Home Meter Accuracy

(Samuel Ashford) #1

Sano’s CGM post got my attention. Lately I’ve been experiencing elevation in BG and am starting an extended fast this morning. Tuesday or Wednesday I have a CGM (Abbott Freestyle Libre) coming in the mail. I will use the two week sensor and monitor.

My concerns in general are with instrumentation. My day job requires the use of instrumentation and gauges, and in my 12 years of experience dealing with measuring hydraulics, I have seen numbers all over the map, and have only seen extremely tight numbers with extremely expensive instruments.

Then there’s the additional complications of other variables occurring all over the map. And these variables are often never the focus. It’s only “OMG! My glucose is high!” or, “Yea! My BG is down this morning!” We’re pretty emotional people, I often find, and really look forward to celebrating the wins. But how accurate is the instrument? How much do we know about what’s going on in the bloodstream day in and out, hour to hour? What’s really happening? Can we tell this from a single BG reading at 5:30 am? Or just after fasting?

I have probably pricked my finger into the thousands, and have tons of data on spreadsheets to this regard. I’ve done many days on an hour to hour reading N of 1, to see what peaks and valleys happen, and on what fuels.

Here’s my concern. How can we really know that our meter is “accurate”? What is the gold standard? I’ve seen myself, for instance, that the TrueTest (and I guess, the True Metrix also) brands run considerably lower than my Bayer Contour Next EZ meter, at about 10-15 mg/dl lower. That’s important in terms of percentage. If my Contour Next is 110, then the TrueTest is 100. Which is “correct”? I’ve also seen some N=1s on the internet of diabetics measuring multiple meters, of different brands, with numbers all over the spectrum, to the point that it’s kinda dizzying.

So if I start having high numbers with my Bayer, I could just bury my head in the sand, grab the TrueTest, and truck along enjoying “lower” numbers for days. (laughter). A few days ago, I wanted to test the matter, and bought a second Bayer Meter. Two identical meters were off 10 mg/dl from one another. What to do? Do I buy a third?

It leaves me with the feeling of not much confidence in our meters. Or, perhaps to say it another way… at the end of the day, how do we really know what we know?

Some would say it doesn’t matter, that +/- of that range is insignificant. But it’s around 10%.

I suppose many would say that we should compare it to the “day of” while doing the blood draw at LabCorp. So, we’d say the lab with the centrifuge is the “standard”.

I’d love some feedback on this, because since we have no way of measuring insulin at home, with a fingerprick, the unfortunate surrogate is BG. Actually not a great surrogate at all, because one can have reasonable BG with high insulin. But nevertheless, it’s what we have.

So I guess I’m throwing out for discussion this whole topic of meter accuracy, and looking for some good ideas and feedback.



I’m also concerned about home meter accuracy and in the post below, I used my dual-use (Glucose and Ketones) Nova Max Plus to evaluate my response to liquid sucralose and the values were very inconsistent.

I recently ordered another dual-use meter, Precision Xtra with testing supplies, and waiting for it to be delivered so I can repeat my experiment and hopefully get more consistent results.

(Gloria) #3

I didn’t have time to do an exhaustive search of the https://www.fda.gov/ site, but I did find a “510(k) SUBSTANTIAL EQUIVALENCE DETERMINATION
DECISION SUMMARY” for the Nova Max Plus Blood Glucose and β-Ketone Monitoring System. http://www.accessdata.fda.gov/cdrh_docs/reviews/k091547.pdf

This compared the Nova Max to the Nova Max Blood Glucose Monitor and
Precision Xtra Advanced Diabetes Management System. There is precision and linearity in the latter half of the document. I’d have to do a deeper dive to see what acceptable differences are allowed. For a medical device to be sold in the US (whether made here or not) there would have to be an FDA filing. I don’t know the extent of the documentation.

The other thing to consider is that some of the variability may come from manufacturing of the test strips.

Happy hunting.


Do what the professional engineers do…use a gauge R & R study to determine correctness of measurements.

(Gloria) #5

I have been thinking about this and am not sure that one could control enough device and metabolic variables and collect enough data points to answer @Kairos question.

So I did some more hunting of my own. This is what I found.

A Review of Standards and Statistics Used to Describe Blood Glucose Monitor Performance

Standards That Specify Less than 100% of Data

International Organization for Standardization (ISO) 151974 and Clinical and Laboratory Standards Institute (CLSI) C30A5 have been written for SMBG and POC systems, respectively. C30A cites the ISO standard for acceptance criteria.

International Organization for Standardization 15197 states that for minimum acceptable accuracy,

“Ninety-five percent (95%) of the individual glucose results shall fall within ± 0.83 mmol/liter (15 mg/dl) of the results of the manufacturer’s measurement procedure at glucose concentrations ≤4.2 mmol/liter (75 mg/dl) and within ±20% at glucose concentrations >4.2 mmol/liter (75 mg/dl).” (The 95% limits are not confidence limits. They are percentiles. This means that the requirement is for the 95th percentile of the distribution of the differences to be less than the limit stated.)


@Fiorella, is the link below an accurate description of the study?

@kairos,is this something you can perform with your meter and report the results?

Anyone, can you outline a simple protocol directly targeted at testing home meters using this as a basis?


I will soon have access to 3 blood glucose meters, 2 of which will also measure blood ketones, and the only thing I can think of is just sticking a finger, measure, clean, squeeze more blood, measure again and repeat on the next finger while logging the results at T0, T30 and T60, but that doesn’t reference a known standard and even if I tried to do this within the timeframe of going to a lab and getting a serum glucose and comparing to my measurements from capillary blood, I don’t think that would be good enough given the delay between phlebotomy and eventual testing.

