Glucomen Day CGM Review

[Lucyr]

So I decided to give this a go whilst it was on offer bringing the starter pack in line with the cost of libre. Since there’s not much info out there on it, I thought you may be interested to see how it goes.

Probably not the ideal start as my bgs were high today since I forgot lantus last night. I have a libre 2 with a couple of days left on, which I’ve left a couple days overlap for comparison.

I tried the meter first, at 3:40
Accuchek bg 14.2
Glucomen bg meter: 16.4 (double checked, 16.5)
Libre 2 14.8
All of those too high to calibrate so I waited till later on, but wasn’t a great start. Within tolerance but Glucomen bg meter further away than libre2 from accuchek meter.

I calibrated it later on, 4:40, the bg readings were still different at:
12.6 accu check bg
14.2 Glucomen bg
12.8 libre 2
I calibrated to the accuchek reading not the Glucomen reading. Slightly defeats the point of them sending you a meter that can calibrate automatically, but I don’t yet trust the bg meter

6:40 Before tea, the first real chance to compare CGM readings
Accuchek BG 7.8
Glucomen BG 8.7 (seems more reliable at better bg levels)
Libre 2 8.4
Glucomen CGM 8.0, but with two down arrows which continued dropping down to 6.2 at 18:50
It was right that I had been dropping, I’d done loads of correction earlier, but libre 2 had a horizontal arrow and accuchek bg 10mins later was 8.6 so Libre2 still winning at this point.

I will add some more, but should make a disclaimer at this point. We all know sensors can be less reliable to start with, and here I’m comparing a libre2 that has been in a while, with a brand new Glucomen sensor. However, they claim the needle free insertion means less trauma, and the sensor starts when you insert it so you can’t do the out it in one day early thing you can with libre2, you’d have to use the readings on day 1 I think.

 

[Lily123]

The libre seems to be more accurate but like you said in the disclaimer the Glucomen hasn’t had time to settle

 

[Lucyr]

Roughly 2.5 hours after tea, I was feeling high again, and the Glucomen was wanting another calibration, so time for another comparison.

9:10
Accuchek BG 12.2
Glucomen BG 13.0
Libre2 13.9
Glucomen cgm 17.4
I calibrated it to the accuchek bg

Day 1 conclusions: This is further out than a libre would be on day 1, let’s hope it settles down overnight (and that my bg improves!) otherwise it may be a long month / may find out what their support service is like!

 

[Lucyr]

Day2 and we’re off to a much better start

Fasting readings at 6am:
Accuchek BG 7.3
Libre sensor 6.7 ->
Glucomen sensor 7.3 ->

 

[Docb]

Can I remind of something to bear in mind when interpreting these data? It is that there is an inherent variability in the reading you get from a finger prick - if you recall my ten finger test suggested a 95% confidence interval of +/- 1 on the mean value from the 10 readings. The effect of this is that the finger prick has real limitations when using it to "calibrate" another system.

This observation is in no way trying to denigrate the system, far, far from it. I am just suggesting that care is taken in interpreting the data in order to get the best from it. For example, there is a good case for suggesting that the differences within the readings for each set of data are not statistically significant.

Congratulations @Lucyr on what you are doing. You are properly testing something out - keep putting up the numbers. I am sure this sort of data will exist within the organisations that make the devices but there is no way anybody outside will get to see it because it will make their PR claims look a bit thin.

 

[Lucyr]

Can I remind of something to bear in mind when interpreting these data? It is that there is an inherent variability in the reading you get from a finger prick - if you recall my ten finger test suggested a 95% confidence interval of +/- 1 on the mean value from the 10 readings. The effect of this is that the finger prick has real limitations when using it to "calibrate" another system.

This observation is in no way trying to denigrate the system, far, far from it. I am just suggesting that care is taken in interpreting the data in order to get the best from it. For example, there is a good case for suggesting that the differences within the readings for each set of data are not statistically significant.

This is a good point that i probably haven’t made clear. Whilst I agree that the differences so far are probably within the allowed tolerance, what I’m not looking to do here is say whether the readings are statistically significant / acceptable etc.

