Comparing CGM systems: Dexcom's G6, G5, Libre and Medtrum's A6 compared

[tim2000s]

Thought people may find this interesting. It's a comparison of CGM systems in terms of accuracy and variance.

https://www.diabettech.com/cgm/the-cgm-royal-rumble-4-systems-face-off-one-will-win/

And if you want the abridged version, see below:

 

[Scott-C]

Well, it's interesting that libre put on a such a good show.

I've had a few sketchy sensors, but the recent ones over the last few months have been fairly solid, noticeably different to those from earlier in the year, to the point that reader scans are matching pretty closely to calibrated xdrip+.

I wonder whether they've made any changes to their calibration/manufacturing techniques.

 

[tim2000s]

Well, it's interesting that libre put on a such a good show.

I've had a few sketchy sensors, but the recent ones over the last few months have been fairly solid, noticeably different to those from earlier in the year, to the point that reader scans are matching pretty closely to calibrated xdrip+.

I wonder whether they've made any changes to their calibration/manufacturing techniques.

They mentioned in the talk of the Libre 2 release that they are changing the calibration technique for the Libre 2 to be factory calibration per sensor instead of per lot. It wouldn't hurt them to do that with the Libre 1 and not announce it while they rolled it out across plants.,

 

[Scott-C]

They mentioned in the talk of the Libre 2 release that they are changing the calibration technique for the Libre 2 to be factory calibration per sensor instead of per lot. It wouldn't hurt them to do that with the Libre 1 and not announce it while they rolled it out across plants.,

Yes, that makes sense (although the little rascals can still be a bit flaky on the low side).

 

[tim2000s]

Yes, that makes sense (although the little rascals can still be a bit flaky on the low side).

Agreed - what's going to be interesting is running the same type of test with Libre using a MiaoMiao & Bluecon vs G6 with the standard xDrip algo then the LibreOOP algo, which is using the Abbott library from the LibreLink software.

 

[db89]

Agreed - what's going to be interesting is running the same type of test with Libre using a MiaoMiao & Bluecon vs G6 with the standard xDrip algo then the LibreOOP algo, which is using the Abbott library from the LibreLink software.

Are these algorithms documented anywhere for further reading? My Google-fu let me down when I started looking at the options the other week.

 

[tim2000s]

The xDrip algorithm is basically a linear calibration model with an offset in the first few days of use to counter trauma effects.

The OOP algo is the secret sauce in the Libre reader and LibreLink apps. At this stage, I don't think anyone's decoded the algo itself, just figured out how to use the library it's encased in.