• Guest, the forum is undergoing some upgrades and so the usual themes will be unavailable for a few days. In the meantime, you can use the forum like normal. We'd love to know what you think about the forum! Take the 2025 Survey »

Why compare BGM and CGM Values?

B

Bill_St

Well-Known Member
Surely what really matters is the Medical Results - not the readings from either

A large Swedish study demonstrates this - This is the comparison that really matters -
The values from each device type is immaterial- what is important is the results to Health
These are not small differences ….

IMG_7080.jpeg
 
Last edited:
Interesting.
Noting in passing that this may be a little out of date, as for example the Libre and Libre 2 were intermittently scanned but the latest software for Libre 2 and 2+ gives continuous reading and there is usually no need to physically scan.

However a graph of BG continuously over day and night is more effective at understanding BG levels than just finger pricks.

IMHO as a T2 who self funds Libre 2+.
 
I think the point of the slide you show is that with a CGM one can see the behaviour in between what would otherwise be fairly widely spaced finger prick readings. I imagine the presence of alarms may well also help to avoid people from running high. It's a shame they don't give a bit more data about average number of finger prick readings for that group of people.

The other unconnected reason to compare them is to make sure your CGM is actually reflecting something close to reality and if not get it replaced.
 
Bill_St said:
Surely what really matters is the Medical Results - not the readings from either

A large Swedish study demonstrates this - This is the comparison that really matters -
The values from each device type is immaterial- what is important is the results to Health
These are not small differences ….

View attachment 70121
Click to expand...
Do you have a reference for this study, please? It looks to be from some time back.

I would have thought that testing - ie knowledge of what our blood glucose levels are and what affects them - is not an end in itself. It's what we do with the information gained that counts.
 
“Why compare BGM and CGM Values?”

How I do it is I work with the BGM to calibrate the Diabox app for the CGM?
Which actually don’t need much fettling on that front.

What happens consistently within reason for me.
… is the estimated calculated HbA1cs with the BGMs & CGM. Come home 15% (sorry edit not “lower.”) higher than the lab result.
The BGM reading always seems to lag around 20 to 30 minutes behind the algorithm estimated on a Libre 2.
However, the BG meter will tell tale after a rise from a hypo a good 30 minutes after a sub 3.5mmol low. (If it happens?)

You just gotta know how to read the “runestones…”
 
Last edited:
I agree with the value of seeing what is happening between finger prick.

However, it would be interesting to compare these results in a few years time.
I wonder if CGM being relatively new and used more by people with greater engagement with their diabetes.
It is possible, once CGM becomes even more common (and no longer needing to scan) the engagement may decline.
 
@Bill_St Can you provide the source please? Some of us would like to read the paper before commenting.

It matters a lot if this graph is in a peer-reviewed paper in a respectable journal
or a glossy brochure from a company trying to sell their merchandise.
 
KennyA said:
Do you have a reference for this study, please? It looks to be from some time back.
Click to expand...

There is probably no way around that. Since it is tracking hospital admission rates, it would not make much sense if it did extend over a few years.
For example, it is probably too early to see whether the change from manual NFC scans to continuous readings would have any effect in hospital rates.

@Lupf

I don't think this is the full article (pdf is only 6 pages), but there are more details at https://pubmed.ncbi.nlm.nih.gov/39316385/ and

It does say "Support for this study was provided by Abbott Diabetes Care."
 
sgm14 said:
There is probably no way around that. Since it is tracking hospital admission rates, it would not make much sense if it did extend over a few years.
For example, it is probably too early to see whether the change from manual NFC scans to continuous readings would have any effect in hospital rates.
Click to expand...
I was asking for the date of publication, if it existed - some things found on the internet turn out to be junk, and some turn out to be contemporary and reputably published.

According to the reference, the study was published by Diabetes Care and by PubMed in September 2024.
 
Thanks @sgm14 for the link to the source.
The study starts in 2017 and was published recently.

If I read correctly, the isCGM cohort is self-selected ("adults with type 1 diabetes in Sweden who initiate isCGM")
whereas the control sample is not well defined. It doesn't look like some T1s were randomly chosen to get CGMs and the others were told to continue with BGM. I also couldn't find anything on the make-up of the two cohorts, e.g. age profile.

Without this information there could be a significant bias, which could explain this outcome without giving all the credit to CGMs.
For illustration, if you would monitor a cohort of young athletes versus a cohort of senior citizens you would also see a significantly larger occurrence of CVD in the older cohort. While this example might look extreme, self-selection of the CGM cohort will lead to a bias, which I would like to see quantified in this paper.
 
Lupf said:
It doesn't look like some T1s were randomly chosen to get CGMs and the others were told to continue with BGM. I also couldn't find anything on the make-up of the two cohorts, e.g. age profile.

Without this information there could be a significant bias, which could explain this outcome without giving all the credit to CGMs.
For illustration, if you would monitor a cohort of young athletes versus a cohort of senior citizens you would also see a significantly larger occurrence of CVD in the older cohort. While this example might look extreme, self-selection of the CGM cohort will lead to a bias, which I would like to see quantified in this paper.
Click to expand...

Did you see the flowchart of the selection process in the pdf? Whilst it is not completely clear (or at least not clear to me), it does give some details.

But you would probably have to pay for the full report to get the answers to your questions and even then there is no guarantee.
 
Back
Top