Therapeutic control achieved by an anticoagulant clinic may be assessed by calculation of the proportion of time spent by the patients within the prothrombin time target range aimed for by the clinic. Three calculation methods for such assessment have been described by other investigators. The three methods were compared with the use of the database of one anticoagulant clinic (Leiden Thrombosis Centre). Discrepant levels of control were obtained by these methods because of incorrect assumptions in two of them. The most accurate method is the “cross-section-of-the-files” approach, in which the last-recorded prothrombin time per patient is used. The authors believe it should be implemented in each anticoagulant clinic. With the use of this method, the level of therapeutic control achieved by the Leiden Thrombosis Centre for its patients on long-term treatment amounted to 80%. The same method was applied by 37 other thrombosis centers in the Netherlands using a single target range of 2.8–4.8 International Normalized Ratio for their patients. Thirty-one centers achieved a level of control of at least 70% in their long-term patients. Underanticoagulation was observed more frequently than overanticoagulation