Precision is defined as the agreement between replicate analyses of the same material. Within-laboratory precision of the prothrombin time (PT) and INR can be optimized by choosing an appropriate combination of thromboplastin and instrument. Accuracy (true value) of the International Sensitivity Index (ISI), required for the calculation of INR, depends on the choice of international reference preparation for the thromboplastin, and on the instrument used by the local laboratory. Interaction between thromboplastin and each patient's individual coagulation factors influences the accuracy of the INR. The contributions of imprecision, inaccuracy, and interaction effects can be combined to calculate the total uncertainty of the INR using Tomenson's equation. Confidence intervals for the INR were calculated in typical examples. The 95% confidence intervals of the INR around 3 are approximately as wide as currently recommended therapeutic target ranges.