When I uploaded my CGM to Clarity this morning, I noticed a banner link note about Dexcom temporarily removing the estimated A1c statistic from its program, effective February 11, 2017. When I clicked on the link, this notice appeared:
Since we’ve had discussions about this number in the Clarity program recently I thought I’d call attention to Dexcom’s action. I wonder what kind of feedback Dexcom has had.
I also wonder if they intend to replace the current estimated A1c with one that corresponds better with more people. I know for some of us here, the estimate corresponded well with their lab A1c while some, like me, observed a consistent gap in the two numbers.
It’s not a burning issue, but I’m curious. Is anyone aware of any additional context for this action?
I find this confusing. If they give you an average blood sugar then you can just plug it into the ADAG equation anyway. My impressions is that this is exactly what they do anyway.
Probably exactly right and that’s likely why it’s problematic… it’s nowhere near as simple as average bg = this a1c and estimating and presenting it as such is extremely overly simplistic
diaTribe published a report last summer in preparation for the FDA hearing about developing study outcomes beyond the A1c number. That report included a table that correlated a range of BG numbers as corresponding to a discrete A1c value.
The Dexcom Clarity program takes a specific average blood glucose value expressed in mg/dL and equates that to a specific estimated A1c number. How do these two items square with each other?
What I concentrate on is time in range. The A1C estimate is always 0.5-0.7% off. Which is emotionally draining. “I’ll have a 6.2% this time!” Then I see my lab A1C: 6.8% So, I deal with the reality of the trends and try to get the line as flat as possible.
A1c is just one part of the puzzle. Another equally important statistic is standard deviation. It certainly is possible to have a 6.5 A1c with a standard deviation of 100. That would be someone who is wildly out of control. If his doctor only looked
at A1c this patient would get a big ‘atta boy!’
I know that Dexcom Studio have the standard deviation statistic, but in my brief foray with Clarity I don’t recall seeing it.
Due to a move last year and I went nine months between endo appointments and the related A1c tests. For the first time I bought two Relion A1c kits from Walmart. I used one in June and one in September. They yielded 6.0% and 6.1% results, while Clarity estimated 5.4%-5.6%. In December, the doctor office point-of-care A1c test came up at 6.0%. I think the Relion intermediate numbers last year were credible.
This doesn’t surprise me. My experience has been that my actual A1C from the blood test is at least a half-point or more higher than what Clarity indicates. At my last appointment, I jumped from 6.1 to 6.7, and Clarity had me consistently at ~5.6. I know, however, that others report their A1C’s are much closer to the Clarity estimate. I am T2 and am generally in range (<150) around 90% of the time.
I’m afraid I’ve always seen the Dex A1C as kind of a feel-good number, yeah. Half-point below blood test is about right for me too. But it is still useful for comparative purposes. It’s consistently been giving me <6 but blood test is equally consistently 6.1-6.3 or so. Recently Dex came up 6.1 which is kinda depressing–means I may not be awarding myself a smiley sticker at my next endo visit in a week or two. Pretty sure it’s due to the holidays, plus a really nasty cold that, along with the weather, kept me getting out on my bike for nearly a month–that really screws things up. :-/
I’m not sure it’s a statistically valid measure though as the distribution is skewed, not symmetric. Also, being, say, 60 points below normal at 20 mg/dL means something very different than being 60 points above normal at 140 mg/dL. But still, I agree the stddev does provide some useful information as it gives a ballpark sense of how much variation there is in the numbers.
Thanks for the heads up. I like The Clarity reports very much and have had good success with my a1c’s correlating to them. I’ll be anxious to see what they tweak.
