I want to convert my 30day average numbers in my spreadsheet to predicted HBA1C %'s - can anyone give me the formula?
xyzzy
Well-Known Member
- Messages
- 2,950
- Type of diabetes
- Other
- Treatment type
- Diet only
- Dislikes
- Undeserving authority figures of all kinds and idiots.
=(BG*18.05 +46.7)/28.7
Gives old HbA1c as a % and BG is the cell with your average BG
=(HbA1c-2.15)*10.929 converts the HbA1c % to the new IFCC HbA1c
I actually do a weighted 30 day and 90 day avg and guess that 80% comes from 30days and 20% from 90 day.
Gives old HbA1c as a % and BG is the cell with your average BG
=(HbA1c-2.15)*10.929 converts the HbA1c % to the new IFCC HbA1c
I actually do a weighted 30 day and 90 day avg and guess that 80% comes from 30days and 20% from 90 day.
Grazer
Well-Known Member
- Messages
- 3,115
Just remember 2 things:-
1) the formula works for an average glycator. Some people have a higher HbA1c than others from the same average blood glucose level - it can be quite considerable.
2) For it to work, uyou must have a correct average blood glucose reading. This isn't just adding up your readings and dividing by the number of readings - you don't know what's happening in between those readings. Some people use a separate formula to work out the average - that's a bit of mguesstimation though, and if they're accurate with it uit's more luck than judgement half the time. I use a mathematical model from graphs I construct, and I'm normally within about 10% of the right answer when I get my lab HbA1c - for me, that's within 0.5 to 0.8 on the old %age measurement.
1) the formula works for an average glycator. Some people have a higher HbA1c than others from the same average blood glucose level - it can be quite considerable.
2) For it to work, uyou must have a correct average blood glucose reading. This isn't just adding up your readings and dividing by the number of readings - you don't know what's happening in between those readings. Some people use a separate formula to work out the average - that's a bit of mguesstimation though, and if they're accurate with it uit's more luck than judgement half the time. I use a mathematical model from graphs I construct, and I'm normally within about 10% of the right answer when I get my lab HbA1c - for me, that's within 0.5 to 0.8 on the old %age measurement.
xyzzy
Well-Known Member
- Messages
- 2,950
- Type of diabetes
- Other
- Treatment type
- Diet only
- Dislikes
- Undeserving authority figures of all kinds and idiots.
Yes Grazer is right(ish) imo.
While just taking a 30 day average should get you quite close to increase accuracy you could mimic the testing regime of the 1000 odd diabetics who were used to build the HbA1c model. They tested 7 times a day doing a wake up then a before and two hours after every meal regime. As my wake up and before breakfast are the same I do a before and +2 hours after plus a +4 hour before bed. On top of that there is then a weighting to be applied. Effectively for even more accuracy you should be taking a 30, 60 & 90 day average and using a proportion out of each (can't exactly remember) to work out an overall weighted average. The more readings you have the better. My son who told me all this and implemented in his own spreadsheet reckons he gets within +/- 0.1% of his true HbA1c. However as Grazer said everyone is different so obviously those initial 1000 people didn't all register the same results! What they did was plot their scores and then did a regression analysis to work out the line of best fit on the data. So if you are close to that best fit line and mimic the same testing environment you should get very close to your actual HbA1c. If you are further away from the best fit line then you wont be so accurate no matter what you do. The best is to get one real HbA1c and then feed any discrepancy back in to your own formulas as an additional "me" weighting.
I don't do the 30/60/90 weighted average as to me my diet is pretty constant and my average levels are now only falling very slowly so for instance my 30 day average is now 5.2 and my 90 day average is now 6.0. In a few weeks time my 30 and 90 day average will likely end up the same so if we assumed they end up at 5.2 then it doesn't matter what percentage of the 90, 60 or 30 day average I take unless in the future I go mad and ignore my control for a period of more than a few days. So at the moment I take 20% of the 90 day and 80% of the 30 day as a good guess. Before I got to having 90 days of readings I took 80% of my 30 day and 20% of my overall average. Will see how good a guess in early April!
While just taking a 30 day average should get you quite close to increase accuracy you could mimic the testing regime of the 1000 odd diabetics who were used to build the HbA1c model. They tested 7 times a day doing a wake up then a before and two hours after every meal regime. As my wake up and before breakfast are the same I do a before and +2 hours after plus a +4 hour before bed. On top of that there is then a weighting to be applied. Effectively for even more accuracy you should be taking a 30, 60 & 90 day average and using a proportion out of each (can't exactly remember) to work out an overall weighted average. The more readings you have the better. My son who told me all this and implemented in his own spreadsheet reckons he gets within +/- 0.1% of his true HbA1c. However as Grazer said everyone is different so obviously those initial 1000 people didn't all register the same results! What they did was plot their scores and then did a regression analysis to work out the line of best fit on the data. So if you are close to that best fit line and mimic the same testing environment you should get very close to your actual HbA1c. If you are further away from the best fit line then you wont be so accurate no matter what you do. The best is to get one real HbA1c and then feed any discrepancy back in to your own formulas as an additional "me" weighting.
I don't do the 30/60/90 weighted average as to me my diet is pretty constant and my average levels are now only falling very slowly so for instance my 30 day average is now 5.2 and my 90 day average is now 6.0. In a few weeks time my 30 and 90 day average will likely end up the same so if we assumed they end up at 5.2 then it doesn't matter what percentage of the 90, 60 or 30 day average I take unless in the future I go mad and ignore my control for a period of more than a few days. So at the moment I take 20% of the 90 day and 80% of the 30 day as a good guess. Before I got to having 90 days of readings I took 80% of my 30 day and 20% of my overall average. Will see how good a guess in early April!
xyzzy, have you ever been tempted to buy an A1CNow meter and test your own HBA1C ? I know we get the test free on the NHS but it's only every 6 months and quarterly would be nice.
Grazer
Well-Known Member
- Messages
- 3,115
There was a thread on the A1CNow meter, and I recall I think that most people found it quite a bit different to the real lab result so didn't carry on with it (it's quite expensive I believe per test). There was one poster who used it almost weekly I think.
xyzzy
Well-Known Member
- Messages
- 2,950
- Type of diabetes
- Other
- Treatment type
- Diet only
- Dislikes
- Undeserving authority figures of all kinds and idiots.
swimmer2 said:
No - I did read about the home kits but they looked expensive and not particularly accurate. I'd be more inclined to buy Cholesterol home test kits but again a recent review showed they weren't that accurate either!
I thought you could get quarterly HbA1c tests or have I finally found something my surgery is actually good at! I got told I couldn't redo my HbA1c until 3 months as the local hospital would reject doing it as it would be too soon. But have the form for the blood test all ready and waiting.
Actually going to bide my time for a couple of weeks. I could do the test now as 3 months just passed yesterday. I want to try and make sure I end up with a score less than the evil Nurse Nightshade so I can rub her nose in it! Bit of an ask as I'll need to drop from 11.3 to sub 5.4 but going for it. Seems to be one good thing of having a wife whose gone all ULC on me :lol:
