Statement of Ltc. Deborah Birx

based on the wide variability of the viral burden assay (an
observation with which Dr. Redfield agreed when Dr. Deborah Birx
first raised it in the 15 SEP meeting) he reasserted his conclusion
that the data suggest an antiviral effect of the vaccine."
One of the conclusions of this meeting was, based on Dr.
McCarthy's analysis relating viral burden to immunization, that
viral burden data could be interpreted as safety of the rpgl60
vaccine; however, I do not recall any changes in procedures
specifically delineated. At a subsequent NIH AVEG meeting on this
same date, Dr. McCarthy stated that there was no relationship,
other than random, between immunization and viral burden.
"Overselling the gpl60 Phase I data may date back as early as
the New England Journal of Medicine publication (N Engl J Med
1991:324:1677-84). Oral Presentations of the phase I continuation
data over the last year and a half have repeatedly shown only a few
selected "home run" patients. The data from the entire study group
has not been well represented by these selected patients."
Approximately half of the presentations regarding e
ase I trials have been made by me and I have always presented the
entire cohort. I did so at three national meetings in the past six
months and there has never been any question of overselling or
misleading information. Dr. Hendrix was not present at any of
these meetings and he has not requested a copy of any of these
presentations. We have presented all our data for scientific
scrutiny at to other groups, including Dr. Fauci and his research
staff.
"Data analysis has been either sloppy or, possibly, deceptive
with use of inappropriately chosen "control" groups, unorthodox
statistical methods that abuses the data to come up with the
desired CD4 trend conclusions..."
This is a very serious allegation for which there is no basis.
As a scientist who has participated in at least 90% of the
parametric and non-parametric data analyses, I would not call t
tests and the Wilcoxon test "unorthodox". It should be noted again
that CD4 data have not been presented since the June 1991 NEJM
article where the CD4 data were referenced only in terms of safety.
It should be noted that at the Amsterdam meetings, the four papers
which preceded Dr. Redfield's, reported either stabilization or
increases in CD4 cell counts during the studies over 9 to 18 months
of followup.
"...failure to include appropriately performed analyses that
fail to support the desired conclusion and badgering of
statisticians and colleagues by Dr. Redfield, sometimes
successfully, to agree to data analyses against their better

6