Question about SD

[Carmen]

Could someone enlighten me, please? How do you calculate the SD (standard deviation)? You download it from your meter, or you calculate it every day?

I've come to realize that a good A1C is worthless if not coupled with a good SD.

Thank you.

[payam7777777]

I've come to realize that a good A1C is worthless if not coupled with a good SD.
very true.
we used to calculate sd using a scientific calculator. just enter the numbers into the calculator and it gives you the sd. meter softwares do that too, i guess. but i think it's not really that necessary to calculate sd very often, what is more important is having an understanding of what sd is and that we have to minimize the fluctuations in bg as much as we can.

[Riley'sMom]

If you have excel on your computer you can enter the blood sugars in a column, then use the function tool to calculate standard deviation for you.

or type '=stdev(A1:A11)' where the letter A is the column that the blood sugars are in and the numbers 1 and 11 are the first and last cells in the column that you used.

[CC'sMom]

How do you calculate the SD (standard deviation)?

I've come to realize that a good A1C is worthless if not coupled with a good SD.



I never heard of this before. What is standard deviation?

[Seans Mom]

Freestyle management software will give that information to you when you download. If you're using the freestyle meter.

[payam7777777]

What is standard deviation?
sd is a mathematical representaion of the fluctuations in bg. the more you fluctuate the higer your sd.

for example consider the two series of numbers:
50,250,60,260 and 100,200,110,210
they both have an average of 155 while the fluctuations in the 2nd seris is much less than that of the 1st series hence the 2nd one has a smaller sd.

i read in one of John Walsh's presentations that a study showed that a constant bg of 300 is less harmful than a bg that fluctuates between 100 and 300!

The standard deviation is the most common measure of statistical dispersion, measuring how widely spread the values in a data set are. If many data points are close to the mean, then the standard deviation is small; if many data points are far from the mean, then the standard deviation is large.
http://en.wikipedia.org/wiki/Standard_deviation

[mph]

sd is a mathematical representaion of the fluctuations in bg. the more you fluctuate the higer your sd.

for example consider the two series of numbers:
50,250,60,260 and 100,200,110,210
they both have an average of 155 while the fluctuations in the 2nd seris is much less than that of the 1st series hence the 2nd one has a smaller sd.

i read in one of John Walsh's presentations that a study showed that a constant bg of 300 is less harmful than a bg that fluctuates between 100 and 300!

The standard deviation is the most common measure of statistical dispersion, measuring how widely spread the values in a data set are. If many data points are close to the mean, then the standard deviation is small; if many data points are far from the mean, then the standard deviation is large.
http://en.wikipedia.org/wiki/Standard_deviation

Thank you for clearing this up (for me, too)! It took me 3x to get through statistics class in college...........:o It may as well have been Greek.

Any advice for "how" to reduce the sd? Where is the best place to try to even out the #s first? Because with hypos and breakfast spikes, we fall into the 60-300 daily category.:(

[payam7777777]

oooooooops!

[payam7777777]

It took me 3x to get through statistics class in college
same here. really.

Any advice for "how" to reduce the sd?
im afraid i dont exactly know what to say but we benefitted from testing the 1 hour post meal bg. that's of course us. i should also add that checking the 1hr post meal is highly addictive!

Where is the best place to try to even out the #s first?
my guess would be the spikes. what has worked for us is that we have a relatively very small basal (less than 1/3 of tdd) thus our boluses are bigger relatively and work like superbolus.

we fall into the 60-300 daily category.:(
oh i' truely sorry to hear that. i'd suggest you test more often. again it's very addictive but over the time it gives you a clearer/better understanding of what's going on... i mean it helps you 'learn' better from the patterns. that's pretty much all i can say. i wish i could be more helpful.

[Carmen]

Thank you all for your responses.

Payam - do you mind telling me what is your target 1 hour after meals? Is it 10 mmol (180 mg) or less?

Thank you again. I think that I need to keep a close eye on the SD from now on!

[payam7777777]

our target 1hr pp is anythin between 140-180.
if we're >180 at 1hr pp we give another extra bolus of .1 to .5 or more units deoending on the number. experience showed us that whenever we're >180 at 1hrpp we're headed to hit 200-250 at the 2 hr pp so we always do the extra bolus when we're higher than 180. it's like (and it's true that) when bgs are higher than 180 the insulin becomes less and less effective.
if we're < ~140 at 1hr pp we give some honey to Ali and thus prevent a bad low from happening because after one hour about 60 percent of the bolus is still active and it means a big (~80 point) drop in bgs which we cant afford with a bg of < ~140.

[twodoor2]

This is probably my all time favorite question from CWD!!!! I love statistics!!

Just remember a low standard deviation is not an indication of good numbers. If you have a low standard deviation (in our case anything < 50 is good, < 25 is outstanding) , but the overall numbers are high, then the correction factor is probably not correct, although the food dose might be. Case in point: If you have a day that has blood sugars at each check in the 250's +/- 10 points from that value, the numbers are bad, but the standard deviation is very low. We've had days with a standard deviation in the 20's, but really high numbers. Not a good day!!

If the standard deviation is high (> 50) , but all the numbers are back in range at meal time (when all IOB is gone from the last dose), then you have too many spikes, and this may be an indication that the basal might need a tiny adjustment, or the GI or fat intake is too high.
We calculate the following statistics each day with our spreadsheet. Here is an example of the following. Note that the % correction of TDD is 1%, and the standard devation is 48. This was a good day. Sure, there are some spikes, but it is difficult to remove all of them (especially at breakfast).

% basal of TDD=47%
Total Bolus=4.75
Total Carbs=98
% corr of TDD=1%
# TESTS = 7
Avg BG = 160
min BG = 75
max BG = 225
std dev = 48

%of #'s in range
57%

We use Excel to calculate the standard deviation through the the built in function STDEV(). We actually round it to the nearest integer. This is our example, where it calculates all the blood sugars recorded in the range of cells C384 to Z384.

=ROUND(STDEV(C384:Z384),0)

The actual mathematical calculation is a bit too time consuming to do it by hand or with a basic calculator. A function (via a calculator or a spreadsheet) is the way to go as far as calcuating it.