@mohe0001 this one is for you, mostly, and me, a little.
In Dan’s article he looks at the kind of Dexcom CGM wipe outs I’ve observed and commented on for years. The article is certainly tl;dr, even I skimmed it. Yet he has set down a mighty piece of evidence, researched and backed up by his own original work.
I’m posting this topic because, while boring and technical, it really is important. I reached the same conclusions myself a couple of weeks ago; I made the same point (in a private communication via my health providers website) that spastic CGMs are dangerous.
I’ve had some similar concerns about the jumpy BG data from Dexcom G7.
I don’t consider the G7 data jumpiness to be dangerous though with the Omnipod 5. The Omnipod algorithm moves at a snails pace with the micro bolus, effectively smoothing the data by it’s slowness. Maybe the bolus calculator could become dangerous if it used a momentarily bogus BG number?
Here’s a sample of the jumpiness I’m referring to:
I’m curious if others think that this degree of jumpiness on Day 3 would be bad enough for a replacement from Dexcom product support. (I kept the sensor going and it settled down by Day 5).
Just a couple of days ago I was driving back from New Jersey to Boston and my Dexcom G6 went wonky, informing me that my blood glucose was so low it was unmeasurable. This inaccurate information set off alarms on watch, pump and cgm. This went on for 2 hours or so.
No big news. We’re all used to this. For me, it happens more than 50% of the time about 2 hours after a new sensor is inserted. The sensor is then fine for 10 days. But the first 4 hours are often very rough.
Has anyone seen any documentation from Dexcom on this? I haven’t.
Personally, I think documentation is important. It’s a minor irritation if you know about it, but could lead to bad treatment if users don’t know to ignore it.
I view the Dexcom G6 as reliable enough yet it still needs a sober adult in the room if it is positioned as the basis for realtime insulin delivery as with an AID system. My Loop DIY allows me to switch off the closed loop and I’m not shy about doing just that if the CGM is not earning my trust.
First day jitters are one thing but longer-term delusional dives and excessive elevated and inaccurate excursions test my patience. I’ll allow a few but I’m fairly quick to pull the plug on an ill-behaved sensor. I just physically remove the bad actor and let the software eventually report a “failed sensor” so I keep replacement police from reflexively denying my claim.
I’m seriously looking into the Eversense 365 as the solution when my G6 supply runs out later this year.
G6 would definitely be considered “delusional” (genius term) for the first 12-24 hours for me if I didn’t pre-soak. This is driving some concern for me with G7, but sounds like I can let it sit for 12 hours at least before truly starting it.
@Terry - how do you go about the process in your 2nd paragraph above? I’m reading that you get the sensor to error itself out?
Eversense 365 looks interesting, but the idea of having adhesive on my skin in one place for a solid year seems like a lot. My skin isn’t that sensitive, but I think it might not fare well being covered up for a year. Can’t wait to hear about others’ experience.
@JessicaD – If a sensor is beyond a reasonable chance of redeeming itself, I physically remove it and wait on the software to report an inevitable “sensor failure.” This is logged as a hard failure and there’s no dispute with customer service. (This tactic was suggested to me by another TuD member. It works and makes my life easier. I don’t abuse it.)
Then I go ahead and begin a new sensor session. I may have already inserted a new sensor. When my sensor lasts the entire 10 days, I presoak it 16-20 hours in advance.
