Hi Again Marps,
Wow, that is a large variance. When I first started w/ the Dex, I saw that kind of variance. Since then I have learned how the Dex works with my body and can get better performance out of it.
I only calibrate twice per day if I can minimize the time my BG goes over 160. I find that this prevents the calibration from drifting, which for me seems to occur at BG of approximately 160 and greater. If my BG does go above 160 for extended periods of time (such as a failed infusion set or illness), I will calibrate 3 or 6 times that day to try and re-establish the calibration accuracy. As I stated previously, for the most accurate calibration, I will try to wait until my BG is stable (such as less than a 10 point change over 30 - 60 minutes).
I also will restrict my calibration to BG roughly 50 - 150, since I have my alarms set at 70 and 140. I find that a more narrow calibration range gives better accuracy within that range. My opinion is that calibrations at, for example BG of 200, will decrease the accuracy of the calibration when my BG goes back to the euglycemic range, which also happens to be the range where I want to spend the most time. The downside to this approach is that I tend to have poor accuracy at BG over 160, since I never calibrate there, but I accept this since my goal is to minimize the time at these levels. For accuracy at BG > 160, I resort to fingersticks until my BG comes down and then I can rely on the Dex again.
Another thing I do is when I see a calibration reading that differs more than 15 - 20 points from the Dex, I will do one or two more additional fingersticks and average the results before I enter the reading in the Dex. Remember, most BG meters are rated at +/- 20% accuracy. IMHO, I feel the average may be closer to the true reading than a single fingerstick. I don’t have the link right now, but Freestyle published a report where they show that their meters can obtain +/- 10% accuracy more than 90% of the time. For this reason, I switched from the OneTouch to the Freestyle Lite.
Again, my opinion is that when everything is working right, a 5 point or less difference is darned good enough for me and will not make an appreciable difference in the amount of insulin I take. So despite what the manual says, if I minimize hyperglycemia and continue to see < 5 point differences during calibrations, I do rely on the Dex reading without a confirming fingerstick.