Freestyle Libre sensor Serious inaccuracies- Help us improve this!

I was going to give up on the Libre until t was suggested I used the Glimp app.
This allows you to calibrate it against finger pricks and I found it much more accurate that the LibreLink app.

However, I do not use the readings on the Libre to dose from. The value of Libre (and any CGM) is not a finger prick replacement: it is the trends and history. These have enabled me to adjust my insulin dose and timing to get a much more consistent control with less peaks and troughs.

I use the Xdrip+ app and as @helensaramay has said you can calibrate the app with finger prick readings and it improves the accuracy of the Libre. I am nearly always withing 10% of my BG readings with the Libre anyway so its an added bonus

I too have found it to be pretty accurate. Sometimes it's a bit off, but I haven't seen any changes in HbA1c since using it.

I've found that sometimes they are more accurate than other times, but once I work out how far out it is it stays about the same throughout its 2 week life, my current one seems to show about 1.5-2mmol/l below what my BS is, but its consistently shows that, my previous two were good and were generally less than 0.5mmol/l out and it was consistent through its life.
In the time I've been using the libre (since June last year) my Hba1c has improved a lot, it went from mid 50's (and those mid 50's were I think by luck entirely) to 44 .

But this is nothing new.

Libre is at least 15 mins behind bloods and if your not sure about what Libre is telling you then finger prick all the time.

Using Glimp, xDrip or Spike helps because you can calibrate with finger pricks but even that is not fool proof.

Yes Libre is good for Trends and can certainly help manage things but I have had it where it is telling me 15.8 yet blood prick said 10.5 also had it where it has said LO for nearly a week but xDrip was telling me 4 or even 5 plus confirmed with finger pricks.

Libre is great tool but in some respects still in development and I also find that depending where in the body you stick it can also vary on how close it is what the bloods say, and another thing is, from reading around, the sensors themselves can be hit or miss.

Agree with what people have said above. I like to have the Libre to check trends, but I don't rely on it's own reported values...

I have been using the libre for nearly 10 months, there has been the odd sensor which gave me high readings, but as noted above it does have a delay on it so if you are doing a finger prick at same time you won’t get the same readings especially if you have not long eaten and your bloods are being affected but sensor hasn’t caught up.

However when I wasn’t on the libre and ways changing between different meters, I used to test on both to start with and the variation in readings was very high, but when you read the info they send you there is a difference in calibration so really none of these are 100% accurate.

For me I have put my confidence in the FreeStyle Libre and the benefits far outweigh the slight glitches.

For what it's worth, even the pinprick machines are quite inaccurate - I tried 4 different models and on the same puncture (pinprick), obtained a 25% variation in readings. So I agree that these machines are only really useful for trends and not for accurate measurements.

I have been using Libre for some years. It has helped me massively reduce HbA1c levels. Concurrence with BGM is normally good, when it is not, I check with another BGM. It is generally the BGM reading which is inaccurate, not the Libre. I am also a member of the FaceBook Abbott Freestyle Users Group. With nearly 20,000 members it naturally has many reports of Libre inaccuracy but with over 10,000 active members, is this surprising?
It is also noticeable that many of these reports come from “new” users.
When people report errors, we sometimes get many more saying that it is accurate for them - the normal silent majority whose accuracy is good.
I do have concerns with using BGM to “correct” Libre using 3rd party apps. While this may seem a good idea if Libre has a one sided bias, how do we know that it is not the BGM that is biased?
Some people do take great care to use one of the better reported BGM and take multiple fingerpricks to “calibrate” their app but how many do? Is not one of the main reasons behind the factory calibrated Libre to reduce fingerpricks, not to increase them in order to attempt to recalibrate with a 3rd party device?
One major source of Libre error is from tissue damage at insertion. It is clear from test results that accuracy is much lower on the first day after it is put in. Many people avoid this by attaching the sensor the day before it is needed. The day before it is activated. Pre-insertion.

TuviaDror said:

I have diabetes type I. Has HBA1c around 6.4% Was using the Freestyle lite for the several years or so to monitor my blood glucose.
Started using around 6 months ago the freestyle libre CBGM sensor
I discovered it reported BG levels lower than what they actually were, hence misleading me to not give bolus
The end result was my HBA1c got to 7.4%!!! Terrible :-(
Its all reported in the following two youtubes.
Pls view and act accordingly so that we can improve our monitoring. Its in our hands!

In order to help Abbott and the community understand the magnitude of the problem, I prepared two short YouTubes.
If you put in You tube Tuvia Kutscher you will find the two movies.
The links are

1. (the short movie)
2. (the long movie)

Click to expand...

