Administration of prophylactic low-dose subcutaneous heparin to prevent postoperative deep vein thrombosis is expensive, entails treating many patients unnecessarily, and causes some side effects. By using a predictive index a population of patients who are at particularly high risk of developing postoperative deep vein thrombosis may be identified preoperatively. Prophylaxis was given only to these patients, resulting in an incidence of deep vein thrombosis of 3.8% compared with 16.1% in previous studies in which no specific prophylaxis was given. By limiting prophylaxis to the group of patients identified by the predictive index as being at high risk of developing postoperative deep vein thrombosis results may be obtained that are as good as those expected from treating the whole population. Thus many patients are saved from exposure to low-dose subcutaneous heparin.