Statement of John F. Brundage, M.D.

STATEMENT
OF
John F. Brundage, M.D.
1310-1500
8 December 1992
COL Dangerfield began our meeting with a review of the
informal 15-6 investigation process and reviewed the tasks for
which he is responsible to determine the facts. He then reviewed
the allegations that had prompted the investigation.
I was asked to describe the events that I was involved with
prior to the meeting in Amsterdam. I recalled that, several days
before the weekend I was to leave for Amsterdam, I returned to my
office to find LTC Redfield sitting there. He said ne had gotten
the data that would be the basis for his presentation in Amsterdam
and explained his interest in presenting a statistical analysis.
I told him the data set was small, and it was not gathered in a way
that permitted rigorous statistical analysis. I then went through
an approximately 30 minute discussion on statistical methods to
compare two groups of data. I suggested that a comparison of mean
values would probably not be appropriate and the p value from such
a test would not be valid. Unlike a t test, non-parametric tests
could be used when data were not from a normal distribution and his
data did not appear to be so distributed. I told him rank tests
were commonly used for such data.
LTC Redfield asked how to do a rank test. I walked him
through a simple explanation, and I xeroxed a textbook section
that gave step-by-step procedures for rank tests and gave it to
him. He then organized his da... did the rank test. I then
explained the Chi square and Fischer's exact tests; we did some
samples using his data arra2gd in 2 x 2 tables until I thought he
understood what I had said. I did not get the impression that this
was the end of his analys t. In essence I gave LTC Redfield a
short and very basic refresher course in statistical tests. I told
him I thought it would be a mistake to emphasize statistical
comparisons or p values. However, I thought it likely that someone
would ask him. I suggested that he should simply respond that he
had done a rank test and give the result. Once again, I cautioned
against emphasizing statistical comparisons or p values.
It was my impression, when he was working the data in my office,
that his sole concern was with the data - he did not seem to be
selecting patients to present nor did he seem to be overly
concerned with how the answers came out. He seemed more concerned
with learning and applying the procedures for doing the various
tests.
~I attended his presentation in Amsterdam and I did not feel it
was inappropriate. Because of time limitation, he went over the

11111111111111111111111111 'liiiillllill
5571095.0466.015 AI