Print

---

Dear friends,

I need to clarify a couple of points in my message on statistics.  First of
all, the Red Cross still tests all donated blood for AIDS.  What they no
longer do is notify the donor of positive test results.  This is because 10
out of 11 positive results turned out to be false.

The other clarification should be cleared up by the following:

The relationship between population size and false positives.

When you are doing a statistical study you must keep in mind the percentage
of target subjects compared to the general population.  If the target group
is less than one percent of the general population is quite possible for
false positives to be greater than true positives.  Such is the case of
testing for the AIDS virus.  Approximately one million people in the U.S.
are HIV-positive, much less than one percent of the general population.
Therefore, when AIDS testing was applied to the general population, 10 out
of every 11 positives turned out to be false positives.

Now this doesn't mean that everybody in the U.S. was tested, but percentage
of blood donor volunteers who were HIV-positive was about the same as it was
for the entire population.  So even in a small sample group you'll have the
same problem.

In a private message, a question was asked about the percentage of people
with Parkinson's who were misdiagnosed.  I do not know of any clinical
studies that have been made on this subject.  The only thing I can say for
certain is that false negatives (saying it was not PD when it was) far
outweigh the false positives (saying it was PD when it wasn't).  One
confounding factor is that, in a sense, those being tested for Parkinson's
disease are pre-screened.  A person is not going to be tested for PD unless
they exhibit some of the symptoms.  Pre-screening eliminates a lot of the
subjects who not have the condition so the ratio of subjects who have the
condition is much larger.



Bruce
57/9
Sinemet, Mirapex & Tasmar