Patient-specific errors in agreement between International Normalized Ratios measured by a whole blood coagulometer and by a routine plasma-based method

Abstract

We applied a new statistical method to improve comparisons between systems measuring prothrombin time (PT) by splitting disagreement into systematic errors, which can be eliminated, and random errors, which can not. We found that the disagreement between International Normalized Ratio (INR) measurements based on plasma and whole blood was significantly patient-dependent.