« Diabetes Feeties. | Main | Guest Post: For All the Ladies. »

The Glucose Meter Shuffle.

I have several One Touch meters, a Freestyle one, and a Dexcom continuous glucose monitor that I consult on a regular basis.  (Not usually at the same time, but I have been doing multiple checks recently.  More on that below.)  I also have an Agamatrix meter and an Accu-chek one, somewhere in the diabetes cupboard in the bathroom, only without any strips that aren't expired. 

And I have a lot of anxiety when it comes to glucose meters.  The variability of these machines makes me crazy in the head, and it caused me a lot of grief when I was pregnant, because my blood sugar goal range at that point was so tight and so specific, and any variability was huge for me.  (I shared some samples of wonky results in this post.) 

In the last few weeks, I've been doing some experimenting with my meters, inspired by at-home research conducted by Stacey at The Girl with the Portable Pancreas.  It's an messy process, which includes testing with several meters at one and trying to grab a photo.  It's also messy because it involves the people at One Touch (because they provided me with some spare strips after I spoke with them about my meter issues and I have a running dialog with them about this problem), my insurance company (because I had to work with them to get a partial prescription covered for a batch of Freestyle strips), and my patience (because I'm short on that due to the aforementioned). 

So far, I've noticed a few trends:  The Freestyle meter almost always runs higher than my One Touch meters.  I'm not sure if that makes it more "accurate" or not, but it is consistently higher.  Thing is, there isn't a pattern I've found yet when it comes to syncing up with the Dexcom.  For example, the photo below has the Freestyle much higher than the One Touch meters, but my Dexcom was at 98 mg/dl with a down arrow, and I felt low.  So what am I supposed to trust in that instance?  If I go with the One Touch results, I'm potentially treating that number.  If I follow Freestyle, I'd happily get in the car and drive at that number. 

So close, yet so freaking far.

The One Touch meters are usually close to one another (like in this photo):

But are they close to my actual blood sugar?

But are they close to my actual blood sugar? When I was pregnant, I saw One Touch meter results that seemed lower than I felt, and when I consulted with the Dexcom, I was prompted to recheck.  Then I'd see a higher number on the meter.  Without the Dexcom, I wouldn't have ever second-guessed my meter.  Part of why I wanted to conduct this meter comparison now was to see if I could duplicate those variances and then document them, but I haven't seen the same problems since running these tests.  (I know it sounds weird to be frustrated by a lack of problems, but it's like when your car makes a noise when you're driving around town, but refuses to make the same noise when the mechanic is standing there.  How can they fix what I can't show?)

I know there is an "acceptable" 20% margin of error, but how would you even know to double-check your meter?  This issue matters to me, and it matters a lot.  These glucose meters are the only tools I have to monitor my blood sugars, and I make treatment decisions based on their results.  I need them to be consistent, and accurate.  If I'm treating highs that aren't high, I could end up with a serious hypoglycemic event.  And if I'm treating lows that aren't lows, I'll end up running higher and that, in the long run, will hurt my body.

What is my "true blood sugar?"  Is it the result from the machine at the endocrinologist's office (which I believe is an old school One Touch)?  Is it the lab result from a venous sample?   Is it from any meter I can use at home?  How can I trust any glucose meter fully, knowing these variabilities exist and are FDA accepted?  What can I (we?) do to tighten these results?

How can I feel safe?

[Disclosure:  I have a relationship with Dexcom. And Animas.  But my relationship with diabetes takes precedence over any business relationship.  So while my bias is exposed, I need to also say that my bias is highest when it comes to making sure I'm healthy.]


I HATE the variations!! I've posted mine several times, but the most recent one and the worst one was about two weeks ago.

It was about ten o'clock at night, and I'd just helping my father downstairs (already, we should be thinking my blood sugar was a bit higher than usual, because helping him with his e-mail stresses me out). I came upstairs, tested, and was 56.

Now, I have hypoglycemia unawareness, but usually, by the time I hit the fifties, I feel a bit off. Maybe not shaky or hungry or anything like that, but I feel as if I can't understand what's happening around me. So I tested again.

