I'm a newbie, I just got my G4 2 days ago, and I was wondering if adding more test strip data makes it more accurate, or if simply sticking with the twice a day inputs is about as good. How accurate (%) do you find the G# between say 70 and 150? Also are there accuracy issues low or high?
Hi there! I've been using a Dexcom for awhile now. I had a Dexcom SevenPlus (the old system) and then switched to the G4 right when it came out (around the new year). I actually just talked to a Dexcom technical support guy on the phone yesterday about this very subject! He said that with the new G4 system, calibrating more often is totally fine, and it will help the system "learn your boundaries". With the old SevenPlus system, they told us not to calibrate more often, because it would confuse the system, but with the new G4 it actually helps it be more accurate. He said you can calibrate up to 6 times a day, and he suggested calibrating at least every time it prompts you and then once before you go to sleep at night (that way you don't accidentally miss a calibration during the night). He also said that a good rule of thumb is anytime the dexcom is more than 20% off what your finger stick is, to go ahead and calibrate it. Then, give it about 10-15 minutes to readjust and it should be more in line with your meter readings. Also, remember that because the Dexcom is reading interstitial fluid, most of your readings will be on a 10-15 minute lag time from your finger stick anyway (because a finger stick reading is taken directly from the blood). So, keep in mind, that if your blood sugar is dropping or rising very quickly, it might take the Dex awhile to catch up. OK, so that's what I got as "official" advice from a Dexcom tech support guy yesterday.
Now, in my personal experience, the new G4 has always been pretty accurate in the "normal" and "low" ranges, but doesn't do quite as good of a job in the "higher" ranges (usually about 140. But, that's just what I've noticed, others might have other experiences. Good luck! It takes a bit to figure the system out, but I love mine and have found it so helpful!
personally, having had the G4 since December I usually only enter the two prompted test strip readings. sometimes when I believe the readings are off, it’s because my glucose is dropping or rising faster than the G4 can keep up with. (just what Erin explained). frankly, the readings have been very accurate with just twice daily test strips entries. don’t make life too complicated. you’ll do fine.
I understand the lag issue, but I'm also finding (day 2) that in the lower range (below 100) the Dexcom seems to run 15 points lower than actual BG. I've got 2 good regular meters (different brands) that are pretty accurate. Not slamming on the tool since it is extremely nice to just look down and have a good idea where my BG is.
Also, it is possible to change the range of the graph that is shown on the screen of the G4? Ideally having it go from 0 to 200 instead of 0 to 400.
I calibrate as little as possible. After the first couple of days it is not unusual that I don't calibrate for the rest of the week. As long as the G4 is within 10 points of the BG meter I don't calibrate.
I calibrate whenever it’s off by more than 20 points AND I don’t have reason to believe its going to catch up. My Dexcom is accurate between 80 and 120, but I find it almost invariably INaccurate above 150. It’s kind of frustrating, actually, how often my Dexcom is off, considering how everyone exclaims how accurate it is.
Hi,
I’ve had the dexcom for 2 months… Just diagnosed 2 1/2 months ago… According to dexcom reps fastest approval ever… Had it shipped about a week after leaving ICU ( recovering from DKA) … Doesn’t hurt I’m a doctor, have great insurance , and have an endo doc who trusts my research into the care of my new disease. Have read some obscure research articles on cgms and some esoteric dexcom info… My personal experience and that based on research articles suggests the more we calibrate the better ( as long as calibration done during steady state situations )… I recalibrate 4-6 times per day and find it common to walk around during day or wake up in the morning no more than 1-2% off my ping glucometer! I have also noticed that the dexcom reps are nice People better trained at selling the product than answering technical questions. Side note : the glucometers we trust our lives to have a +/- 20% “acceptable” error rate… If my glucometer reading is really off my dexcom I’ll recheck and I routinely find my dexcom is close to the second and/ or third test reading ! Frustrating to live with such variability in our " gold standard " daily devices but good to know that the dexcom is probably more accurate than we give it credit!
I test 8-10 times a day and calibrate my G4 every single time (as long as the arrows aren't going straight up or down).
I used to only calibrate 2-3 times a day, but wanted to see how much it would change the precision by calibrating all the time... and I've found that it does seem to help a bit (but more importantly it doesn't hurt it at all).
Tony, I was thinking that about steady state situations being best for calibration. Last night I woke up a few times and I had a straight line at 119-120. Very frustrating since it is day 2 of using the meter (and a little high), but it was within 5 of my test in the morning and dropped when I took a little insulin. So apparently that straight line was pretty accurate. Overnight numbers were a big reason for wanting the device.
