Freestyle Libre tolerance

[Radders]

I’ve been a Freestyle Libre user for a few months and have been getting low readings on the last three sensors. Realising that they are not measuring the same thing as a meter I have tolerated this but when my latest sensor told me I was at 5 when I was actually at 10 on the meter I realised that there may be something wrong.
I phoned Abbott and gave them a list of discrepancies between the two readings and they said they will send a new sensor.
I asked what the tolerance for error is and they told me 15%, so now I am kicking myself for having persevered with the last two! I did ask if they could compensate me for those but the best they would offer was free postage. I get that anyway thanks to the voucher codes!
Just posting this in case anyone else is putting up with inaccurate readings as I did.

 

[Northerner]

My sensors (when they have worked!) have all shown around 1 mmol/l below fingerpricks, which is fine for me because it means that I am running my levels slightly higher than when I was just using fingerpricks - well, it makes sense to me! 🙂

 

[everydayupsanddowns]

Yes I think it's worth contacting Abbott where sensors are consistently reading significantly different to BGs. Shame tey couldn;t do anything about the other 2 sensors

However... there are caveats to that because, of course, Libre is not reading blood and has an algorithm to try to reduce the 10-15 minute 'gap' in response of interstitial fluid to BG values down to nearer 5 mins. If your BG is changing, results between fingerstick and Libre are much more likely to vary (but may be much more similar to what BG would have been reading 5-10 mins ago.

There are also inbuilt inaccuracies with meters, of course - so there's no real way of knowing which of the values might be closer to 'reality' without double-checking everything on several different meters

Abbott expect the Libre to have a MARD (mean absolute relative difference) of something like 10-12% from lab values. When I've looked at some of my own data a reader can easily read 2+mmol/L out from fingerstick a few times and still fall into that MARD range if lots of the results are much closer (as mine always are). Like Alan, I would generally expect a sensor to read around 1mmol/L different or less most of the time. With bigger differences likely at higher values, and much bigger differences likely when BG is moving rapidly.

I'm sure you know all of this already @Radders - it's more for people reading who haven't come across the Libre yet.