How accurate is your Dex?

Replies to This Discussion

Permalink Reply by Rickst29 on June 30, 2009 at 11:44am

Nilufer, please see "love handles!" which I just wrote for Mike (immediately above).

▶ Reply to This

Permalink Reply by Jan on June 30, 2009 at 7:53am

We just stared DN's first sensor. Dex Rep did not like area she chose to wear it, but DN does not want to wear on arm/stomach because of bathing suits. So any problems might be due to placement of sensor. When her BS is stable (no insulin or food digesting), sensor is accurate. It has been inaccurrate when eating (alarming for low and she was 160) also high alarm beeps at 160 and she is already 220. So it will take a while for us to get the hang of it. She spikes high and fast after eating, so we will definitely have to prebolus ALL meals, not just breakfast. We slept through high alarms. Have to get her wireless mike hooked up. Dex just vibrates and gives two short beeps, does not keep beeping like Minimed. So thrilled that the insertion process was virtually painless. That was our main gripe with Minimed sensors. She will be able to wear it, since it does not hurt. Hope to figure things out within a week, month or so. Love the Dex.

▶ Reply to This

Permalink Reply by Rickst29 on June 30, 2009 at 11:38am

Good, you're ALREADY learning about high-bG events which you didn't even suspect before adding Dexcom! (I.e., how fast and high she spikes during and immediately following meals).

For most people, Dexcom is nearly always very "late" in tracking increasing bG. But alarming "low!" when she is at 160, that shouldn't happen with a good Sensor. Was she sitting on the Sensor? When butt-cheek-located Sensors are inserted too far down, and the wearer sits down, both blood flow into the area and flow of bG into ISF gets disrupted by all the pressure. (Sitting on, or nearly on, your butt-mounted Sensor stops the local blood flow. Even though the big arteries don't become totally squashed down, the tiny capillaries and small feeding vessels right next to the Sensor DO get squashed, and blood flow decreases dramatically.)

Cells around the Sensor continue to absorb Glucose from the "local" ISF for energy, but that ISF doesn't get Glucose replenished by blood (because the blood flow is curtailed, and local blood is similarly "sucked dry" already). AFAIK, that's about the only way a "good" Sensor can give a badly erroneous "low" reading. The same error can occur by rolling over and sleeping on it; do not automatically feed an in-bed "low alarm" without checking bG to make sure that it wasn't one of these "squashed location, stale ISF ran out of Glucose, but only locally" incidents.

For most people, Dexcom is much better AND much less delayed in tracking downward bG movements than it is at upward bG movements. But that 160 versus 220 is normal for such a strong and sharp spike. pre-bolus helps, but pre-loading fat and/or reducing the carb load might be even better tactics to avoid those spikes.

pre-bolusing is certainly called for, and maybe the carb/fat/protein balance needs to be addressed too. When a meal is going to be carb-heavy, I try to load up the fat first (salad oil, nuts, or cheese). Or, I drink wine, by itself, before the carbs start-- so that the alcohol-provoked drop in bG counterbalances the sudden rise form the carbs. But the timing of that balancing act is very sensitive, I often need to lunge for sugar tabs in the middle of the meal if I miss it even slightly.

Pre-bolusing will help match the timing a bit better, but for most people, no amount of timing can match the sudden, really strong surge which a carb-heavy, low fat meal will create. Pre-loading some fat, or largely switching away from such carbs entirely.

For example: Mashed potatoes. It's just a pile of useless, empty calories (until you add a bunch of really unhealthy, saturated fat stuff to MAKE them taste like something.) Unless you need empty calories to gain some weight, just let 'em go.

▶ Reply to This

Permalink Reply by Jan on July 5, 2009 at 5:49am

We had tried Minimed sensors with my niece and they were a pain, plus inaccurate a good deal of the time. Just started the Dexcom and I think our readings are great about 70 percent of the time. When BG is rising rapidly or dropping quickly (when you have the arrows), I do not find Dexcom is in range with the fingerstick. However, whenever we see one of the up or down arrows, we do not expect them to match because there is a quick rate of change. Just seeing the directional arrows will guide us. For instance post-prandially, if we see a lower BG value than it should be at the time, with a down arrow, that will signal to us that she may need more carbs. Up arrows at night, we check, give correction and will add a temp basal on top of the correction. We are still newbies and feeling our way around. I'm sure we will get the hang of it within a month. I do not expect BG numbers to match all the time or be close; fortunate to have the trend arrows, for times when BG is changing rapidly.

▶ Reply to This

Permalink Reply by Renegadez on July 8, 2009 at 12:58pm

Rickst29 So never enter data into Dexy on the upward trend? Just test and bolus and wait for Dexy to give a flatline straight arrow? Then test and confim and enter new dtat finger stick into Dexy? So Do Not enter every finger stick into Dexy?

▶ Reply to This

Permalink Reply by Rickst29 on July 9, 2009 at 11:26am

Yes, exactly!

NEVER perform "enter bG" during a period of rising bG. Just don't do it, it wrecks the calibration curve.

▶ Reply to This

Permalink Reply by Renegadez on July 9, 2009 at 11:29pm

So do you think it adds to Dexy's accuracy to add additional BG entries if they are added @ the proper stable even arrow times?? Like to add more reading then just when she asks for calibrations or every 12 hrs is all she needs? I was under the impression more BG's = more accuracy.

▶ Reply to This

Permalink Reply by Rickst29 on July 10, 2009 at 1:21pm

Yes, this DOES help accuracy, for two reasons:

First, it pushes down the priority "weighting" of the older, stale readings.

Second, when the Sensor is nearly new, the "curve" changes very quickly. Your subcutaneous tissue does mount a bit of a response, even though the wire surface is titanium. (It's a response to the puncture damage and pressure from the wire, not so much a "foreign invader" immunological response.) Conditions change rapdily during days one and two, waiting a full 12 hours between pokes on these days is (IMO) not optimal.

There's another issue which might call for more pokes: If all of your recent calibration readings are at nearly the same bG, you should try and enter a couple which are closer to the edges of of your normal range. Entering a whole bunch of calibrations at 105, 111, 102, 106 creates a "solid" data point for bG readings near those levels, but it can't set an accurate slope for the curve.

If you find yourself with level, stable readings right around 80 mg/dL, you really, really want to enter a calibration before grabbing at a sugar tab, regardless of how recently your entered a calibration value at 100+. This really helps a lot to set the right "curve" for Dexcom to create reliable readings at low bG levels.

Similarly, if you find yourself with a level, stable reading a bit high (NOT post-meal, those aren't stable!), then you should enter it before correcting. If you've entered at least two two calibrations under good conditions, separated by at least 30 mg/dL, AND one of them is at the low side of your normal range, I think that's enough for for every day after days one and day two. Day one and day two need more (especially day one).

The bottom line is, the QUALITY of the calibration readings is much more important than the number of entries. It only takes two readings to set a really good calibration, but they need to be at least 30 mg/dL from each other, and they need to be done during ultra-flat periods. (The "level arrow" isn't sufficient, you've probably noticed that the "level arrow" can remain present even during a slow rise in bG.)

HOWEVER: A lot of redundant readings aren't helpful. So, if you entered a high-quality "106" calibration only six hours ago, and you've just done a new meter readings and get "102", don't put it in-- it will merely lower the priority of other recent readings with better dispersion through your bG range, while adding no new information at all.

▶ Reply to This

Permalink Reply by Renegadez on July 12, 2009 at 11:04pm

Ok this sounds like good solid info I will definitely try to use this theory. thanks very Much!

▶ Reply to This

• First
• Previous
• Next
• Last