Seven hundred and thirty-four patients were included in a prospective study for incidence of clinical venous thromboembolism under prophylaxis with either heparin, aspirin or warfarin and for the expenditure of hospital resources. Thromboembolic complications were more frequent (P less than 0.02) and hospital costs clearly higher in the low-dose heparin treated patient group compared with the aspirin and warfarin groups. There were no distinct differences between aspirin and warfarin treated patients neither in results nor in costs. However, carefully monitored treatment with warfarin with Thrombotest always less than 0.20, appeared to be the most effective prophylaxis in patients with hip fractures. In conclusion we now use aspirin as general prophylaxis in orthopaedic patients, and warfarin in patients with established risk of thromboembolic complications.