I did do calibrations in the set posted here. But, all sensors startup with factory calibration. Calibration where the star is.
2nd 50 values have no calibrating. I didn’t post those. Not sure if there’s a difference in accuracy. Will need to look.
Confucius had sun dials! A quick internet search (which I should have done before posting the comment) doesn’t attribute the saying to Confucius, but references it as Segal’s Law. No idea who Segal was…Thanks for the feedback.
1 Like
@Bryan, thanks for sparking my interest. I read up on the history of timekeeping which I otherwise would not have done. I am not a good spark generator. I mostly rely on sparks provided by others.
1 Like
@mohe0001, before the first calibration the meter read higher. After the calibration the G6 read higher. This means that the high G6 bias is a result of the calibration. I think a more meaningful study would be one without any calibration.
Ok, thanks for the recommendation. I’ll start there. I have a lot of hope that calibration might help.
I’m not familiar with your description of “destroying the factory calibration” or even the concept that there’s some sanctity to it.
Can you elaborate, please?
If you don’t calibrate the G6, then you get a MARD that is on par with the best meters. Calibrating is a gamble. If your meter is more accurate than the factory-calibrated G6 and you are good at calibrating, then you end up ahead. If your meter is less accurate than the factory-calibrated G6 or you are not so good at calibrating, then you end up worse than staying with the factory calibration. By “destroying the factory calibration” I meant: Calibrating and thereby telling the G6 to forget about the factory calibration in favor of your input.
Dexcom G6.pdf (481.0 KB) Meters.pdf (422.7 KB)
I think I’ve noticed that - bad cali making things much worse, but sometimes it makes things better. Thank you for the clarification. Super helpful. My meter is not ultra accurate, so perhaps this is a problem. This really helps clarify how this works.
I was looking at meter accuracy for ReliOn Prime here: https://www.diabetestechnology.org/surveillance.shtml
Good to see a 2nd source, though, because mine is right on the boundary between being labeled ok and being labeled crap.
What about you, @YogaO? Are you a calibrator or a non-calibrator? I’m sure we all know HOW to cali correctly, but whether we do or not in the moment, might be up in the air. I’m still on the fence about it.
Thanks for the explanation, but I think you’re making an unfounded leap with regard to “factory calibration.”
Dex uses an algorithm to determine the BG value based on what it’s reading in my interstitial fluid. If I need to calibrate the sensor because of a large deviance from a fingerstick, I’m not destroying the algorithm, I’m providing more information so the algorithm can give me more accurate readings based on my interstitial fluid.
I would posit that this is verified by the fairly rare times I have to calibrate and the subsequent many days of accuracy afterwards.
Now, can a more accurate meter help me make a better call on calibration? Certainly, and a more accurate meter gives me a high level of confidence in the results. However, even the most accurate meter still can produce a wildly inaccurate result, right?
@mohe0001 - I rarely calibrate, and these days I rarely ever do a fingerstick. It’s to the point that I’m considering stopping my monthly strip subscription. I mainly do fingersticks when I’m seeing results that don’t match my expectation for the moment. For example, I had to search for my meter tonight because my BG was on a plateau between 140-160 for hours, and I thought it should have dropped. The fingerstick was off by 2 mg/dl, so no calibration there.
I always calibrate if the results are beyond the 20/20 rule, but otherwise I leave the G6 alone. I think I may have done my last calibration about a month ago.
1 Like
I may have missed someone else making this point, but I only calibrate if I know my blood sugar is not changing rapidly. It seems to me the value at my CGM could be a little different due to different location in the circulation system than the value in my finger. It could also be different due to the delay between measuring and calibrating if blood sugar is changing fast.
I always calibrate the first or second day of a new sensor, I always check it about 5 days into it and usually it’s okay. Then I recheck when I restart one or if a reading feels off. Sometimes I have to calibrate it again, sometimes not.
I don’t know if I’ve ever tried to ride it out the first 24 hours to see if it gets accurate later. The calibrating works well for me.
@YogaO, I agree with you that calibrating with a more accurate value does good only.
It is possible that there are ways to ‘bypass’ the calibration algorithm because I have seen it as an option (maybe on the G6 Dex receiver???). I have wondered if doing that might improve accuracy for me. If you don’t enter the code written on the paper covering the adhesive, then it reverts to something like G5 or G4 behavior where it REQUIRES manual calibration periodically. I keep meaning to try that.
Whats the 20/20 rule? 20% off the manual stick? Values within 20% have been rare for me.
I think I may have missed the boat on all these conversations since I was such a late upgrade from G4. Thanks, you guys, for walking through this again. Your probably sick of sensor accuracy conversations. They have probably been going on for years without me keeping up.
