Glucose Meter Accuracy: (Please Vote) Is 20% close enough for you?

Please vote in the poll below, even if only to let me know that I’m being overly anal-retentive.

I only learned recently that the industry standard level of accuracy for blood glucose meters is to be within 20% of lab results. I learned this after a doctor’s visit where I tested my blood glucose (BG) via my home blood glucose meter within a minute of my blood being drawn for lab work. I noted my BG as 82 and thought, good, right on target.

Anyway, this time, I was careful to check very close to the time that the blood was drawn and equally careful to note that particular reading. Again, my meter said 82.

My lab result was 69! This is a difference of just under 19% (((82 – 69) / 69) * 100) = 18.84%, though perhaps I should just round to 19 since I don’t really have that level of precision in the actual readings.

I was very upset by this, and not just because my doctor warned me about low blood sugars.

If my meter is consistently 19% higher than my actual lab results would be, then it is affecting my tight control of my diabetes. After all, at 82, I do nothing and am happy to be in my target range. At 69, I’d probably have a snack or at the very least suspend my basal rate for an hour, more likely the former.

So, I called the manufacturer (Abbott) and was told that the industry standard level of accuracy is within 20%. I have since told them repeatedly that I do not care about the industry standard, only about my diabetes care. I have recommended that they go against the standard, pick a tighter standard, e.g. 5 or 10 percent and then advertise that their meter is more accurate than the competition.

I believe this would be a huge selling point. Do you agree? Please vote below.

Do you have a brand of meter that you believe is more accurate than 20%? If so, please let me know which one below. Mine is a Freestyle Flash. They have already sent me a replacement Freestyle Lite. Depending on whether I use the test strips that came with it (about 15% higher than my old too high meter) or the extra strips they sent along with it (generally within about 5% of my current meter), I get very different results. I don’t know if I can trust this through multiple batches of test strips over a period of years. I had hoped that when I called back and told them about the batch that came with the meter they would tell me that those strips were recalled. No such luck!

On the up side, this is the only meter I can find that will operate at 40 degrees Fahrenheit, meaning that I don’t need to wait half an hour in a ski lodge for the meter to warm up sufficiently to do a reading. I like the meter a lot … or would if I could get more accurate results out of it.

Anyone know of a better meter? Have you tested it along side your lab tests? How close was it?

Thanks for any input.


33 Responses to Glucose Meter Accuracy: (Please Vote) Is 20% close enough for you?

  1. Funny you should be experiencing a similar difficulty with your meter.
    I use Roche’s Accu-chek Aviva Nano and as a Senior most of the cost is paid by our Canadian Medicare system. Lately I have been getting higher readings than I thought should be registering. Instead of 8.5 to 9.0’s I have now gone to the 10.0 to 11.0 levels.
    So I bought another meter of the same type to see if the results are repeatable. And they are. So my body has changed a bit and I will need to make a new appointment with my family doctor.
    I am currently controlling my Type 2 Diabetes using Metformin, Jenuvia, with several other additional medications regards water retention, high blood pressure, and cholesterol, etc.
    On top of that I am in a special drug research trial of the new medication called Ranirestat. It is a 26 month double blind study in which we are not told if we are given the placebo, the regular dose, or an enhanced dose of Ranirestat by Easai? I know that I am getting the drug, because I can smell it in my urine!
    This trial is a very wide final round investigation in N. America to see if it blocks the bad metabolites from the hyper glucose condition. Their aim is to gain approval from FDA and Health Canada. The drug is thought to block the process which produces Peripheral Diabetic Neuropathy leading to retinal damage and blindness, loss of skin nerve sensitivity, etc.
    I attend regular quarterly physicals at the Research facility of the University of Toronto at Toronto General. They keep very close tabs on how I am doing. I know from their tests that these meters are not that precise nor accurate.
    Your meter readings must be on a different scale or different system, I suspect.
    Now that I have the two meters and use the same test strips but from different batches and using the correct code chips, and am getting good repeatability between meters. Normal non diabetic readings should be from 4.5 to 6.0 mmol/L first thing in the morning and two hours after a normal meal.
    Your thoughts on this are welcome. [I am a 71 year old male Cdn. bearing DNA traceable directly back to Bloody Mary, Queen of Scots who also was a Diabetic!…through an illegitimate branch of the Stuarts!]

