# Posts tagged 'glucose meter test'

RSS Feed**You can check the precision of your blood sugar meter and even possibly the accuracy. **

Here's the method Dr. Richard Bernstein uses to enable you to check the precision (reproducible results) of your blood sugar meter and even possibly the accuracy.

As you probably know, most of the blood sugar meters on the market, according to studies that have already been published, are grossly inaccurate. I see the best accuracy, especially around normal values – so I consider normal values around 80, except with children, in the 70s, and how do you do that? Well, one could say, “Let's look at the control solutions or test solutions that the manufacturers provide with their meters.” The trouble is that, nowadays, most of the manufacturers are not providing normal controls – it has to do with some FDA guidelines that don't make sense and they're doing the minimum that they have to do to comply with FDA regulations – so they usually have control solutions that would read somewhere around 30 or 40 milligrams per deciliter, on the one end, and another solution that would read around 300 milligrams per deciliter. Now how do you now, measure for accuracy? I'm not interested in precise accuracy at a blood sugar of 30 or at a blood sugar of 300; I want it accurate where most of my patients' blood sugars are, which would be, let's say, between 80 and 100.

So how do I get this, and how do I go about testing? Well, what we do is we get hold of the manufacturer's control solutions to a meter, and we use 3 parts of the low control, which might be 30 or 40, as the middle of the range, and 1 part of the high control, and we put them in a little plastic squeeze bottle or in a glass dropper bottle that has a tight seal 'cause you don't want evaporation – evaporation will make the concentration of glucose go up – and this mixture will have a blood sugar of somewhere around 100 – it might be 108; it might be 113; it might be 97 – and you could calculate what that value would be by taking the middle value of the low control – so let's say the low control is 30 to 70: you can get the middle value by adding 30 and 70 and dividing by 2, and you get 50. And if the high control is 200 to 400, the middle value would be 300, so we would take 3 parts of the low control, so 3 times 50 would be 150, and we'd add in 1 high control, which would be 300, so that adds up to 450, and we've taken a total of 4 samples, 3 and 1, so we divide 450 by 4 and you get something like 112. So this solution should read at 112 milligrams per deciliter. So if the solution is accurate, if the manufacturer's provided you with accurate control solutions, then it's a good test of the accuracy of the meter – it should read that value that I just told you, of 112.

Now, just as important or more important, is the precision or reproducible results. What if you did 10 tests in a row, would they all read 104? And what we find, with most of these meters, is that they're scattered over a wide range. In fact, we found that one of the most popular-selling meters, a meter that many of the insurance companies require for diabetics, instead of the most accurate one, reads chronically low and may read very low or very haphazardly. So that particular meter, you might think you're going to get a reading of 104, and you get a 160 or you get 50 – that's how wide a range of readings you get with some meters – and, therefore, you do 5 or 10 tests in a row, you get a load of different numbers. On other meters, like the one that we recommend and we've mentioned it before, we get the same value over and over and over again, plus or minus one or two milligram per deciliter. So that shows consistency. And then if our reading is almost on the nose with the value of our control solution, we can be fairly confident of the accuracy, but we're assuming that the manufacturer made accurate control solutions. So the only way to be absolutely sure of the accuracy is to do a blood sugar at the clinical lab, while they're drawing blood from your arm, and see how it compares.