My aim is instead to share whether the level of accuracy is enough to work with as a user, and to compare where its best for me to spend my money. Sensors are pretty individual so any results I share are valid only for me and are not in any way a recommendation for or against others using either system. What works for me won’t work for everyone else etc.

 

[helli]

I did notice that the sensor requires daily calibration, which may be slightly annoying

My Libre set up with xDrip allows me to calibrate. I don't have to but it gives me peace of mind to calibrate at least once a day (usually first thing in the morning when my levels are stable).
One or two finger pricks a day is far preferable to the 10 pricks I was doing before the Libre.

 

[rebrascora]

I hope you will still continue to share your results though as it is interesting.

 

[Lucyr]

So here's some more results from today. Glucomen definitely improved today, both were inaccurate at times but as you'll see in bgs below, they were more evenly matched on their inacuracies. Sometimes the direction arrows seemed a bit slower to catch up on the glucomen too, e.g. double down arrows when i had stabilised after a fast drop, but generally much more comparable today. BGs are still running high, i hope to get a more normal day of bgs for comparison soon.

6am (Fasting) - Glucomen wins
Accuchek BG 7.3
Libre sensor 6.7
Glucomen sensor 7.3

12:50 (Bg was rising even though pre lunch) - Neither really win
Accuchek BG 10.4
Libre sensor 8.8
Glucomen sensor 7.0, didnt say 10 until 25 mins later

2:00 (an hour after lunch so had been rising) - Libre wins
Accuchek BG 11.3
Libre sensor 11.8
Glucomen sensor 10.7

5:30 (Before tea and ages after corrections, bg should be stable) - Glucomen wins
Accuchek BG 8.0
Libre sensor 9.6
Glucomen sensor 7.8

6:30 (One hour post fish and chips...) - Libre wins
Accuchek BG 11.3
Libre sensor 11.8
Glucomen sensor "temporary data interuption". Came back 20 mins later at 10.7

9:20 (4hrs post fish and chips, should be stable) - Libre wins
Accuchek BG 12.1
Libre sensor 11.9
Glucomen sensor 11.4
It's correction time... need to get these BGs sorted out!

Day 2 summary: Glucomen has definitely improved on day 1, and comes in really close to BG for several of the checks, more often those where BG is stable. I feel like libre keeps up a bit better when bg is changing.

 

[everydayupsanddowns]

These are really interesting posts @Lucyr

Thanks so much for sharing your experiment. Would you be happy for me to move this thread to the ‘pumpng and technology’ section so that people asking about the Glucomen CGM can be referred to it?

 

[Lucyr]

Yes of course Mike. Here’s yesterdays bgs for day 3. I also had a call with them and learnt a few useful things, I was calibrating before breakfast but from lunchtime yesterday I’ve changed to calibrating it before lunch as before breakfast isn’t the most steady of times of day for bgs since we can naturally rise. Also need to make an effort to stay well hydrated which I don’t normally do.

FBG 8:30am
Accuchek bg 7.0
Libre sensor 6.7
Glucomen 7.1 (overdue calibration so calibrated to 7.0 after this)

9:50 (1hr after 2 slices fruittoast so peaking now)
Accuchek 8.5
Libre 8.0
Glucomen 6.7 (it didn’t actually have me getting to 8)

12:30 pre lunch
Accuchek bg 8.3
Libre 6.7
Glucomen 6.4
Wondered if the accuchek bg was out since both sensors agreed and I should have dropped from breakfast insulin!

15:25, nearly 3 hrs post lunch so should be dropping
Accuchek bg 5.4
Libre 6.2
Glucomen 5.4. It gave a predictive low alert (you’ll be under 3.9 in 15 minutes if this trend continues) so I had a snack as this was 2 hrs into an 8 hour drive. This feature is something I really like and a plus over the libre.

17:45 Pre sandwiches and another 2hr driving check
Accuchek bg 7.9
Libre 8.3
Glucomen 7.9

20:10 2hr driving check and 2hrs after white rolls that I under bolused for
Accuchek bg 13.9
Libre 16.2
Glucomen 13.6

Day 3 summary: A real turnaround since day 1 and I’d say Glucomen is as accurate as the libre now for my spot checks. Both Libre and Glucomen had one reading that was further out than the other, but both equally useable. Today was the first day I switched on any alerts (wanted to get confidence in the readings first) and predictive low feature for me is a bonus point for glucomen.

