March 17, 2010

What is Accuracy?

At some point in learning to live with diabetes comes the discovery that the number on the blood glucose meter has range of plus or minus twenty percent. That tends to horrify people.
The FDA is having hearings this week on meter accuracy. Good friend of the blog Ellen Ullman is testifying for patients. Yay Ellen! LY/MI!
Ellen has me thinking about accuracy. My thoughts didn’t fit in a tweet. (I am a wee bit wordy - I know it is a schock) She asked at what cost more accuracy; higher co-pay, bigger meters, more blood, and slower response time, all of the above?

All that begs the question that I am really trying to come to grips with: Would tighter variance standard for meters, +/- 15% or 10% result in better health outcomes? Or are there other issues in the whole of T1 care that present greater opportunities for better care?

Well first off what is better care? Lower A1C? Less BG variance? I think most folks would say well it is both low A1C and less glucose variance. Will a tighter margin of error in meters achieve those results?

I suspect that there are many variables in diabetes that are not with in the 20% margin of our meters. How about carb counting? Consider a bowl of spagetti. Are the sources of my information, portion control and actual consumption within 20%. Is the actual sugar content of a tomato that was used to make the pasta sauce that tight? The blend of wheat in the spaghetti? Did the manufacturer vary the high fructose corn syrup to achieve the brand’s signature taste because the tomatoes weren’t as ripe? Is that food label really that accurate? I have no idea.

How about the impact of my activity? Did the effect of the gym on BG vary? Do habits about getting to the gym?

Is the insulin fresh? Did it get too hot? Is the set working well? Air in the tube? How accurate is the amount of insulin drawn into a needle? How is the stress? What is the weather? How early did I pre-bolus.

Assuming all of that is dead on the money, how often am I testing? Those numbers are a point in time. They have no indication of any change in the direction of BG or the rate of that change at the time. And if those numbers vary from the goal is that variation the result of a bolus issue or a basal issue?

Head ready to explode yet?

Is there anyone on God’s Green earth (and I will grant you there is extra green today ‘cause it is St Pat’s - I don’t thing the green die matters but the beer is another story) who has all this figured out to +/- 20%?

So my two bits are this: Accuracy is promoting better outcomes. In that definition, I wish everything varied less.

I think good (a word that means highly automated, very unobtrusive, dirt simple and probably working though a wireless connection between all diabetes devices) analytical tools that can incorporate the whole scope of life with type 1, including the information from meters, could have a more positive effect on outcomes in living lives’ with type 1 than accuracy range of +/- of 10% vs 20%.

The real magic IHMO is in living lives’ with type 1, not dominating lives with diabetes.