It takes at least 1/2 hour of activity to “freshen up” the ISF around a squashed Sensor. During that time, you should NOT enter the conflicting bG readings as calibration points, because the Sensor doesn’t yet have “valid” voltages from “good” ISF. You’re distorting the curve… just use those bG readings to guide your first hour of activity, and calibrate Dexcom later.
Turning back over to the proper side during sleep time is a much slower recovery, because the tissue surrounding Dexcom is totally inactive; your breathing RATE isn’t all that much slower than active waking time, but the STRENGTH of the muscle actions is much less. YMMV, of course; for you, the entire night might be unrecoverable after the initial squashing mistake.
I know that your outstanding A1C implies very little time spent above bG 140 mg/dL, but to widen the “reliable” range, you should be sure to enter calibration figures when that does occur. (At the time it levels off in the Dexcom, not while it’s rising. For me, this occurs 15-20 minutes after the bG increase has ended.)
If the upward/downward trends are STILL wrong, after you’ve stopped feeding it “bad” calibration figures at times when you’ve got particular reasons to expect inaccuracy (e.g., bed-squashing), then this site doesn’t work for you: Try elsewhere.
Your insurance/financial situation might make a 3-month test relatively painless, from a financial perspective. But I would be watching the calendar like a hawk, to be absolutely sure that I notified Dexcom that it’s failed to work within the 30-day money-back trial period. Notify first, and tell them that you’ll try some other sites if and only if they fax you a statement guaranteeing your right to a full refund for an additional 30 days beyond the standard policy.
This is because many people have switched from Dex to Navigator, and gotten much better results. After you get your money back, or assign the refund back to your insurance, you can try the other one. (Abbott and MM don’t have such a refund policy; that’s why everyone should try Dexcom first.) If Dexcom doesn’t work for you, don’t apologize-- I know darn well that it’s certainly NOT your fault-- you’re obviously not (ahem) inexperienced in Diabetes Management, or clueless about Math and Science.
Is the Sensor Pad getting loose around the edges before it fails?