« Finding Your Voice Online. | Main | Their Words. »

A Big Ol' Discussion About Meter Accuracy.

Using this image again because, damnit, it's a nice one.During the course of the Roche Summit, we had a big ol' activity about meter accuracy. I've written about meter accuracy before - actually, it was way more of a spastic rant - and I have remained frustrated about the lack of accuracy that's in the meters we use every day.  So when the Roche people had us talking about meter accuracy and our perceptions and expectations, I had plenty of opinions.

We were at tables of about eight people apiece, and each table gathered together to decide how we wanted to lean.  Our choices were 5/30, 10/20, 15/15, and 15/10, meaning that we could chose between a percentage of  variability on the low end of the blood sugar range (the first number) and a percentage on the higher side of the range.

EDIT on 7/8:  I was wrong.  Thanks to Amy at DiabetesMine for clearing this up.  The numbers stand for "number of points your meter might be off on the low end, and percentage of total error on the  high end." 

As a diabetic who was diagnosed as a kid, and also as a recently-pregnant diabetic, I had a very, very tough time deciding on an accuracy pairing.  Would I rather take the hit when I'm low or high?  During the course of my pregnancy, I was aiming for a fasting blood sugar under 90 mg/dl.  With a percentage of error in the double digits, was my 80 mg/dl really an 80, or was I dipping into the 60s?  Or the 100s?  My decisions for both ends of that small spectrum are different, and the results if my meter is "off" could be really problematic.

In an ideal world - and my table picked this as our initial ratio - we'd want 0/0.  Total accuracy.  But I know that's not possible, based on the constant fluctuation of a diabetic's blood sugar and the limitations of the technology.  For me, I was leaning more towards heightened accuracy on the lower end, because the lower the number, the scarier a "mistake" could become.  Also, with my goals set at 100 mg/dl and not 300 mg/dl, my hope was that I'd be dealing with a more accurate meter because my blood sugars would be running tighter.  Also, also, I was more willing to take a sharp blood sugar tumble from a high versus from a low. Further to fall, I guess.

My vote was for 5/30.  

And then I felt this wave of frustration, because the idea of having to choose heightened accuracy at one end of the scale or the other made me pretty fired up.  If you can be more accurate on both ends, why wouldn't you be?  Oh wait, is it the cost?  Could it be that once again, affordability comes into play?  Our group talked about different strips for children (more accurate on both ends of the range) and for pregnant women, but if I were to mention that concept to my mom, she'd say, "You're still my child, and I'd want you to have the most accurate strip."  There were comments that a more accurate strip would be a more expensive strip, and insurance companies (who barely want to cover four strips for a type 1 diabetic, never mind the 15 a day that I blow through) would fight endlessly to deny coverage due to the cost.  I understand that business is business, but isn't life still ... life?

At this point, I'd just take a meter that gives me the same result twice in a row.  (Because in the last 23 years, I've seen some wicked variability that meter companies should hang their heads in shame over.)  I'd even allow it to be 5 points off in either direction.  Nothing gives me more confidence in a meter result than seeing it line up with my Dexcom.  And nothing shakes my confidence more than seeing results that vary - wildly - mere minutes from one another.  With those kinds of results, I have just as much faith in a Magic 8 ball.  (Ask again later.)

What plays into your decisions on what is "acceptable" for accuracy?  Is it once your child sleeps through a low?  Once you find out your pregnant?  Once you have a seizure?  Once you correct down a 300 mg/dl and fall into the 20s?  Would you be wiling to pay more for a more accurate strip?  Would insurance companies be willing, or would they ever see the benefit of covering preventative care now instead of treating complications later?  What is your definition of "accurate?"

And if you had to select one of those accuracy ratios, which would you choose?


I also want to voice some frustration for the "4 strips per day" rule. Every time I have an appointment with my endo, I have to refill my strips Rx because my insurance won't allow a year's worth of 200 strips per month. So, if I'm testing 6-7 times a day on AVERAGE (meaning I have days where it's less and days where it's more because of . . . well . . . freakin' DIABETES!!!), then I definitely want it to be accurate.

But if I had to choose, I would be OK with 15/15 because it would be the same at both ends and I wouldn't have to think, "Is 200 considered high or 300?"