  2. Roight Munkee,

    Yes. There are different scales for blood glucose monitoring. In the U.S. (and possibly only in the U.S.), we use mg/dL instead of mmol/L.

    Here’s s good conversion chart on a site with a tremendous amount of good information regarding continuous glucose monitoring, for any who use CGM. I do and have had a lot of back and forth with the person who created this site.

    For other languages, go to the home page and choose the language you want. Then navigate through his well laid out links on the left.

    Good luck with the drug trial. I have very little knowledge of type 2 diabetes treatment. The disease is completely different than type 1 but has similar symptoms.

    I do know that many people in the U.S., including my father, now do treat type 2 with insulin for tighter control. If you’re not taking insulin and are having trouble, you might look into this. Insulin, unlike other diabetes drugs, has no real side effects.

    While it’s certainly true that type 2 diabetes is not caused by too little insulin, and in fact many type 2 diabetics have much more insulin than should be needed coursing through their veins, more insulin does still lower blood glucose and can be effective at helping to manage type 2.

    The biggest downside to treating with insulin is that, like me, you’ll have to worry about, recognize, and always be prepared to treat severe hypoglycemia. I always carry glucose tablets with me and also have a glucogon emergency kit that my wife knows how to use. If you get one of these, the hope is always that it will expire and be thrown out as a complete waste. Thus far, I have never needed this and still continue to buy them and especially travel with them, just in case.

    Thanks for the input on the meter. It makes me feel both better and worse to know others have the same issue. Perhaps if we get the word out and let the meter companies know that they would be rewarded for making a better meter, they will listen.

    • Mr. Fusion says:

      I remember many years ago when my American fiancee looked at my Canadian blood test results. Needless to say they were foreign to her.

      She had to find a conversion chart to find out I would live.

      It is my understanding that drugs must also be within this 20% of nominal. Anyone have anything on that?

      • Mr. Fusion says:

        Maybe I should mention my now wife is an RN who at the time specialized in internal medicine.

      • That’s funny. Regarding the meds, I’m not sure what the tolerance is. I believe it applies to generics because the brand is the standard against which the others are judged. A friend of mine had an issue with a very low dose medication he needed where the generic was consistently worse than the brand name because even a little bit off on such a low dose mattered. There can also be differences in the “inactive ingredients” that may matter more to some people than others.

  3. I did not mention that I was recalled for some extra special testing a week ago. There is a new machine which reads/measures the number of nerve bundles in the retina. This machine has an ocular piece which touches the eyeball through a gel sort of like how an ultrasonic transducer would, but has a Red light strobe. If there are 3 or 4 nerve bundles observed in the various selected fields, then there is no evidence of Peripheral Diabetic Neuropathy present, but less than 3 is not good news. I was told I have 4 in most of the fields scanned and none less than 3. The retina is the first place in the human body where diabetic damage can be detected.
    Other tests used include measuring the ability to detect fine vibrations and touch sensations on the skin near the toenails, plus measuring the nerve conductivity of mild electric shocks. Those conductivity tests are quite uncomfortable plus they cause you to jump/jerk, but are allegedly non injurious. One has to be somewhat masochistic to keep going back to that testing!

  4. Sounds like your testing in Canada is far more comprehensive and high tech than here in the U.S. (So much for all of the hype on that bad Canadian medical system providing less care than in the U.S.) Did they perform that test on your eyes just as a precaution? Or, had they already detected some level of retinopathy and wanted to check more carefully?

    My eye testing has mostly been having a specialist dilate my pupils and look inside with bright lights and lenses. I have had my optic nerve photographed once and even had my entire retina photographed for a baseline years ago. But, such testing is not normally performed here without any complications having been detected.

    As for neuropathy testing, my doctor just touches my foot in various locations with a filament and asks when I feel it. Sometimes he puts a tuning fork to my ankle and asks when the vibration stops.

    I do get an A1C every 3 months as well as some other routine blood work. Once a year, he is more thorough and also checks for albumin in my urine.

