“I don’t take a corrective bolus if I’m at 180 or lower. My doctor and I believe that the enemy of good control can be great control… meaning if you’re at 170 and correct, then you often drop down to 70, and then treat and bounce back up. It becomes a yo-yo effect.” -Kris Freeman, on Diabetes Mine
I made a comment on the blog post
Talking with guys like Kris and talking through the numbers and what effects certain things have is probably the most useful information out there — since guys like Kris keep phenomenal records (and I don’t mean logging every BG), I mean exactly what went in the body and what went out the body and adjusting.
I’m wondering how Kris actually feels about BG accuracy on his meters — reading his comments about not correcting anything over 180. I have had the same opinion for the past decade — mainly because of the inaccuracy of the meters. In his example: 180 could actually be 140, and if you correct to 110 — you’re dropping down to 70 (add exercise on top of that and you end up bonking hard).
The tighter control you have the harder it will become to have tighter control without bouncing around (is a 65 actually a 90?) — from what I see as poor accuracy from the meters.
June chimed in with:
I agree with Cary above, though, who warns Kris that a 180 reading on a meter could actually be as low as 140. In the almost 38 years since I was diagnosed with Type I, I have had the worst control – yo-yo-ing all over the place, BG-wise, when I tried to keep my numbers TOO close to normal.
The published data that glucometers really only have +/-20% accuracy becomes concerning if you are looking for a target range of 80-110 mg/dl. Since it is far too easy to over bolus on a “false” number.
In anemic subjects in the intensive care unit (ICU), we noted up to 30% error in glucometer results compared to the laboratory value. A 30% error from 80 mg/dL yields a true value of 56 mg/dL, a clinically significant difference.
As part of the JDRF artifical pancreas project, I would hope would eventually want to address this issue as part of their pipeline. Although, they seem more concerned about combining the CGM with Pump to be more self-aware (reducing huge high and low numbers).
CGM versus a more accurate glucometer? They are for totally different purposes — for looking at trends the CGM is more effective. For a point in time analysis – I want the more accurate glucometer.
From a technical stand point, why can a normal glucometer have accuracy of less than 5%? This is a note to LifeScan, Roche, Bayer, Abbott, et al — I will pay hundreds of dollars in cold hard cash for a significantly more accurate glucometer (rather than the 5 different $20-to-$200 meters that I pay[paid] for in test strips). I’d make a gamble that there are others who would be on board for reducing lows and reducing highs by simply increasing the accuracy of our equipment.
Cary,
You are absolutely correct– I too would love to see a more reliable blood glucose meter. I have found (with a number of different, insurance-subsidized LifeScan meters) that point-in-time measurements can vary widely; I’ve seen measurements of 220 and 140 within seconds of each other, with a third measurement somewhere in between.
Clearly there are several complications involved in accurate measurements– admittedly, sometimes the fault is mine; I can only assume the most extremely off-base measurements must often be due to food residue, lotion, etc. on my hands. But I would be very interested to know, and to see studies of, variation independent of user error. Do you know of any?
Additionally, I have heard that meters are less accurate at the ends of the spectrum as well (that is, when blood sugars are very low or very high). Does this align with your experience? Do you (or your readers, or LifeScan, Roche, Bayer, Abbott, et al) know what causes the inaccuracy?
Karmel,
I am enjoying your wish list on A Sweet Life. The difference between clinical results and real work are very different. In the real world, I do not wash my hands before testing, I may have hand lotion on, etc. But even if we lived in a clinical bubble, the accuracy is still not that good. I am still trying to find infomation from the meter companies why the variance in the numbers exists — the machine, the test strips, the air? If a scientist (or PR rep for that matter) wants to chim in — it would be much appreciated.
In my personal experience the %-variance does get a little wider if you are looking at higher or lower numbers.
-Cary
http://asweetlife.org/karmel/blogs/type-1-blogs/my-diabetic-wish-list/5383/
Hi. I will be presenting at this FDA/CDRH meeting on March 17, 2010. (http://www.fda.gov/MedicalDevices/NewsEvents/WorkshopsConferences/ucm187406.htm ).
Would be so kind as to complete this brief and confidential survey http://www.surveymonkey.com/s/bloodglucosemeters on how you select and use home blood glucose meters and ask your friends, relatives and others you know who have diabetes and use meters to do the same, I would sincerely appreciate it. It’s to be used to present data at that FDA/CDRH meeting. Thank you so very much. . I am trying to gather data from patients with all types of diabetes: Type 1, type 1.5, LADA, type 2, gestational, monogenic…. It should take less than 10 minutes to complete.
All the best,
Ellen
Ellen,
From the options on your survey and the topic in question, it looks like we are on the same page as to where glucometers need to go in the future. Smaller is better, more features are welcomed. But if it was the same machine we’ve been using for the past 5-10 years, just more accurate — that would be huge!
-Cary
[...] friends of diabetics– what’s on you Diabetic Wish List? (Shout out here to some of the recent discussions and requests I’ve seen concerning more accurate glucose [...]
Cary,
You are absolutely correct– I too would love to see a more reliable blood glucose meter. I have found (with a number of different, insurance-subsidized LifeScan meters) that point-in-time measurements can vary widely; I’ve seen measurements of 220 and 140 within seconds of each other, with a third measurement somewhere in between.
Clearly there are several complications involved in accurate measurements– admittedly, sometimes the fault is mine; I can only assume the most extremely off-base measurements must often be due to food residue, lotion, etc. on my hands. But I would be very interested to know, and to see studies of, variation independent of user error. Do you know of any?
Additionally, I have heard that meters are less accurate at the ends of the spectrum as well (that is, when blood sugars are very low or very high). Does this align with your experience? Do you (or your readers, or LifeScan, Roche, Bayer, Abbott, et al) know what causes the inaccuracy?