Yeah, so I might get a reading of
Sensor 13.8
Libee Finger Prick 13
Glucometer 12
No idea which to take
Joe, when you're in the low teens, all of those numbers are basically the same.
If I recall correctly from earlier posts, you're an engineer, so you probably expect a degree of certainty in measurements.
Precision is a relative term when it comes to measuring bg. For various biological and electro-chemical reasons, there's a lot of "wooliness" in the subject, but, once you get to grips with it, it's perfectly possible to make rational decisions about what's going on.
In an ideal world, I'd like my bg to be around 5 as often as possible, but I know that, given the amount of variables involved, that's wishful thinking, so I'm ok with it floating around between 4 to 7.
What's that got to do with libre inaccuracies? What it means is that I've got three zones I pay attention to. 4 to 7 is fine, below 4 needs attention, so does above 8. I'm not interested in the actual numbers, just generally which ballpark I'm kinda in. It's not a case of saying I'm 5.5, more 5 ish.
Big differences in libre/meter tests are important at lower numbers because of hypo risk, and I'm not going to pretend that those don't happen with libre arround the 4 to 5 mark, but when they're happening around the mid-teens, I wouldn't expect the numbers to be the same, because I don't think the libre factory calibration is based on being accurate at such high numbers.
The numbers you've given in your example, between 12 to 13.8, I would be paying absolutely no attention at all to which one is correct, because all I need to know, if I was in that situation, is that they're generally around 13, therefore too high, and none of them is in the 4 to 7 zone, I'd just be saying all of them are much higher than I'd want, so I'd be thinking about a correction dose. The 1.8 difference wouldn't make a huge amount of difference to the amount of the correction dose - that can be tweaked with a few biscuits once back in range.
How calibration works plays a major part in this. None of these things, neither dexcom nor libre, will be stable over a wide range of bg. If you use dexcom, the general advice is to calibrate while stable. If you do that while in say a 4 to 7 range, it'll be calibrated for that range, but that really doesn't mean that that calibration is going to hold true if you're floating around in a completely different range in the mid-teens.
Libre is factory calibrated. It's meant to mean it'll be accurate all the time without doing the dexcom twice a day gig but we all know that that kinda doesn't work. I'm not a scientist, so I have no idea at all about Abbott's factory processes, but I suspect that they've calibrated based on users being around 4 to 8, because that's where T1s should normally be, and not being in rhe mid-teens, so it's no surprise that there's going to be some wild flyers the further out of range you are: if it's calibrated for 4 to 7, that calibration won't work for 10 to 15, so there will be inaccuracies.
We've talked on other threads about kitting libre out with blucon to make libre a cgm. Calibration will play a part in that. The sensor is sending a number to the app and the app then turns it into a bg reading. Calibrating is taking a bg reading and telling the app my bg is x right now so when you get a number from the sensor, that's what it means in terms of actual bg. If you do that when bg is flying around all over the place, it'll get confused, and it'll also get confused if you're telling it how it looks when you're at 12.8, compared to telling it when you're at 5.6. I suspect that libre is set up to read well when in a general 4 to 7 range, but not so much when in the teens.
I'm honestly not sure whether that's made anything clearer or more complicated, but all I know is that, after a bit of experience of using these things, you can take what it's telling you, make a few adjustments to account for the inherent uncertainties of them, and then say, aye, ok, I'm now going to do x,y,z based on it, without any serious risks to health.