embolism averted would exceed $80 000.

Conclusions
After hip replacement with conventional perioperative antithrombotic prophylaxis, oral anticoagulation administered for 6 weeks is effective in preventing recurrent DVT and symptomatic pulmonary embolism, unless the bleeding risk is very high. More-over, this strategy is less costly compared with stopping antithrombotic prophylaxis at the time of hospital discharge. Alternatively, ultrasound screening is also effective, minimizes the risks of bleeding, and has a low marginal cost-effectiveness ratio.

Arch Intern Med. 1996;156:1661-1668