Libre inaccuracies

[barrym]

This is a bit worrying.

Yesterday I felt a bit low just before lunch so did a quick scan with the app and it showed I wasn't low, 5.7. I was suspicious so tested with my AccuCheck Mobile and that showed 7.8!

I was a bit busy yesterday with visitors so apart from feeling worried over what confidence I could have in the readings I didn't do any further tests.

Today I've done three tests using the app, reader and meter.

In every case the app and reader have been identical.

In every case the meter has read higher, two out of the three significantly so.

Meter Reader App
8.2 7.2 7.2
6.7 6.4 6.4
6.7 5.6 5.6

I suppose reading lower is a 'safer' option than the other way around, but taking the 8.2 v 7.2, that's the difference to me taking a correction dose or not.

I spoke with Abbott and they're sending some strips for their reader so I can do three comparative tests using their kit. Fair enough. But this particular sensor is already a replacement for one which missed a low (discussed in a previous post). I went low quickly, took remedial action, and got it back up. I captured the low on the meter, but it never appeared on the Libre graph. So that was the Libre reading higher OR as I surmised at the time, the testing algorithm smoothing out erroneous readings and so fast changes/corrections like that were not getting included. Just my theory.

All in all I'm a bit concerned. Confidence definitely dented.

 

[Deleted Account]

I would be very surprised if the app and the reader showed different readings - they are scanning the Libre sensor for the latest stored value, not actually taking a reading.

As for the different between the Libre and your finger prick, this was something I experienced when I trialled the Libre and did not bother pursuing the Libre. However, it has become very popular which suggests either
- there is something in my physiology which means my arms give a significantly different reading
- my BG is constantly changing during the period I scan. This may be due to exercise or constant snacking and, as a result, the delay with the Libre was too far out.
- I was unlucky with the two sensors I had and they were wrongly calibrated
- I was paying too much attention to the value and not concentrating on the arrows which seems to be what other spend more time looking at
- I am less trusting of the Libre and perhaps it was my finger prick that was less accurate.

 

[Bluetit1802]

Out of my 9 sensors, 8 have been lower and one has been higher than my AC Mobile.
I don't inject insulin, so it doesn't matter for me.

At the beginning of a sensor I do several comparison checks (remembering that the sensor runs about 15 minutes BEHIND the finger prick blood.) Although reading lower/higher, I find each sensor to have been consistent in how much higher or lower it reads, so I calculate the average difference and do some mental adjustments when I scan. I always finger prick first, wait 10 to 15 minutes, then scan.

The upward and downward arrow should tell you which direction you are heading in - rising or dropping and how quickly. That is the thing to watch when you scan.

 

[Diabeticliberty]

I don't wish to alarm you but the Libre rarely if ever reads accurately compared to finger prick testing. If you want a real highlight of this take a look at the predicted a1c reading on the first page of the Libre reports generated by the software download. My last actual HbA1c was 46. My Libre predicted 34. A lot of the Uber Libre Club suggest that there is a time lag factor between a finger prick reading and a Libre sensor reading. Some suggest 10 minutes some suggest 15 minutes. I ran a series of control tests a couple of weeks after adopting the system and found that, at least in my case time lag is just nonsense since the Libre and my AccuChek Mobile results actually fluctuate higher and lower than each other. There is no constant margin of discrepancy just wandering results. There is no time factor involved. The results rarely if ever meet up and certainly not after a 10 or 15 minute time lag. Sorry but there it is and best you know really.

 

[donnellysdogs]

If you ever take the time to complain to abbott, they dont go on readings at the time of your scan. They want blood readings 15 mins after the scan.

The sensors are not reading the same "matter" so there is bound to be a difference.

They are useful for indications with graphs more than relying upon the actual readings...

They are an indicator only...

 

[Mbaker]

This is a recurring theme on many threads, it might be that the device needs a means of adjustment to the individual. Last week I was having readings in the 2's, but knew this was not accurate. 2 days ago I had to literally pull the sensor off early as I was getting a burning sensation where the sensor was located, to the point of severe discomfort - I think I will not get another of this type.