Would you believe my meter said I was 89? And then 92???

I called OneTouch, they sent me a vial of strips, offered to replace my meter (and lose all my lovely data on my UltraSmart? I don't think so!), and that's the "end".

But you know what? It's not the end. I'm sick of triple checking numbers because I don't trust my meter. I'm sick of losing precious strips because I'm scared. I hate that I doubt every low, and feel the need to double-check, just because it's happened before when my meter has been dead-wrong.

And what also gets me, like what you said, is that you never know when you need to double-check. I could be walking around, happily in the nineties, or so I think, and then feel terrible ten minutes later, when I'm really in the fifties and crashing. I almost went to bed one night with a 65, because my meter only read low once in the three times I tested. Well, guess what? After eating a couple of cookies, the buzzing feeling in my teeth went away.

We NEED better accuracy. How about we start a petition to the FDA to tighten up their guidelines? Every person who tests their blood sugar more than once per day would gladly sign up on that!

Wow what a great post I never thought of this before. Thanks.

As someone that uses multiple meters this post really has me thinking about my meters. I have my Dex, a Ping which is my main meter, a One Touch Mini I take with me when we go out and I want something a bit smaller than the Ping, and another Mini in my desk at work...just in case I forget my Ping at home.

I read a piece on control solutions being a waste. That person suggested testing at the lab when you get your blood work done, saving the result, and then comparing it to the numbers on your lab workup when they come back to you. Be interested to see how each of these meters would hold up to a lab. Just saying...

I've thought about doing my own experiments, too. I've notice my accuchek, which i do them in a row, seem to come up to pretty close to the same thing. With my freestyle, I tend to get more varying numbers in a row. I think the accuchek may be more acc-ur-ate. But, I use the freestyle because it does with our omni...

(Long time reader, first time commenting!)

I definitely feel your frustration, but at the same time I'm glad to hear I'm not the only one getting wonky results!

I use the Bayer Contour Link (it comes with the Med-T pumps in Canada), but only because it "talks" to my pump. If my sugars are on the lower side, it seems fairly accurate compared to how I feel. But if I creep above 10 (180 American), it seems to return a completely random number. Personally I think it's a P.O.S., I hate it, and worse I don't trust it. That's the worst part.

One day I tested at 24.1 (433) and knew it was wrong. There was no way I was that high. My second try came back at 16.8 (302) and finally I got 10.2 (183). This was all within the space of 60 seconds. That is a HUGE difference! The correction bolus for 24.1 could kill me if I was actually 10.2.

I took it in and told my nurse educator about it, so she gave me a new one and sent it back as a complaint to Bayer. But the new one acts the same way. It's ridiculous.

i have decided that at my next BG lab test, i am going to test myself, using my 3 different meters (all one touch) and see what the results are. i too have tried this with varying results, but not as far out as yours!

First, thank you for the mention :) Second, every time I think of this I get crazy. Since switching to the Freestyle Lite, I haven't had any inconsistent back to back results. And although that makes me feel better, it doesn't make me 100% comfortable with its readings either. I just don't get how this is "allowed". Seriously, for our lives depending on our BG at any given moment, how could there be any doubt as to what that BG is? At any given moment?

TOTALLY hear you! I had to switch from Freestyle to OneTouch, and got different results every time (Freestyle seemed to run high for me too). I also will get different results on any of them, depending on the mood of the meter. The meter people just send me control solution and/or new monitors (interestingly enough, the control solution was higher on the Freestyle too, what gives?). My doctor sympathizes, but can't do much. It's terrifying to not trust your meter.

The stated 20% margin of error is way more than that sometimes. My fingers suffer! Something must be done to ensure better accuracy! Over and under dosing can't be good.

I did a test of three meters vs the lab when I was first pregnant. I carried out three tests on each meter - 3 separate drops of blood each one used once on each meter. Took about two minutes in total and was done within 2 minutes of the lab draw.

