NB - it is not worth complaining about UK doctors' surgeries. There's a whole machine aimed at deflecting complaints

A good description of HBA1C appears here, but I will attempt to give my own description below.

Your HBA1C is your

**glycated haemoglobin**. Haemoglobin is found in your red blood cells - they're the cells that carry oxygen through the body and basically enable us to function. Your blood also carries glucose (sugar) around in it, and over time, it combines with the haemoglobin somehow (I don't understand how it combines, but that's a whole level lower than I need to go). Glycated Haemoglobin is just a measure of these "combined" cells. The act of combining takes time, so the HBA1C is seen very much as an average reading over time. To give a context, "time", here, is 1½ or 2 months, and of course it is an ongoing process. So, from one day to the next, you wouldn't expect your HBA1C to change much. As you might expect, the HBA1C number is a ratio of glycated cells to other cells.

If you're not a diabetic, your body produces insulin and combats the sugar in your bloodstream, and the amount of this glycated haemoglobin is at quite a low level - below 42 mmol/mol is thought of as "normal", although there's a small range above this (up to 48 mmol/mol) which is informally labelled "pre-diabetes". I say "informal", because it isn't a real diagnosis - it just means that the likelihood is that you'll go on to develop diabetes. When someone

*is*diabetic, there's more glucose sloshing around, more combination, and therefore that ratio is higher - speaking from experience, my HBA1C has at times been over 100 (it's a lot lower than that now).

HBA1C is quite important because this is the value used by clinicians to make diagnoses, and, if necessary, to determine what medication you have.

Anyway, having spent all this time describing HBA1C, I don't bother with it much. Instead, I prick my finger at least every day. This is something I can do in the convenience of my own bathroom, rather than having to go somewhere and give a syringe of blood. The glucometer measures everything that much more directly - how much glucose there is in the blood at a certain point in time. That's the key, though,

*at a certain point in time*. It's a spot-value.

The other thing is that the amount of glucose in the blood is not only dependent on whether you are diabetic or not, but on what you've just eaten. When you eat (or drink) something, your body takes time to digest it, so you might have raised sugar levels for some time afterwards.

The way I try and get around this is by measuring my blood sugar every day when I get up. Because I do it after a night's sleep, I know that anything I might have eaten the day before should have been well-and-truly digested, so I'm measuring a kind-of baseline value. Then I write (i.e. copy) it down in a spreadsheet. This is why I like my glucometers to connect with the computer in some way. Anyway, I end up with a bunch of spot-values - I'm almost 900 so far - and looking at past readings I'm able to see how my sugar is behaving over time. I just use straight statistics (i.e. available out-of-the-box with something like Excel) to get an idea of what my sugar does over long periods. I generally look at the last month, so the same kind of timescale as the HBA1C, and work out the average value and a measure of its variance called the standard deviation. I remember back to when I studied maths, and these values are universally-accepted ways to measure things. So I end up with average values which are as useful as the HBA1C values.

But I've pretty much mentioned all this before, in other posts, I think. I used to take readings every day, then look at periodic readings - every 3 months, say - and calculate this average value, to give me an idea of how my long-term sugar varied.There were two parts to this - identifying which values to use, and actually doing the calculation. Doing the calculation itself was trivial, but identifying which cells to use, I did manually. The trouble is, months have a variable number of days, and in any case, sometimes I've taken several readings per day, so "a month" might be 28 readings, or it might be 100 readings!.

But this weekend, I put my programming head on and developed a macro (small program) for Microsoft Excel which worked out these calculations for me. I mean, I've programmed stuff for Excel before, but 10 years ago and, even then, only to solve a specific problem. Fortunately, little had changed, other than menus etc. to find my way in. So, the result? For every data point, every day, I'm able to calculate, automatically which readings are within that last month, i.e. which readings to use and to discard, then to when calculate statistics. For every reading. So I end up with this:

The average graph, obviously I don't want that to go to zero, else I'd be dead! But it'd be nice to get down to the kind of numbers I was at before ever I was diabetic. Of course, I can get lower numbers by taking more and more insulin, but I'd rather keep the insulin dose as low as possible, so it is a halfway house. I'm hoping to be more ambitious in the future. The SD measures how much this average varies, and ideally I'd like this to get to zero. But I know how different food can cause fluctuations in the sugar (which, let's face it, isn't going to change), so I'm not sure whether that will improve at all.

I'm not be completely finished yet, though. In theory, these values should just get calculated every time I open the spreadsheet, but quite often I'm opening it and seeing:

so there may be a little more work to do.

I think what I need to aim for is something that can be calculated when I say so (for most readings, that'll only be once) then just store the value in the cell.

I think tha's it, done. I just introduced a button which, when I click it, fills the necessary columns (with numbers as opposed to a formula), so the only time the stats get calculated is when I say so. I'll have to push the button every time I import data from the glucometer, but I can live with that.

ReplyDeleteIndeed, now that I'm calculating all these stats, I can see that there tend to be mid-30s readings every calendar month, so just over 1 per day.

ReplyDelete