Meter Reading Quality - or is it?

Go Back   The Diabetes Forum Support Community For Diabetics Online > Diabetes Forum Start Here > Diabetes Forum Lounge

Diabetes Forum Lounge The Diabetes Forum lounge is the more social area of the community. Please feel free to drop in, and discuss anything that’s on your mind. Use this area for off topic discussions, making friends and being social.


Like Tree7Likes

Meter Reading Quality - or is it?


Closed Thread
 
Shared Thread Thread Tools
Old 02-12-2012, 10:33   #1
Moderator
 
John.in.France's Avatar
 
Join Date: Sep 2010
Location: Arleuf, France
Posts: 5,468

Member Type
Type 2
Diagnosed in 2010

5971 likes received
1054 likes given
Default Meter Reading Quality - or is it?

On the forum many of us present our morning meter readings and we all treat the figures as absolutely reliable despite the paperwork which comes with the test strips showing that (in my case), only slightly less than 60% of readings are specified as being within 5% of reality.

Just out of interest, I've knocked up a little graph of my numbers in 2011. The blue dot represents the actual reading and in 2011 and the red blur, the range allowed.
The Diabetes Forum - Building a New Community to Serve Diabetics Online - John.in.France's Album: My Graphs - Picture

Minor point - the graph uses French units - grams per litre. Just ignore the decimal point and the numbers will make more sense to most of you!

I averaged 85 on my meter. My lowest was 65 and my highest 100. In reality, that meant I might have gone as low as 54 or as high as 125.

However, in fairness I doubt my figures were as wildly out as that. Even now 54 would give me hypo signals (which I didn't have) and I'm happy to say that on the one occasion (in the afternoon) when I was at 125 (not sure why), I felt distinctly uncomfortable confirming Jenny Ruhl's suggestion that hyper feeling do start lower as you get used to normal values. Using the Nathan formula my HbA1c figure of 5% correlates reasonably with the figure on the meter (suggests 5.1%). Using the DCCT formula, my 5% fits with 97 - 12% higher than my meter but within the margins allowed. Which is right? I don't know and don't really care. The meter is doing its job of showing my pattern reasonably well.

So, anyone who is trying to cross check meters against each other, or who worries when they get a number 5 to 10 points higher than they expect - remember this picture and stop worrying!

John


jwags, Shanny and silvertiger like this.
__________________

View John.in.France's full Diabetes Forum profile here.

Last edited by John.in.France; 02-12-2012 at 11:22.
John.in.France is offline  
Old 02-12-2012, 12:54   #2
hans
 
hans's Avatar
 
Join Date: Feb 2012
Location: Sweden Southern district (län) Skåne (Eng. Scania)
Posts: 1,411

Member Type
Type 2
Diagnosed in June 2011

1387 likes received
658 likes given
Default

Quote:
Originally Posted by John.in.France View Post
On the forum many of us present our morning meter readings and we all treat the figures as absolutely reliable despite the paperwork which comes with the test strips showing that (in my case), only slightly less than 60% of readings are specified as being within 5% of reality.

Just out of interest, I've knocked up a little graph of my numbers in 2011. The blue dot represents the actual reading and in 2011 and the red blur, the range allowed.
The Diabetes Forum - Building a New Community to Serve Diabetics Online - John.in.France's Album: My Graphs - Picture

Minor point - the graph uses French units - grams per litre. Just ignore the decimal point and the numbers will make more sense to most of you!

I averaged 85 on my meter. My lowest was 65 and my highest 100. In reality, that meant I might have gone as low as 54 or as high as 125.

However, in fairness I doubt my figures were as wildly out as that. Even now 54 would give me hypo signals (which I didn't have) and I'm happy to say that on the one occasion (in the afternoon) when I was at 125 (not sure why), I felt distinctly uncomfortable confirming Jenny Ruhl's suggestion that hyper feeling do start lower as you get used to normal values. Using the Nathan formula my HbA1c figure of 5% correlates reasonably with the figure on the meter (suggests 5.1%). Using the DCCT formula, my 5% fits with 97 - 12% higher than my meter but within the margins allowed. Which is right? I don't know and don't really care. The meter is doing its job of showing my pattern reasonably well.

So, anyone who is trying to cross check meters against each other, or who worries when they get a number 5 to 10 points higher than they expect - remember this picture and stop worrying!

John

Well, I don´t understand all of what you are trying to explain. I´d probably catch more of it if it was written in my mother tounge.
But obviously there are some margins of error, but in the long run those divergences will have less impact if you have a long series of figures. Elementary statistics or have missinterpreted your message?

hans is offline  
Old 02-12-2012, 13:22   #3
Moderator
 
John.in.France's Avatar
 
Join Date: Sep 2010
Location: Arleuf, France
Posts: 5,468

Member Type
Type 2
Diagnosed in 2010

5971 likes received
1054 likes given
Default

Quote:
Originally Posted by hans View Post
Well, I don´t understand all of what you are trying to explain. I´d probably catch more of it if it was written in my mother tounge.
But obviously there are some margins of error, but in the long run those divergences will have less impact if you have a long series of figures. Elementary statistics or have missinterpreted your message?
No, I don't think you've missed the point.