My lab result was 7.1 (128). The One Touch meter I tested came out significantly lower - average was 5.7 (103). The Freestyle meter had the closest average to the lab number - 7.2 (130) but the most variability in the three results making that up. The Contour USB (my regular meter) came out at an average of 7.4, but the three numbers were very close together.

I've always suspected One Touch meters to read a little on the low side. The couple of times when I've been most surprised my A1c wasn't lower was when I'd been using a One Touch as my main meter. And a number of side by side tests against other meters have always shown it a good bit lower. It may be a safety measure - it's certainly safer in the short term to treat a low that's actually higher than a high that's actually lower. Who knows?

I've come to the conclusion that I can't worry too much about it, because until meter companies act on it - if they can - there isn't much I can do. If I'm concerned about an individual result, I usually re-test. For the last couple of years I've had a policy of always rechecking any number higher than 14 (250) to avoid the risk of unnecessarily large correction boluses. I'm afraid of missing lows, but then I think I'm more likely to miss a low by simply not testing at that time than by testing and getting an erroneous result. It's one of the many reasons I love CGM!

This is a great post, Kerri. Thanks for sharing it. I am very interested in hearing more about the feedback you get from the meter makers. This is so maddening! There's also the added confusion of wacky info from the CGM to factor in!

Are you using the same poked finger? If so I have heard that really matters. I can't imagine poking 4 times to get a sample for each meter, but you should try that. I know what you mean though when you test and you feel really low, but it says something like 65. Today that happened, I took some glucose tabs, and then went on with my day, 1 hour later I was 57, so I ate the rest of the glucose tabs, then suspended my pump. Yikes, so frustrating. Hey can you write a post about continuous glucose monitors for people with pumps that don't have them? I want to get one too but I don't understand them or how they work.

To echo what Emily said, different fingers will yield different results so that will add another variable to your testing unless you are testing each meter with the same finger poke. Plus, I've heard the longer blood stays on one's finger the higher the blood sugar result will be (up to a point, I assume) Since there are so many variables involved in this whole process of checking our sugar, I completely ignore it and stick with one meter until it does something completely crazy-then I replace it.

I totally agree with the frustration with this issue!! Thanks for making a strong point about it! My daughter was diagnosed at only 9 months, and the smallest measurable dose of insulin (0.5 units before we went to diluted) would drop her up to 300 points. We had to wait until she was >350 to treat her. The 3 different meters would give us numbers that were sometimes more than 80 points different. Terrifying!!! And no sympathy from nurses/docs- simply, " that is how the meters are", and get over it. Great post!

This is so bizarre because I JUST got off the phone with the Lifescan folks. I am asking that the One touch meter that came with my pump be replaced because it is constantly way lower (as much as 40% lower) than the other meters in my house (and I have like 6 or 7 of them floating around). Anyway, the lady on the phone told me that meters are allowed to be 20% off from lab results and 30% off from the readings of other meters!!! Seriously!?!?! How the fructose am I supposed to make treatment decisions based on such variable information.

About 2 weeks ago, I treated what I thought was a low based on the reading from my pump meter. Turns out that I probably wasn't low at all; my BG ended up in the low stratosphere as a result. I knew this meter had a problem, but when I compared results the next day, I was shocked to find that all my other meters were mostly in line with one another, while the meter that came with my pump was about 30 points lower.

I know the APP is a priority, but I'd be a much happier camper with some better meter accuracy. D has too much guesswork in it already; I am beginning to wonder how many of my highs and lows are simply due to treating lows and highs based on false data. Grrrrrrr....

These variables drive me nuts!
One of the reasons we stopped having Bean test with her 'extra' FreeStyle meter at school and exclusively us the one built into her PDM.
Even sometimes the difference from strip to strip...like the other day we were putting in the two BGs for the new dexcom sensor and one strip gave us 297 and the other 264. UGH!!

I am doing a little test right now too. Nova vs. One Touch compared to the Dex. A few people mentioned it in the comments above, so I will say I use the same blood drop for the tests (requires a bit of an extra squeeze!)

