# Re: [IP] Converting HbA1c's to average BGs

```>This only converts hbA1c TO your average bg levels. There is no way
>to do the opposite. You would have to have, literally, a continuous
>record -- minute by minute -- to get enough data to build and
>equivalent function that would get you close to the hbA1c value.
>email @ redacted

Michael, I agree with your basic conclusion (calculating expected hbA1c from
meter readings is difficult) however my reasoning is a little different (and
I am still trying to do it).

You are correct that the important number is not the average of your meter
reading. Instead what you need is a time weighted average (a very
sophisticate time weighted average in
fact).

It is obvious that a value of 350 mg/dl for 8 hours would have a greater
effect on the hbA1c than the same 350 for a 30 minute spike. With enough
meter reading you can probably deal with this issue fairly accurately.
Certainly 3 or 4 meter readings a day are not adequate to account for this
issue. Furthermore, as you said, a continuos record would be the best data
to deal with the issue. However, I believe that 10 -12 reading a day (which
a lot of us pumpers are taking) may be enough data for a reasonable
approximation. In fact, in light of the second (greater) problem more than
10-12 reading may not substantially improve the resulting accuracy.

This greater (i.e.. harder to deal with) problem is that the time weighted
average must account for the fact that 350 mg/dl for 3 hours today affects
the hbA1c much more than an equivalent 3 hours at 350 mg/dl 2 months ago
will. This is because the high today (on the day the hbA1c is measured)
exposes all of your hemoglobin to the high levels of glucose while only a
portion of the hemoglobin that was in your blood 2 months ago is still there
now (for the new hbA1c measurement). In other words the time weighted
average has to account for the fact that glucohemoglobin is continually
being removed from your system and the longer you wait to  take the
measurement the less important any previously generated glucohemoglobin
becomes.

Unfortunately, the "proper" time weighted average to use for this
calculation is not readily apparent. I've searched the literature on this
issue and most of the models that people talk about appear to be only
educated guesses. That is actually not at all surprising because the factors
that effect the rate of removal of hemoglobin from one's system are probably
very diverse and very individual.

With all that being said, I will continue to "play with" my numbers to try
to generate an "average" blood glucose value with is useful for predicting
my hbA1c level. After all spreadsheet calculations are very cheap and as you
can probably tell I obviously have more free time on my hands than I should!
----------------------------------------------------------
for HELP or to subscribe/unsubscribe, contact: HELP@insulin-pumpers.org
send a DONATION http://www.Insulin-Pumpers.org/donate.shtml

```