There’s an algorithm SW and there’s calibrating the sensor, but there’s not a calibration algorithm. The G6 algorithm won’t “revert” to the G4/5. Omitting the code will force you to calibrate every day, but it won’t change the underlying SW.
The 20/20 rule is the same that you had with the G4. Only calibrate if your fingerstick meter reading is 20% higher or lower than the CGM value.
For ex: fingerstick = 150, CGM > 180, then calibrate.
If your fingerstick is under 70 mg/dl, only calibrate if the value is more than 20 points.
For ex: fingerstick = 65, CGM < 45, then calibrate.
If I have a lot of values that are requiring me to calibrate, then that tells me I either have a bad sensor, or I have placed a sensor in a bad location. Regardless, I’m going to call Dexcom and get a sensor replacement.
I also found that I had to move away from the areas where I placed my G4 sensors in order to get consistently good G6 results. The G6 SW is more sensitive, I had to let my G4 areas rest and heal.
My 50,000 foot view of the G6 is that when it’s bad, it’s really bad, and when it’s good, it’s really good. There’s no middle.
3 Likes
I see.
I remember the occasional complaint like this re: G4. I’ll keep trying.
1 Like
I hear a lot of disparate information from various sources on when and how often you can/should calibrate the G6 along with whether or not calibrating too frequently will ‘destroy’ subsequent accuracy. So I Talked with one the Dexcom Patient Care nurses/CDE’s for about 20 minutes. Some of the things she told me -
-
20/20 is indeed their guidance on what should trigger calibration. (I’ve read 30/30 in some places - including TuDiabetes). This is based on G6 clinical trial results showing that G6 correlates with whole blood measurements within this band.
-
It doesn’t hurt to calibrate if your difference is less than this. Don’t expect calibrating with very small differences will do much good at all.
-
They recommend using Contour meters (I don’t really understand why). Following their guidance of washing your hands prior to testing, not using alternate sites and entering the data promptly is very important (I’ve been doing this since I got my G6 and was very surprised how much difference washing my hands made)
-
They do not recommend calibrating more than 3x per day. (Speaking for myself, If I had to do this above the 20/20 limits, I would ditch the sensor!)
-
Their general experience is a single 20/20 calibration should align things for several days or even the balance of the session (that’s my experience)
There is quite a bit of good info near the end of G6 manual based on their clinical trials, including accuracy, sensor life, pain, etc. broken down by adult and pediatric patients. Very interesting!
The difference between their Patient Care support and tech support is really striking on anything complex. I had to wait quite some time on hold, but worth it for things like this!
7 Likes
You were right, @YogaO. 20/20 rule is key.
No one had ever mentioned that to me before.
I called the Dexcom nurses. Huge breakthrough.
Thanks for talking through this, everybody.
I dont understand why Doc didn’t mention this.
Must not know the rule.
1 Like
FWIW, 30/30 was the initial guidance for the G6, but I think somebody @ Dex decided that was actually unrealistic from a practical perspective, so they reverted back to the 20/20 rule.
I might note that in the current FAQs on the G6, calibration for “out of range” readings doesn’t exist, period.
@mohe0001 - I will reiterate that your doc is either giving you incorrect guidance, or you are talking past each other.
Just one example: If the Dex is more accurate than fingersticks and should always be believed over them, why does Dex specifically direct users to perform a fingerstick if the CGM results don’t match what you’re feeling, and then calibrate if necessary?
Or this: If the Dex is more accurate than fingersticks, why doesn’t Dex mandate that the code be used in every instance, and that if you lose the code, that Dex will replace the sensor, because it’s so damned accurate compared to a fingerstick. (Ok, that was TWO examples!)
IOW, what a crock!
The nurse specifically addressed my question on the phone - is error based on the sensor or the manual stick as the gold standard? As has been the rule throughout all of human history - the manual is the standard for the sensor. Beyond that, lab is the gold standard for the manual. We might think of it like this:
Lab value: Gold standard
Manual Stick: Silver standard
Sensor value: Bronze standard
After reading all 57 comments (as of 12/17/19), I’m curious about the age/experience of folk’s endocrinologists relative to their focus on A1c results vs. CGM readings for “time in range”.
My 40-something endo likes graphs but also listens to report of whacko CGM readings, and ultimately judges my history by my A1c and general state of health. By her report, most of her patients are noncompliant “and take so much time”, so she likes easy. (Unfortunately, when things are easy, she starts looking for zebras, but that’s another tale!)
By the same token, my late sister’s compliance could only be “read” by her A1c because CGMs weren’t widely available and she couldn’t do meter readings because of severe scleroderma in her hands. Her 70-something doctor coaxed and coddled her into compliance and was as delighted when she hit 7 as she was! (She was mildly developmentally disabled.)
YDMV, and doctors will vary, too!
4 Likes