Ah, youth. :)
("In MY day, we had to bleed ourselves with a knife onto a pile of sticks covering five chicken gizzards pointing to the north, turn our backs and GUESS what our blood sugar was!!! Now get off my lawn!")
The inaccuracy you mention is actually in the spec - the G4 really is less accurate at higher glucose ranges. That was a design choice, because the necessary tradeoff it provides is much higher accuracy in the low glucose range where accuracy is critical. If you're high, you're going to fingerstick in order to know how much correction to take, so as long as the CGM tells you "high," it has basically done its job adequately. On the other hand, when you're low, it's much more important to know exactly how low and in a timely fashion. The accuracy of the G4 is unbeaten in that range. Ergo, the tradeoff was worth it.
I tend to cluster-calibrate after insertion - the first day or so I'll calibrate at every fingerstick (except double arrows).
I find that after day two, If I enter a value whenever the Dex says I'm level and I happen to check, plus the two requested, I'm in pretty good shape.
I know it is day 3 in my case, but my G4 readings (below 110) have been a consistently 15 points lower than my fingerstick readings. Today I've been darn close to flatlining it, and my dexcom puts out 85 when I'm 100, 71 when I'm 87 (finger stick). It isn't like I calibrated it with a lower number, so I would think the variability would sometimes be above my FS meter and sometimes below, but not consistently lower.
I have 2 reliable standard meters (calibrated the G4 to only one of them) and one is consistently 6 points lower than the other, so I think this was programmed in. Now on the plus side, my G4 is very consistent in its numbers, but it's a PIA while I figure this out.
yes, adding more data is good and will make the G4 more accurate. Just be sure that you add readings when the arrow is level and not when your blood sugar arrow is going up or down on the receiver.
yes, I love to see that flat line. It will get lower, give it a week or so.
Yeah, I see what you mean. Most of the time, it's okay, but it's so frustrating when I see that my CGM is telling me I'm at 180, and I test and I'm at 240. It's like, thanks for nothing. Additionally, it tells me I'm low about 20 minutes after I've felt and treated the low. However, I'm not abandoning ship for the exact reason you state - it's too useful for overnight lows.
I also get frustrated because I was hoping for it to be really useful when as I'm training for a half marathon. Yesterday, I did 8 miles, and the readings were so off that I way overtreated a low and ended up at 270 when I finished.
I love it when it's on target, but it causes actual problems when it's not working well.
I look at it this way -- if I'm going to take a reading anyway, why not enter it into the g4? it takes a few extra seconds for an overall process that takes a few minutes.
I'm finding it to be reasonably reliable, but my BG is consistently 15-20 higher when it says I'm between 70 and 130. Pretty much all the BG values I enter are 15-20 points higher, so it's either programing or my unit is bad.
Remember the dexcom measures interstial glucose not blood levels so there is/ can be a 15-20 minute lag time… Its not uncommon that when i recheck my blood levels they match almost exactly with numbers from 15-20 minutes later on dexcom… Sometimes when numbers are really way off I recheck with my glucometer x 2-3 ( remember industry standard allow ± 20% error in glucometer readings!!)… A few times the initial crazy dexcom number was consistent with numbers derived from second / third blood glucose reading !!
Also when exercising vigorously the dexcom may be 15-20 points off ( usually lower … Again there is lag time plus when exercising there is decreased blood flow to extremities and peripheral tissues … E.g. Belly fat) … Not uncommon for me to get 50-60 reading 1-1.5 hours Into workout yet when I test I’m in 70-90s … Unless I hit 40-50 on dexcom while working out I ignore knowing its most likely artificially low .
TC
I understand that, but when I enter 5-6 BG readings into my system over a day or two after start up with every one higher than the Dexcom output, it consistently is 15-20 points higher than my BG, the issue isn't delayed BG readings or variability (lack of precision), but rather there is probably programmed inaccuracy because they aren't willing to risk someone dying from a hypo. When I'm flatlining within the 90-120 BG range) for 2-3 hours and testing my BG over that time period with my standard meter, the Dexcom is consistently 15 points lower.
As for the "industry standard of +/- 20% error....that is what is permitted by the FDA, but the bulk of the readings from good meters are probably around +/-5%. I have 3 reliable BG meters right now (was just sent one with my pump), and do multiple finger stick tests as well.