« SEO-riously? | Main | Health Care Reform: How Does it Affect People with Diabetes? »

Accuracy in Glucose Meters.

Accuracy would be nice.I've been diabetic for over twenty years, and when I was diagnosed in 1986, my first meter was an Accu-Chek (this post has some pictures of these old diabetes meters).  The strips were large, absorptive pads that soaked up the blood drop and changed color after I wiped the blood away with a clean cotton ball.  The color strips were either stuck into the machine and read after 120 seconds, or the color pads were compared against the guide on the side of the strip bottles.

Comparing colors?  Total guessing game.  But waaaaay back in 1986, glucose meters were FDA "OK'd" to be off by a margin of 20%.   That means (if my math is right, and don't trust me that it is because I'm still a Morrone) if my meter said "100 mg/dl," I could have been anywhere between 80 mg/dl - 120 mg/dl.  For me, particularly at this stage in my life, that's "almost low" or "almost needing a correction bolus."

Wouldn't you think we've come such a long way since then?

Oh, but we haven't.  

The standards for glucose meter accuracy haven't been tightened up since before my diagnosis.  Sure, meters have gotten fancier and faster and they come in a variety of brands, colors, and sizes, but they aren't more accurate than the ones I used when my fingers were being lanced for the very first times.  (And this was proven to me when I was at the ER the other night - the nurse busted out a Lifescan meter that was older than dirt to test my blood sugar.  It was their approved hospital meter, and the results were the same on that meter as they were on my current Lifescan meter.)

Scott wrote a remarkable post about the issue of meter accuracy yesterday (click here to read it) and Kelly also hit this nail squarely on the same day (her post can be found here).  And Bennet questions what accuracy is in the first place.  And after reading their posts, I'm inclined to comment as well. 

I base EVERYTHING I do, diabetes-wise, on the information provided by my glucose meter.  I base my insulin-to-carb rations on those results, my basal rates, my food intake, and my exercise decisions on those numbers.  I calibrate my Dexcom using those numbers.  EVERYTHING.  Hell, maybe not even limited to just diabetes stuff.  I won't get into my car and put the key in the ignition if my meter shows a result that's too low, and I am reluctant to eat more than a salad at dinner if my glucose result is way too high.  

So when the numbers being provided by my meter aren't as accurate as my dedication and determination deserve?  I get a little fired up.

Over the course of my diabetes life, I know my meter has thrown some wonky results.  I'd have to ask my mom if she ever saw any crazy results, but I know since monitoring my own disease, I've seen plenty.  Over the last fifteen years or so, I've used meters from Accu-Chek, Agamatrix, and Lifescan, and in the last few years, exclusively from Lifescan, due to what my insurance covered.  And for the most part, I never second-checked any of my results, unless I really felt that they were "off" in accordance with my physical symptoms.

But since using the Dexcom, I've been double, and sometimes triple, checking results.  Take, for example, the other day when I received the following results:

Incident One:
3:56 pm  86 mg/dl
3:57 pm  159 mg/dl
3:58 pm  164 mg/dl

The only reason I double-checked that first result was because the Dexcom had me at 170 mg/dl, not 80-ish.  I know the CGM isn't for dosing or treating, but it's pretty friendly for trending and "gisting," so the discrepancy caused me to double-check the result.  And, for the record, my hands were freshly washed with soap, dried, and clean.  The lancet was new (I know - dumb luck on that one).  And I wasn't eating or exercising at the time.  So what gives?  86 and 164 are very different numbers.  That's waaaaay more than 20%.  Even I can do that math. 


Incident Two:
5:51 pm  110 mg/dl
5:51 pm  203 mg/dl
5:51 pm  229 mg/dl

This is not the first time this sort of thing has happened - I've blogged about it before and have linked to several of those posts throughout this post.  But now, things are different.  I'm in a constant state of blood sugar panic these days, worrying that not only are these results (accurate or otherwise) affecting my health, but how are they impacting my currently-being-built baby?  Would I be double-checking these results if I wasn't pregnant?  If I wasn't worried about BSparl, would I be willing to use three test strips (aka almost $3.00) per glucose check per sitting?

The FDA recently had meetings about this very topic: glucose meter accuracy.   (CWD mom Ellen Ullman actually spoke at this meeting, representing the diabetes community.  Go Ellen!!)  And Scott had an excellent synopsis of one of the most disturbing statements made at this meeting:

Excellent point made by Scott.

Not good enough for me.  I'm expected to maintain an A1C of 6.5% (or less) throughout my pregnancy.  Diabetes complications, no matter how hard we work to advocate against the guilt, are still ascribed to the patient, not the disease.

