Cookies

A few days ago we had a going-away party for a colleague. I'm not exactly sure how it happened, but it fell to me to set up the cookie tray during the party prep...

"Pssstt."
"Huh. What?"
"Watch this," they wispered.
"Watch what?"
"We'll make the diabetic put out the cookies."
"Hehehe... Cool. It's fun to watch him twitch in self-restraint."
"Yeah...Hehehe. I can see the veins start to buldge on his forehead already."

...Setting up the cookie tray entailed nothing more than opening the pack of cookies and arranging them on a tray.

Nothing to it, right?
Nothing fancy, either.
Just your standard-issue Chips-Ahoy!
Just spread them out around the perimeter.
And put some Reese's Peanut Butter Cups in the center.
To make it look nice.
That's all.

I haven't had a sweet since before thanksgiving.
Let me tell ya: I almost stuffed eight

Kinda Like This

close

of those little devils in my mouth while no one was looking.

But I didn't.

I went the very next day to have my blood drawn for my A1c.

I've made it.
No cookies, no cakes, no pies, no ice cream.
No candy, no doughnuts, no pop tarts, no brownies.
None through the holidays, none so far this year.

I'm pretty damn proud of myself.

Now, I sit and wait for the results to come back
(note to self: make an appointment with endo to get A1c result).

But I'm pretty impressed also with the remarkable effect this no-sweet diet has had on my blood sugars. See here!

Here is the table of averages from November:

November Stats
days	mg/dl	# tests	tests/day	S.D.
7	149	89	12.7	62
14	149	182	13.0	62
30	143	402	13.4	60
60	155	799	13.3	67
90	154	1,229	13.7	68

And now here is the data as of this week:

February Stats
days	mg/dl	# tests	tests/day	S.D.
7	136	88	12.6	57
14	132	202	14.4	56
30	130	436	14.5	51
60	139	859	14.3	61
90	137	1,335	14.8	60

Pretty cool, huh?

My average blood sugars have fallen about 15 mg/dl - pretty cool.
My standard deviation has also decreased by about 5 mg/dl - also pretty cool.
I'm testing slightly more frequently than usual.
I've been slipping upward the past couple weeks.
My 30 day average was the best in both periods.
My 60 day average was the worst in both periods, too (odd).
I know that the 60 day average this time included the holidays, but I have no idea what was going on back between September and October to make my averages drop like they did (all I know is that Halloween certainly was no help!).

Based on these averages, my A1c SHOULD come back in the sub-6.0 range. But if I've learned anything from my previous experiences, I know that I don't really lie on the regression line.

Perhaps I'm an outlier.

Erratum

I originally wrote this post at 1am on Friday night. When I woke up on Saturday morning, the first thought in my head was: "Of course I don't lie on the regression line. In fact, it's completely possible that NO ONE lies on the regression line. We could ALL be outliers (though, technically, that couldn't be true)."

Somehow all my statistics training was already fast asleep at 1am.

The regression line is just an estimated linear relationship between average blood sugars and A1c results across many individuals. The relationship is unique for each individual. Like in the graph below, each dot represents the unique relationship between average blood sugars and A1c results.

There's clearly a pattern in the relationships: higher average blood sugars correlate with higher A1c readings and lower average blood sugars correlate with lower A1c readings. But the diagonal line is just a best linear estimate of that relationship. It's the simplest way to say "ON AVERAGE your A1c reading is equal to [{average blood sugar + 77} / 36]" (which is approximately what the regression results from the study I'm familiar with found).



Although, I could very well be the organge dot (in which case, perhaps I really am an outlier).

close

My money's on 6.6.
I can almost hear the clicking of a roulette wheel and a bouncing marble looking for its home.