  5. Not so. I was nominated for this research guinea pig position by my family clinic. This is all funded by the Drug company. I get $50.xx / visit for gas, parking, lunch, & only break even.
    This new machine is one of only a few available yet. They are specially calibrated for the Drug Mfg so that all of their data is validated for the FDA + Health Canada.
    Google Dr. Vera Bril, PhD. Professor of Neurology for insight into her research which is in conjunction with Easai.
    I was also getting those filament tests and a vibrator in place of the tuning fork, but the new one is adjustable and recordable.
    Along with the same blood tests you mention, they always run a cardiac EKG, B/P, Urine sample, and I have to fill out an attitude test that I call a fudge factor test! I don’t know how many thousands are in this program continent wide, but I think I recall there are 250 at the Toronto location. Our medicare does not pay for hardly any of this if at all. I did have an MRI which may have been on the Province’s tab, but one of my visits entailed something not paid for by Easai, soon after the Gov’t health care audit idiots sent me a questionnaire to ask if I had seen Dr. Bril’s office on such and such a date, if so, do not reply! Isn’t that a brilliant way to run a bureaucracy?

  6. Ah … I have not been lucky enough to be part of any studies yet. I have asked to be, on occasion. Generally, the companies do not seem to respond.

    Is there anywhere on earth where a bureaucracy is run as smoothly as … say … a clown show?

  7. mfg says:

    yea! It’s like reading a shotgun blast at 50 yards on a 4 x 8 sheet of plywood. the closer you are to the point you aimed at the better the percentage. The further away you measure from the point you aimed at the greater the % error. But if you average all the measurements your overall % is damn close. problem is at least 2 fold; we as consumers are not willing to pay for greater accuracy, therefore any free market company will not sell any meters that are expensive. the big companies don’t listen to the normal person who uses the meters, they listen to the marketing experts who crunch the numbers from a general public and determine an average on what is acceptable. it does not set well with me either. type 1, 4 years plus, virus initiated, medtronic pump with sensor. sensor is greater than 20% accurate , more like 40 % and at least 20 minutes behind meter readings. all the professionals i talk with tell me to watch the trends. problem is after one reading of extremely low bg at a greatere % than 20 i will no longer care.

    • I think some of us would pay more for greater accuracy. I don’t know how much more it would cost to make a more accurate meter and test strips. I’ve never paid for a meter. I can usually convince them that the money is in the test strips. Years ago, they always gave away meters to get the test strip sales. Now, I have to work a bit to convince them.

      I don’t feel too guilty about convincing them to send me a meter. After all, the test strips are $0.78 per strip. At 8-12/day, I’m giving them about $2,847 a year in business. They can afford not only to give me one $19 meter to get that business, they can give me two so that I have a backup when I travel to remote locations in developing nations with no easy way to replace such items.

      Seriously though, at almost $3K/yr, I think there would be incentive to make a more accurate meter if they believed that would get my business. Hell, I’d even pay a buck a strip for that, which would increase their take by 25%.

      I really don’t think it’s so much an issue of cost as it is that they’re all blindly following some silly industry standard and are standing by the useless statement that their meter is no worse than anyone else’s. The problem is that it’s also no better.

      Why spend the money on product differentiation by features when what quite a lot of us really want is a more accurate meter.

      How often do I use the 14 day and longer averages? Rarely.

      When I do, do I modify my treatment as a result? No.

      How about the alarm clocks? I never ever use them.

      More accuracy at the expense of some of these useless systems? I’d switch meters in a heartbeat.

      As it is, the main reason I picked this meter above others is not even one they advertise. Mine works at 40 degrees F. That means I don’t have to warm it up for half an hour after it comes out of the inside pocket of my ski jacket. And, I can use it when hiking or canoeing in cooler weather in spring and fall. They don’t even advertise that.

      Imagine how quickly people would switch to a meter that gave readings within 5 or 10 percent instead of 20, even if it cost more.

  8. I bought a second meter because I thought that the current one was starting to fail. I have been running double testing on each once a day in the morning B4 breakfast [brunch!] If some one wants to do a statistical analysis on the results I will post them later.
    At this point I think that the meter is working fine, ad that my Pancreas/Liver may have changed. I will eventually take this to my family doctor for his review. He expect that he will possibly make an adjustment to the main prescription, but I want enough data for it to be credible and convincing.

    • Are both meters the same brand? Are the results typically close to each other?