I'm working very hard to take care of my diabetes, and to have my efforts impacted by industry apathy isn't right. 

To echo Scott's call to action, I'm urging us all, as members of this influential diabetes community, to visit www.regulations.gov and search for Docket No. FDA-2009-N-0604 to leave your comments.  Tell the FDA that 20% isn't good enough, and we deserve every chance to take the best care of ourselves.  

I don't care if my meter comes in 15 different colors and can juggle knives - it needs to be ACCURATE

Raise your voice.



Well said, Kerri. I left a comment on that docket.

I recall that Accu-Chek meter, too. How far have we really come?

The basics will always be the basics: test, bolus accordingly.

We rely So much on the numbers!! 30ish years and accuracy is still crap? That's sad. And wrong.

Kerri what a great post! It infuriates me when I'm calibrating my dexcom and I see just how vastly different my numbers can be with no more than 30 seconds between testing. I was complaining about it to my endo at my last appointment. Meters have improved so much in the last 20+ years we've both had diabetes you would think they could improve this accuracy.

As the father of a type 1 diabetic son, I have researched the +/- 20% accuracy. The meters are required to be more accurate in the BG range that would be considered 'normal' low to high (something like 30 to 250) and the +/- 20% accuracy is for the much higher highs. That is why I have told my son, that if his BG is over 300 (happens more than I would like) that he is to reduce his correction bolus to about 50-75% because the in-accuracy might make him go from high to low, but reducing the dose might still keep him high, but into a range where the meter is more accurate.

Good post... The funny thing is that some doctors (read:mine) still think meters are more accurate than my Dexcom. As your post points out, I've gotten to the point where if my sensor is fresh and working well, I have much more faith in it.

That statement certainly doesn't make any sense. If the people who rely on glucose meters for their LIFE are people with type 1, not type 2, then clearly we should be focusing on the needs and desires of the population who more readily use glucose meters on a regular basis. If I am the devoted customer of Lifescan, as opposed to a type 2 diabetic, then *I* am the one they should be listening to. Putting the input of the minority customer above the loyal customer just doesn't make any sense. It's bad business.

Great post, Kerri, I will do my part today.

I posted a comment as well.

The doctor mentioned in your post was the doctor that saw us in the hospital when my daughter was diagnosed eleven years ago. We only saw him once after the hospital stay and knew based on how he handled a newly diagnosed Type 1 7 year old child that we needed to get another doctor, and we did just that...enough said....

Geez, they really need to fix this problem. And besides, even though "the vast majority" is type 2, aren't type 1's required to check bg's WAAAAY more often than type 2s? If so, they should be making BOAT-LOADS more money off of type 1's than 2's just in strips alone, further proving that they NEED to make them more accurate.
I'm going to look up the link now. Thanks!

Last week at my endo I had to have fasting blood work drawn. When I woke up my BG was about 128, so I drove to the doctors, let them jab me, then checked with my OneTouch Ping meter to see if I was still OK to drive in to work before eating.

It had me at 96 - hey, no problem for a 15 minute drive, I was still 90 when I got in and ate.

Great, right? Except according to the lab results I got back at my appointment that afternoon, I was really 62.

96 is 62 + 50% inaccuracy, not 20. I would certainly not have driven if my meter had said that...

Have you ever noticed too that high readings take longer to appear on the screen than low ones? What is that about? I blogged about it once (http://www.irunoninsulin.com/?p=1169) but no answer - and I swear it happens every time... a low BG pops up quickly, but if the meter's taking too long, I already know I'm high - its as if counting the high BG literally takes longer because its a big number. Have you ever noticed that? Call me crazy...

Thanks Kerri for this post- I'm already on it!
Being diagnosed in 1988, I remember (and have pictures@) of the old glucometers, still have them (somehow I think they will be a nice museum exhibit someday!)
I am shocked at how many times, just as you showed, that my numbers are more the 100 off in a mere min. apart of checking! Diabetes is stressful enough, let's not have the 1 thing that is supposed to be {relatively} easy, be the 1 thing that throws us off track everyday! If I could show my frustration over this-- GRR! :)

Anyways, thanks as always for making me look at things and just make me want to do more!


After using a CGMS on our daughter 24/7 for the past 3 years I can tell you that we have lost a lot of trust in the accuracy of our meter results, (various brands, it makes no difference).
We retest frequently if the numbers are off and in about 1/4 retests it's the meter... not the CGM... that is wrong.
I'm so grateful that we have the CGM as a guide because a dosing error on a child with 1/250 correction factor has to be precise!

