Wednesday, July 7, 2010

When 3 + 5 = Cheese.

How am I supposed to attain "tight control" with equipment that only works sometimes?

Last night, I put in a new CGM sensor. (I say "in", because while the transmitter rests on top of my skin with some adhesive, underneath that is a small wire that stays under my skin, as long as I'm wearing the sensor.) I let New Jim calibrate for two hours while I was sleeping, got up at 1:30 am to do the two finger sticks required to calibrate it, and went back to sleep.

Since I've gotten up, Jim's results have been all over the place. I've already done 8 finger sticks today, and it's only 10:30 am.

Usually, Jim is very accurate. I would say he's typically within 10 - 15 mg/dL of what my meter comes up with when I do a finger stick. Today, he's been as much as 50 points off. He's also told me I'm trending upwards, when I was actually falling. Later, as a test, I did two finger sticks 20 minutes apart, and verified those results against what Jim was telling me. My finger sticks were 84 and 83, respectively. Jim, on the other hand, thought I had dropped from 124 to "LOW". (A reading of "LOW" means that I'm below 40, on the CGM.)

This is pretty frustrating. I actually ended up just shutting down the receiver, since the results have been so wrong. Jim has been vibrating and beeping all morning, and very unnecessarily, since his results have been consistently off. I would really, really like to throw him out the 3rd floor window of my office building - but instead, I'll just keep testing every 45 minutes or so, and hope that I catch everything.

Days like this make me wonder, though - how does anyone expect a person in my position to gain "good" control - much less "perfect" control? The tools we have available to us can be inaccurate and unreliable. Did you know that blood glucose meters are only required to provide results that are within 20% of a laboratory standard, 95% of the time? 20%! That's ridiculous! How is it that an iPad, WiFi, and GPS systems exist - or any other modern technology you want to name - but 30 years after home glucose monitors became widely available, we can't better accuracy for our health? C'mon, ISO and FDA - you can do better than that. And not only can you, you should.

The continuous glucose monitor is no different, in that a large amount of error is "allowable".  According to my Dexcom Seven Plus User Guide, a clinical study showed that for the range of 40 to 400 mg/dL (which is the full range of the CGM), only 76% of readings were within 20% of the laboratory's glucose readings.  That's a whole lot of room for error, people.

Let's go back to the 20% margin of error for glucose meters results - the very results that I'm using to make decisions on food, insulin, and so many other things.  Here is why inaccuracy is so dangerous:  if my meter reads 100, that means that in actuality I could be either 80 (on the low side of normal, and would probably convince me to eat something) or 120 (the high end of normal, which might cause me to take a corrective amount of insulin).  Let's take this a step further and say that I test, and the meter spits out 300.  If I see this, I'm following my corrective formula of 1:20 over 100, which means I'll be taking 10 units of insulin to bring myself back down to 100 (hopefully).  But what about the margin of error?  Accepting the 20% possibility, I could really be 240 (in which those 10 units of insulin would likely bring me down to 40, which is frighteningly low), or I could be 360 (and those 10 units would only bring me down to 160 - still out of range).

I don't believe that my good health is too much to ask for.  I'm willing to do the work, but that effort can be fruitless if I don't have tools that work properly, and work accurately together. 

As Mark Twain (or someone else, equally wise) said:  The man with a watch always knows what time it is.  The man with two watches is never quite sure.

1 comment:

  1. It is pretty scary how crude the tools we have are. Don't get me wrong - I'm glad we're not still peeing on color-code sticks. But still, they are far from perfect, aren't they?

    ReplyDelete