The Dexcom G6 cgm system for the vast majority of people provides far more information then any consumer grade (home) BG meter (fingerstick).

No single solution work best for everybody. Choice is always important.

Dexcom has a page that gets into relevant clinical trials that have been run. This may be worth the time to read through and see which of this (if any) would be helpful to prove your point. As well, the information and clinical trials available may help you to fine tune the point you wish to make or perhaps word it slightly differently.

Dexcom: Clinical Studies: Evidence for Benefits of CGM Use


Because of the way my blood sugar spontaneously jumps around in response to temperature changes, changes in stress levels, rising and falling insulin resistance, and who knows what else, the overall dosing imprecision is so massive that getting an additional 1.087% accuracy in blood sugar measurement would be swamped by all these other factors.


For me, my Dexcom is more accurate. But I have chronic anemia (since childhopod) and the meter that is covered by my health plan reads higher than reality - reports have been published by the Diabetes Tech Sciety. SO I was treating highs that I did not experience, and ignoring pending and actual lows. Lots of major crashes. If I could recall the definition of MARD the Dexcom G6 MARD is 9% nad most meters is 20%. I don’t think the 20% margin that is acceptable to the FDA is all that great, and the fact that meters are not periodically re-evaluated after FDA approval makes me angry


There’s a basic misunderstanding here; the reason Dexcom did that news release is because it suggested that the G6 met the current US requirements for the fingerstick blood glucose meters we all use. It is THOSE requirements that allowed 5% of the test results to be out of range and “therefore” quoting 4% out of range was, in marketing speak, “better”. It was, after all, marketing, not fact.

At the time in question the FDA guidelines were met by fingerstick meters that produced 19 out of 20 readings within 20% or 20mg/dl (whichever was GREATER) of the correct value.

In practice every meter sold in the US and not recalled was significantly better than that because the international standards organisation had revised their original 20% standard from 2003 to 15% in 2013 and the Europeans adopted this in May 2016. The 2013 standard still uses the 5% rule - 5% of readings can be outside the range - but it also introduced a 1% rule around the 2003 20% range; only 1% of readings can be outside the (original) 20% number.

So the G6 hit the market at just the wrong point for the marketing droids; I assume the Oct 2016 press release was rushed because they did not have the final results. Nevertheless I can see why a nurse might make that assertion because the G6 was, APPARENTLY, better than the FDA guidelines for fingerstick meters that no manufacturer was following any longer. Hey, my mother was a nurse and she was also very intelligent but she gave me some bad advice along the way, along with all the good stuff.

The truth starts at page 289 of the G6 user manual.

John Bowler


Ok, @John_Bowler, I guess I’ll have to further explore this. I didn’t know that FDA had changed the rules. Have you seen the data from Diabetes Technological Society? How do you respond to that. I’m not 100% sure that I understand.

Can anybody confirm that fingersticks are no longer required? I think I read that online somewhere and it sounds so ‘wild’ that I would like someone using the equipment to confirm this.


Here’s the text from the G6 “using your G6” guide; the only thing that anyone reads given that the user manual has 360 pages (sides):

"Chapter 4: Treatment Decisions

With the Dexcom, you can make treatment decisions without using your blood glucose (BG) meter (meter).

But sometimes you must use your meter instead of the G6. And other times it’s best not to treat, just watch and wait."

Then (the * * * indicate elisions by me for conciseness):

"4.1 When to Use Meter Instead of G6

Rely on your BG meter for treatment in these situations:

o G6 doesn’t snow both a number and arrow.
* * *
o Your G6 readings don’t match your symptoms.
* * *"

There’s nothing that I can find in that guide (I can’t speak for the user tome) indicating that treatment decisions should only be made if the G6 is calibrated except that it suggests that if the G6 is off from the meter (my G6 is currently running 20mg/dl higher than the meter) you should calibrate. In fact I know that the G6 is running high because it violated the second condition at one point; the G6 indicated 80mg/dl at one point my blood sugar was well below that. The meter did show 60mg/dl at that point although the meter (Omnipod PDM + FreeStyle strips coded at 16) is notorious for being low.

I admit that the DexCOM words above are both ambiguous (“you CAN”) and barely English (sentences starting with “But” and “And” and other cringe-worthy phrasing), but all the same the meaning seems clear.

This is my first sensor on the G6, I’m going to run it UN-calibrated until it self-destructs or fails, about 9 days on average, then I’ll try calibrating the next one. The only issue I see is that it took maybe a day to stablize; it initially read exactly the same as my meter (which probably reads low) so calibration in the first day would have been pointless.