I would like to do a bigger test but not enough to pay for strips out of pocket. And I'm not going to keep getting a new prescription each month to try everything out. Insurance sure doesn't make it easy to find out what works best for us!

This is how I tested meters for my 8 yr old--I didn't just test 4 meters against each other, but also against themselves in repeat tests to see if they gave the same most consistent results. I did the lifescan that came with her MM pump, the accucheck aviva, a freestyle lite, and this other one that was supposed to have this new technology--don't remember the name. I tested each one twice at the same time to see if the same meter gave me the same results. it's one thing for a different meter to be 30 points higher, but if the same meter is 30 points higher 1 second later then you may as well toss it. I made this big chart with averages and all sorts of data--D folk are such data geeks! I did not use my dd as the only guinea pig but did my own bg too. The aviva and the life scan were the most consistent. Love the aviva but love love love have bg go automatically to the revel so I stuck (no pun intended) with that.

I have 2 meters, but only use one at a time, because if I did as you all do with multiple meters, I would drive myself crazy.

That said - I recently participated in a study where my endo was testing a new CGM-like device over a 4 hour period. They tested via blood (to the lab), and I compared my OneTouch UltraSmart meter to my Dexcom to the lab (centrifuge) value. The results were kind of awesome. You know who came closest to the lab? DEXCOM, baby! The meter was consistently higher than the lab, and Dexcom was right in between, sometimes as close as 5-10 off. I think that's pretty impressive.

Now, that all being said, I will usually feel low before Dexcom tells me so, and when my BG gets below about 85, Dexcom is not as accurate (to my meter or to how I feel).

Take with a grain of salt, please. :)

Slightly off topic and don't be aghast, but we test our 2 type1 kids almost exclusively on the forearm. Our endo and diabetic educator can't stand it! We have compared fingertip and alternate site testing a bazillion times and it's always really close. Our a1c's always reassure us that our records and the lab work are consistent. I can't understand why the docs are so concerned with our lack of accurate readings when we/they don't even know how
accurate the meters are to begin with. Sheesh!

Hmm, funny how this coincide with an experience I had last night. I'd walked the dog for about 1 h after dinner, and tested when we got back. I used my main meter (Precision Xceed), which gave me a 1.4 (25). I did not believe that, so I tested again using the same outlet and got 2.5 (45). Despite not feeling low (hello partial hypo-unawareness), I went with the latter reading, because I expected to be on the lower side due to the walk as well as a pre-dinner site change. Seemed to work as I was 6.2 (112) at bedtime.

Maybe it's time for a comparison between meters again. I don't have any One Touch meters, but a FreeStyle Lite, an Accucheck Mobile, and the Precision Xceed. When each of them were first used, I compared their results to a result from another meter using the same droplet for both tests, but I haven't done a full comparison for a long time.

WOW... what an eye opener for me! I've always been a one meter at a time PWD. ..and a Freestyle Mini at that! I was wondering if the machine was running high, as it's been quite a few years since I got it. I got the wrong strips last time I filled a prescription, so I just got a new Freestyle - something. (I don't like it as it doesn't have a backlight.) Anyway, once in awhile I would check it with a random sample at the lab... but I have no idea how they check blood glucose, so like you said, how accurate is that?!

There are times I feel crappy, test, and I'm at 7.2 (130) but I KNOW I'm low. Sooooo frustrating!

I should probably pick up an armload of monitors and do some experimenting of my own!

Thanks, Kerri...

ugh dude, i don't know what/how to trust. i used to occasionally test twice in a row, when i got a high or low i wasn't expecting. lately i've been testing 2 or 3 times in a row all the time because i'm clearly getting some off numbers. therefore, i'm gonna run out of test strips early this month so gonna have to test way way less for the next 2 weeks). i know i should stop using up so many test strips, but it's hard when you want to correct (with insulin or food) and you know you can't trust the result!

yesterday i tested and was expecting to be mid to high 100s and the meter said 80. i didn't believe it and tested again, 160, and then 162. if i hadn't retested i would've ended up 300+ after i ate. and last week my meter said 297 before bed and a minute later said 203. if i'd corrected the 297 and gone to bed it would've been BAD news.

one thing... someone told me that a lot of the variation/problems are with test strips, not meters. i've noticed this to be true, and get worse and better boxes of test strips. i use one touch and i bet your variation within different one touch meters is actually a test strip issue & not a meter issue. i bet if you tested on the same meter 3 times in a row you might have variation too.

