Yeah, exactly Terry. Trying to glean meaning, at all, from the presentation of a single sample as representative of "normal" in a scientific paper was a heavy lift.
I suppose I couldn't argue with the notion that a single individual represents the "ideal normal", but, clearly, there is a larger range of "normal" than demonstrated by that single individual.
Before I had diabetes(that I knew of) or hypoglycemia(before diabetes) I sometimes had bg in the 60's. I also read that women frequently have lower bg. I think the range is a bit wider than we think for pwod. I like that acronym btw :)
I also have read since being on insulin that insulin needs are the lowest at 3am, I often see a drop then and I would think pwod may have that too.After many years, and many awful overnights, my CGM is set to alarm at 80 to avoid problems. If I could insure that my BG would flatline at 70 all nite-OK, but I cannot. I have had a lot of really bad overnights. Personally, in a perfect world, I want to sleep at 100-120.
I have been encouraged to increase my A1C levels, and have been pretty unsuccessful, as I am completely trapped in the numbers. My endo, PC physician and the folks at the Joslin Clinic tell me that research shows that any A1C below 7 is OK and keeping it between 6 and 7 is best. I have had docs refuse to see me with a 5.1 A1C, as that is a very scary number. The lows can be devastating, hard on the brain, and really nasty.
I know this idea is controversial here, as numbers are....important. I just do not want to buzz all night because my CGM is on high alert, be visited by the paramedics, or put my husband into an awful and really difficult overnight. Honestly, I just want to sleep.
Hi Terry, the software is accessible for everyone. It is our Glucosurfer project. The real challenge is to export the data from the pump and to import it via upload to your personal log within the Glucosurfer.
You are correct about the yellow area and its percentiles. The AGP is really an amazing tool. The folks that worked on this concept have done a good job to reveal the weak spots regarding I:C or a mix of I:C and basal rate adjustments for pumpers. Because we have only one very powerful server at hand we calculate the AGP for every day and every user and save the data to our database. We decided to use a 30 day window for these calculations. This way we wanted to reduce the negative influence of illnesses or holidays. The AGP should help to identify the hours that might benefit the most by adjusting the I:C or basal rate. In a 14 day window it might be too soon for these adjustments. In a clinical setup it might be helpful to be more aggressive - the AGP is aimed at clinical personal. But for the Glucosurfer we wanted this adjustment process to be smoother.
Your analysis is also correct. Some lows in the afternoon created some rebounding highs. But there is more to this because you see the influence of my last holiday here. We walked 90km on the beaches of the Baltic Sea in rough to nice weather. I had to reduce my dosages significantly and the physical actitivy reduced the speed of digestion causing some late spikes around 18:00 to 19:00. The activity has also ignited the liver to increase its dumps of glucose. I am still very happy that the A1c was just 6.2 afterwards. Our system calculated 6.1 and this shows how good our predictions can be - at least for the inner workings of my body.
I have not fully utilized the AGP yet. But with the current rise in temperatures - our Whitsunday has set a new all time record - I am about to make some adjustments based on the AGP. I will see how this will turn out.
I think your software and the way the software itself handles data is fantastic.If I left you with another impression, I do apologize.
If it were written into a smartphone app, I'd definitely use it.
Is it?
To me your remarks did not come across as negative. I am still thinking about what "normal" really is - a valid point you have made. In a way our tools - the meters and cgms shape our reality - even for researchers that rely on their output. As an example cgm devices might apply algorithms to suppress outliers they measure. On top of that different methods of analysis will come to different conclusions as to what "normal" looks like.
The Glucosurfer is an App but it needs the Glucosurfer server for analysis. The difference to many cloud based services is that your data is truely anonymous to us. We do not collect any personal data so we can not connect the information we get from Gooogle, Apple or Microsoft with any account in our system. Since I have coded all of those Apps in my spare time there is always room for improvements. For example the reliability of the Android app is weaker than that of the App for Apple or Microsoft. Or the import of Animas xml data is much more reliable than that of Medtronic cvs files. Still I think we have something to offer that is different from other "more professional" offerings. Hopefully it is enough for your specific needs. But even if not - just let me know what we should improve in your opinion.
