Primary care anticoagulant management using near patient testing

Analytical performance of a point-of-care device in monitoring patients on oral anticoagulation with vitamin K antagonists

BACKGROUND

[Please check the following sentence for clarity: "Point-of-care devices measuring international normalized ratio have clinical appeal, reports of 'off-label' in-hospital/primary care use report improved time to intervention/dose adjustment."]Point-of-care devices measuring international normalized ratio have clinical appeal, reports of 'off-label' in-hospital/primary care use report improved time to intervention/dose adjustment. We evaluated the accuracy and precision of a device for such multiple patient use compared to a reference laboratory.

METHODS

The point-of-care international normalized ratio result of patients on oral anticoagulation at the Vascular Surgery clinic was compared to the reference to check for statistical and clinical correlation. This was a prospective case-control study design with sample size calculated for sensitivity of 87.5%, precision 5% and desired confidence level 95%.

RESULTS

There were 168 patients tested; 55% were male, the mean age was 45.4. Sixty per cent were in the target international normalized ratio range. Tests were done for statistical and clinical correlation. The international normalized ratio range using the point-of-care device was 0.8-7.5 (reference lab 0.8-10), mean international normalized ratio was 2.22 ± 1.6 (point-of-care device) compared to 2.46 ± 1.3 (reference lab). The mean absolute difference was 0.79 ± 0.92 and the mean relative difference was 8.1% ± 1.03. Data was analysed using a Bland-Altman plot yielding a mean of 0.738 (standard deviation 0.92). Concordance between the tests was 75% with r2 = 0.52 on linear regression. Using an error grid plot, excellent clinical correlation was seen in 63.8%. In 5.4% major corrective action was needed but potentially missed if relying on the point-of-care device.

CONCLUSION

The accuracy and precision of this point-of-care device is moderate. It may have potential utility only where access to a reference lab is difficult.

Assessing agreement between point of care and pathology laboratory results for INR: experiences from the Point of Care Testing in General Practice Trial

AIMS

To assess the level of agreement between international normalised ratio (INR) results obtained from pathology laboratories and point of care testing (PoCT) devices used in a general practice setting.

METHODS

INR pathology results were collected from multiple pathology laboratories and CoaguChek S PoCT devices over a 6 month period. Agreement was assessed using both clinically relevant agreement and the Bland Altman method.

RESULTS

Analysis was based on 1664 dual measurements collected on 417 patients from 26 general practices across Australia. The percentage of dual measurements satisfying the expanded and narrow agreement criteria were 91% and 89%, respectively. The mean difference in results and the 95% limits of agreement depended on the average INR result: mean difference = -0.30 + 0.08 x average; 95% limits of agreement = -0.30 + 0.08 x average +/- 0.77.

CONCLUSIONS

The current study provides further evidence that PoCT is an acceptable alternative to pathology laboratory testing in a general practice setting. The Bland Altman method is a useful and flexible tool for assessing agreement. Limits of agreement should be reported in future method comparison studies to assist clinicians in patient management.