I agree with the variability of meters. Just the other day I felt shaky and tested to see that I was 217. I knew that it wasn't correct to I tested a minute or two later to find out I was 78. Wicked. I understand that BG's are (usually) changing, but some level of accuracy needs to be attained and right now I believe that we aren't at a reasonable level. I would side with you Kerri and go with 5/30, if I had to choose something "reasonable." I would much rather be a little higher than I thought that a little lower than the meter reads.

I am currently extremely outraged over this issue. I am 22 weeks pregnant and since I am testing a LOT more these days, I have noticed a TON of discrepancies. Last night is a good example that is still fresh in my mind. Before going to bed my meter said 161 (Omnipod - freestyle). That didn't make sense to me so I checked again. The same meter said 101. I pulled out my One Touch and that meter said 132. I tested twice more on each meter and got 112 (Omnipod) and 124 (OneTouch). How in the heck was I supposed to decide how to calculate my Dexcom? I think I entered in 120 and went to bed. (I must also mention that I washed my hands before testing so there was no residual sugar left over from the cookie I had eaten earlier).

And don't even get me started on my issues with my fasting numbers while pregnant. I aim for 100 or below because these days with my meter problems, who the heck knows what I really am anyways.

17 more weeks and I can kick this OCD to the curb.

I'm a type 2 taking both Lantus and NovoLog, and also on metformin AND Januvia (crazy, I know.. we're working on it). I'm also a little OCD… I test more than some type 1's I know… and I hate the discrepancies! Personally, I am more afraid of the accuracy being off on the high side, because I don't want to correct for what I think is a high and end up falling low. Just yesterday I tested (with clean hands) at 193, which seemed odd, and moments later on a different finger I got a 131. If I had corrected for the 193 I would have been crashing!
But no, I don’t think the insurance companies are willing to go for preventative care. You’ve seen, along with so many others, how hard it is to get pumps and CGMS’s covered, even in situations like pregnancy, where its more important to have the right information at hand. My doc put me on Januvia, which I love, but then I got a letter from my insurance company telling me they don’t actually cover it, unless there is a proven 3 month failure of metformin. Basically I have to get REALLY sick for 3 months straight to prove that I should take this medicine. Well thank you very much insurance company!

At various points, you said "points" and "Percentage", and I think my answer would be WAY different depending

If we assume POINTS, yeah, 5/30 is the way to go, BUT if we are talking percent?

Lets just say were talking 5/30 percent

at 75 points, you'd be +-3.75 point - I like that, but say at 250, you'd be +- 75 points!

Now, say you picked 10/20 - it's 7.5, and +-50

at 15/15, you're 11.25 (starting NOT to like that lower number spread) as your 75 is now 64-86 and the high is +- 37.5. and really, there isn't too much difference (speaks a type II) between 212 and 287

so, if we're talking percentage of reading, I'd TEND to lean towards 10/20 vs 5/30

Keep in mind that 30% off on the high end means waaaay off. A 300 might actually be a 390 or a 210, so either you're going to way over correct or way under-correct...

Sure over/undercorrecting for a high is less dangerous, since you'll have an opportunity to test later. But still that kind of discrepancy just isn't acceptable. (On the other hand, being off by 5% if your bg's 60 means only a point or two. As long as you know you're low, you're going to treat that low pretty much the same without really caring whether you're in the high 50's or low 60's...)

But yeah, the consistent responses here are that our meters are sometimes much more than 20% off. I can't tell you how many times I've started to correct for a reading in the 300's before realizing I didn't feel high, re-testing and finding out I'm under 100. Sometimes I don't even realize until after I've bolused and feel myself plummeting, test 15 minutes after bolusing and find I'm at 50 and dropping. (Freestyle/OmniPod seems to be really bad at giving those false highs.) And that's where it gets really, really scary...

This just turns my stomach thinking about it. And reading the previous comments made me even sicker.

My last trip to the endo, the CDE threw boxes of test strips at me when I told her how limited I was. She told me that I should be able to test whenever I want. Why can't insurance companies think that way?

I don't think I can pick a ratio. They're all bad. Can I just get a pancreas transplant? Will that be cheaper for all of these companies in the long run?

I would definitely have to go with wanting the meter to be more accurate for highs rather than lows. But this is coming from someone who has NEVER had a problem feeling her lows. They wake me up in the middle of the night and I often feel my blood sugar falling even when it hasn't technically gone in the low zone yet. So for me, lows aren't as scary. But I know that not everyone is like this.

