How much is normal for the Dexcom to be off from the BG meter? I just started my very first sensor yesterday so I am still in the first 24hours, but it seems like the sensor readings are 20 points or more off of my meter results. Does it help to continue entering meter readings to the dexcom to help it be more accurate? Or should I just do the 2 per day and hope the accuracy improves? Any help is greatly appreciated.

Laramie, WY

Views: 2177

Replies to This Discussion

Everyone has their story. Here is mine. I have been wearing a Dexcom 7+ for almost two years. My self taught rules are:
1) Do a finger stick and tell Dex with every bolus.
2) If Dex wants a finger stick, do it and enter it.
3) The more finger sticks the better. Get a person with a degree in statistics to explain the math.
4) Do not do two or more finger sticks in 10 minutes or less except for the initial start-up double stick.
5) Always test fingers prepared by the book. Non-finger sites (arms, etc) risk dilution with sweat when you are hot, hypo, or sick and therefore are NOT ACCURATE.

1) My readings are close to BGs most of the time. For me, close is ±10% for each and the values should overlap. For example 120 and 100 are good (120 - 10% = 108 and 100 + 10% = 110, therefore, this example overlaps by 2).
2) Keep in mind, a BG is blood and the CGM is testing interstitial fluid - get a good grip on this and you will understand the reason for the time delays and other reasons the values may not be dead on every time. It is the TREND to work with for control.
3) When in doubt, contact tech support or the local Dexcom Clinical Manager. Learn from the people trained to know.

Hope this helps.

Folks, I just found TuDiabetes yesterday, but I've been a DexCom (Seven Plus) user for over two years, and T1DM for over thirty years (see my page).

Disclaimer: I am NOT a physician, nor a nurse-practitioner, nor a diabetic educator. I'm actually a self-employed IT guy with a background in helping build rocket ships. (I was a dual-major in college--Aerospace Engineering and Automatic Control Systems Engineering). I CANNOT and WILL NOT guarantee the accuracy of anything I relate here for any purpose other than helping others find their own paths toward less uncertainty in managing their diabetes.

That said, I can corroborate much of what others have described here. Basically, it's a mistake to view the precise number on the DexCom or on a finger-stick meter as absolute gospel. (At one time I had a graph from a clinical study on the DexCom that showed its accuracy performance over the range of BG levels it can detect and report, and that indicated--I believe--that its performance was superior to other CGMs on the market, especially in the low ranges--if I find it again, I'll post it here--DexCom does have some information about clinical results on their web site: [click here]) So we all have to rely on a combination of inputs to make decisions about carbs and insulin. Having the DexCom is not a panacea, but it sure is a vast improvement over its lack (corroborating Mike's comment).

For me, the single thing the DexCom has added that is valuable beyond measure is the ability to report BG rates--how fast is my BG level rising or falling. If you grok elementary differential calculus (...and if you don't, don't worry, just ignore what I'm about to say...), what we're talking about is the time-rate-of-change of the blood glucose level -- dBG/dt -- also known as the slope, or the shape of the curve. Even if the immediate BG value on the DexCom is off from what a clinical blood test would indicate in the same instant, the recent shape of that curve is still pretty much the same. And I, frankly, am way more anxious to know the trend than simply whether or not I'm at a low (or a high) value this moment. Not to diminish knowledge of current BG values in any way, it's just that if you know you're heading downward, but not there yet, you can do something about it (I know I'm preaching to the choir here). No tool before the DexCom has ever had this capability (though I imagine other CGMs have similar functions).

Sensor site. I have several small benign lipomas (fat tumors) on my abdomen. Once in a while I manage to get a DexCom sensor stuck into one. The lipoma tissue is not vascualized like ordinary dermal tissue, so the interstitial levels of everything, including glucose, are not what the DexCom is designed to work with. When I do hit a lipoma, I usually wind up having to toss the sensor and start a new one a few hours after it warms up (corroborating Richard157's comment).