Yes, there was a trade war initiated in 2013 as a result of the European threat to adopt the 2013 ISO standard. It started with some interesting data from Germany which was adopted by various Xenophobes in the US; the original data had more the appearance to me of an attempt to discredit a particular German manufacturer. Nevertheless the adoption of the ISO standard in Europe was delayed until 2016. It’s not clear to me when the new FDA recommendations go into effect; there is implication that they are apply to all new applications but that’s not explicit. Anyway, those recommendations certainly don’t apply to CGMs and the G6 certainly doesn’t meet them; I don’t think it even meets the previous (20%) recommendation though it might if the results weren’t bucketed.

The FDA approved the G5 for treatment decisions according to DexCOM (the link DexCOM give to the FDA web site is to a publicity document, not a formal approval letter). Here’s the diatribe article, it seems correct to me (given that I haven’t tracked down the formal approval document):

The FDA publicity stuff on the G6 is here but it only comments on the approval for closed-loop systems:

The BG meter FDA recommendation from 2016 explicitly applies only to patient use BG meters; everything else including CGMs is explicitly excluded. The above link shows that, at the time, (May 2018) the FDA had no formal guidelines for accuracy in CGMs when used in closed loop systems:

“[T]he FDA is establishing criteria, called special controls, which outline requirements for assuring iCGM devices’ accuracy”

Use of the present tense is intended to imply the corollary “the FDA does not have criteria…” Ah, English, don’t you love it.

John Bowler


G6. No fingersticks.

(unless you have reason to doubt the dexcom)


Ok, so I looked at the User Tome (extra cup of tea, 6g sugar, .5IU). There’s an authoritative statement from Dexcom:

“Making Treatment Decisions Using G6
You can use your G6 reading and trend arrow to make treatment decisions – like
treating for a low or dosing for a high. See Chapter 11 for more information on
treatment decisions. With G6, there’s no need to take fingersticks to calibrate the
system or for treatment decisions (as long as your symptoms match your G6 readings).
This can reduce the pain and burden of excessive fingersticks (Aleppo 2017) and
reduce potential errors due to inaccurate calibration.”


Thanks, @John_Bowler for sharing the information!

My reply isn’t directed toward your post or anything of that nature, but rather … what you said finally clicked for me … the reason I was able (and many others) to get a CGM in the first place was because I wasn’t able to tell highs from lows or anything in between.

So, that leaves me to wonder, what about those people, who aren’t symptomatic? Like I mentioned, it isn’t really directed toward you, but rather for everyone to think about for themselves.


FDA previously provided approval to bolus from the G5 readings so this has been around for some time. The G6 continued the bolus from cgm as well as added FDA approval for no calibrations.

We have used the reading from the cgm UNLESS the symptoms DO NOT match the cgm. This was for the G5 (actually we were doing this well prior to the FDA approval) as well as currently on the G6.

Obviously everybody should do what works best for themselves.

We started using Dexcom with the Seven Plus system. We have always found the Dexcom data to be highly reliable.


@John_Bowler, Thanks for sharing so much info, John. It is super helpful and well documented. Much appreciation.


G5 and G6 have Nightscout integration? They are considered IoT devices? Anybody stumbled upon a way to turn off internet functionality in the hardware?

I’m going to a big computer security conference at the end of the month. I’m a little worried about using my wireless devices (although I am probably just paranoid). There will be lots of IoT hacking at the conference and I will be far from home. Should I go back on syringe insulin, just to be safe? Or, is this paranoia, self destroya, in your judgement?


Yes, I use a G5 and I am happy that I no longer need to test prior to bolusing, but I’m still calibrating and my G5 (as was my G4) is very accurate.

My point was in regards to those who are not symptomatic. If I ever do upgrade to the G6, for my own purposes/reassurances, I will undoubtedly still test with a glucometer (and calibrate) at least twice a day.

I can understand where no finger sticks definitely helps with children and those who are squeamish to such and even for those who don’t like too!


If the G6 is accurate then it simply becomes pointless for the fingerstick.

I totally get doing the fingerstick at first. It is a big change. And based on various posts, it would seem that some people just do not do well on the Dexcom for whatever reason.

But if you have accurate numbers on the Dexcom and after a period of time on the G6 with repeatedly seeing the fingerstick in agreement with the Dexcom then it may simply become pointless.

I don’t think squeamish is a factor in the slightest. Hopefully my posts did not imply that as certainly that would not have been my intent.


Makes me want to upgrade since I DO still have to provide a number of manual BG reference points to corroborate the sensor data. I am, personally, squeamish about communication with the internet in newer models.


This is a blogpost I wrote earlier this summer that describes where I’m at with trusting my Dexcom. it’s not perfect but nothing with diabetes is.