Plus, just mathematically the difference isn't as great for lows as for highs. For example, if both ends of the spectrum offered 10% accuracy and you test at 80, you are somewhere between 72 and 88. Not a huge difference. But if you test at 300, then you are somewhere between 270 and 330. Both are still bad numbers but that is a 60 mg/dl DIFFERENCE!! That's a difference of 3 units of insulin in my correction bolus. Nope, still not acceptable. So why bother with the tighter accuracy for lower numbers? Not to mention that it's kind of disappointing to me that the best they are offering for accuracy in the high range is 10%.

For me, lows aren't fun but they are easily fixable. As someone who is carrying a child and afterwards will have to live with the complications (none yet for me, thank God) of diabetes, it's the highs I'm scared of. It's the highs that are teratogenic for growing fetus' and it's the highs that create the complications (and the drastic swings but really if you aren't ever high, then you probably aren't having the swings).

Sorry for the rant but this topic gets my goat. I won't even go into the frustration I'm having now that I check my blood sugar an average of 12 times a day and my insurance only wants to cover 5 strips a day. Grrr. . .

Rant over. ;-)

You know my two words on this little activity. Same initials as Blood Sugar.

Accuracy need to improve constantly. No sitting with last decades technology.

I was big on the 15/15 or 15/10 in my group. 30% for blood sugars above 75 mg/dl is just too much. (I don't think they emphasized the point enough that 75 mg/dl was the cutoff point between the two percentages.)

30% could still mean you're low if you're sitting at 80 mg/dl...or you could be stable in the low 100's. And 30% at 150? Could mean the difference between a small correction of insulin or being stable in the low 100's. And 30% at 300 could mean a significant difference in a correction dose.

Yeah, I'd love to see 10/10. I know that not all technology is perfect, though. Getting enough meters and test strips out to everyone who needs them - there's absolutely no way to guarantee perfection.

I can't tell you how many times I've tested and felt like it wasn't right...tested again to find that it was WAY off from the first number...tested a THIRD time...and then took the average of those 3!
I'm extremely blessed to have great coverage so I get a TON of strips each month...but there are still months that I end up buying out of pocket...spending an additional $100!!!

To learn more about accuracy and why it's challenging, this article Glucose Meters: A Review of Technical Challenges to Obtaining Accurate Results (PDF) is a useful read. It's from the Journal of Diabetes Science and Technology.

I am 27 weeks pregnant and have been frustrated over the same issue. My OB is VERY adamant about by fasting being under 90 and my 2 hour tests to be back under 120, but I have had to tell her multiple times that I really don't know what's real when I test. I am testing 8 - 10 times a day currently and have also done the two-meter trick and have almost always gotten a different result. it is SO frustrating, especially when you think you could be doing damage to your unborn child by not correcting appropriately!! Accuracy needs to be moved up the priority list immediately in my mind, as even after I give birth, I still want to be as tight as possible so I can see my little guy (and any potential future little ones!) grow up...

Our endo told us to stick with one kind of meter because the results can differ between them. I've seen this when we switched from the Freestyle to the Ultra Touch. We always had a problem with repeatable numbers from the Freestyle. With a high reading, the same blood drop could vary as much as 75 points, and when dealing with the small amounts we are with my son, that's unacceptable. What I want is a meter that varies no more than 5 points when using the same blood drop.
As for test strips; I had my endo write us a scrip for 8-10 strips per day. I refill that like clockwork so I can build up a safety stock in the event we lose the insurance, or they decide he should only test 4x/day.

I guess I am not really understanding the variables, but all of them seem..well..kind of dangerous in a way. And if they are asking you which way you want the INaccuracy, that means that have to target that with accuracy. So if they are going to use accuracy to get to the INaccuracy, why not just use it to get the numbers correct? Does that make sense at all?

I am doing an experiment with my kids this coming week. (you can find my post about it "Experimenting") I am doing a second test at least 3 times a day for 3 days. We will be reacting to the first test like normal, but I am going to take down the second BS number and make notes on what bolus we would have taken off of it or any other changes we would have made. I think it would be interesting to show how much of a difference it would have made over the 3 days. It won't show the whole truth, because we won't know if the kids would have dropped dramatcally or if they went high...but I think not showing the whole truth is the reason for all of this. Care to join us with our experiment?