 

[Diabeticliberty]

With respect a blood glucose reading taken from whichever 'matter' you choose to take it from is not 'bound to be different'. If it then what is the point of testing? Alternatively I will stop buying sensors and start relying on 'Basil' my psychic tortoise to tell me what my blood sugars are.
Perhaps they (Abbott) might be well advised not to go offering diabetics of all types predicted a1c results when their system is to be used as an 'indicator only'

I am sorry for the flippant element to my reply but accurate glucose results are important to me and to all of us. We should not just casually explain away inaccuracies as if they don't really matter all that much. Accurate results are what we are paying for. After 33 years of being type 1 I will not adjust my insulin or lifestyle based on what in essence is a very rough indicator. The 'Emperors New Clothes' element that comes with Libre never fails to bemuse me.

 

[barrym]

Thank you all for taking the time to reply. However, nothing there I didn't already know from reading previous posts. But I must say, in my free trial earlier in the year the sensor and my meter were close, and what I hadn't mentioned is that I tested this particular sensor when I fixed it and all three were identical!

Before I bought my first batch I carefully considered all the 'evidence' here and compared it with my trial experience and concluded that it worked for me. It's the inconsistency that I find unacceptable, between sensors and seemingly during the life of a sensor.

If this is as good as it gets then Abbott should say what the margin for error is.
As said above, this is a medical device and should perform to the standards of meters otherwise it is no use to me. As for arrows, I rarely see anything other than horizontal, despite what the graph shows.

 

[donnellysdogs]

No machines, even blood testers are 100% accurate-they are all allowed tolerances of error...

As a sidepoint, did you realise that you may not even get a litre of petrol from garage forecourts? Again they are allowed tolerances too and its best to fill up your car slowly??

Also our bodies can react quite differently under stress and for example... in hospital pre my mastectomy reading I had a cgm in. Accurate is superb with it. In hospital with stress it was completely inaccurate to my bloods.
After op I put a new one in next day and perfect again. My DSN even came upstairs to see me to say "see we got you safely to op" -trouble was, the cgm readings she was seeing were inaccurate and when she arrived I was going iow (my biggest fear)....

If unhappy with accuracy of libre do all the things they require to complain and if you are not happy with their reaction you can always report to MHRA as problems with medical equipment.

Unless you take time to report things to MHRA then unhappiness from people will always continue...

 

[Diabeticliberty]

No machines, even blood testers are 100% accurate-they are all allowed tolerances of error...

As a sidepoint, did you realise that you may not even get a litre of petrol from garage forecourts? Again they are allowed tolerances too and its best to fill up your car slowly??

Also our bodies can react quite differently under stress and for example... in hospital pre my mastectomy reading I had a cgm in. Accurate is superb with it. In hospital with stress it was completely inaccurate to my bloods.
After op I put a new one in next day and perfect again. My DSN even came upstairs to see me to say "see we got you safely to op" -trouble was, the cgm readings she was seeing were inaccurate and when she arrived I was going iow (my biggest fear)....

If unhappy with accuracy of libre do all the things they require to complain and if you are not happy with their reaction you can always report to MHRA as problems with medical equipment.

Unless you take time to report things to MHRA then unhappiness from people will always continue...

In all fairness it is not the inaccuracy that causes me concern. It is the complete inconsistency with which the Libre performs. If there was a margin of error which was constant them I would be far more happy than the constant wandering that my own Libre constantly indulges in. I used to believe that it was just me as nobody else much seemed too bothered. I then met up with some other users face to face and discussed with them their various experiences. Theirs were not wildly different to my own. In fact they were alarmingly similar. Even the most ardent Libre die hards who for reasons which I still cannot reason will never hear a word said against their precious luvvy system go a little bit 'hear no, see no' when asked about the predicted A1c from the Libre and their actual HbA1c from their previous diabetic review. My own rantings and ravings however ring quite hollow as I have used a Libre for nearly 2 years now constantly. I kind of defeat my own objections. 31 years of hammering my own fingertips 15 tines per day with lancets had rendered my left hand into something like Freddy Kruger. At least now I have my beloved fingertips back.

 

[donnellysdogs]

Yep, n when you've had a double mastectomy ad lymph node removals and nobody can give you a true idea about whether having a libre stuck in arm could cause lymphodema or not, its another worry.......

