The State of TuAnalyze, Jan-2011

Thanks for that! Now I can’t look at the map without having the Raiders Theme Music going in my head! :slight_smile:

heh heh heh. I have a couple of hats too!

Quick! Pass Me the Whip!

I’ve been thinking about this. A1C have been around for awhile and BG testing has been around for a while. I think the key question isn’t necesarily a ‘goal’ for either one of them but using them to communicate with each other, so the short term testing that leads to decent long term results can be targeted and taught. There’s lots of threads about ‘endorks’ and frustration with the medical community but I think that it may sort of be on us to tell them what to tell us so that those of us who are struggling for one of the millions of reasons that we struggle can find solutions? I’ve read a few books (Dr. B. TLAP, Pumping Insulin and one of Sherri Colberg’s books…) and, while they are all interesting, I feel like there’s a solution out there that could be ‘codified’ more readily than it is but people are all sort of on their own page and working with their own limited medical/ insurance resources which seems to be a chore in itself.

I think the A1C project is great but I think the next level would involve synthesizing A1C data and BG logs. Which is a miserable task in my opinion. I hate logs.

I hope yoou realize that this noble effort at communication is really statistically insignificant? When the sample is so small, one very good or very poor reading will really skew the results. Is there published lab data that might better represent the population? The data is further skewed by the fact tjhat individuals can choose to report or not report. Many may be reluctant to report poor results and very willing to report success. I find it hard to believe that the averages are as good as you show.

We are aware that even if the entire community were to report their numbers, it would not be representative of the entire diabetes community, only of the TuDiabetes members which, by definition are a self-selected group: people who choose to join a social network (can afford to do so, etc.) and participate on it.

This is why we need as many people from the community as possible to participate: so we can paint the most complete picture of the community that we can. We had some discussions about this very topic when we launched:
https://forum.tudiabetes.org/topics/tuanalyze-is-here
and shortly after that:
https://forum.tudiabetes.org/topics/why-do-you-sharenot-share-your

I also wonder if lab ranges vary – I have heard from people that had their A1c measured at different labs and got quite different results (around 0.5 difference). I was told that A1c’s should not be compared across labs…

Good point. Let me bring this up to the research team’s attention. Perhaps we could tag A1c entries?

I do not think this is necessary. A new wordwide standard for the A1c is on its way. In Germany this new figure is called A1c (mmol/mol). Starting this year Germany and the UK will show both the old and the new A1c on lab results. The International Federation of Clinical Chemistry (IFCC) has developed this standard to make the numbers more comparable from country to country and from lab to lab.
http://www.clinchem.org/cgi/content/full/50/1/166

The TuAnalyze Team should prepare for that. They need to offer two input fields: Ac1 (%) and A1c (mmol/mol). I think the A1c (mmol/mol) can be converted to A1c (%) so both numbers can be shown in the same graph.

Great feedback, Holger.

Interesting! I’ll watch for the change in Hungary too.

Two things.

First, I find this project not very useful because it does not include the type of diabetes that each of us has (i.e. Type1, Type2, Type1.5). It should be trivial to include that question when Tuanalyze is collecting information. As a T1, I see no benefit in comparing my A1c to Type2’s who have a very different treatment regime. And I imagine T2’s would feel the same about comparing their A1c’s to Type1’s. You could break it down even farther if desired: i.e. chart T1 by decade of diagnosis; chart only T2 that are using insulin or T2 that are not using insulin; etc. If you’re asking us all to enter information, the more information entered the better (since it is all anonymous anyway).

Second, I was looking through my old doctor records, and found my oldest A1c went back to 1983. But as has already been suggested, the “normal” range for tests differs wildly from test to test when looking at A1c’s that far back. “Normal” ranges on the tests I’ve taken range from “5.5-8.5” to “4.0-7.5” to “4.0-6.0” to “4.2-5.9”. Clearly the A1c’s from these tests are not comparable. So the suggestion that you collect values way back in times seems of little value.

Hi Jag,
About the first point, the diabetes type data is not asked for in the A1c part of the application. But it is information that is asked for in the survey:
http://tud4.us/TuAnalyzeSurvey

As for the second point, it is very good feedback and I will make sure to share it with the research team.

This is a great conversation and I thought I’d chime in on the points raised so far.

First, A1c is a great and relatively (relatively) simple measure for self-reporting about diabetes health status. A1c can be used to describe a person or population. It is optimal to have other information together with A1c – including information about diabetes type, age, treatment, self-care and treatment etc. As Manny points out, these other measures are included in the survey so we can characterize A1c and diabetes in the community. It is more complicated to map and display real-time stratified data (so A1c by type…), and we look forward to developing toward this goal.

Second, despite noting that A1c is relatively simple to measure and self-report, the measurement and meaning of A1c are not unchanging. The heads up about global measurement conventions (from Holger) is appreciated. As are Jag’s comments about needing to understand A1c by type and over time. The app collects A1c in % for this reason, but we have color coded the interface to correspond with today’s ADA guidelines (so we are using today’s lenses to look backward).

Third, the interrelationship of A1c and BG is of interest (to speak to Acidrock23) but reporting about BG is very complex over the course of a day, week, month as it’s tied to many dynamic factors so any one value has to be linked to many others measures; for this reason, diabetes care teams struggle with printouts of BG (reams of data points, nothing to guide understanding). An ambitious goal to link these.

Fourth, Dickengel raises questions of statistical significance and these are crucial to understand especially around any measures of change or comparison. Data in the slide deck note when a change is “significant” meaning outside the realm we would readily ascribe to chance.

Fifth, variability by lab is possible however even US national health surveillance systems rely on patient reports for health metrics including A1c values. We can all imagine a day when lab values will be uploaded directly into a system like this but we are not there yet.

