My understanding of the 20 percent variance is- a test of 200, Could actually be 160 or 240. I have done two test on the same drop of blood and seen 10 percent differences. I have never seen 20 percent, personally.
Kim, I saw you ask if this was "right", and yes it was!
A couple of thoughts on the 20% variance thing...
I don't have data to prove this, just a hunch. But I think if the meter is going to be off by up to 20%, I prefer to think that it's going to be off
in the same direction when/if it's off. In other words, the variance can be used to rationalize lots of things. Like if you get a 200, a 150 and a 250. You could technically say "well that's pretty flat" when it looks anything but flat. That
could be a 200, a 180 and a 200 if you fudge the figures + and -. It could also be a 160/180/300, or a 240/120/300. You can add 20% or subtract 20% however you want, and it will give you three completely different shapes to your curves.
The first one - 200/180/200, looks flat, and looks like the dose should be increased.
The 2nd one - 160/180/300, looks like the insulin wore off early, and the BG just kept rising... make you think the dose needs to be increased.
The 3rd one - 240/120/300 looks like a pretty good curve, and you might just be leaving the dose alone.
But so do the original numbers of 200/150/250.
You can drive yourself insane if you tinker with the numbers, and they'll eventually "fit" what you thought to begin with.
So when I look at numbers and variance, I figure if one was low by 20%, it's more likely that they're all off by about the same amount, and in the same direction.
Also, from what I read on the board, it's generally believed that the lower the reading, the less "off" the meters tend to be. Like they might be off 20% in the higher BG ranges, but the low numbers you get (like under 100) are more likely to be accurate and the degree of variance is less. Again, no data to back that up, but it seems to be what lots of people think.
IMO, the only numbers that you
must take at face value are the low numbers. Like maybe 60's or below, and definitely under 50. You have to assume, for safety sake that those numbers are right on the money, because there's not much margin for you to be wrong and keep kitty safe. If you ever see a 40, assume it's real, and that you need to boost them up. You may never "see hypo symptoms" and many cats go down to 30 and you wouldn't believe it from the way they are behaving. But when you get to the point where the symptoms are obvious, like staggering, glazed over eyes, falling over..... at that point it could be too late, and kitty will need immediate vet emergency treatment.
Other than that, the numbers are just numbers, and aren't critically important. It's the patterns that matter, the shapes of the curves over time.
Carl