I also think it is a test strip issue with the One Touch. In one 90 day shipment I was getting "Error 4's" (randomly - sometimes 3 in a row - sometimes 1 in 10) and Lifescan didn't know why. They replaced 50 strips. And the Lifescan rep was quite concerned that my strips were "cut wrong" - something I didn't realize since the strips are so crammed (50 in one vial) together. I thought that was the way they were supposed to be. (Stuck together on the long side.) I just happened to mention that I had to pull the strips apart to get just one out of the vial. He said that should not be the case. And I had been doing that for a long time with previous shipments, too. He said I was the first one who had mentioned it to Lifescan. He said the fact that they were "cut wrong" should not affect the result - but really - why shouldn't they just be cut right in the first place?

Thank you, Kerri, for opening this up this discussion. If enough of us speak up, maybe someone will listen eventually and something will change.

I ran an experiment last summer, because I was blaming a higher-than-I-wanted-it-to-be A1C on the inaccuracy of my primary meter. My endo. told me to bring both meters (primary and secondary) to the office, and while blood was being drawn from my arm I was busy squeezing blood onto two tests strips (Freestyle Lite, and what HAD been my primary meter, the One Touch ULTRA Link.) Both meters came in pretty close (within 10% of each other) while the lab result (yes, I know this was VENOUS blood) showed a glucose level that was higher than either fingerstick. Wish I still had the exact numbers, but that was my own personal way of checking the accuracy of my meters.

I'm using an Aviva and a Medtronic CGM. And not seeing much of a correlation, even after accounting for the CGM lag. The CGM has missed 3 lows in the 50's over the last week and a half. And I didn't feel one of them (only caught it because it was time to calibrate) and needed help for another. :-( Yes, I would be willing to skip the CGM (technology is still very primitive) if I had a meter that I KNEW was accurate. But at least the CGM gives me fairly accurate trends, and can show me when I need to test. But can I trust the test results???

This research statistically corroborate the trend you noticed ("The Freestyle meter almost always runs higher than my One Touch meters"): www.skup.nu

BTW, my personal experience (10 comparisons) is consistent with this trend as well.

Great post Kerri.

I've always felt that meter variations are a small issue in the big fight of diabetes, but I have to say that after seeing your pictures I feel it's a bigger issue than I've thought.

I always double check with the same meter before I give Justin a large correction. It bothers me that the 2nd test is always far from what I got the first time. I go with the lower number to correct, but there should be more reliable results for something so important.

A friend on the CWD mailing list had a daughter that was part of the Navigator trial and I remember her telling us about the requirements for BG testing. They were required to wipe the finger with a swab, get out a blood drop, wipe away that blood drop and use the second blood drop for the test. For better accuracy. So, the question I have is this, is everyone cleaning their fingers before testing, and what if you tried the second blood drop thing to see if there was more accuracy? I know we are not very good about cleaning off my daughter's fingers-just poke and go, so sometimes I worry about the reliability of her readings.

So I have a question that some what has to do with accuracy... I know you're supposed to wash your hands before you test, but lets face it i don't. But i was wondering about baby wipes. Do any of you use those before you test? Do they screw with the accuracy? Do you have any better suggestions? Cuz i test in class a lot and i don't wanna leave to wash my hands if i can avoid it. What are your thoughts?

I got my first meter in 1978, nearly 20 years after dx. I have never had more than one meter at a time; my current one is 6 years old.
That solves the variance problem for me.
I also wonder if it's a bit of a false concern. I use my meter to confirm what experience and habit tells me my BG should be. I'm almost always right (I don't use a CGM.) Rather than use the data as a starting point for management, I use test results as validation. And yes, there are occasional surprises but not often.
I have a very good, very well test routine that has worked for more than 50 years. It's liberating because it erases nearly all worry about technology and accuracy. In my mind, this is the best way to deal with numbers that don't match. Or ... you could toss all your meters but one and go with that.