Glucose monitors have been around for what...??? 28+ years?!? (I just had my 30th D-Anniversary July 1st. I had a meter shortly after being diagnosed...)They should have 100% accuracy by now! Everyone knows how much we dole out for a box of test strips.

I've had type 1 diabetes for 46 yrs. and seem to have more problems with lows than highs, so out of that hat, I'd choose 10/20. But, I'm neither a child nor pregnant. Real, true accuracy would be a nice thing, though, but I'm not holding my breath. I also love it when my Ping meter and Dexcom agree, or at least very close.

I would go percentages and say 5-10% margin differential while most of the time especially with freestyle, yet again I always double check my numbers if they do not correspond with what I am feeling at that particular moment.
Accuracy is a mere luxury if you ask me coz you will never be at 100%, I guess it depends on the particular individual what they are comfortable with.
Now I am a type 2 and do not have to bolus but when I test and the numbers are optimal I might let it ride for the motivation factor, one reading does not make or break my self management. If the highs go on for a few more tests then I adjust.

Right now we are at 20/20 right? Why can't we go 10/10? Freak...any other way is not acceptable. I wasn't expecting this to make my blood boil like it is, but these are my childrens lives here. Their well being. Asking me to choose is like asking me which arm I want to cut off. Sure, the right is more important to me...but without the left, life would be just as difficult. All the options are unacceptable. It's a joke they ask us to choose. I choose for them to improve accuracy on BOTH ends.

I would go for more accuracy on the low end. As we know, hypoglycemia gets you in trouble quicker than does hyperglycemia (at least in my own experience of 38 years of Type I!)As for strips and insurance ... I partially blame the meter/strip companies. How do they justify charging $1.00 for something that costs them 5 cents to make? I'm referring here to the cost of one test strip!

Being one of the only parents of a child with diabetes at the summit, I tried to voice my concern to my group. I actually talked with a Roche rep at more length to relay to her why I disagreed slightly with the group. (Though their points were completely valid, we just have different needs and perspectives.)

My child weighs 50 pounds. If her blood sugar is high and the meter is way off, I could be giving her a huge amount of insulin that she may not actually need. The 1 unit difference between her actually being at 250 versus 350 could send her very low.

That being said, if the meter comes up 80 (meaning it could be anywhere from 60 to 100), I am going to keep an eye on her, maybe give a snack, and definitely recheck in a little bit to make sure she's not going even lower.

But what if the meter comes up 70? She could be 50 which is very low for her. Or she could be 90, in which case I wouldn't be giving her quick-acting carbs at that point. But you have to treat based on that number that comes up.

And any time we get a crazy high number, we always wash hands and double check. I am always amazed when there is a gigantic swing.

Of course I want accuracy at both ends and there is no good answer.

(And I was going to write a post about this, but I guess I'll have to reiterate and expand upon what I just said!)

Well this is a hot topic for me right now considering I spent an hour on the phone with our meter company today. My daughter is 7 and honestly could have died due to inaccurate readings during the night last night. I almost corrected a reading of 499 during the night but luckily I was coherent enough to retest her and she was actuallu 50!!! I can't stand to think about what could have happened if I corrected that and went back to sleep! I want accurate readings! This is my child's life!

I didn't completely get the explanation, but thought they'd said it was Percentage on the Low End and Actual Number Points on the High end (so a 5/30 would be within 5% accurate on Lows, and within 30 points on Highs. In Leighann's group, I personally also felt like I'd rather have the accuracy more fine-tuned at the High end versus the Low end. For a specific reason: we were talking numbers greater than the 75 mg/dL, and it was balancing the eating to correct versus the taking insulin to correct. I see the dangers more with Overcorrecting Highs, especially at bedtime when you may take too much insulin and then go to sleep and not have the ability to wake up and catch a Low. If it's Low, you may eat something but won't be in as dangerous a spot as you're still going up based on the food eaten. That's why we ended up at the 15/10 ratio - within 15 points on the Low end (above 75 mg/dL), but within 10% on the High End.

What a frustrating exercise!! Not pregnant, i'd probably lean towards 10/20 if the above are the only choices. But I'm currently 25 weeks pregnant and right now i would vote like you 5/30 since i'm much more likely to be low these days than high.