I too was just in the hospital and they used a lifescan meter. It said 122 and my Accu check aviva said 108. My accu chek may not be 100% accurate, but it is extremely consistant- within 5% on retest for me. I believe that the Accu Chek and Wavesense are more accurate than the lifescan run of meters because they use more filters in their strips. Also, it's not so much the meters as the strips that need improvement.

Kerri-thanks for letting us know about this. I left a comment. It is extraordinary that the progress made on "pretty colors" and such seems to have taken priority over a variance allowance in accuracy. Surely medical science and devices have improved enough to lower that limit to 5-8%. That certainly seems reasonable though under 5 would be preferable.

It is frustrating when they are not accurate and all you can get from the manufacturer is the ol' strip wasting control tests to verify. If my results and how I feel don't jive, I'll do 2 or 3 at a time to verify.

Tight diabetes management requires predicting trends and it is not acceptable when we have to use unsure data to make our decisions! We need a sort of "blood sugar GPS" to tell us where we are, which way we are headed and how to safely arrive where we need to be!

I agree wholeheartedly with you Kerri! Also, they fail to mention that over half of TYPE 2's also use insulin! I will be sure to make a comment asap! Thanks for bringing this to our attention.

Great post. Completely agreed with you Kerry. Yesterday I had:
1:09 pm - 67
1:09 pm - 172
1:10 pm - 66

I've made my comments, and I too have seen a way bigger that 20% difference between my son's readings literally seconds apart. I have found his One Touch meters to be more consistent between each other and multiple reading than the Freestyle, but more accurate? I just don't know.

I never noticed THIS big of a difference, but I've also never tried a detailed experiment. But I want to say, above all, yes, I hate the guilt this disease comes with, and I want accurate meter results for my own sanity!

Wow...I had no idea. I barely EVER double check my results and that's because I hate wasting my test strips, and I run out without doing that.

I am going to run over to that site asap!!!

I routinely do what you did and that is to retest when the first meter result does not roughly match the CGM report. Often I find that tests two and three match the CGM but that test one was aberrant. I don't get such aberrant results often but when I do it makes me wonder why. Maybe now that the meter manufacturers know that many of us are "testing" their meters against our CGM results, they will on their own pay more attention to accuracy.

This has really made me stop and think.....thanks for posting this. Have posted my own response over on my own blog. Am really quite disgusted, as I had no idea that the % of difference was so high...

I remember panicking when pregnant and my meter read 17!! I called my endo and asked why I was still alive...seriously... 17????!!!! He said that at really low numbers, the results are highly inaccurate. So what, I was really at 35 or something. Good grief!

done and done!
I am so offended!!! Grrrr!!!!!

OOOOO You've reminded me of some real stress a while back. I contacted the manufacturer of my test strips when I was seeing a discrepancy over and over again between this canister and another somewhat older but not outdated one. The newer ones read higher and coincidentally had a wider range of acceptable accuracy 91-122. HUH? My blood pressure was going off the charts while trying to get good service from the One Touch Ultra folks. According to the customer service people, I was wrong every single step of the way. "You must not retest with the same wound even if it is still bleeding." "You must not retest without waiting 15 minutes." "If you are under stress" (could they hear the rising strain in my voice?) your numbers will not be accurate." By that point I was furious and decided the best course was to hang up and get a monitor from a different company at my next office visit. Never once did they ask me the serial number on the canister. It was apparent the company policy was talk around the problem and never check it out.

I haven't read all the replies, but I wanted to say that as a Type 2 I do need a more accurate meter even though I'm diet/exercise controlled (no oral meds or insulin). When I go to workout I often test first. If my sugar is good or a little low, I make sure to eat first. If it's on the high side, I my go straight to the treadmill. One day I tested twice with the same poke. There was a 25 point difference. That was the difference between a snack or no snack. Those that say that we don't need an accurate meter because we're not on insulin are probably the same that think that Type 2s shouldn't test at all or no more than 1-2x/day. ARG!

Thank you so much for posting this topic. It's something that has bothered me for many, many years.

When I told my doctors about it, they said, "it's your fault, you're not using the meter correctly."

I was wondering what good is a CGMS if you have to confirm using your fingers and they give a reading of 100% off? How are you supposed to make an intelligent decision on such different numbers?

Wonder if a walk in DC would do any good?

Wow--great post. Will def. post a complaint on the site you provied-thanks!


I agree things need to be accurate, I count on it. This situation has happened to me many of times. However I have had a scary situation, where I correct the high inaccurate reading and as it turns out end up panicking and trying to eat and drink juice as quickly as I can to try and reverse the bolus I just gave myself for the inaccurate reading.