      • Yes, and yes. Accu-Chek Aviva Nano

        Old Meter New Meter Date Time
        9.5/9.9 9.5/10.1 2011/09/15 09:15
        9.3/10.0 10.0/10.4 2011/09/14 09:45
        9.7/9.9 10.0/9.7 2011/09/13 08:30
        9.9/10.6 10.1/10.3 2011/09/12 08:00
        9.9/9.8 9.4/9.9 2011/09/11 08:15
        9.7/10.4 9.8/10.2 2011/09/10 08:15
        9.1/9.3 9.8/9.7 2011/09/09 10:00
        9.8/10.1 10.0/10.2 2011/09/08 09:15
        10.1/10.5 10.1/10.0 2011/09/07 09;15
        10.2/10.8 10.3/11.0 2011/09/06 09;15
        10.1/10.5 10.8/10.9 2011/09/05 08:30
        9.2/9.7 9.0/8.8 2011/09/04 10:45
        10.5/10.2 10.4/10.3 2011/09/03 09:30
        10.2/10.3 10.0/10.3 2011/09/02 08:15
        10.1/10.0 2011/09/01 09:30
        8.8/9.3 9.8/9.2 2011/08/31 11:00
        10.0/10.0 10.4/10.2 2011/08/30 08:30
        9.8/9.4 9.2/9.7 2011/08/29 80:00
        9.7/9.9 9.9/9.8 2011/08/28 09:30
        bought 2nd meter, began dual testing
        9.9/10.6/10.7 2011/08/27 09:30
        9.4/9.6 2011/08/26 07:45
        10.3/10.9/10.0 2011/08/25 09:45

        [in early July, ranged in low 8.0’s to mid 9.0’s]

        -daily testing is after being awake for 1/2 to 2 1/2 hours prior to having 1st coffee and morning pills, and any food.
        Blood pressures all below 140/79 @ 61 and weight stable @ 258# + – 1

  9. So much for making a neatly spaced data chart! Word Press compressed each line into that “lump”!

    • It came out OK in my email. I can make something of it. I just don’t understand the data. Why does each reading look like two numbers with a slash between them? When I do a reading I get one number (per meter when doing side by side by side testing).

      • I made it into an HTML table. It’s a bit more readable now, but wordpress still has these very narrow columns and your comment is already indented. So, not much room for a table. I’m calling it good enough.

        So, what are the numbers separated by slashes? Were you doing multiple readings on a single meter and getting such different results?

  10. I had separate vertical columns for each meter and two readings on each meter, then date and time.
    Prior to acquiring the new meter there is a case or two of triple readings on the old meter!
    ie, 9.9/10.1/9.7 as an example.
    All of these were spread across the page in neat vertical columns, but the Word Press software removed the spacing and jammed each line to the left in a compact mess!

  11. The only error in the re-posted chart is for 2011/09/01 where both readings are on the old meter. I was interrupted by a phone call that morning if I recall correctly.
    I know that these readings may be higher than the medical lab. results. I attended a quarterly physical for the drug company on 2011/07/27 and found out at the special visit a month later that the lab result was 8.4 versus my morning readings of 9.0 & 9.2 some 3 1/2 to 4 hours earlier. In the mean time I had my normal breakfast then drove the 38.7 miles to Toronto General, then get to the 5th floor testing rooms. Hence that 8.4 is not comparable a morning fasting. When the next fasting visit is scheduled, I will try to remember to take my own testing meter and do my own to get a comparison to their values. They take 2 readings, one before and one after I take the Ranirestat pills to measure the blood levels of the drug 1/2 hour after ingestion.

    • Corrected.

      Interesting test, before and after. Sounds similar, but for different purposes, to a glucose tolerance test.

      For meter accuracy testing whether the reading is before or after meals or medication shouldn’t matter as long as the test is performed within some very short time (call it epsilon) of the blood being drawn for lab work.

  12. swellendam bnb…

    […]Glucose Meter Accuracy: (Please Vote) Is 20% close enough for you? « The Misanthropic Principle[…]…

  13. bill says:

    consumer reports just did a revied in nov 2011. All the walmart relions rank good and accurate but they didn’t do the target meters! I guess I will be getting the reli-ons. the strips for the walmarts and the targets are about .35 each here in baltimore. MUCH cheaper than the other strips.

    other monitors at $1.00 per strip is just shameful. I’ve read the strips costs pennies to make.

  14. bill says:

    sorry. they did a review in nov 2011.

    and the strips are .35 a piece if you buy 100 of them. I have both meters relion and upandup. The meters are cheap it is the strips that are expensive!

