I don’t know why I didn’t think that would be true for Dexcom. I read a lot of research on sensor technology, and they say there’s a great deal of inter-subject variability in regards to accuracy. I guess it’s just very apparent with Medtronic, but the vast majority of people who I’ve spoken with pitch Dexcom as being spot on. I used it for a year and remember it being that way (but also know better than to trust my memory farther than I can throw it). 
I just responded to another that I forget this can be true even for the beloved Dexcom. Medtronic gets blasted for their sensor difficulties so it’s easy to see, but sensor performance, according to what I’ve come across, can vary person to person. Have you guys ever tried anything BUT Dexcom? I called it “beloved” back there which sounds so nasty, but I really like the product and miss the reliability. I don’t think I’d ever dose based on a sensor only because of all the bad ones I’ve come across, but maybe I just haven’t found the right one. 
You still need to be the Bg “Sensei” ( per @Eric) to know when to dose or need to finger stick. There is no way around it.
1 Like
I think like everything diabetic you adapt and decide when to dose off the Dex. We have not tried the MDT system, although their latest sensor is probably adequate.
WIth that said, there are times when we don’t dose off of it, like when it is changing quickly, but when the Dex trend is relatively flat then we always are able to dose off of it. Once you start using a sensor again, I think you realize when it is right to dose with it.
That’s funny. And sounds about right. I’m currently trying to learn Medtronic Guardian language where a 110 after I’ve eaten means 150 and a 135 means 194. If I’ve been moving around, a 110 means means a 78 and a 135 means a 112. But there’s also those periods where 110 means 110 and 135 means 135. It’s the English language equivalent of homographs… only way more common.
I have the exact same language translation for my Dexcom. They are somewhat similar. Probably like being able to speak either Romulan or Klingon.
To me, it’s not a matter of how accurate it is. The key issue is whether the CGM is trustworthy. I.e., do I get blindsided with a seriously wrong SG that tricks me into getting into a bad glycemic state? I really don’t care about an error of 20mg/dL because if my target is 100 and I’m steady at 80 or steady at 120 that’s ok with me. And if my BG is moving and I correct to flatten it out around 100 (SG) still there’s no harm. I had a Medtronic sof-sensor (the one before the Enlite) that convinced me everything was fine with a slow decline from around 87, when in fact I was around 42 and slowly declining. If driving I could have crashed. There were other instances of similar treachery. Not the Dex (G4 or G5.) I have experienced huge errors from the DEX, but only when it already has shown me that it’s not to be trusted (by jitter and gaps in the graph that indicate a poor or failing sensor.) In that case, if it matters, I’ll check with a fingerstick. But when the Dex is running right (according to the graph and the reasonableness of the match at calibration time) I find it trustworthy. Even if the Dex is getting a little squirrely, I’l still dose off of it because if I take too little insulin I’ll get a high alert and that will cause me (or LOOP) to add corrections early and often; similarly if I take too much insulin I’ll get a low alert and take glucose to move back to range and level off.
For me, it’s a “sometimes” kind of experience, not common, and pretty much always either during the first day before the sensor has settled, or at the end of the sensor when it is starting to fail (as signaled by jitter in the graph.)
Of course Dexcom has big data from the G5 in the field, so they would have statistics on how bad things get in practice, both within and across the users. But I’ve never seen those kinds of results published: presumably they are considered highly proprietary. I do hope the data informed the design and internal testing of the G6 sensor and algorithm. Have any of the more open data collectors (such as Nightscout) published any bulk statistics?
Dexcom says the accuracy isn’t quite as good in pediatric users, and gives this reference Improved Accuracy of Continuous Glucose Monitoring Systems in Pediatric Patients with Diabetes Mellitus: Results from Two Studies - PMC So one could decide not to dose off of the CGM, or one could decide to watch the CGM closely after dosing and surf — intervene early to correct a trend that would lead the BG out of range. Either decision seems reasonable to me. I do find it surprising that his Dex is “more than 30 off for more than 50% of the day.” I wonder what’s the typical circumstance. Is it systematic, like lagging during fast moves, or just randomly off by arbitrary amounts in any direction. In that latter case, the Dex would be almost worthless.
As has been referenced, studies in CGM’s indicate there is much more variance in pediatric patients so no, we’ve not considered moving from Dexcom. We don’t have a problem at all with the variance…The most important aspect of the CGM (for us) isn’t being able to dose from it…it’s being able to have a general sense, 24/7 of what his BG’s are, and more importantly, how they’re trending. When he gets older, I’m pretty sure they’ll get more in line with what other people experience with the Dexcom. I have zero issues with the Dexcom! I attribute our lack of accurate data, to our sons age (been wearing the CGM since 2 years of age and is now 4).
