OBJECTIVES: To quantify direct health care costs of oral vitamin K antagonist anticoagulant treatment in patients diagnosed with non valvular atrial fibrillation (NVAF). We examine whether a correlation exists between poor INR control and direct health care costs. METHODS: We designed an observational study. We revised the clinical histories of patients diagnosed with NVAF and treated with OATs at the Fundación Jiménez Díaz (FJD) between 01/10/2009 and 30/09/2010 (N=1,257). We collected INR value, number of visits due to INR control, type of anticoagulant (warfarin or acenocoumarol), hospital admissions due to complications and other current medication. The cost of all drugs was assumed to be the official price. Expenditure in INR control was calculated using the costs of all health care resources involved in the procedure. We used DRGs for the cost of each hospitalisation. The cost of a hospital visit was calculated using four scenarios, using actual invoices or using analytical accountancy methods. All costs are expressed in 2010 euros. RESULTS: The monthly average number of visits per patient was 1.17. Direct health care costs are in the range of 423,695€ and 1,436,038€. The average cost per patient varies between 392€ and 1,341€. The average cost of those patients with an INR within therapeutic range in 25% of the visits was 441.70€-1,592€. When INR was within therapeutic range in 25%-50% of the visits the average cost was 512.37€-1,703.91€. INR within therapeutic range in 50%-75% of the visits represented an average cost of 400.80€-1,375.74€. When INR was within therapeutic range in over 75% of the visits the average cost decreased to 305.23€-1,049.84€. CONCLUSIONS: The main part of direct health care costs of treating NVAF patients with OAT are due to hospital control in a specialised area and the high frequency of the visits. There is an inverse relationship between good INR control and direct health care costs.