At the same time, choosing 5/30, i can't help thinking that a 30% error range on a high number (how high is high anyway???) is WAY too much. It means 200mg/dl could show up as 140 or 260. How are you supposed to make treatment decisions based on such a wide range?? I'm going to read the article posted in another comment. I'm hoping it will explain why tightening accuracy at one end of the spectrum, automatically loosens it at the other.

My vote would be for 10/20, 15/15, or 15/10. 30% off on the high end is totally and utterly unacceptable. End of conversation for me. I'm an adult type 1 and take very little insulin--for me a 30% variation on the high end could cause disastrous results from giving unneeded amounts of insulin for an inaccurate number. A variation of even 20% on the high end is nearing unacceptability for me (though that's what it is currently).

My ideal would be 10/10, but given that may be expecting too much, I would be satisfied with 10/15 (tho that's not an option given).

Like everyone else, this issue just makes my blood boil. Yes, I would pay a (small) premium for more accurate strips, even if my insurance company wouldn't.

My gut reaction is that we shouldn't have to settle for any inaccuracy...but since reality bites, I'm going to have to go with less accuracy for lows and greater precision for highs. It is my experience that almost every time I get a number above 250, I will get big variations if I try to verify that initial test. If I take the first result at face value and correct, I almost always get low later. Is it because my correction ratio for high numbers should be adjusted, or is it because the meter is throwing out crap?

Just last week I rang in at 387 (ouch). I tested immediately after to verify and got a 352. A third test came back at 313. I corrected for the third number -- SEVENTY POINTS LOWER THAN THE FIRST TEST -- that's like two units of insulin too much I would have taken! I still ended up low later. (To clarify, I don't have problems with hypoglycemia when I'm correcting highs under 250.)

Like a previous poster pointed out, I do tend to feel my lows in enough time to deal with them. There isn't much difference in a 30 or a 50 for me; they both feel awful and require attention. What makes me very upset is knowing that a poor meter reading may cause me to have a low later.

I wonder why they can't make two different types of strips: one accurate for lows and one accurate for highs.

You know, all of these great comments help me process this situation.

It is a lot to think about, and I think there is a lot that we don't fully understand, AND there is a lot of misconceptions out there.

The times are different now. We've come a very long way in the last 30 years, and maybe I'm not all worked up about this because of that perspective.

But with that being said, there are some really great points in these comments. They've got me thinking.

Our insurance company is from heaven and pays for everything except our insulin co-pay. We get as many strips as my daughter needs, delivered to our front door. (Go Medical Mutual of Ohio!) How lucky are we?? Since we have so few other costs, I would definitely pay more for more accurate strips for my daughter.

Ugh. I want 0/0 too. Or if there must be a reasonable margin of error, I feel like something like 2/2 would be reasonable.

If I read correctly and the low # inaccuracy is points (first #) and the high # inaccuracy (second #) is percent, I think I would have to go with 10/20 or possibly 15/15. I want to know when I'm actually low - 75 is a fine place for me to be. 60 is not. But 30% inaccuracy on the high end would be totally disastrous in terms of giving correction boluses (20% is bad enough). If that is in percentage but my correction ratios are still in points - that magnifies the problem! So if I'm 350 and could actually be 245 or 455, that is HUGE - for me a difference of 3.5 units (4 vs 7.5)!

0/0. Period. For *everyone* regardless of insurance plan, or age. To the insurance companies: why not?!

Remember when the car makers said they couldn't put in seatbelts because they would be too expensive? I do and they were lying. The makers of our meters are lying. They don't care about our health just their ungodly profits. The ADA says all meters are accurate, thanks. I just got a rather large drop of blood and was able to do 3 tests with my One Touch Ultra Mini. I got 138-116-122 on the same drop of blood. And of course that is less than the variation they are allowed to call accurate. The medical industry is lying as usual and they controle the FDA and the ADA and the're not doing it for your own good.

Corrected Spelling...

Remember when the car makers said they couldn't put in seat belts because they would be too expensive? I do and they were lying. The makers of our meters are lying. They don't care about our health just their ungodly profits. The ADA says all meters are accurate, thanks. I just got a rather large drop of blood and was able to do 3 tests with my One Touch Ultra Mini. I got 138-116-122 on the same drop of blood. And of course that is less than the variation they are allowed to call accurate. The medical industry is lying as usual and they control the FDA and the ADA and they're not doing it for your own good.

Post a comment

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