39 Comments
If you use a CGM, on a scale of 1-5, how satisfied are you with the accuracy of your CGM?
You must be logged in to post a comment.
© 2023 T1D Exchange. All Rights Reserved.
Every once in a while I get a sensor that never stabilizes over time and is not ‘trustworthy’. This is probably less than 5% of the sensors inserted.
I often have to pair attention to my symptoms of hyperglycemia rather than the sensor readings. There have been times when I just fineshed a meal and I keep getting low blood sugar alerts. I then use my blood glucose machine and the results between the sensor and the blood glucose machine are not even close. It’s very frustrating. I have calibrated a few times but that is a long time consuming process as you have to do it often and within a certain time frame. Who has time for this?!
Use Dexcom G7. I checked satisfied. I think it’s pretty accurate but loses connectivity daily so not very satisfied.
Jane,
I saw my Endocrinologist last week and asked her about the G7. She said that it does not connect to the pump yet. (?)
I supposedly use the most accurate CGM, but it is routinely far beyond 15% different from a finger stick even when calibrations are advertised as not required.
Again, it’s worth noting that glucometers themselves are not all created equal. The required accuracy is like +/- 15 % of the reference value and not all of them meet even that standard:
https://www.diabetestechnology.org/surveillance.shtml
far too many times I have to verify my Libre2 with a fingerstick, very disappointing
On a scale of 1-5, I said, “Very Satisfied.” 🙂
Honestly, I don’t do finger sticks to check my CGM accuracy. About the only time I do finger sticks is the during the two hour warm up for a new sensor. I have trusted that my Dexcom G6 reading are accurate, and have been very pleased with the results. My recent A1c was 5.4. Things are going well with my Tandem X2 pump and Dexcom sensor system.
I’m sure this is something that is just one-of-those-things that varies from person to person. I am not thrilled with the arrows on the Dexcom, LOL…an arrow can be straight at 75 and 3 minutes later I’ll be 56. The sudden drops are frustrating. YES, much easier to live life with a CGM than without!!
I learned recently that taking high-ish doses of vitamin C can make my CGM inaccurate. I had started taking 1000 mg/day about 2 months ago (pretty hit and miss), and as soon as I quit, things stabilized. GO figure! I knew that Tylenol would do it. Does anyone have another med or supplement that messes up their CGM readings?
I believe the Tylenol warning has been ameliorated to “Don’t take more than the recommended amount” since the G6 but I still avoid it. I don’t want to give the thing any excuse to screw up.
Dexcom G6 – Very satisfied!
I know the G7 has a smaller profile but I’m in no hurry to upgrade.
Been using Dexcom CGM since the G4 and it has steadily improved. It’s still not perfect, measured against fingersticks, but I’m “satisfied.” Flakiness in the first 24 hrs of a sensor is still a thing, though to some degree ameliorated by the practice of “pre-soaking” (inserting your new sensor 12-24hrs before the expiration of the previous one). This is particular concern as more people move into using CGM-connected AID systems.
it’s also important to note that we think of our consumer BG meters as the fact-checkers for our CGMs, but depending on the model they can be pretty far off too. IIRC they are only required to be something like 80% accurate to clinical measurements, and if you ever do two fingersticks at the same time, like one on each hand, they aren’t going to match exactly either. When I’ve done this just for kicks I’ve seen them be as much as 10-15 points off from each other. Which may also reflect the fact that glucose isn’t distributed perfectly evenly throughout your metabolic system.
Having a specific number for a thing creates an impression of pinpoint accuracy that isn’t a fair reflection of reality. This is why actual scientific results have error bars.
I am using Dexcom 7 for the last 6 months or so, I am fairly satisfied but sometimes I find readings a bit too different compared to some finger prick I do at least once a day mainly before I go to bed so if its more than 1.5 mmols of a difference on my Dexcom 7, I realign the Cgm reads.
I’ve been very lucky to have accurate, consistent results from my Dexcom G6 sensors throughout their 10-day lifetime. I say “accurate” because fingerstick results from separate fingers usually vary more than the difference between CGM results and fingerstick results and because my CGM results line up very well with lab results. Such a gift in terms of peace of mind!
oppisite with me..
I am on my second G7. The first one worked great. The 30-minute warm-up time is wonderful. This second one had to be calibrated, by a lot. Its first reading was around 50, but meter gave me back around 150. That was yesterday, it was reading comparably to the meter this morning.
The Control IQ algorithm requires the pump receiving accurate data from the Dexcom G6. Well, the G6 sensor is ,more often than not , significantly inaccurate during the first 12-24 hours when the readings are compared to a BG meter (Contour Next).
CGM inaccuracies during the first 12 hours are one reason I haven’t switched to a pump. There’s been 2-3 times during the first 12 hours of a sensor when I feel I’m going low and a finger prick confirms it, but my sensor says I’m at 200. I always think how dangerous it would be to have an AI controlled pump under that circumstance. It only has to happen once for it to be a disaster. I would be interested in hearing how pump users avoid this.
I agree that CGM inaccuracies could mess up insulin deliveries from a linked pump. Luckily the link can easily be turned off temporarily so the pump uses preset basal and bolus settings if you have concerns about that.
