Saw an awesome post on alt.support.thyroid today on the concept of the "normal" range on lab test results.
http://groups.google.com/group/alt.support.thyroid/browse_thread/thread/a5ee2a2b8f6fe91f?hl=en Poster: kgrhoads at alum dot mit dot edu, Oct 22, 9:28 AM
"When doctors talk "normal" ranges, they seem to be unaware (or have totally forgotten) that the "normal" here is the technical statistical usage, being the Gaussian distribution, a.k.a., "normal" distribution, and has ABSOLUTELY NOTHING to do with "normality" in the common usage sense. A value can be within the Gaussian reference range for the lab and be highly ABnormal. "
"Now, that does not mean that this value is ABnormal. It just means that being within the reference range is NOT PROOF OF NORMALITY in the common meaning for "normality". "
"Personally, I believe the use of "normal" for labs should be abolished, and only "Gaussian" should be used. That way people are not being set up for misinterpretation of the meaning of the lab values. "
Reminds me of one of my favorite quotes from my brother: "I'm right, because I proved my point with math."
1 comment:
This is somewhat related to how the US Recommended Daily Allowances were arrived at. Intakes of various compounds were set at the minimum level that prevented any obvious form of disease.
Note the careful wording there: "Obvious." They had no way of knowing that borderline/low levels were not the cause of other biological mishaps. They just knew that low Vitamin C caused scurvy, so they set the RDA at the level just sufficient to prevent scurvy.
Me, I'd rather see distributions segregated by population. I'd rather be closer to the mean blood values of normal-weight, happy, well-adjusted people who can run a mile without dying than of the general population. I've seen the general population, and am somewhat unimpressed with their average level of conditioning.
Post a Comment