A comparison between six- and four-week intervals in surveillance of oral anticoagulant treatment