Dexcom accuracy

I would like your thoughts on the accuracy of dexcom. I have been using dexcom for at least 6 months and find it helpful only during basal testing. I just don’t know what to make of it.

Sounds like you have a different experience with Dexcom accuracy than my wife and I have with ours. Most of the time our Dexcoms are so close to meter readings that I am continually am amazed and pleased as is my wife

When my bg is within normal parameters, the dexcom is usually pretty accurate for me. But I find that it reads lower by a fair amount when I drop below 80 or so. I get a lot of false alerts for going low when I am between 70 and 90.

1 Like

I found that it is most accurate for ME when I have it on the love handle all the way on my side rather than being closer to my belly button.

I’ve worn every one of mine (almost 4 month’s worth) 9-12" from my bellybutton. I don’t calibrate too often except when I’ve worn one way past official times, and it’s becoming unreliable. If I do 3 cals 15 min apart and the sensor fails to maintain correct numbers, then I pull it (I’m talking about when it’s way past 7 days)

I have done calibrations ‘on time’, and sporadically when I’m very stable. Doesn’t seem to affect either way unless I do it right when I start to drop or climb.

1 Like

Our current Dexcom G5 sensor has been in place for 11 days (one restart). Abdomen location. Calibration this morning using a Contour Next One meter was off by 1 point from what the Dexcom was reading.

A typical sensor lifetime for use is about 12~13 days. Past that and the numbers start to get obviously erratic.

Similar to @Hammer, we find greater accuracy when within our normal range. As the range gets further away, the accuracy drops. However, we are typically not calibrating in those further ranges, so I don’t find it surprising that the Dexcom is more accurate in the range for which we are entering the calibrations.

We do not get many false low alerts at all.

I started with the Dexcom 7+ system eight years ago. Then I switched to the Dex G4 and the accuracy improved a lot with the 505 software upgrade several years ago. While I know that these devices are not as accurate and precise as science lab instruments, I am happy that they are good enough to make good diabetes decisions.

In fact, I am using the Dexcom G4+Share CGM to provide glucose data to my automated insulin dosing system, Loop. For me, I would much rather live with my Dexcom CGM than without. You still have to thoughtfully use Dex glucose levels and know when to double check with a meter as well as physical symptoms.

I realize that there’s a wide variety of personal preference when it comes to diabetes devices and treatments. Some people like the extra data while others see it as excessive and a nuisance.

If your CGM makes your life harder rather than easier maybe you need to take a CGM break and see how that feels. I feel like I’m driving blind during the weekly two hour warm-up data blackout. I’ll often punctuate that two-hour period with four or five fingersticks. Our diabetes and our personal styles vary.

Before your first CGM, how many times per day did you test? W/O CGM, my average has been around 17 for years. Even when dieting like I am now, I love having the G5 to keep me even more level than I’d be just doing finger sticks.

I averaged about 14/day until the last couple of years, even with the CGM. With the FDA approving use of the G5 data for treatment decisions, I finally decided to back off the fingersticks. Even though I use the G4, the accuracy is similar to the G5. Some days now, I only test once or twice but I’m not hesitant to quickly ramp up if the need arises.

True. Same sensor. Same algorithm. Dexcom only submitted the G5 to the FDA for the additional label change.
(Big difference IMHO for G4 vs G5 is battery life and transmission signaling differences)

The (not yet released) G6 will be a different sensor and a different algorithm that has been shown to have increased accuracy over the G4/G5.

I also found the G5 receiver screen, with its white background, to be too bright even for brief overnight glances. I’m always amazed when these usability gaffes surface.

1 Like

I find the (old) Dexcom Receiver great as a nightlight when going up the stairs in a dark house at night. (Really - I am not making this up.)

If you are looking for usability gaffes, (not that I want to bash Dexcom but…) you will not have to look very hard if you give the (new) Dexcom Touchscreen Receiver a test drive.

1 Like

I use it as a wearable nightlight when going back and forth from bed to bathroom. :slight_smile:

I think that the white background of the G5 was in response to some of us who wrote blogposts about how we couldn’t read the G4 receiver screen in bright sunlight.

My 2013 blogpost about the G4 screen:

Now that I am using Medicare-provided sensors and using the G5 receiver, I hold it over the side of the bed and tilt it up slightly so I can see the number. Otherwise it would be like shining headlights in my husband’s face. He never complains about the Dexcom interruptions of our sleep, but sometimes he is too busy snoring to hear the Dex…

Since things like background color are an easy software fix, I’m surprised that Dexcom doesn’t let us choose. Or change it from day to night. At least the new touchscreen receiver is dimmable, but I can’t believe that you have to scroll to enter BG calibration numbers. To me this receiver pretends to look like the phone, but unfortunately doesn’t act like the phone…

But although I maintain the right to complain, Dexcom has changed my life for the better and I appreciate Medicare coverage of the device.

ha ha ha
Very good description in a nutshell !!!

1 Like

I also complained about the Animas Ping screen’s readability in sunshine. That would explain about Dexcom’s shift from a black to a white background on the display. Something as simple as a day/night viewing option would mitigate that problem.

1 Like

Mine is frequently very off but it also Accurate a lot of the time and the trends are usually right. The other night it said 75 I felt low though and tested and it was actually 51. sometimes it’s over 100 points off so I do have to be very careful and I always use a finger-stick. I would never rely on it for dosing etc unless I had just tested five minutes before and it was close to my blood finger-stick. Dex is life saving for me and I couldn’t live without it however for me it’s not reliable over fingersticks. I don’t know how it got FDA approved for insulin doses. That’s dangerous.


I find that since I have been calibrating in the am before breakfast and
just before bed it has been more accurate. I think that’s the trick to

D’Etta Meloche

Do you ever test your bg about 15 minutes after getting a reading that doesn’t match the sensor reading? When my readings don’t match, if I test a bit later, due to interstitial readings being “behind”, they often correlate super-closely even when the G5 and fingerstick readings don’t match when looked at at the same time.
Well, you aren’t going to get ANY readings when you are asleep, from fingersticks, so the Dexcom has a great value proposition if you consider that. :slight_smile:

Well if you think that’s incredible, tell me how the FDA approved the Precision meters back in the mid 90
s? Or Sof Sensors? For the vast majority of users, the Dexcom system is accurate enough of the time, and more accurate than other systems, to be a breath of fresh air in a sea of inferior products.

1 Like