External quality assessment of prothrombin time: The split‐sample model compared with external quality assessment with commercial control material

Abstract

Objective. CoaguChek S is a point‐of‐care, whole‐blood, prothrombin time monitor. The purpose of this study was to compare two different methods for external quality assessments of CoaguChek S. Material and methods. In the traditional external quality assessment scheme, commercial control material was sent to office laboratories and the results were compared with a method‐specific target value. In the alternative external quality assessment (the split‐sample survey) patient samples were analyzed on CoaguChek S at office laboratories, and venous blood samples from the same patients were analyzed at a hospital laboratory using an assigned comparison method. To obtain comparable performance criteria for the two methods, the limits for “good”, “acceptable” and “poor” performance evaluation in the split‐sample survey had to be expanded because of uncertainties in preanalytical factors and the comparison method. Results. In the traditional external quality assessment the total imprecision (between‐office and within‐office) was 8.0 % at the low level (1.6 INR (International Normalized Ratio)) and 10.5 % at the therapeutic level (3.4 INR). In the split‐sample survey the total imprecision was 12.3 % at the low level (2.1 INR) and 10.7 % at the high level (3.0 INR). Seventy‐five percent of the participating office laboratories were characterized as “good” with the traditional external quality assessments, whereas the corresponding number was 73 % using the split‐sample model. Conclusions. Available commercial control material for CoaguChek S is different from patient samples. This study demonstrates that split‐sample survey is achievable, and is an acceptable alternative to traditional external quality assessment for point‐of‐care prothrombin time monitors where appropriate control material is difficult to obtain.

• CoaguChek S
• external quality control
• international normalized ratio

Acknowledgements

We express our thanks to Heidi Eilertsen at Akershus University Hospital for analyzing the patient plasmas using the comparison method in the split‐sample survey, and to Per Hyltoft Petersen for invaluable comments on the manuscript.