Objective - A health economic evaluation of two alternative treatment settings, inpatient care and outpatient care, for acute deep venous thrombosis. Design - A randomised multicentre trial in a defined population in regular clinical practice. Setting - Hospitals and related health care centres in the Jönköping county council in Sweden. Interventions - Patients were randomised to either an inpatient strategy (n = 66) or an outpatient strategy (n = 65) using low-molecular-weight heparin, dalteparin, administered subcutaneously once daily and adjusted for body weight. Subjects - Of 224 eligible patients, 131 entered the trial and 124 completed the economic part of the study. Main outcome measures - Direct medical and direct non-medical costs during a 3-month period. Results - Total direct costs were higher for those in the inpatient strategy group, i.e. Swedish Crowns (SEK) 16 400 per patient (Euro 1899) compared to SEK 12 100 per patient (Euro 1405) in the outpatient strategy group (p < 0.001). More patients in the outpatient group received assistance when they returned home. Few patients in either group reported sick leave. There was no difference in total number of days between the two groups. Conclusions - Total direct costs were significantly lower for the outpatient treatment strategy for deep venous thrombosis compared to the inpatient treatment strategy. No significant difference in health impact could be detected. Deep venous thrombosis can to a greater extent than previously be treated in primary care, safely, at a lower cost, and in accordance with patient preferences.