Evaluating the Accuracy of Home Glucose Testing
Ever wondered about the accuracy of home glucose monitors? Clients have. In a recent Omnibus survey of 1,002 adults with type 1 and type 2 diabetes, respondents listed "wondering if the test result is really accurate" as the third top barrier to testing. According to some experts, the method most diabetics rely on to the keep their condition under control may not be accurate enough.
The inaccuracy may have something to do with the absence of industry standards to regulate the accuracy of a home monitor. In fact, a proposed international ISO standard calls for an accuracy level within 20 percent. The American Diabetes Association (ADA) maintains that accuracy should not fall below 5 percent. Most home monitors fall somewhere between those two numbers.
According to a 1999 study by authors Raimund Weitgasser, Brigitta Gappmayer and Maximilian Pichler, modern advances have improved testing outcomes, but the numbers still fall short of the goals of the ADA.
"I would say in the past one has looked upon accuracy when it comes to monitoring at home as something that you really perhaps didn't need because you just wanted to know if you were high or low," says Joakim Hagvik, global marketing manager for HemoCue's glucose product line, Sweden. "But now since you're actually treating yourself, you are actually basing your insulin dose on those numbers … If you measure glucose more often and you pay more attention to the meters, then you will have a better outcome, you will stop the progression of the occasion of diabetes and perhaps even live longer."
Another study conducted in 2001 by James C. Boyd and David E. Bruns, department of pathology, University of Virginia Medical School, Charlottesville, Va., found that "glucose meters that meet current quality specifications (outlined by the FDA) allow a large fraction of administered insulin doses to differ from the intended doses." The effects of such dosage errors on blood glucose and patient outcomes have not yet been determined.
"(The authors of the Boyd-Bruns study) showed that if you have an error of let's say 5 percent, which is what ADA says … it will have no dramatic effect on the treatment," says Hagvik. "But if you go up to let's say 10 or 15 percent error, or even worse a 20 percent error, then actually you would have quite a number of cases where you will have a gross error in insulin dose. You will actually mistreat the patient or the patient will mistreat himself or herself."
Hagvik says that different meters will give different readings and strip lots will also cause variances. Advances in technology that have reduced the amount of sample and sped up the results may have also compromised accuracy, he claims. "Of course, when you do that … there's a tradeoff. As soon as you go down in volume, you also then will have to sacrifice on your precision. … Let's say if you had an orange and you had some juice left on your finger just before you did the test, if you have a really small sample, that sugar will affect your results even more. There is a tradeoff here. … And also this thing with speed because to get such fast results you need to extrapolate so you don't wait until the reaction is completely finished."
Fortunately, user errors that contribute to inaccuracy can be addressed. Meters can be calibrated to match the lab, which according to the FDA still produces "the best way to resolve any questionable result and the best sample from any sick patient." Physicians can also observe the way the meter is handled and then give clients advice on getting better results.
Problems and Solutions for Potential Errors
A diabetes educator is another source to help teach clients how to better manage their condition. A recent Harris Interactive survey, commissioned by the American Association of Diabetes Educators (AADE), showed that 69 percent of people with diabetes feel very knowledgeable or knowledgeable about managing their condition, but only 59 percent have worked with a diabetes educator.
Problems and Solutions for Potential Errors
The FDA offers the following some common problems and their effects on meter glucose readings.
Problem: Sensor strips not fully inserted into meter
Result: False low
Recommendation: Always be sure strip is fully inserted in meter
Problem: Patient sample site (for example the fingertip) is contaminated with sugar
Result: False high
Recommendation: Always clean test site before sampling
Problem: Not enough blood applied to strip
Result: False low
Recommendation: Repeat test with a new sample
Problem: Batteries low on power
Result: Error codes
Recommendation: Change batteries and repeat sample collection
Problem: Test strips/Control solutions stored at temperature extremes
Result: False high/low
Recommendation: Store kit according to directions
Problem: Patient is dehydrated
Result: False high
Recommendation: Stat venous sample on main lab analyzer
Problem: Patient in shock
Result: False low
Recommendation: Stat venous sample on main lab analyzer
Problem: Squeezing fingertip too hard because blood is not flowing
Result: False low
Recommendation: Repeat test with a new sample from a new stick
Problem: Sites other than fingertips
Result: High/low
Recommendation: Results from alternative sites may not match finger stick results
Problem: Test strip/Control solution vial cracked
Result: False high/low
Recommendation: Always inspect package for cracks, leaks, etc.
Problem: Anemia/decrease hematocrit
Result: False high
Recommendation: Venous sample on main lab analyzer
Problem: Polycythemia/increased hematocrit
Result: False low
Recommendation: Venous sample on main lab analyzer
This article originally appeared in the June 2006 issue of HME Business.