Calibration frequency. When I first got the DexCom, I was so in the habit of doing a finger-stick at every meal, that I was calibrating it every time I did one. Then my endocrinologist reminded me that the DexCom specs only require two calibrations per day. Duh! Personally, I have not noticed any difference in the DexCom's accuracy (to the extent, of course, that I have data to judge by) between 2-a-day and 5-or-6-a-day calibrations (re jrtpup's comment). The comments about not calibrating when BG rates (up or down) are high is also something I can corroborate.

Reusing sensors. I've experimented with doubling the sensor's nominal duration (going two weeks). Mostly, it seemed to work fine, although in the second week the adhesive tends to come un-stuck around the edges. But I've also had a problem with not hearing the low alarm when I'm asleep, and that plus the variations in accuracy that do occur made me decide to go back to 1-week durations per sensor. Your mileage may vary.

So technically, are there inaccuracies in the DexCom (and for that matter, in your BG meter)? You bet. Is having the gizmos better than not having them? (<--Note: rhetorical question).
Welcome to the family James!

I've gotten up to 19 days from a sensor; the most I've heard of is 31 days. I find the second week even more accurate than the first. The lifting adhesive issue is easily solved. I use a product called Opsite flexifix. It's an extremely thiin film. I get the 4" rolls and cut a small strip for each side of the dexcom adhesive. The stuff lasts forever, through water, sweat, etc. I've had the entire dexcom adhesive lift, and the sensor last another week because the flexifix held so well.

I, too, had trouble hearing the alarm during the night. I have it set to vibrate and alarm, and sleep with it under my pillow. No way can it NOT wake you up!
Dig this. I called DexCom tech support to confirm what I already knew (as do we all)--there's no volume setting on the DexCom. I did, however, get a suggestion couched in a stern "...this is not something we officially recommend..." disclaimer: Put the receiver in a water glass on a firm surface (dresser, night table, or in my case, window sill above the bed head), with the alarm set to both beep and vibrate. It is much louder vibrating the glass!
LOL I think I'll keep it under my pillow rather than risking broken glass. I have a feeling my terriers would consider it a chase-able object!
wow that is great. it says it shuts off after 7-days, I hope that is wrong. I'm gong to try wearing mine longer than the 7-days and see what happens.
It does shut down after 7 days; it will warn you first. Go to the menu, stop sensor the start sensor. It has no idea you haven't put in a new one.
I found with the trial week the system was off a lot, seeing the same thing now, about 20 points is pretty good assumption( for me at least it is ). That doesn't bother me too much anymore, I like knowing what direction I am going in so I can take action, especially if I am dropping quickly.

I only enter the BG when it asks for it, during trial I entered more readings and it didn't seem to make a difference.

First 24 hours is not a good time to judge, mine went up to over 400 but my true reading was only in the 170's.
My son's almost always goes up to "high" on the first night he wears it, followed by a couple of hours of ??? and then dead on readings soon afterward. Dexcom said it was possibly because he's so slim that it takes more time for the sensor to "settle." I have no idea - I don't really care since we know it happens and can plan for it. : )
mine just went off, BG low, can't even read it. Meter says 69, Dexcom says 43. That is a lot. It seems to be like that for me tons of times.......
I inserted my first Dexcom sensor yesterday (Sept 1 @ about 7pm) then, today, had a lot of questions with the tech support about problems with the system being off by a significant amount. I believe, now, that I had probably managed to put the sensor into my abdominal muscle; on reflection it felt like an intramuscular injection when I did it and it continued to fell that way.

Nevertheless, I established the accuracy, at least according to Dexcom. Above 80mg/dl the Dexcom and the fingerstick reading must be within 40% of the average (20% error in the Dexcom, 20% in the BG meter), below 80 they should be within 40 (20 mg/dl for the Dexcom, 20 for the meter.) Of course this is an almost useless inaccuracy; it's the limit. Outside that limit something is broken. In any case both readings should be within 20% (>80mg/dl) or 20mg/dl (<80mg/dl - well, probably <100mg/dl for obvious reasons) of the correct value.