I am Diabetic patient since 1995, and I use one touch/ ultra2 meter for testing my glucose, due to my work nature (in medical laboratory) I used to do double check on my readings between my One touch and our clinical chemistry analyzer (HITACHI ) and both shows almost same results with difference in 4-5 units (as stated earlier 20% error) which I consider normal for the fact that there are no identical readings in the world either from meter, or analyzer. and with comparing to my state at the reading time, I find it accurate. so this is what I experienced during 16 years
if you need discussion my email maharith@hotmail.com

So glad to see this post...it's a source of frustration in our house between myself, my daughter and husband. Have noticed these variances too and worry about over or under treating our daughter all the while weighing the affect on her little fingers and big feelings of being poked and prodded. We resorted to our own little system for checking: wipe with an alcohol swab, prick her finger, draw blood, wipe with clean tissue, get another drop of blood and then apply to testing strip. We have found that when we do it this way it is closest to her CGM. But really...w Who knows!!!

Awesome post and totally true. I can check 5 times and get 80-150 within those 5 times. It's agonizing.

If your actual blood sugar is 250, your meter could read 200 or 300 and be within the 20% margin of error--crazy!

I'm not a diabetic but am looking into meters for a friend.

I read a post somewhere by a doctor that the implanted blood glucose meters are not as responsive as a finger prick test - results are delayed 15-30min due to the location of the implant compared to the very responsive results from fingertip blood.

I wish I could cite a reference but I bet the Dexcom manufacturers have more data. Failing that, there is sure to be some research on the subject somewhere.

Hope this helps!

I am a diabetic, have given up on meters. As a statistician, I also find that the meter makers are not providing the necessary/sufficient information on variability. They should report, the within and between subject variability of each meter. I guess, their issue lies with the variability (batch) of the strips. In the absence of this information, perhaps a warning like "This meter reading is just a pigment of imagination -- don't get worked up about it. You (and your healthcare provider) are donating money to keep some people (other than you) alive. Thank you for your contribution"

Using religious fairy tales asThe whole IDF "PR" and its endless reiteration by the MSM is so tedious, especially when they blatantly ignore the actual

I have tried testing both units with the same Blood and get wide results. What I don't under stand is that last summer I went to the Ultra test , it was giving me the lower numbers. Now the Contour is giving the lower numbers . So I end up giving my self more shots then I go way to low, and have the sweats ,so you have get some sugar into me ASAP . What to do . I test my self so many times with both units its not funny. Like I said I will use the same blood sample and get way big results.

Ok I use 100+ test strips a month and my insurance has a co pay. I was using a Bayer Contour USB and a OneTouch UltraMini and both have a high price for their strips. I found CVS TrueTrack have the lowest cost strips so I got a CVS TrueTrack meter. I a comparison test between the three meters and the TrueTrack was consistently 21% lower readings 150 vs. 120. So I returned it to CVS and got another one to test. This time the tests were all over the board and done within thirty minutes! This is so frustrating.

First Test:
412 One Touch UltraMini
511 Bayer Contour USB
487 CVS TrueTrack
19% H/L Diff

Second Test:
441 One Touch UltraMini
454 Bayer Contour USB
324 CVS TrueTrack
29% H/L Diff

Third Test:
416 One Touch UltraMini
569 Bayer Contour USB
405 CVS TrueTrack
29% H/L Diff

Fourth Test:
386 One Touch UltraMini
332 Bayer Contour USB
354 CVS TrueTrack
14% H/L Diff

(I realize the numbers are too high, but that’s why I am testing)

I have found a difference up to 20 points also. I tested with 2 different meters both from the same finger site. Why are we even testing if the results are going to be that far off? To me its a waste of money buying the monitors and strips if they are not going to give accurate results. I have to pay for these as I have no insurance coverage for them.

Post a comment

(All comments are moderated. Thanks for your patience!)