As you say, the more data you have from the same meter, the more reliable will be the pattern it shows.

On the forum we have had several discussions where members were trying to cross check one meter against another. The results they obtained were worthless!

Others were worrying when they got two readings in quick succession showing a 10% difference. Again, a worthless comparison!

Sorry my English tends to become verbose when I'm ranting!

John

__________________

View John.in.France's full Diabetes Forum profile here.
John.in.France is offline  
Old 02-12-2012, 13:33   #4
Senior Member
 
Join Date: Jun 2009
Location: North East Ohio, USA
Posts: 4,368

Member Type
Type 2
Diagnosed in 2007

1774 likes received
1507 likes given
Default

I do think meters just give us a glimpse into what your bg is that second. If you took it a minute earlier or later you could have had a completely different number. I do think some people get totally obsessive about that number, but forget all the variables. I have started to cut back on testing especially in the morning. I feel the 2 hour number after meals is a much more important number.

__________________
diagnosed Feb 2007
metformin-2550
low carb organic diet
3 T of Coconut Oil daily
moderate exercise- mostly walking 3-5 miles, tennis - 4 hours per week, weight training, hiking on weekends and vacations 7-10 miles

HbA1c 5.3 3/11
HbA1c 6.1 5/12

Age 62
jwags is offline  
Old 02-12-2012, 13:44   #5
hans
 
hans's Avatar
 
Join Date: Feb 2012
Location: Sweden Southern district (län) Skåne (Eng. Scania)
Posts: 1,411

Member Type
Type 2
Diagnosed in June 2011

1387 likes received
658 likes given
Default

John!
Ok, I think I get the point. Also found the Nathan formula explained on the net. Verbose? Had to look it up for a translation. Just a bit verbose, but I guess you have an academic background and of course that has an impact of the exposition. But I don´t care.

hans is offline  
Old 02-12-2012, 13:46   #6
Moderator
 
David Burke's Avatar
 
Join Date: Dec 2011
Posts: 4,319


3729 likes received
Default

So to put it simply:

All he did was take a whole bunch of readings and confirm that the values fell within the manufacturer's specifications. And cross checking some of the more extreme values against how he was feeling he found more correlations.

An interesting experiment and data collection.

I have my own observation, if I may. Considering my job entails performing lots of trend analysis, I'm not sure you came to to correct conclusion. When I see your graph I see your readings are within the center of the accepted range from your meter's specifications. I would expect this when someone applies tolerances to readings. The only way to truly verify your hypothesis would be to take a known value of BG , say 100 mg/dl, and compare multiple readings to the state meter accuracy, say 20%.

From what I see, you just posted your own trends. Trends which I'm JEALOUS of!! I wish mine were that flat. I'm ranging from 230-52. Admittedly I'm experimenting with foods again so that is acceptable...for now. Still, I want mine to be that stable.

Just my observation and conclusion based on why I see.

Sent from my iPhone

David Burke is online now  
Old 02-12-2012, 13:58   #7
Moderator
 
John.in.France's Avatar
 
Join Date: Sep 2010
Location: Arleuf, France
Posts: 5,468

Member Type
Type 2
Diagnosed in 2010

5971 likes received
1054 likes given
Default

Quote:
Originally Posted by David Burke View Post
So to put it simply:

All he did was take a whole bunch of readings and confirm that the values fell within the manufacturer's specifications. And cross checking some of the more extreme values against how he was feeling he found more correlations.

An interesting experiment and data collection.

I have my own observation, if I may. Considering my job entails performing lots of trend analysis, I'm not sure you came to to correct conclusion. When I see your graph I see your readings are within the center of the accepted range from your meter's specifications. I would expect this when someone applies tolerances to readings. The only way to truly verify your hypothesis would be to take a known value of BG , say 100 mg/dl, and compare multiple readings to the state meter accuracy, say 20%.

From what I see, you just posted your own trends. Trends which I'm JEALOUS of!! I wish mine were that flat. I'm ranging from 230-52. Admittedly I'm experimenting with foods again so that is acceptable...for now. Still, I want mine to be that stable.

Just my observation and conclusion based on why I see.

Sent from my iPhone
You're quite right - these figures are simply showing that over the year my blood glucose remained in a range consistent with my HbA1c results and that my meter seems to be consistent.

What is its margin of error? Who knows, and I'm not prepared to invest in the effort needed to confirm it using the approach you suggest.

These figures are fasting values - the only ones the quack seems to care about. Like JWAGS, I think the post prandial 1 and 2 hour figures are more important to us as indicators of how we are coping with meal carbohydrates. However, given the attitude of the French medical service to supplying Type 2s with strips and the fact that over the last eighteen months my OH has managed to nail my diet down pretty well, I've decided that I can accept the risk of settling for the fasting test on its own for now.

