AIM: To compare the International Sensitivity Index (ISI) of the Thrombotest reagent used with a steel ball coagulometer (KC) to the ISI of the same reagent used with the manual (tilt tube) technique. METHODS: The study was carried out by eight laboratories using their own KC instrument and method of testing. All laboratories used the same batch of Thrombotest to determine the clotting times of fresh blood samples from 20 local healthy volunteers and 60 patients on long term oral anticoagulant therapy. KC clotting times were plotted against manual clotting times on double logarithmic scales. Orthogonal regression lines were calculated to assess the ISI. RESULTS: In two laboratories the ISI of the KC method was lower than that of the manual method; these differences, however, were 2% or less. In the other laboratories no clinically important differences were observed between ISI values obtained. However, the clotting times determined with the KC methods were shorter than the manual values. CONCLUSIONS: The ISI of Thrombotest determined with the KC methods was very similar to the manual value. Therefore, use of the ISI value supplied by the manufacturer without adjustment is justified. The mean normal prothrombin time, however, must be determined locally.