[Previous Months][Date Index][Thread Index][Join - Register][Login]
[Message Prev][Message Next][Thread Prev][Thread Next]

Re: [IP] One Touch Ultra and varying consecutive readings - comments?

Hi, Catherine.

>Example: I tested ... and my BG read 340. I didn't feel like I was 340 but 
>I gave a correction bolus. I tested BG again for kicks right away and the 
>meter read 280. I immediately checked again and the reading was 310. I had 
>a tough time eating enouh to fix that gaffe.

I don't see a problem with any of the numbers. A bg analyzer is accurate to 
+/- 20% of a lab analyzer on any one sample. Thus, if you had a good lab 
analysis that measured 300 mg/dl, acceptable home analyzer results could 
vary between 240 - 360 mg/dl. That does not mean that the results *will* 
typically vary by that much, just that they *could* vary by that much.

Most results will be clustered around a number central to the lab value, 
but not necessarily identical to it. For instance, my meter might show, 
after five successive analyses, an average of 320 mg/dl, with values 
ranging from 280 to 330 mg/dl.

In this case, you had three values: 340, 280, 310. So what do you know? You 
know that your bg is elevated. You also know that you'll need a correction 
bolus. If your carb-to-insulin ratio is 50 (1 unit lowers your bg by 50 
mg/dl), then the difference between the highest and lowest bg readings 
would translate to a difference of 1.2 units of bolus. I wouldn't have the 
slightest problem myself "eating enough to fix that". All it would take for 
me is about 15 g of carb.

So what to do? Well, one option is to target for a slightly higher bg when 
correcting from a high bg. That is, instead of, say, trying to get from 
around 300 to 120 in one fell swoop, you could aim for 180 instead. That 
way, you account for any possible inaccuracy in your analyzer by staying on 
the safe (high) side.

>She said the percentage of error between 280 and 340 is %17 (don't know 
>how she
>came up with that number).

340 - 280 = 60
60 / 340 = 17.65%

That is, 280 mg/dl is 17.65% lower than 340 mg/dl.

>2) when immediately retesting, use a diiferent finger, as the first drop 
>of blood is the truest. If you
>squeeze a second drop from the same puncture you risk contaminating the 
>blood drop with other tissues, therefore varying the result.

I must admit, I'd never heard that before. I occasionally retest using the 
same puncture. I have _not_ noticed large variations. YMMV.

>I started using this meter because it was supposed to be one of the best. 
>I really like the meter, but I've never had this wild variety in test 
>results before.
>Anybody else noticed this?

I recently posted a comparison of the One Touch Profile to the One Touch 
Ultra and found the two to be comparable. (You can find it in the mailing 
list archives here: 
) My Ultra unit tended to show lower values than my two One Touch units, 
differing by about 8%. However, I did not compare accuracy -- I do not know 
if the Ultra results are closer or further from the "truth" than the Profile's.

I find that the Ultra tracks my bg as well as the Profiles. I have not been 
surprised by low readings when I felt high or vice versa. I would NOT be 
shocked by results that you saw. It would motivate me to conduct a simple 
study -- I'd test twice in a row for a while and see if the analysis was 
precise (reproducible). If not, I'd study my *technique* first. What's the 
effect of washing my hands? Of leaving my hands moist? Of using shallow or 
deep lancet punctures? Of using more or less blood? Of changing the way I 
add the blood to the Ultra strips? Of using the same lancet site on two 
successive strips?

There are a lot of sources of variation that are under my control. I'd rule 
out those sources first before focusing on the meter.


regards, Andy
for HELP or to subscribe/unsubscribe, contact: HELP@insulin-pumpers.org
send a DONATION http://www.Insulin-Pumpers.org/donate.shtml