AIM: To evaluate the accuracy of the approximate linear formula of International Normalised Ratio (INR) imprecision by formal mathematical analysis. METHODS: Using probability theory, an exact formula for the coefficient of variation (CV) of the INR was derived. The CV from the approximate formula was compared with the CV from the exact formula for INR determinations between 1.0 and 10.0 with International Sensitivity Indices (ISIs) between 1.0 and 3.0 and prothrombin time ratio CVs between 1.0 and 10.0%. RESULTS: When the ISI equals 1.0, the approximate formula and the exact formula are equal. When the ISI is more than 1.0, the approximate formula overestimates the exact CV, but by less than one hundredth of the exact CV in the parameter ranges studied. The approximate formula is most accurate when laboratories achieve excellent prothrombin time measurement precision and use sensitive thromboplastins. CONCLUSIONS: The approximate formula provides a simple means for estimating the imprecision of the INR and is sufficiently accurate to warrant its use in clinical laboratories.
AIM: To evaluate the accuracy of the approximate linear formula of International Normalised Ratio (INR) imprecision by formal mathematical analysis. METHODS: Using probability theory, an exact formula for the coefficient of variation (CV) of the INR was derived. The CV from the approximate formula was compared with the CV from the exact formula for INR determinations between 1.0 and 10.0 with International Sensitivity Indices (ISIs) between 1.0 and 3.0 and prothrombin time ratio CVs between 1.0 and 10.0%. RESULTS: When the ISI equals 1.0, the approximate formula and the exact formula are equal. When the ISI is more than 1.0, the approximate formula overestimates the exact CV, but by less than one hundredth of the exact CV in the parameter ranges studied. The approximate formula is most accurate when laboratories achieve excellent prothrombin time measurement precision and use sensitive thromboplastins. CONCLUSIONS: The approximate formula provides a simple means for estimating the imprecision of the INR and is sufficiently accurate to warrant its use in clinical laboratories.