The problem, as other people have suggested, is that both devices are inaccurate. I know my Bayer Contour can be off by approaching 10% because I've done multiple readings that differed by about 20%.

What I saw today, however, was not within the acceptable limits. Initially at 6AM I woke up with a low blood sugar and the fingerstick gave me a BG of 59, however the Dexcom said 129. The difference of 70 is well above the permitted 40 and, as soon as I entered the fingerstick reading, the Dexcom went into Sensor Error #1. Later, at 3PM, the Dexcom reading was going rapidly north, reaching 200 at 3:11PM, but the fingerstick said 103 at 3:13PM, a difference of 97 versus the permitted 61. Once again the Dexcom went into Sensor Error #1 as soon as I entered the number. It never recovered (and Dexcom are sending me a new sensor).

Well, my second sensor is now 6 inches higher up my abdomen, where there is more subcutaneous fat. The last lady I spoke to suggested that I should be able to pinch up at least 1/2" of skin+fat (IRC), but at my first site I only had 3/8". I also took care to pinch up more skin before injecting the sensor the second time. With any luck I won't have the same problems this time round!
The first 24hours is the worst and can have differences like that. Interstitial tissue does not track with faster changes seen on

As for entries from fingertips, I enter throughout day as
i take them. Some folks argue you shoulf only enter the two. I disagree and prefer to add the few extra updates and keep unit
tracking better. Generally , I find better tracking on second day forward and usually within 10 units. That said though, should there be a
fast change you will see differences up to 20 points as unit moves to catch up on the interstitial tissue where sensor is.

I always had bad readings on the recommended gut location and better performance on arm. Why they picked gut as official
position truly escapes me.

It really is a bummer somedays as here you are with your$ 20 dollar hand held with $ 1 dollar strips versus the $900 dollars in latest trick electronics with a $ 75 probe and logic should dictate that the expensive new guy should have gold plated readings. Mounted on your finger tip - possibly but currently nope.

But overall the reading technology upgrade is extremely helpful as one gets familiar with the new technology and its limitations.
and spot checking against finger tip readings and bearing in mind interstitial tissue cannot track fast changes at fingertips.

Once a sensor is settled in after 24 hours, readings do improve but I watch closer with handhed on startup of a new sensor.




From the Diabetes Hands Foundation blog...

DHF Joins Diabetes Advocacy Alliance

Diabetes Hands Foundation is incredibly honored to join the Diabetes Advocacy Alliance, an organization with the drive and potential to affect a powerful, positive impact on diabetes and healthcare policy. Diabetes Advocacy Alliance is a 20-member coalition of leading professional Read on! →

Helmsley Charitable Trust Renews Support for DHF

HELMSLEY CHARITABLE TRUST GRANTS SUPPORT TO DIABETES HANDS FOUNDATION FOR FOURTH YEAR  Funding in 2015 to support major transitions in programs and leadership at Diabetes Hands Foundation BERKELEY, CA: February 18, 2015 – The Leona M. and Harry B. Helmsley Read on! →

Diabetes Hands Foundation Team


Melissa Lee
(Interim Executive Director, Editor, has type 1)

Manny Hernandez
(Co-Founder, has LADA)

Emily Coles (Head of Communities, has type 1)

Mila Ferrer
(EsTuDiabetes Community Manager, mother of a child with type 1)

Mike Lawson
(Head of Experience, has type 1)

Corinna Cornejo
(Director of Operations and Development, has type 2)

Desiree Johnson  (Administrative and Programs Assistant, has type 1)


Lead Administrator

Brian (bsc) (has type 2)


Lorraine (mother of type 1)
Marie B (has type 1)

DanP (has Type 1)

Gary (has type 2)

David (has type 2)


LIKE us on Facebook

Spread the word


This website is certified by Health On the Net Foundation. Click to verify. This site complies with the HONcode standard for trustworthy health information: verify here.

© 2015   A community of people touched by diabetes, run by the Diabetes Hands Foundation.

Badges  |  Contact Us  |  Terms of Service