Is there enough confidence in the sugar solutions we previously used to check the home meters?


Yes, quality digest magazine has a good reputation in industry for publishing articles you can trust and use.

(Samuel Ashford) #8

Bill, thank you for the input.

As to your question, I would love to take this up - the R & R Study - and report back, in hopes it could benefit some of us. However, there’s a lot on my plate with my day job, and it will have to wait a little while.

Currently, I’m in an extended fast in order to resolve issues with elevating BG, and with elevating trigs. I am logging every little thing to see what affects the glucose, and that’s taking up a lot of time at the present.

(Angelina ) #9

My gadget is a reliable brand from Australia, measuring Ketones and Glucose. I prick my finger and sometimes use up to 6 glucose test strips and they all come back with different levels, same prick, time delay of probably 10 seconds and the difference ranges from 3.8-4.8, thats a bit much - any opinions here? I have another Glucose meter - doing exactly the same, which makes me question my results now! Anybody else having this experience?

(Samuel Ashford) #10


I have been pricking my finger for years. I have seen many, many fluctuations among many meters, and at many different times of day and phases of life.

I think the key words here are variables and trends.

What I mean is that there are many, many variables; meters (quality, age), seconds and minutes between testing, capillary vs. venous vs. interstitial fluid (location of draw), time of day, relation of draw to food intake, digestion and sleep, etc. And, there is the importance of trends. What is the timeline like? What does history reveal? What are the peaks and valleys like? Are you averaging between meters 4 - 5 mmol for fasting glucose, or is it higher, like 6 - 7?

I have witnessed huge swings in the most accurate retail meters (and I own several) over seconds and minutes. The current meter I trust the most is the Bayer Contour (and Contour EZ). The reasons are two:

  1. First, it was found to win out in a study done by the US NIH a while back (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3876374/).
  2. And second, when I recently had a blood draw at a local lab here in the states (LabCorp), the tech let me have some of the venous blood on my Bayer meter. The lab’s centrifuge results came back 1 mg/dL higher than my Bayer meter. Pretty close.

Currently, I’ve finished a 132 hour fast (pretty excited about the results; intend to post them soon on the forum). During the fast, I used an Abbott Freestyle Libre Flash Glucose Meter (an on-demand interstitial meter) and compared it to my Bayer. I saw all kinds of results. Sometimes the Bayer was 10 mg higher; sometimes the Libre was higher. Now, these are different locations of the glucose in the body. But suffice it to say that there was never a point of absolute certainty. I think that’s the chief lesson here. We will never find pinpoint precision because the movement of glucose in our bodies is always in FLUX. It’s a moving target. And personally, I don’t think it’s meant to be “hit.”

I think what you’re doing is fine. I would suggest, that as much as time and money can afford (based on your personal needs - are you T2 Diabetic? is it therapeutic or interventional in nature?), to experiment and tweak your meter use to your unique personal situation. I commend you and think you’re on the right track.

BTW - for any interested - I will gladly share a link to my Google Photos album of all my meter results. I would be very happy if it helps anyone.:smile:

(Jacquie) #11

Totally agree! Over the 2.5 years I’ve been testing (mainly blood glucose, rarely test ketones), I’ve collected three meters - Bayer Contour, Freestyle Precision Neo (the Precision Xtra in Canada), mainly to measure ketones and Accu-Chek Aviva, my present favourite for BG.

This discussion makes me think of Segal’s law. :slight_smile:
“A man with a watch knows what time it is. A man with two watches is never sure.”[1]

(Angelina ) #12

thanks for the study, very interesting
Quote: this sponsor–investigator study, conducted at a single site in the United States (Bayer HealthCare LLC, Diabetes Care, clinical trial facility in Mishawaka, IN)
just saying…it does make a difference to me, knowing where it was performed, not saying it was rigged, obviously

on the downside: I want an Abbott Freestyle Libre now ;)I guess 3 watches are better than 1 LOL
I think an ongoing reading would be a valuable INDICATION when you are fasting! Therefor would love to see your graph

BTW I’m a keto enthusiast, not pre or diabetic, managing my cancer with this kind of lifestyle.

(Michael Wallace Ellwood) #13

I was a bit like that when I bought a 2nd bathroom scale!

(George ) #14

Here are the results from a couple of my own tests using the same model blood glucose meters, strips from the same batch and same blood drop. ReliOn Ultima by Abbott (walmart) and Precision Xtra by Abbott.
I went on to order the MediSense control solution just like the instructions say to do.
the “expected results” range is 68 - 124 mg/dl.
It’s not very comforting to see such a wide range when testing for accuracy.

Blood glucose meter newbie
(Bacon for the Win) #15

apples to apples. Use the same one and don’t make yourself crazy over it.

(Samuel Ashford) #16

Well, this thread has gone way farther than I expected. Thanks for the images, George. Here’s an interesting one of my many images. It’s two Bayer Contour EZ meters. Tested with the same drop of blood. I love it. Notice, of course, the date and time stamp.

So much for the myth of certainty.

As NelleG said, don’t make yourself crazy. It’s a needle in a haystack.

(Bacon for the Win) #17

just to add another variable to this equation…

I tested with the strips hubs has with his meter. It was the last one so I went and got another vial. Sealed, in the package, expired two years ago. Numbers were almost identical. So why would I throw away who knows how many vials of test strips? We have a big box of supplies because the spouse uses an insulin pump. I’d really love to see if his pancreas is making any insulin at all, but that’s not for this thread.

I do have strips with current dates so I can continue to compare when I use an outdated vial. I bet they’re good for a long, long time.