Thank you for posting this! I don't have diabetes, but something called congenital hyperinsulinism (CHI). It's a genetic disorder and causes severe hypoglycemia. Children born with CHI run a high risk of developing disabilities and developmental delays as a result of repeated and severe hypoglycemia. Obviously, people with CHI need to use blood glucose meters to diagnose and treat hypos. The problem: blood glucose meters become even MORE unreliable when it comes to diagnosing low blood sugar. The industry argues that, for most diabetics (i.e., type IIs), it's really just important to identify the high blood sugar, as that's what leads to the complications. They seem to overlook the complication of DEATH that people with Type I (or CHI) can experience if they miss a low blood sugar.

I am in total agreement that I want my meter to be accurate. I want to know, before I get in the car and drive, that my BSL is 85 and not 55, which can mean a world of difference for me and other people on the road.

Done! Thanks for giving me the opportunity to make my voice heard.

The only time I've been in the hospital for my diabetes since the first year I was diagnosed was due to meter error. I tested four times in a row, with washed hands and proper technique. I got two numbers in the 70-range and two numbers in the 250-range.

Beyond commenting on the FDA document, call your meter company and let them know you want more accurate results. I called mine, and I could tell they were expecting me to say that I wanted flashing bells and whistles, not a more accurate meter.

Thank you for this excellent post about a problem faced by all diabetics Type 1 and Type 2 alike.
I am a Type 2 diabetic for the past 26 years and always had meter accuracy problems despite which meter I have used, and believe me I have tried in these 26 years at least 2 dozens different meters.
then a few years ago David Mendosa in his excellent blog and site for Type 2 diabetics presented a meter which I considered unique in accuracy The Swedish made Hemocue 201.
here is the link to David's site:
It is an expensive meter, with some drawbacks like its size and the fact that the strips need refrigeration. But I have been using it for the past 5 years as a benchmark checking the accuracy of my meter (a Bayer Breeze 2)particularly when I have highs or lows and the Hemocue 201 proved superior to all the others as I checked them in Hospital and Laboratory environment.
I don't know how practical my solution is but had worked for me innumerable times.

I just left a comment. Yes, 20% inaccuracy is not acceptable! Thank you always for sharing.

I posted about this too a while ago ... http://www.act1diabetes.org/2010/01/19/blood-glucose-readings/. I really like your point about the 20% margin of error being the same as it was over 20 years ago. How could that possibly be???? I left a comment, thanks for posting the link.

Thanks for your post. I do agree with you. it needs to be accurate for we pay it accurately when we bought it, right? :o)

I agree - and that doctor is wrong. Type 2 diabetics DO need to monitor their results closely, especailly as type 2s are diagnosed younger nowadays - I'm type 2 and my doctor wants me as close to 6.5% as possible. The closest I've gotten in 6.8% I need a meter that is accurate.

Thank you.. I am recovering from aorto bifemoral bypass and struggling with issues because of long standing diabetes that was poorly controlled. I was just recently referred with the dexcom seven. It seems so much is available that we often don't know about or are limited to. I am not happy with Obamacare but that's another story. Just really glad to have found your blog.

Why are there so many different meters? I deal with people who are changing them very frequently

Thanks for the post - I commented too. I think people that think such high variances are okay have never experienced a high or low...

The incidents with the CGM and the freestyle meter happened as well with me... really odd and frustrating. I had to ask my mom to do some checks on herself a couple of times to see what the deal was. Its frustrating and confusing....and not a relief necessarily when I am spending my second semester at college 2 hours away from my momma

As with other products that are defective (e.g.,contact lens solution with eye-burning thimerosal ), Big Pharma will eventually correct the problem, but rather than do a recall or admit the mistake, they will sell the repaired version as "new and improved" and charge a higher price.

What I find most interesting of all is that no one has questioned the accuarcy of the test strips. How come?

"I don't care if my meter comes in 15 different colors and can juggle knives - it needs to be ACCURATE."

Yeah. Accurately said. The right measure is a need especially when we are talking about cure, eh.

I know this is an old post, but I just need to get this off my chest and. I live in Norway (Europe),- here the official guidelines state the following (my translation): "95 % of tests done by diabetics (meaning not in the lab) shall be within ± 25 % when bls is ≥ 75 mg/dl and within ± 18 mg/dl when bls is

The rest should be: "...under 75 mg/dl..."

Another thing while I´m at it; I found out today that there is an international website where they publish the results form testing meters. The new Bayer Contour XT with the the new test strips named Contour Next has come up with great results. Check out www.skup.nu.

Post a comment

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