A short story about paying—literally—for my mulishness.
See also: My first post on Beeminding diabetes.
Over the course of three days this past week, I managed to derail all three of my Beeminder diabetes goals. On February 21st, I derailed on my mean blood glucose and standard deviation goals. On the 20th, I was on the wrong side of the yellow brick road for both goals, although I was just barely over the centerline on my standard deviation goal. Then on the 21st I indulged in some unwise1 midnight snacking that sent my blood glucose through the roof:
This spike ruined both my mean blood glucose and standard deviation for the day, and since I wasn’t on the right side of the road to start with, the road didn’t auto-widen, and I derailed.
But that’s not the real story here. What’s remarkable is how even though I’d just taken a $10 hit from derailing on those goals (paying out $5 for each), I also derailed on my goal tracking the percentage of my readings in the hypoglycemic range two days later.
There’s a sense in which this almost makes sense since an increase in hypoglycemia is always a risk when you’re trying to improve your blood glucose control. In fact, derailing on my mean and standard deviation goals did provide the kick in the pants to get me to focus on getting back to the “tight” blood glucose control that I try my best to maintain. On the 22nd, I recorded a very good mean and standard deviation:
This tight control continued into the 23rd, and I actually started clinging to the low end of my target range. Since the 23rd was a Saturday, I didn’t set an alarm and slept late, waking up around 11:20 a.m. to my Dexcom showing several hours of readings hugging or sitting just below the 70 mg/dL line2. (My meter read 76 mg/dL when I tested upon waking.)
I’m not much of a breakfast eater, so after waking up I just settled in front my computer and started working on a project. Over the next several hours, my blood sugar drifted down and down, and even though my Dexcom (still in “Vibrate” mode) was buzzing at me every five minutes, I mulishly persisted in ignoring it because I was so engrossed in my project3.
And here’s the kicker: the percentage of my readings under 65 mg/dL on the 23rd ended up being 27%, which is exactly one percentage point higher than the day before. Since my road auto-widened the day before to accommodate the sudden spike in low readings4, I would have been fine as long as I stayed at the same level or had a lower percentage of low readings. The fact that I derailed by one percentage point might have been frustrating if it weren’t for the blatant and obvious fact that I could have easily prevented the derailment if I hadn’t been such a stubborn idiot, happily hacking away (with a lovely hypoglycemia-induced headache brewing at the base of my neck) and ignoring my Dexcom. So, in the end, I am absolutely happy to have paid Beeminder another $5: I utterly and completely deserved to be punished for such idiotic behavior5.
I was going to say that such midnight snacking was atypical for me, but then I took a look at the sparkline for my #snack hashtag where I track my carbohydrate consumption on your.flowingdata.com, and it made a liar out of me, as it’s clear that my most frequent carb snacking time is between 9-11 p.m.: (Alas, self-tracking is a double-edged sword, with not all truths about ourselves discovered thereby being truths we’d actually like to believe about ourselves…) ↩
This means that my Dexcom had been alarming off and on for several hours before I woke up. I didn’t hear it because I accidentally left my alert profile on the “Vibrate” program. (Audible alerts always kick in below 55 mg/dL, but I wasn’t that low.) The ability to program different alert programs for different times of day is a design feature I am desperately hoping that Dexcom integrates into their next generation system because I like to use the discreet “Vibrate” or “Soft” profiles most of the time, but I would love to have the “Attentive” program come on automatically every day at midnight. ↩
I was f—ing around with this AppleScript [JEB, 2016: Resource behind link has disappeared 😢] to import my Kindle Highlights into DEVONthink. I don’t actually know AppleScript, and the script didn’t do exactly what I wanted “out of the box,” so it quickly became that kind of half-frustrating, half-fun and completely addictively engaging task that I’m sure many other hacker types are well-acquainted with. (And eventually: I gave up and started writing my own Python script to do what I actually want to do.) ↩
The spike in low readings on the 22nd was partly due to the fact that several hours of data were missing that day because I had to
insert a new Dexcom sensorreboot my Dexcom sensor. ↩
And I’ve written the whole story up here for a little extra public shaming. ↩