The what? Nine CGMs, all worn at the same time, to push the boundaries of the idea that comparing them properly needs side by side data.
Which CGMs are you testing?
By any stretch of the imagination, nine CGMs is a lot to fit onto locations that are considered appropriate by the manufacturers. Fortunately, the form factor of most has reduced enough to make this easier.
Here’s the full lineup of sensors started simultaneously:
Dexcom One+
Dexcom One
Libre3+
Libre2+
Yuwell Anytime
Caresens Air (non-calibrated version)
Syai Tag
Glucomen iCan
Microtech Linx
Observant readers will notice that there are a few missing from the list. Those I chose not to include are:
Medtronic Simplera (last trialled in March 2024)
Dexcom G6 and G7 (hardware is identical to One and One+, as is performance)
Medtrum Nano CGM (no change in hardware since the last trial in 2024)
SiBionics GS3 (not yet available in the UK)
Eversense (Currently no UK distributor and prohibitively expensive due to minor surgery requirement for installation)
The main area of interest is the five new sensors that were heavily marketed at ATTD 2025. The question is whether they can live up to their marketing?
How are you testing them?
Using a variation of the traditional Diabettech CGM test procedure™
Primary testing model
I take 8-10 fingerpricks per day, distributed over the day, approximately 1.5 to 2 hours apart. Additionally, if a sensor says hypo or I feel hypo, if possible, I fingerprick outside of schedule and check.
Where this test will differ subtly from previous experiences is that it is based on what real world use of the systems would be. This means that I fingerprick, then take the reading displayed on the device at the time. The reason for taking this approach is that users in the real world do this. If I am planning on dosing for a meal, or doing exercise, I look at the number now and the arrow direction and make a decision from that. I don’t think about what my glucose level was five minutes ago.
In this model, I am testing the effective, actionable data. This is the core of the test.
This data will be analysed using a MARD versus the fingerprick data, which has all been obtained using a Contour Next system.
Secondary testing model
In the secondary model, according to each of the apps in use, I will be able to pull historic data points. These will be extracted and the data compared side by side to look at the Time-In-Range metrics for each sensor. I may also try and run more quantitative analysis of the lag times and variance in rise and fall, depending on the quality of the outputs.
Other observations
Given that there are a number of sensors here that haven’t been widely seen in the wild, I’ll try and highlight some of the features that appear in the apps, to try and offer a better insight into what the manufacturers are trying to offer.
Even at this preview stage, there’s already an observation about the amount of waste these things create….
Anything else?
At this stage, no, but if there’s anything you’re interested in, mention it in the comments and I’ll see what can be accommodated!
As we are at the peak of a British summer – are any of the CGMs affected by the hot weather – i.e. if you’re sat in the sun for a while are the readings any more or less accurate?
I assume my Dex One+ is ok (as I have no alternative), but no idea is hot and cold conditions might affect the readings
really want to know your comments of all CGM
Wow your a brave man ! ! !
RE: the last comment regarding waste. I use the Dexcom G6, with a 90-day sensor.I’ve read comments from G7 users about “how great its all one piece!” And then I read about the tariffs from the US White House, the rare metals in the transmitters being just that – rare. I question the wisdom of tossing a transmitter in the trash every 0 days. It’s not that hard to insert the transmitter into a newly placed sensor
Oops, typo. Next to last sentence should read “every 10 days”
I’m not suggesring you get up in the middle of the night to do BG readings, but any info about susceptibility to compression lows (and any other environmental impacts on accuracy of the readings) would be of interest.
I’d be interested in if and how each sensor responds to compression (e.g. does it read low?), and any other potential environmental/contextual causes of variation in accuracy.
I’m not sure that there’s a good way to test that. It’s fairly easy to observe when it occurs, indeed both Dexcom One and Libre3+ have data that suggests that they have both endured compression lows during this test. I’ve seen nothing on the other arm though, so they are perhaps less likely to show this.
Hi! Really interesting.
What were the results of the comparison?
Thanks
The waste created IS concerning, although it seems that G7 (one/one+) is smaller than the applicator for G6 which is HUGE by comparison, I understand the need to ensure that the sensor is applied correctly, but back in the day (I know, I am old) you could load an insulin needle into a device that would then “give” you the shot with the press of a button. Why can’t there be an applicator that you re-use with the new sensors/transmitters? Less waste might also translate to less cost in the long sun, including all that extra plastic going to the landfill! I’m very interested to see the results of your sensor samba!