If the numbers for my fasting, or my HbA1c start to slip, I reserve the right to re-instate post meal tests!

John

__________________

View John.in.France's full Diabetes Forum profile here.
John.in.France is offline  
Old 02-12-2012, 14:00   #8
Senior Member
 
Join Date: Sep 2011
Location: SW Arkansas
Posts: 3,646

Member Type
Type Pre
Diagnosed in 2008

3635 likes received
1231 likes given
Default

Your graph is showing your measured BG (blue dots) and the +/- expected range of error (red shaded area). Am I wrong or are you always graphing error based on your meter reading always being in the middle of the range of error? if meter reading is 20% high that should plot at the top of range and vice versa for 20% low

Is 20% error +/- 10 or +/- 20?

__________________
Told I was Type 2 lean on 2/13/12.
a1c 8.8 (8/18/11) 7.1 (11/8/11) 6.6 (2/13/12) 5.5 (5/22/12) 5.4 (8/20/12) 5.9 (9/18/12) 5.5 (2/28/13) 5.5 (4/22/14)
Douglas Labs Ulta Preventative X Multi-vit
Extra D, E
Seal Oil
L-Arginine
Sodium and Potassium Bicarbonate
Magnesium Citrate
Apple Cider Vinegar
Biotin
59 YY

Love the LCHF diet. The cheese goes well with my whine

updated 5/5/14
mbuster is offline  
Old 02-12-2012, 14:05   #9
Moderator
 
John.in.France's Avatar
 
Join Date: Sep 2010
Location: Arleuf, France
Posts: 5,468

Member Type
Type 2
Diagnosed in 2010

5971 likes received
1054 likes given
Default

Quote:
Originally Posted by mbuster View Post
Your graph is showing your measured BG (blue dots) and the +/- expected range of error (red shaded area). Am I wrong or are you always graphing error based on your meter reading always being in the middle of the range of error? if meter reading is 20% high that should plot at the top of range and vice versa for 20% low

Is 20% error +/- 10 or +/- 20?
My reading of the leaflet supplied suggests that the meter reads within 20% of the truth. I've added an explanation of the assumptions to the caption since the text on the graph itself is difficult to read.

So, I have my reading and it might be as high as 120% of the truth or as low as 80% of the truth. So a blue dot on 80 gives me a red line range of 67 to 100. Where on that line the truth lies? I don't know.

John

__________________

View John.in.France's full Diabetes Forum profile here.

Last edited by John.in.France; 02-12-2012 at 14:12.
John.in.France is offline  
Old 02-12-2012, 14:23   #10
Moderator
 
David Burke's Avatar
 
Join Date: Dec 2011
Posts: 4,319


3729 likes received
Default

Exactly the problem. We don't know where the true value is. We know an approximate range. We can deduce from the readings which foods, diets, portions, etc cause adverse reactions, which is the underlying goal. But when it comes to things like insulin dosing, OMG is it off!! I could be right on, use too little and stay high, or overcorrect and go hypo! All because of the starting error of my meter. And las week I did all three. Same meal, same dose of nph. But because starting level of BG was off I didn't get right correction. But that's for another thread .

I think someone needs to develop a better way to calibrate the meters. The test solution itself has too wide of a range, and as John pointed out, my idea is cost prohibitive, especially when meters are essentially free.

Perhaps what we need IS the universal test strip. An ISO standard would be adopted and every manufacturer of test strips or meters would have to conform to the ISO. Perhaps it's not reinvent the wheel, just change the tires, to get better performance.

Sent from my iPhone

Shanny likes this.
David Burke is online now  
Closed Thread

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
HbA1c test inconsistent with meter reading jwags Diabetes 9 03-15-2013 09:59
For weight loss, quality of food trumps amount: study fgummett Diabetes Diet and Nutrition 2 07-02-2011 05:47
BG reading of 416 after the gym?? dallascowboy Diabetes Complications 19 02-03-2011 02:57
BG meter reading and beans dit Diabetes 5 08-02-2010 16:56
Handy Insulin Case - Quality Leather Case for Diabetics and Diabetic Supplies HandyInsulinCase Diabetes Medication and Supplies 1 01-16-2010 19:58

LEGAL NOTICE
By using this Website, you agree to abide by our Terms and Conditions (the "Terms"). This notice does not replace our Terms, which you must read in full as they contain important information. You must not post any defamatory, unlawful or undesirable content, or any content copied from a third party, on the Website. You must not copy material from the Website except in accordance with the Terms. This Website gives users an opportunity to share information only and is not intended to contain any advice which you should rely upon. It does not replace the need to take professional or other advice. We have no liability to you or any other person in respect of any content on this Website.


All times are GMT +1. The time now is 05:20.




Powered by vBulletin®
Copyright ©2000 - 2014, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO