A randomized trial comparing INR monitoring devices in patients with anticoagulation self-management: evaluation of a novel error-grid approach

Background In addition to the metrological quality of international normalized ratio (INR) monitoring devices used in patients’ self-management of long-term anticoagulation, the effectiveness of self-monitoring with such devices has to be evaluated under real-life conditions with a focus on clinical implications. An approach to evaluate the clinical significance of inaccuracies is the error-grid analysis as already established in self-monitoring of blood glucose. Two anticoagulation monitors were compared in a real-life setting and a novel error-grid instrument for oral anticoagulation has been evaluated. Methods In a randomized crossover study 16 patients performed self-management of anticoagulation using the INRatio ® and the CoaguChek S ® system. Main outcome measures were clinically relevant INR differences according to established criteria and to the error-grid approach. Results A lower rate of clinically relevant disagreements according to Anderson’s criteria was found with CoaguChek S ® than with INRatio ® without statistical significance (10.77% vs. 12.90%; P = 0.787). Using the error-grid we found principally consistent results: More measurement pairs with discrepancies of no or low clinical relevance were found with CoaguChek S ® , whereas with INRatio ® we found more differences with a moderate clinical relevance. A high rate of patients’ satisfaction with both of the point of care devices was found with only marginal differences. Conclusions A principal appropriateness of the investigated point-of-care devices to adequately monitor the INR is shown. The error-grid is useful for comparing monitoring methods with a focus on clinical relevance under real-life conditions beyond assessing the pure metrological quality, but we emphasize that additional trials using this instrument with larger patient populations are needed to detect differences in clinically relevant disagreements.