Recently I wrote an article about how day-to-day highs and lows can even out over time. I mentioned at the end that this was only an average across 21 days, not the 60-day statistics that I usually consider (including when comparing with my HbA1c blood tests).
Well now those 60 days are up, so here’s an update.
So how does that compare with the 3-week data?
| 21 days | 60 days | |
|---|---|---|
| TIR (3.9-10) | 97.8% | 96.7% |
| TITR (3.9-7.8) | 91.8% | 90.1% |
| TISR (3.9-6.7) | 79.1% | 78.1% |
| TBR (< 3.9) | 1.6% | 2.2% |
| Time Actually Low (< 3.5) | 0.3% | 0.6% |
| CV | 21% | 22% |
| eHbA1c | 5.3% | 5.3% |
As you can see, at a detailed level the answer is “not as good”. Most numbers dropped off slightly.
However I’m not worried about this. The term “still good enough” does come to mind.
Because of course I did not stop experimenting. I have continued to experiment with some different insulins, exercise regimen, etc and these have resulted in a few more lows than before.
Plus those 60 days did include the “festive” [or “feasting”] season of birthdays, Christmas, New Year’s, etc.
There have also been the usual vagaries of chance and equipment reliability.
Reliability
Over that time there have been various interruptions to my automated insulin delivery. Including a failed CGM sensor that involved long stretches of occasional fingerpricks while the replacement sensor was settling in. And some significant periods where Bluetooth connectivity with my pump was off. But overall the AID was able to execute decisions 93.6% of time, which means there was an average of 1 hour and 32 minutes every day where the pump just had to run by itself (either continuing for a while on its last instruction, or falling back to default delivery rates.
There was one night where the Bluetooth link to the pump was offline for about 6 hours, starting shortly after I went to bed. So the pump soon reverted to its default rates. I do usually have alarms that would kick in in these situations, but I’d temporarily disabled them as part of an experiment.
It does seem my insulin profile is set up at a reasonable level, as my BG maintained a fairly flat line throughout, ending up at only 6.6 mmol/L. If the system had been able to continue working on the prior evening’s food, I expect it would have soon brought me down closer to my 5.1 mmol/L target, but I was happy it had at least remained stable!
When I did notice the problem I reset the pump and went back to sleep. When I got up 2 hours later I was down to 5.0 mmol/L…
What now?
Life of course goes on. And Every Day Is An Experiment. The fourth anniversary of my last meal announcement is almost here too.
I do have another set of blood tests coming up in a few weeks (as part of starting participation in yet another clinical trial). That will of course trigger another review of my statistics (including seeing how close the HbA1c result is to the recent CGM estimates).