For today and the next couple of days, the libre ends so it’s onto just reviewing glucomen. I won’t post comparisons of as many readings as above, as I’m comfortable enough with the accuracy now and going to reduce the finger pricks over Christmas. I am going to switch on some of the alarms ( low as below 3.9, high as above 13.9 and predictive low as will go below 3.9 in 15 mins, won’t switch on predictive high for now) and will report on whether I find those accurate / useful.

 

[Lucyr]

Here’s some of the alert options. You can choose for each one how often it repeats, and set each one seperately. I can’t see any options to choose the sound, I think they might just use the phone sounds but will find out that when they go off!

 

[Lucyr]

The history shows you the alarms it would have given even if you have the notifications switched off, so for yesterday it would have accurately warned me of being above the 13.9 when I was coming down from huge tea spike, and a projected low alert 10 minutes after I fingerpricked 5.4

 

[Lily123]

The Glucomen seems to be not to bad but also don’t forget that the accu chek has to be within 15% of the actual number so the Glucomen could be more accurate than it seems

 

[Docb]

Out of interest I have plotted your results above. A couple of interesting observations. In general the three measurements follow each other quite well and, although the Libra generally reads a bit higher than the Glucomen, it is not exclusively so.

Quite what the numbers mean, and their value in terms of judging the reaction to them, is open for discussion! My instinct is that the differences between the various measuring methods is more to do with statistical error than anything systematic.

 

[Lucyr]

Yes I agree, thanks for the graph. The readings have continued to be close, and I’ve reduced fingerpricks the last couple of days as confidence has grown.

I’ve had alarms switched on for the last couple of days for high, low and predictive low. As with the libre the high readings are within a point or two of bgs, not identical but always had a fingerprick in double figures when it’s gone off so a good reminder of when I’ve forgotten to bolus for something. I had the predictive low alarm go off once when I wasn’t low but I was eating toast at the time. As I’ve said before, I really like being able to set the low alarm to go off when I’m actually low (below 3.9) and have the predictive low alarm go off when I’m at risk of going below 4 in the next x minutes. It means no alarms going off if I’m 5 and stable but alarms do go off if I’m say 5 and dropping. I think the dexcom has this available? Libre doesn’t have much options for alarms though so a definite benefit.

The one downside at the moment is the alarm noises themselves. It seems to just use my phone system notification sounds. So I get the same noise for Glucomen alarm notification as I do for any other phone app sending me a notification. I’d rather something distinctive so I can tell whether it’s a useful noise or not. The libre alarm sounds were limited but at least different to other phone noises. I’ll have to read the manual for this to see if it’s changeable or not.

 

[helli]

The other thing xDrip allows is for you to set the tone/sound for alarms. I think you can set different sounds for different alarms.
And you can track your insulin on board.

 

[Lily123]

Xdrip sometimes shows amusing predictive warnings: low expected in 75 minutes.

To be honest I turn a lot of the alarms off, especially at night. 6 or 7 alarms during the night is a definite 'No.'

I have to say that Xdrip sounds extremely helpful and fair enough about turning alarms iff at night bu warning you 75 minutes before a hypo???? How???? Wouldn’t your numbers have changed so much in that time frame that warning you that much in advance makes you overly anxious??????

 

[Lucyr]

Well it lost the plot last night. I must have lay on it (feeling ill since yesterday so rubbish bgs and lots of tossing and turning trying to get any sleep) and so there was a gap in the data, but the numbers that came back in the morning made no sense. I put my phone on silent before bed so that I wasn’t woken by alarms. Photo in this one is what the graph looked like in the morning. Had been up for the loo several times in the night but did a fingerprick and was in 11s

 

[Lucyr]

I calibrated and got this message. The first time it came up I thought it said calibrate again after 10 minutes so I waited more than 10 minutes and got a wait an hour message.

After that I tried doing the calibration again and realised it actually said within 10 minutes. So I immediately retested, it must have just wanted me to double check because of the size of the difference.