Thanks for your questions. I hope these responses address your ideas and thoughts.

Elissa Weitzman
Co-Principal Investigator of the TuAnalyze Project

Maybe a better project than collating all BG might be a something along the lines of a worldwide attack on Dawn Phenomenon? To me, that’s the most problematic number as it 1) gets your day off to a bad start and 2) the often alarming numbers can lead you to take wild shots at things. Maybe it’s not a big deal but if even a 25% improvement in AM BG for 50% of the people, and focus on an important number might lead to overall improvement in A1C, some kind of “findings” and progress? This seems like it would be a reasonable project for say a community like Tu Diabetes? Threads or perhaps a group on the topic, results tracking, and progress? Maybe we’d need waiver forms or something though? I still think of Dr. Bernstein going to med school so he could spread his discoveries. At least for me, that one number made a huge difference.

Why should there be any difference in A1c’s between Types 1 and 2? Poorly-controlled people of either type can have sky-high A1cs, and I see that there is a 5% group here on TuD that surely consists of mostly Type 1’s (since Type 2’s are a distinct minority here). The treatment regimen makes far less difference than how well it is being adhered to and how appropriate it is.

Then there is the issue of low vs. high glycators, which is also irrespective of Type. I’m a low glycator – was diagnosed highly symptomatic with BGs mostly staying in the mid-200s and an A1c of 7.1, whereas when my A1c got to 10.7, I was comatose. Yet I have seen many instances of people talking about their A1c’s being well above that, and walking and talking and functioning perfectly well.

If we separated out all the possible confounders, we would each be a sample of one!

Those are interesting points that could perhaps further illuminate the data. I think that the TuAnaylze project links the one test that pretty much everybody flunks, or does a lot of work to pass, but only because you do a lot of work. Otherwise you would flunk too, and derive some meaning from it? Even if your A1C is “ok”, you still are pretty much in a small group and, were you not to have at least a monitor, significant #s of test strips and insulin, you wouldn’t last long. I would like to see more people of every flavor of diabetes be able to set and achieve goals. I think that the way to do that is through communities as, in many cases, we learn that people may or may not be getting what they need from their medical resources.



I think though that Jag is talking about the fact that in the early 80s at least, one A1C might not be ‘congruent’ with another one? I about freaked out when I went to college and went from 4 point something into the 8s but the doc (who wasn’t friendly…) was like ‘no, the tests are different, you need to improve though’ so then I didn’t see the doctor regularly for about 15 years.

I also am not sure I’d agree that sticking to a treatment regimen is actually the important part? I think that’s a problem for a lot of people (perhaps the PWD who aren’t hanging around here all the time?) because they stick to the regimen and it doesn’t work or the level of flexibility isn’t designed into the explanation so they keep beating their head against an impossible challenge? I think that it is critical to teach each person not so much a regimen but how to watch for and react and, eventually, anticipate the changes and get a feel for keeping an even keel?

Ok I am new here and used to achieve good results with out much trouble and no cgm. I was younger and more active. Over time I got older and less active and seemed to care less about the test. Finally something awoke in me and I am paying more attention to the test and using the CGM to help me mangae things. Question though, as you become smarter and learn how bettert o react, does repetitive success at the test get easier? I hope so. How about those with good ongoing test results chiming in.

I think what you proposes is a regimen - the patient should not be trained to do A, b, C but to watch for react and anticipate the changes in order to keep an even keel. Perhaps it is eth ever changing regimen There are so many factors to dial in that this becomes difficult and frustrating for many so they give up. Until we can get a less complex treatment ongoing support and education will be important. For me this board helps serve that purpose more than and Dr or CDE will because you can get many opinions from others who have been through the same issues.

Definitely however there is probably a “middle ground” somewhere between my “don’t ever ask the doctor about anything” and the AMA “don’t change anything withoug asking your doctor” approaches? A lot of people aren’t comfortable making changes but other people (e.g. the Lil Mama thread) are comfortable doing so, asking for feedback, being way ahead of the people providing the feedback and making progress? At the same time, what happens if well, something happens?

I’ve found that while I can sail along and do decently for a while, every now and then I will run into anomalies that require intervention. I could call the doc and get tests or perhaps write a log for a while and maybe learn things or I can just use my sense of what I’ve seen (“highs after lunch” “lows in the middle of the night”) and fiddle around with things which works ok too. It would not work well at all for someone who wants more precise directions from a doctor/ CNA/ CDE for whatever reason because they may prefer to wait 6 months to make changes to their regimen. I don’t think that’s bad unless the “off” is a ton of lows in which case, again, making some changes is more important? It appears that many of the people here are pretty comfortable doing that but I am always leery of suggesting anything to someone with a kid, other than to unplug all their video games, hide the remotes and do push ups with them. Not much about diabetes which is what they are here asking about.

My doc always says “call if you have any questions” and seems open to more communication but I don’t even bother. I suspect that the first thing they’d ask for when I called in would be logs and I’d have to hem and haw around that so that may contribute to my aversion to asking? I dunno, the whole communication thing seems sort of crazy to me, perhaps because the main person I talk to about all this is myself? muwhahahahah… :slight_smile:

Well the person you talk to seems to know what hes talking about. I had an outstanding Dr in TN at Vanderbilt University Medical Center that I met during the DCCT where I first got my pump. His focus was on teaching me how to take care of myself so that I did not need him. I wisj more would have the same attitude. Teach the basics of the things we can adjust and call if you get stuck! Otherwise we have the tools to monitor our progress. But just like avoiding stopping the car to ask for directions, sometimes I prefer to just figure it out for myself. I kinda look at the CGM as a GPS navigation system. It helps guide me down the right path!