Unfortunately this website allows people like you to make wild claims - how does know what you claim is accurate
Need say no more that you think the Libre measures BG - it doesn’t

To an extent, what he says about Libre underestimating hypo level (below 70mg or 4mmol) individual scans is correct.
Abbott do very clearly warn about this and say to confirm with blood tests when near or in hypo levels - which his charts confirm he is.
It is misleading for him to to say they don’t warn you of inaccuracies giving lower values when in or near the hypo levels. They do. Indeed the FDA documents specifically mention this warning.

But much better to be safer and warn of hypo levels when very low - and to tell you to confirm with a blood test - which is what Abbott do.
Most users say that their estimated eA1c levels are very close to their HbA1c blood test results - but most users are not running at HbA1c of 6.0% or 42. Perhaps, as an individual, he is better to use much more expensive equipment or to continue with many blood tests. Most of us wish to avoid that. Most of us find Libre averages very accurate. Which is why over 1,000,000 people are now regularly using Libre and many national health care organisations are recommending and supplying them.

etippelt said:

For what it's worth, even the pinprick machines are quite inaccurate - I tried 4 different models and on the same puncture (pinprick), obtained a 25% variation in readings. So I agree that these machines are only really useful for trends and not for accurate measurements.

Click to expand...

I agree. The finger prick blood test is known to have a variation of + or - 15% from the actual glucose level. This was confirmed by my GP and would not stand up as being sufficiently accurate in a scientific test- that is within the acceptable range of experimental error. How are we supposed to manage our condition with this sort of nonsense. I am hoping to go to the Libre as at least I will get a lot of readings to see when and figure out why the highs and lows occur.

Bill_St said:

It is also noticeable that many of these reports come from “new” users.

Click to expand...

Typical! Blame the user. This machine and it's sensors are widely innaccurate. I've been using it on and off for 2 years and I find it varies from as much as 5mmol out to near on perfect, but the accuracy, or lack of it, makes me quite angry. I suspect it was released on the unsuspecting diabetic public far too soon. And the meter I use is very accurate, before you mention that.

Bill_St said:

r 1,000,000 people are now regularly using Libre and many national health care organisations are recommending and supplying them.

Click to expand...

Hmmm at around £50 roughly per sensor that's a massive profit for Abbott. The more of us they can gull into thinking this machine is accurate the better!

The Libre documentation is pretty clear - that finger prick testing should never be abandoned in favour of relying on the Libre (which measures interstitial fluid, not blood glucose, and runs 10-15 mins behind blood glucose measurements, and is not advised as a way of confirming blood glucose before driving).

So discrepencies between bg and libre readings are inevitable and should be expected.

My experience is that each libre sensor runs a little high or low (usually low), gets more accurate after a couple of days and becomes less inaccurate at the end of the two week lifespan. Each one also tends to be pretty consistent, so it will either run approx 0.5mmol/l high, or 1.5mmol/l low, so I allow for this.

I think that many new users of the libre approach the gadget with unrealistic expectations, and insufficient research. Lets face it, anyone who expects the libre to replace prick testing and function in the same way as a glucometer fits both of those descriptions.

I also think that there are some people whose bodies, lifestyles, expectations and immune systems are simply incompatable with the libre. Their body rejects the sensor filament, or the glue. Just as some people are intolerant to certain foods, or find certain shoe styles unwearable, or have a tendency to get a suntan in conditions that other people don't. I'm sorry that those people don't suit the libre, and I get that they find it frustrating.

As an example, I get a skin rash with almost every type of skin adhesive except the libre. When I started using the libre it was with a clear expectation that I would probably need to discontinue its use, but thought it worth the risk for a trial run. Two years on, still no reaction. For which I a truly grateful. So while I am v sympathetic to anyone who does get skin reactions to the libre glue (because I know what those reactions are like), I really hope that they never change the formulation - because if they do I may find the new formula incompatible with my body and lose the use of the libre.

This is an interesting, technical article explaining how the algorithms in these type of sensors work: https://bionicwookiee.com/2018/11/1...i_MOrmxph0JeiwteF6i2B6lOh4uJKquJpw0EobsKoHv7A

Funny that the abbreviation "BG" is used a lot here, when we all know that's not what it's measuring. I think Abbott are perfectly clear and I find their literature completely non-ambiguous. Succinctly, its 15 mins behind, don't use it for high rates of BG change and always rely on a blood test.

Sure, it's sometimes out, but how do you know your BG meter is accurate - do you calibrate regularly? I don't and I don't reckon many other do either.

The Libre has been life-changing for me; I hardly every finger prick, avoid lots of hypos and highs, can easily test in the night, can test subtly. If you don't like it that much, don't use it and get a CGM instead.