1 Like
Today I read an 82 when I was a 169. The other day my sensor read a 93 when I was a 224. Later that night my sensor showed a 250… I was a 360. Yes, let’s talk about a lack of trust. I test my blood sugar often enough (except for the other day when I experiencing burnout and refusing to do one more finger stick, hence the 360) to KNOW I can’t trust my sensor. This is the Guardian, which is the newest addition to the family of frustrating Medtronic sensors. Of course I do know that with a range of blood sugar like that, the “accuracy” or even its ability to be ballpark is going to weaken, which is true for all sensors (?), but I have to work REALLY hard to get that number to have meaning. It’s hard work and that’s just so that I have a good number on which to put in hard work. Some days it makes me feel like a scientist, running little experiments, finding wonder in the numbers. Some days I just do it. Some days it takes it all out of me. On those days I wish it worked better. I have the 670G, by the way, which I think we’ve already covered. I personally can’t have a system choosing an insulin dose (or lack thereof) based on this. It’s maddening.
He’s 4? Just 4… I don’t know why I thought you were talking about a bigger person. Yes, I do believe his Dexcom will tighten up for him. It really does seem to be what’s true for most. And I’m sorry you and your 4 year old have all of this… and that he’s got 2 years under his belt. It sounds like you all are plugging away and making it work, but that is very hard.
Okay I’m just trying to sort out which of us is that English teacher and which one is the Star Wars fan…
1 Like
I pay $75/month for 3 sensors, but I believe you can get discount cards to get them cheaper - I’ve just been too swamped with life to even look into it. The initial reader outlay was I think $68 or so (and I got a free replacement when my first sensor was off)? When my sensor is accurate, I will sometimes forget to do fingersticks for days…oops. When it’s off, I have to do more. To me, it’s worth the money because I’ve only had 2 sensors be off (I’ve been using this since January), and Abbott quickly replaced one (the other I forgot to ask for replacement). It’s cut down my fingersticks significantly, and improved my control - past 2 A1c’s have been 4.9 and 5.0 (with very few lows), down from 5.3-5.5 range.
Wow your a1cs went from amazing to more amazing. 
i’m Still letting that process…
I do a lot of finger sticks. My fingers would appreciate greatly any kind of reduction at all, I was considering checking it out anyway, but with that information, I think I will. I appreciate your responding! Just out of curiosity, how do you wear it??
That was meant to be a joke… I hope that’s how you took it. I know the difference between Star Wars and Star Trek. I wouldn’t be allowed to live if I didn’t.
1 Like
LOL! I’m still honeymooning - not even on basal though I think I will be sooner rather than later. I also have Afrezza so any excursions above 160 come down very quickly. 
The Libre is approved for application on the back of your arms, so that’s where I wear mine (I know some people have tried other locations, but arms work fine for me for now). It’s a very easy applicator - just press against your skin and the applicator does the work. The adhesive is very strong, so usually I don’t have to put anything over it to keep it in place the full 10 days.
Yeah, that’s just nuts. But the pattern is very clear, isn’t it. Medtronic is trying to save you money on insulin by always telling you to take less than you need.
Back on the Sof-sensors, I developed exquisitely careful calibration technique (4x/day) to try to get some utility from the CGM. All the usual things like careful handwashing and fingerstick technique, plus all kinds of rules along the lines of only calibrate when BG is not changing, and hours away from any prior food or insulin dose, and no food or insulin for 15 minutes after calibrating, and calibrate at a variety of different BG levels but all within the range 75 - 145. And of course the sensor was taped down with a big donut of opsite flexifix to stabilize the tissue around the sensor and help keep the sensor from moving in the skin.
The contrast when I switched to the G4 was exhilarating.
Yeah, I figured you were kidding around.
My kids wore DC shirts to the Avengers movie opening this weekend, just to irritate the Marvel fanboys. 
Out of all of those characters I only recognized iron man. That one I’m not joking about. 
1 Like
I also used Sof-sensors before Dexcom Seven. Sof-sensor frustrations is the reason I found online support groups/DOCs. I learned how to check the ISIG before doing calibration, and also did 3-4 per day.
Dexcom is so much easier.
1 Like
True DAT. And sometimes there just isn’t any reason for weird stuff…in which case I shrug and say ‘eff it’ 
lol…
Now I want some kettle corn! Sorry my reply was so late, had a crazy couple of weeks and am catching up on the good stuff!! Hope you have had a slew of good days. 
1 Like