I was pleased with my usage of libre before cgm funded. I had not had good time with Abbotts Freestyle Navigator years ago so I was concerned... but as a guide it was useful tool.

Have CGM now and wear in bum as nobody had a clue about the lymphodema risk from a sensor stuck in arm for two weeks at a time...

Take your points on board but people should report.. even if the appluances are self funded to the MHRA website....nothing ever happens to improve things unless things are reported properly.

 

[Bluetit1802]

Yep, n when you've had a double mastectomy ad lymph node removals and nobody can give you a true idea about whether having a libre stuck in arm could cause lymphodema or not, its another worry.......

The instructions I got were no needles and not even a blood pressure cuff. I was given a rubber wrist band with these instructions on, in case of accidents and paramedics. However, I soon relaxed about it and have needles, a libre, and blood pressure cuffs as often as necessary and no lymphodema, 4 years on.

 

[donnellysdogs]

My Prof bc and diabetes chap says as lymph nodes were only samples (not full clearance) and clear then they thought bp and tests could be taken from arms but no research data on finger pricks or cgms but they thought low risk!!
(Not their risk though is it??).

 

[Bluetit1802]

My Prof bc and diabetes chap says as lymph nodes were only samples (not full clearance) and clear then they thought bp and tests could be taken from arms but no research data on finger pricks or cgms but they thought low risk!!
(Not their risk though is it??).

The only risk is if any of these things create an infection that the remaining lymph nodes can't cope with.

 

[barrym]

This is getting silly. Just took a scan with the app, a normal post prandial one and got 2.9! I felt fine. Used my Accuchek Mobile and got 4.3, used a strip in the scanner and got 4.4. That felt about right.

I'll call Abbott tomorrow, and I know they'll just offer a replacement. It makes you wonder if these are ever going to work. I have been a huge fan mainly from my free trial, didn't do much double checking. Now I'm using it for real I'm not so impressed.

 

[Fleegle]

It is really difficult isn't it. I find the libre reliable if inaccurate - so like @Bluetit1802 I calibrate on day 1 & 2 and then check randomly. Ok so I don't get away from finger pricks - but I drastically reduce them. I have had some sensors that are spot on - literally finger prick 5.8 - within 10mins scanner 5.8. And sometimes .4 below. When I have had 2's when I am really a 5 I report it. Abbott are getting fed up with me - they have sent me a new scanner after my last complaint and I sensed we were nearing the point where they will no longer replace mine. I have sent about 4 back from a total 10 bought.

The latest problem I have found on the last two sensors is the sedimentary factor. If I sit still watching tele after 20mins my BG drops - if I just shake my are it goes up .4mmol in 2mins. When I do the test - it is about .4 out. Reported to abbott...

 

[barrym]

If it wasn't for the awesome convenience of being able to test wherever, whenever you want, I'd dump it and just do more finger pricking. But I've got in the habit of testing at night. If/when I wake for a pee I test and take appropriate action. I'm out at a rugby match on Saturday, and testing will be so convenient. But what cost?

How have you found upper end inaccuracies? I haven't really checked. But if a 8 or 9 is really 11 or 12 I would be very unhappy.

 

[Fleegle]

I found once calibrated it is linear.

 

[barrym]

I found once calibrated it is linear.

Thanks for that. I shall do some more tests to see if I get the same.

Some of the emerging techs like sugarbeet do a formal calibration with every new sensor. Thought that was ott originally but maybe s good idea.

 

[cott97]

If it wasn't for the awesome convenience of being able to test wherever, whenever you want, I'd dump it and just do more finger pricking. But I've got in the habit of testing at night. If/when I wake for a pee I test and take appropriate action. I'm out at a rugby match on Saturday, and testing will be so convenient. But what cost?

How have you found upper end inaccuracies? I haven't really checked. But if a 8 or 9 is really 11 or 12 I would be very unhappy.

For me I've found it was relatively accurate when I was high but now I am lower it doesn't do the higher end well ie a 6.9 on sensor is often 8.5 - 9.5 and when the libre says 4 it's usually 5.1 so in theory out of tolerance. But if sensor says 6.9 the meter usually agrees...