I’d just like to remark that the scale that is given is not a 1-5 scale, there are 7 options. so is that a 1-7 scale?
After changing from Minimed G3 to Dexcom 6 in February, I am shocked how bad the Minimed sensor was. The minimed sensor was off by 10 to 15 after my morning coffee. During the day it was high or low by 6 to 8 regularly. The minimed control loop couldn’t keep up with all that error. Its like driving a ox cart with rope verses power rack and pinion steering.
The T:slim keeps up and has large error correction bolises with high sensor readings and brings the level back down.
Failure rate on new Dexcom sensor is 25% on internet comments. Mine hit that number easily. Yeah, Dexcom replaces them but what a pain! I am waiting for the G7 but my endo told me that Tandem does not yet have the software done to have the G7 communicate with the pump. This would be a step BACKWARDS for me. I guess I’ll just have to wait. I would have thought that Dexcom and Tandem would have been talking in advance of the US rollout of the G7 so it could proceed without a software delay – I guess not.
Can’t study if FDA hasn’t approved it yet.
With the exception of the first 12 hours after sensor insertion I’m satisfied with the Dex G6 accuracy. The trend arrows tell me more than the numbers.
I’ve been dexcom-ing since the old Dexcom 7 days! Currently usiong the G6 on my TandemX2 pump, using BIQ (CIQ target is too high for me – my target is 80 bg). Anyway, I only fingerstick during the warmup times. Prior to CGM I was a 15-20x a day fingerstick person (I checked the usual times as well as before and after strenous physical activity and every time I was going to drive. – when I was working, I was among first respnders and too many car accidents caused by impaired drivers). Anyway, I was having major bg crashes way too often. This was when there was an obligatory fingerstick when altered to impending lows/highs.
My CGM would alert to a pending low, I did the fingerstick and the meter said anything from 150 – 250! So of course I did a correction bolus. The last thing I needed. The CGM was right. I had the good fortune to be invited to an Endo CME and one of the topics was the issue of accuracy in home glucose monitors. Turns out the meter covered by my health plan (the one that I was using 15+ times a day) read higher than actual on people who have chronic anemia. So I grabbed the hard copies of the reports, learned about the issue with my covered meter. Did my own study, purchased the highest rated meter, and using the same blood sample compared the highest rated meter, the covered meter, and the CGM. Learned that the CGM and highest rated meter were within a few points of each other. Stopped using the covered meter, and now dose from my Dexcom G6. Haven’t had a major crash in years.
Great self-directed research!
Dexcom g6 is great for trends for me, although I usually do a finger stick after warm up as well as double check lows (often not reading as low as I am and not within 20%). Overall, better than fingerpicks all day long.
A lot better than the alternative. Started with chemistry set in 1962. Love all the progress!
Since I took started boiling test tubes at age 4, I have always loved chemistry. Except until I took my first organic course and switched majors. Oh, well.
I checked “very satisfied” with my Dexcom G-6. However, the “compression lows,” which happen at night, are a pain. And I usually do a fingerstick when it reads low, since many times it’s not an accurate low, and I’ve overcompensated in the past.
Very happy and satisfied with Dexcom G6 and Tandem CIQ. Have lived with T1D in my body for over 60 years and am grateful for the AI advances in beta cell function technologies that make life with TID more manageable and so much easier to live as we grow older and wiser. Love it that Tandem system is designed to be upgraded and integrated with Dexcom as more precise AI changes come down the pike.
When I have calibrated my Dex, all was within 5 mg/dl.
Dexcom is the best unless I get blood in sensor.
Hey, switch to Dexcom recently and it’s fantastic
I could be imagining it, but since my supplies are no longer coming from Dexcom, I’ve had more sensor errors and time where there is no reading and I get a message to wait x hours and do. Or remove the sensor.
Used the GUARDIAN (Medtronic) and appreciated it, now utilizing the Libre3. The Guardian always read a little lower than finger sticks, Libre a little higher. Reading the FDA standards, realizing anything made by people is bound for imperfections I’m happy. I clicked “5,” The CGMs gives graphs, trends, statistics, and alarms. To me the alarms are more important than perfection.
The FDA spec for device type QBJ, the Dexcom G6 and G7, for accuracy says, among other things, “) For all iCGM measurements from 70-180 mg/dL, the percentage of iCGM
measurements within +/-40% of the corresponding blood glucose value must be
calculated, and the lower one-sided 95% confidence bound must exceed 99%”
I’d rather see the FDA holding them to +/- 15% and the confidence bound include upper and lower limits and exceed 99%.
Dumb scale; what does satisfied mean? that it is within 3 standard deviations of the result from a home or lab meter? And how are you expected to even know what the expected SD is for a lab result? One sixth of the ‘normal range’ shown on the report? And how are you supposed to compare the two results anyway when it gets below the presumed ‘normal’ where the actual range is not distributed as usual but stops at zero? At least to me, it is not clear what the question means so I answered ‘neither’ (mid point), although you could also argue for ‘very satisfied’ because I continue to use and rely on the CGM and the eyeball distribution of deviations from my home meter appear to be pretty random and fall within reasonable deviations. Whatever that means.