Libre 2 - Phantom high and low

[pawprint91]

Sorry if there's been a thread about this in my absence that I've missed, do feel free to redirect me 🙂

Does anybody else have where they have the high/low alarm go off on libre, only for it to later not show on the graph? This has happened to me a few times recently, and often doesn't show when I've either corrected the high or had a sip of juice/coke for the low. I know the libre is approx 15 mins(?) behind a finger prick so don't know if this has anything to do with it? I have the iphone version of the app, don't know if this affects it either?

 

[Bruce Stephens]

Does anybody else have where they have the high/low alarm go off on libre, only for it to later not show on the graph?

Yes. Not very often but it does happen. The graph is (I think) produced from the 4 times an hour records, but the sensor reads once a minute (and sends that reading out, for the alarms), so I think it can happen that you get a reading that triggers an alarm but (so long as it's only for a short period) it might not appear on the graph.

We know there's also a predictive element to what the apps (and reader) give so maybe that has some effect.

 

[Tdm]

The libre redraws its graph if it thinks its 'got the answer wrong'. Manys the time i went out of range during a post food walk, only to see the peaks shrink once it had time to think about it.
The thing is doesn,'t change are any dots that show a scan.
I don't find dexcom does the same...i have a feeling dexcom does less preducting and says it as it is, which, being from yorkshire, i prefer

 

[rebrascora]

Yes, as I explained in my reply on the other thread, it is to do with the Libre algorithm using previous readings and extrapolating them to predict what your current reading will be, because it is trying to overcome the lag between BG and interstitial fluid glucose. This is fine when levels are fairly stable but can result in erroneous results if levels are changing at uneven rates, changing very fast or changing direction altogether and as @Tdm suggests, it redraws the graph to adjust to what is more likely correct, when it realizes the result it predicted was a bit "overegged" one way or another. ie too high or too low and if you will get a dot hanging above or below the graph line which just doesn't match. Using the reader I don't get dots when I scan but I notice that the graph lags behind real time by about 15 mins to give it time to correct before it plots the graph.
I sometimes wonder if just giving us the actual interstitial reading and allowing us to learn to live with the 15 min delay might be better than it trying to adjust for it and getting it wrong some of the time.