[Previous Months][Date Index][Thread Index][Join - Register][Login]
[Message Prev][Message Next][Thread Prev][Thread Next]

[IP] Re: D prevention

WARNING: The following post refers to statistics, which some consider
to be as repulsive as vulgarity. If you share this view, press your
delete key NOW!

If you've read this far, you're too late. ;-)

Hello, Jim.

Thanks very much for your reply. There's something fundamental about
your argument that I still don't understand. (I don't think it's the
statistical concepts. It's upstream in the data.)

Here's a very brief abstract of what I think you said:

Incidence of Type 1 in the population of families appears to be
randomly distributed. Multiple research projects have not found any
correlation between hereditary factors and the incidence of type 1.

You then posed a few questions, such as:

JH> ... if we select 10000 families at random with at least one
JH> diabetic child, what is the probability that more than one will
JH> have type 1?

And provided the statistical answer as it's known today:

JH> So far, the studies... have not been able to detect a non-random
JH> association between the cases.

Here's my follow-up. Your comments are invited:

Studies show that Type 1 does not follow strict genetic transmission.
But the studies also show that there *is* a non-random genetic
component in the incidence.

One source is the NIDDK (National Institute of Diabetes & Digestive &
Kidney diseases). Specifically, their publication "Diabetes in
America, 2nd edition," available on-line in PDF format here:


I looked at Chapter 8, "Risk Factors for Insulin-Dependent Diabetes."
A particularly relevant section starts on page 167 (actually the third
page of the PDF chapter), "Risk of IDDM in Relatives."

The first paragraph reads as follows:

     More than 80% of cases of IDDM occur in individuals with no
     family history of the disease. However, in the remaining 20%,
     IDDM aggregates in families. The overall risk before age 30 years
     for North American Caucasian siblings, parents, and offspring of
     individuals with IDDM ranges from 1% to 15%... compared with
     rates of < 1% for individuals without IDDM relatives.

If Type 1 were randomly distributed in the population, wouldn't the
risk for those with Type 1 relatives and the risk for those without
Type 1 relatives be _identical_?

Is the risk cited for those with Type 1 relatives incorrect in light
of more recent studies? IOW, are you aware of data that contradicts
what's stated above?

On page 170 (p. 6), there's a statement about onset in identical

      Twin studies have shown that genetic susceptibility to IDDM
      appears to be necessary, but is not sufficient to cause the
      development of the disease, because concordance for IDDM occurs
      in only 36% of monozygous [identical] twin pairs.

If Type 1 were randomly distributed in the population, could such high
frequency in identical twins be expected? Again, are you aware of more
recent data that contradicts what's stated above?

[I'll digress statistically for a moment. If Type 1 were randomly
distributed in the population of families, then it would be safe to
say that Type 1 was independent of family membership. That is, my risk
of getting Type 1 wouldn't change no matter how many Type 1 relatives
I had. If the risk in the overall population of families was, say,
<1%, and my sister had Type 1, my risk would remain <1%. But
epidemiological data shows my risk goes up. Since the population of
families is very large, and the population of Type 1's in families is
still quite large, I can have confidence in such statistics. So if my
risk of having Type 1 goes up because my sister has it, there is
*some* family, i.e., genetic, component.]

If the data cited in the NIDDK reference are incorrect, then I can
understand the statement that "Multiple research projects have not
found any correlation between hereditary factors and the incidence of
type 1." If the NIDDK data are still regarded as reliable, then I
can't understand that statement.

Again, your comments are, as always, appreciated.

regards, Andy
for HELP or to subscribe/unsubscribe, contact: HELP@insulin-pumpers.org
send a DONATION http://www.Insulin-Pumpers.org/donate.shtml