  15. Thanks for the info bill.

  16. sophia says:

    I belive BIONIME SMBG System can provide you accurate reading and precise reading .
    There are number of Independent studies have been carried out, BIONIME not only pass the min ISO 15197: 2003 it is surpass the min requirement and it perfert especially at the low glucose concentration range. ( +-5 mg/DL) can effective detected by BIONIME GLUCOSE METER.

    • Rodnikov Magilovich says:

      Update, I have completed the Easai Trial of Ranirestat, but they are still running others through the program, so have not yet gotten any results. That will be due later this year, I believe.
      Regards blood glucose testing accuracy, at the higher levels, I have begun to suspect that there is an amount of inhomogeniety in the blood, if that is word! My levels have receded to the 8.5 + or – 0.5, so am a little less concerned. Have also now been prescribed a new combination pill called Jenuvamet which is Jenuvia combined with Metformin.
      Easai is a Japanese company with labs in Massachusetts, and production mfg in New Jersey.

  17. Thanks Sophia. I had not heard of that brand before. I will look into it.

    As an update from my side, I have also learned that different labs test blood sugar differently. I think the one my doctor uses consistently shows lower than my meter and have come to suspect the lab more than my current meter, a freestyle lite. Tests across older freestyle flash and newer freestyle lite meters and across the identical meters gives me much more consistent results than any results with my doctor’s office.

    Since my doctor’s tests always show significantly lower than my meter and I have not required paramedic assistance or gotten very low A1C readings using my meters, I am actually suspicious of his lab. I used a different one once and my meter matched more closely. Due to health insurance, I will now be using that other lab every time. I will try to make sure to test my blood glucose very close to the time of having my blood drawn and will compare to that lab results. I’ll post more when I know more.

  18. Rodnikov Magilovich says:

    I too, have seen the doctor’s lab results consistently lower than my meter and I purposely take my reading as close to their sampling time. No idea which lab his clinic uses, but we have a variety of collections sites which all of them are covered by our Ontario Health Insurance Plan if the doctor issues a requisition!

  19. Rodnikov Magilovich says:

    Two update comments:
    -have finished the drug test trial months ago, and there has not been any follow up or reports to me about whether the trial was a success or not! One of the Doctors @ Toronto General when I phoned to ask told me that we may never hear how it went!
    That is just not good enough! [it makes me suspect the drug Ranerestat? is a failure or not effective enough to gain FDA approval]
    -one of the Pharmacy Tech.s told me that [she too is a diabetic and tests herself regularly] just the the way you squeeze / wheedle out the blood drop can change the results from one strip to the next! It is important that if you are double or triple testing, that you test all from the same drop, as different droplets can have different amounts of energy extracted from them just by how much pressure is used to squirt them out onto the skin surface!
    In the light of that comment, then is + or – 20% too tight? And I assume that the thickness of the skin also would have an influence in this case! [avoid callouses, eh!]

  20. Just for the record, I’ve recently started going to a separate lab rather than leaving my blood samples at the doctor’s office. Perhaps, as my doctor suggested, because the blood doesn’t sit for as long before being tested, my meter has been in very close agreement with the lab test on my blood sugar.

    At my last A1C test (where my blood glucose is always also tested), my meter gave a reading of 71 mg/dL. The lab said 72.

    I am more than satisfied with that. It’s not always that close. But, it has been a lot closer lately.

    For any who are interested, my meter is a Freestyle Lite.

    What I have liked for a long time about this meter is:

    1. It works in a wide temperature range. This means, even when skiing, I can pull it out of my pocket in the lodge and do a reading immediately. I have had meters where I’d need to take out the battery and warm it in my hands for a while before I could even turn the meter on.

    2. It is consistent. If I take two readings back to back, or even two readings with different Freestyle Lite meters back to back, the reading comes out almost identical.

    What I was having trouble with was a disparity between lab results and meter results. Now I believe that discrepancy was caused by the delay between leaving the blood at the doctor’s office and the lab doing the test on it.

    So, for those who are concerned about accuracy, I would recommend this brand. I have been using this company’s meters since the Exactech many years ago and have usually been happy with it.

    I’m still not happy that the documented results are only within 20%.

    But, since my actual results are now much closer than that, I’m fine.

  21. ECA says:

    I posted here Early this morning, and its not here??

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: