I’ve been at this for so long. I’m loosing my mind. We are all Tod if you don’t trust your CGM reading to do a finger stick. Sometimes the finger stick matches and sometimes close. But what concerns me is a major difference andd on top of that my main concern with this writing is the multiple difference I get doing finger sticks. I now have to finger stisck 3-4 times and average the reading even though the blood is coming from the same finger. Any advice
Are you using a good meter like a contour one?
I agree with CJ. Start with a good meter like the Contour. Make sure your hands are clean; wash with soap and warm water. Do a second finger-prick if your gut tells you to – I don’t often do this with my trusty Contour Next One.
I rate the finger stick over the CGM reading.
If your glucose variability is low, any food or insulin corrections will be small and low risk. It’s why I minimize total daily insulin and eat low carb.
The problem is that what we want for “close” is always smaller than what we need for close.
Use this to compare your BGM readings to your CGM. Parkes Error Grid | Desmos Its not until your results are landing in zone C that you may need an extra correction or carbs to correct for a dosing decision based on the CGM value. @Yve65 I wanted to share this with you too.
I also agree with CJ on meters and with Terry about rating the BGM over the CGM. I was talking to someone else about meters and came across a Onetouch study they did for the EU. It was a “do the production meters and strips perform in the real world as good as they did in the clinical trial we did to get the meter approved” study. The study showed the meter met the specs but boy howdy was there a lot more variability than in the original clinical trial. So expect 3 fingersticks to be different.
3 Likes
Wash your hands before you test.
Here’s some of my data. Note that I am not doing more than one fingerstick here. But I find them to be pretty close with this machine.
The %Diff column shows the difference between sensor and manual blood glucose readings. A 20% difference is considered sensor failure.
I am using this machine and a Dexcom G6 sensor. This machine requires a script. It’s not OTC.
Let’s look at a different day.
How is the system behaving at noon? Is that a good time to evaluate sensor accuracy?
No, because we see a rapid decrease in BG at noon. See how flaky that sensor data becomes? It is unreliable. The solid circles are manual finger sticks that illustrate a more honest picture of blood glucose during that time interval.
You will always see the sensor become more inaccurate during times of rapid increases and decreases.
You will always see the sensor behave more inaccurately when you are low.
I am producing sensor error all over the place, but that is partially due to the fact that I haven’t adjusted my insulin doses in ages and they are totally wrong. I’m trying to make adjustments quickly in my insulin dosages, so I’m using manual finger sticks to do that, not sensor data.
The sensor data is a mess because I am roller coaster-ing all over the place. Once I make adjustments in my insulin dosing and my BG is more flatline, I can give more credence to sensor readings because a lot of that sensor error falls away naturally once I stop my BG from roller coaster-ing.
My problem in February is rollercoaster-ing from incorrect insulin dosages, not sensor error. If I fix the rollercoaster, then I will fix the sensor error.
By the end of the month, I have fixed that stuff. There’s really no need to look at all those details because system behavior has stabilized.
I can understand where you are coming from, but I don’t share your feelings of frustration. For me, everything is an approximation, and close enough is good enough. That said, I don’t have lax standards.
I build software systems and applications, and by design, they are without error. Still, other systems I deal with are asynchronous, meaning they can run independently and at different times. Occasionally, these differences can lead to inaccuracies, or something flipping back and forth because two systems have differing information. Separately, I work out most days, plus walk 7000+ steps daily. Do I get worried if the system records extra or fewer minutes, or if it miscounts my literal number of steps.
On another level, the absolute number does not matter as much as the trend. I have long used weight scales that incorporate bioimpedance for body fat measurement, and depending on one’s distribution of fat, gender, and fitness, the numbers will vary, but if one measures consistently we can measure the trend, which is more important from a management perspective.
We all have different tolerances for ambiguity, and mine might be a bit higher than most, at least in some domains, but I strive for conciseness in others. My feeling is that you might need to loosen up, if possible, and accept the variation.
3 Likes
Well said. I agree. Trends are important and living a life are important… to much fussing, for me, about numbers, ruins my day.
1 Like