Analytical accuracy and precision of two novel Point-of-Care systems for INR determination

Performance evaluation of the microINR® point‐of‐care INR‐testing system

“…The measurement repeatability analysis (to evaluate precision), conducted as stipulated in section 8.4.4.2 of ISO standard 17593:2007, could also not verify the manufacturer's precision claim, yielding a CV of 7.50% (95% CI 6.02 to 8.97) without variation across therapeutic intervals, despite the fact that this experiment was conducted on fresh capillary blood. Although it is comparable to the CV of 6.6% found recently on capillary blood specimens for the microINR ® device, it is higher than the maximum CV of 4.0% (upper confidence limit = 4.4%) reported in a 2008 performance verification of the (now established) CoaguChek XS system, where imprecision was also calculated from duplicate capillary blood INR values according to section 8.4.4.2 of ISO standard 17593:2007. When the data were examined in detail, it was determined that 5 of the 51 duplicates (9.8%) showed differences between the repeated measurements greater than |0.4| (these differences ranged from 0.5‐0.9).…”

“…Considering that a bias of greater than 10% has been deemed sufficient to warrant corrective action (a criterion also used previously to assess accuracy), the microINR ® 's slight positive proportional bias of 4.4% found here, is acceptable. Van den Besselaar et al also estimated that, had the original calibration values for their study been correct, the bias would have been <3%, which would have been more in line with our observation.…”

“…An initial accuracy assessment using the WHO's International Standard rTF/09 as reagent, conducted by van den Besselaar et al, found an unacceptable mean bias of −13.7%, although the manufacturer had changed the calibration parameters of the device for that particular study. Considering that a bias of greater than 10% has been deemed sufficient to warrant corrective action (a criterion also used previously to assess accuracy), the microINR ® 's slight positive proportional bias of 4.4% found here, is acceptable. Van den Besselaar et al also estimated that, had the original calibration values for their study been correct, the bias would have been <3%, which would have been more in line with our observation.…”

“…The microINR ® POC INR testing instrument (iLine microsystems, Donostia‐San Sebastián, Spain) has been CE mark certified for use by both healthcare providers and by patients for self‐testing . Its methodology has been described previously . Briefly, it determines the INR through monitoring of sample flow along microcapillaries in a disposable plastic test chip, following activation of the coagulation cascade by a recombinant human thromboplastin reagent.…”

“…It is a small hand‐held device which can generate a result in less than 1 minute . Although it has been evaluated for accuracy in Italy and for accuracy and precision in the Netherlands, it has not, to our knowledge, been formally evaluated for test chip and instrument variability, and test chip and instrument failure rates.…”