The accuracy of home glucose monitor, Accu-Check Advantage
Abstract
Objective: To determine the accuracy of the Accu-Check Advantage home glucose monitor, and related factors.
Material and Methods: Data were collected from all subjects who came to the Family Medicine Clinic, Hatyai Regional Hospital, in July 2004 for blood testing. All were asked for blood glucose testing with the blood from their forearms and finger tips. The measured glucose values from Accu-Check Advantage were then compared with the results from standard plasma glucose testing.
Results: There were 144 subjects in this study. Of these, 8.3% were healthy, 2.1% were acutely ill and 89.6% had chronic disease. 99.3% had been fasting at least 8 hours before blood testing. Their plasma glucose values ranged from 68 to 360 mg/ dl. Both venous and capillary whole blood glucose values correlated well with those of plasma (correlation coefficients of 0.975 and 0.981 respectively) The mean of plasma glucose value was 7.3% higher than that of venous whole blood and 5.2% higher than that of capillary whole blood. Using the monitor value at 116 mg/dl as a cut off point to diagnose diabetes mellitus, the venous whole blood tested values had 94.2% sensitivity and 91.3% specificity. The capillary whole blood tested values had 94.2% sensitivity and 90.2% specificity. To evaluate the accuracy of Accu-Check Advantage by using error grid analysis, we found that 99.3% of the venous whole blood tested values fell within zone A and B, and 0.7% within the left side of zone D, whereas 100% of the capillary whole blood tested values fell within zone A and B. From a multivariate logistic regression model, if the plasma glucose level was higher than 180 mg/dl, the error opportunity of monitor values could increase four times by using venous whole blood tested (OR 3.93, 95% CI 1.15-13.45) and three times by using the capillary whole blood tested. (OR 3.09, 95% CI 1.03-9.50)
Conclusion: By using Accu-Check Advantage, we found that both the venous and capillary whole blood tested values had acceptable accuracy. Most values fell within 20% of the control values and were clinically useful. However, the interpretation may be limited in the case of plasma glucose level higher than 180 mg/dl.
Keywords
Full Text:
PDFRefbacks
- There are currently no refbacks.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.