Thanks Holger, I will give it a concerted effort this summer!
"Gluco-normal".
How about gluco-boring?
That makes my D seem exciting... after all people love Roller Coasters!
see where the 4-8 a.m. dashed green line, representative of the 10th percentile, dips below 70I found that fascinating in light of the evil Dawn Phenomena that so many of us battle at that same time of the day.
I don't particularly think standard deviation is adequate at all to measure variation for usAgreed. Binning creates a far better picture of what's going on.
I really like the Dexcom approach in their companion software. Different ranges are binned for how long you're in that range for a set of 24 hour periods (in the case of the G4, 90).
As Holger points out, simple averages can be skewed terribly by a few bad excursions up into the 200-300 range, when you spend 90% of your time 80-100.
Holger, is anyone on the team working on direct download from a G4 receiver?
I'm playing with it myself on Android, so could contribute to the project if you all are interested. There's a 90-day moving window of data at 5 minute intervals -- I would imagine a goldmine for the Glucosurfer software.
If no one's touched this for the project, and you all are interested in some help a long those lines, send me a PM.
I don't know how controversial higher, but still "normal" (at least non-damaging) numbers are.
While there's quite a community of tight-controllers here, I hardly think any of us would criticize any target under 100 over night. If you're prone to hypos over night, I'd shoot for 90-100, and set the CGM to alarm at 90.
Currently we can only process the xml file that can be exported with the dexcom software. A direct upload to the Glucosurfer would be fantastic. Dexcom did a good job with their software. But I think some additional ways to analyze, present and share the data would be useful. I will contact you. Thanks for your offer!
Posted a thread yesterday about how much I love my endo. She's been in my camp from day 1.
It might have worked out differently if I hadn't arrived for that first appointment so well prepared. I had already done extensive research, had a fat file, and after the initial 60 seconds or so of intro, I took charge of the conversation and laid out, in detail, with medical research references, what my plan was.
She's been on-board with my tighter-than-tight goals from the start. While I'm not there yet, my performance over the last year has alleviated any concerns she had, and today I really feel like my relationship with her, my GP, CDE (who I basically never see -- don't need to), et al is as a co-equal member -- not as the subject.
On the other hand, if a patient shows up and starts talking gluco-boring* goals without demonstrating some significant competence in the details of the science, risks, strategies, etc. etc., I can understand how a doctor, with limited time, is going to conclude they're "just another one".
Then go for the least risky treatment proposition.
Yes, I've already got an experimental app running on my Android phone connected directly to the Dexcom receiver via a USB OTG cable, displaying readings live from the device.
The rest is just coding to download the entire 90-day history.
I have this idea to build a little module with bluetooth and a USB port with a microcontroller to plug in to the Dexcom as sort of a "sidecar" module that will then communicate with a smartphone via BT. While the user would still have to carry around the G4 reciever with them, it could be left in a pocket, backpack, purse, etc. and never bothered with as all data is accessible from the phone.
Further, because we have access to the full 90-day history, a smartphone app can be made far more functionally rich than the rather limited feature set on the dexcom reciever.
All I need to get all this done is time. I'm COO at a small tech company here in the San Jose, CA area; I usually have no more than 2-3 hours a night before bed, weekends are dominated with family (wife, 2 kids, 3 dogs, 1 salamander, 2 Guinea Pigs). So the project moves slowly. If it wasn't for that dang salamander... :-)
I'd be happy to contribute to the Glucosurfer project and get some reuse leverage out of the engineering I'm doing. The only caveat is I probably won't be able to put any serious time in to it until about September.
Why did the term "Muggles" cross my mind just then?
We used that term for a while as well
That really is a spectacular looking profile for sure Terry, gluco-normal peeps don’t know how lucky they are if that’s truly representative. My BG tends to run low at night too (I’m not on insulin), I still have some beta cell function and good fastings, just not much first phase kick to control post prandial levels. Thx for the link.
Just don't call us Mudbloods, please.
LOL - Muggles works ! I wouldn't dream of using the other word.