# RE: [IP] Dropping

```Pixie <email @ redacted> wrote:

> Normal is individual, there's no standard "how much of a drop
> is normal in what period of time" hence, there is technically
> no safety amount, it varies from person to person.

< Professional Mode = ON >

Actually, normal is *not* individual.  Normal is usually a synonym for the
mean -- the central tendency of a group of measurements, in this case rate
of reduction in blood glucose per unit of H.  The whole concept of YMMV
depends on the measurements of "normal" (and there are several ways that can
be measured) along with the variability of those measurements.  Technically,
a normal distribution is one in which we see a classic "bell curve" of
responses which can be transformed to the "standard normal" with a mean of
zero and a standard deviation of one.

We also deal with confidence intervals, most commonly a 95% confidence
interval.  That means that in a normal (also called Gaussian) distribution,
any point that falls within ninety-five percent of the *area* of the curve
is confidently deemed to have the same value as the mean.  In a true normal
distribution, this 95% confidence level is equivalent to two standard
deviations on either side of the mean.  That's why we are often given a
*range* of normal values, and told whether our own measurements fall within
that range.

When we calculate "normal" values, we make an inherent assumption -- that
the numbers we see are representative of *everyone* in the class we are
measuring.  That may be all people, or people with diabetes, or people with
type 1, or people with type 2, etc., etc.  IOW, we have to define the basis
for our statement of normalcy.  Normal for a type 1 may not be the same as
normal for type 2, etc.  But (and it's a big but, and very important): these
are not statements of judgment of value -- only statements of the way things
are.  Normal does not mean good or bad.  Normal is morally, medically,
ethically, etc. NEUTRAL.

With some statistical procedures, we are able, by making a couple of
assumptions, to *infer* what might happen at some future time based on the
data we see now.  This is usually expressed as a likelihood, probability,
or, most commonly, a RISK.  For example, there is a statistical association
between having diabetes and heart disease.  So we say that diabetics have
five times the risk of non-diabetics with no history of heart disease of
having a heart attack (myocardial infarction).  That does NOT mean that all
diabetics are going to have a heart attack (there are ways to reduce the
risk!).  What it DOES say is that in the past, diabetics have had heart
attacks at five times the rate of non-diabetics who have no other
indications of heart disease, and, because we can make some assumptions
about the distributions, that the rate we have seen in the past will
continue *IF NO OTHER FACTORS CHANGE*.

To sum up, this is the basic concept behind YMMV.  Not one of us is normal.
We may have measurements that are normal, that is close to the mean (or
other central measure) of the population, but that is just one measurement
out of many that make up normal.  Because that determination is made with
the confidence interval in mind, most often we are given a *range* of normal
and told whether our measurements fall within that range.

If you're still here, thank you.  I hope this is helpful to you.

< Professional Mode = OFF >

Jim Handsfield
Statistician
Centers for Disease Control and Prevention
mailto:email @ redacted OR
mailto:email @ redacted

The opinions expressed are my own and do not necessarily represent those of
the Centers for Disease Control and Prevention, the United States Public
Health Service or any other agency of the United States government.
----------------------------------------------------------
for HELP or to subscribe/unsubscribe, contact: HELP@insulin-pumpers.org
send a DONATION http://www.Insulin-Pumpers.org/donate.shtml

```