Thanks
@SamJB - yes, that seems something like what I was imagining, but I agree with
@Oldvatr that simply taking the area between the horizontal line at 0mmol and the actual curve would not be a good measure.
For instance, by this measure,
10 hours at 5.5mmol = 55, while
9 hours at 4.5mmol, + 1 hour at 14mmol = 54.5
I'm guessing all of us here would prefer the former though
(and yes, for simplicities sake, ignoring for now the actual rise and fall of the curve)
I'm guessing this is what you were suggesting though with your comments at the end about what is a safe level.
This example also illustrates why average glucose isn't really an ideal metric for assessing control. Also, I'm not too clear exactly how closely HbA1c follows average glucose, but it seems it could have this problem too, as Bernstein suggests.
When looking at my libre results, I do tend to look at the areas above the target band as an intuitive measure of how good my control is. This still leaves the question though of where the line above which damage occurs actually lies, and whether the amount of damage is simply linear with height above this line.
If we set this line at 7.0, then in my original example we would get:
60mins * 1mmol above limit = 60pts
vs
15mins * 8mmol above limit = 120pts
I would still question whether this is really a good measure though. Perhaps being 4mmol above target is more than twice as damaging as being 2 above. Also it would give a score of 6pts for spending an hour at 7.1, vs a score of 0pts for being at 7.0, which clearly doesn't make sense.
One of the things that has got me thinking about all this was Bernsteins talk of nondiabetic adults mostly hovering around 4.6mmol, and that the recommended target ranges given for diabetics are often too high. I think he would consider spending the whole day at 6.9 to be damaging, though the metric above would give that a perfect score of 0pts.
His book has definitely influenced me to pay more attention to reducing my 'within range' readings as well as just trying to minimize spikes.
So maybe a better measure would be one that gave all readings above 4.6 some score, increasing with the value above this by some non-linear formula.
Now I realize that this whole discussion might seem a bit like fussing over details, and we can set ourselves reasonable targets without precise measures, such as just aiming to be around 5mmol most of the time, and to never go above 7 for long.
However, I do find personally that tracking small improvements is a useful motivational tool, and I'd prefer that the targets I use for this actually relate meaningfully to the minimization of long term damage that is the real aim.
Also, I've deliberately left hypos out of this discussion, as I think it becomes much more difficult and subjective trying to quantitatively compare the risks/damage of highs and lows, as they are of a different nature.