AIDS Origin: Statistics & Smoking Guns

Thomas Keske TKeske at
Fri Jan 7 22:25:34 EST 2000

This was an older essay on this subject, which I'm
reposting for reference


In the late 1970's experimental hepatitis B vaccine trials
were conducted on gay men in New York and San Francisco.
In each city, the first known AIDS cases were reported within
a couple months of the beginning of the trial.  Many experts
have alleged that there was a direct connection between the
vaccines and the beginnings of the AIDS epidemic.

All men in the vaccine trials were completely healthy
at the beginning of the trial- a requirement for participation.
In some cases, the men came down with full-blow AIDS
within months of receiving the vaccine.  This is possibly
peculiar, because the average incubation period is on the
order of 5 to 10 years.

Government studies, run primarily by those directly
involved in the trials, claimed that the incidence of AIDS
among the trial participants was not really higher than
that of the general population.  Over the long term,
this claim becomes somewhat doubtful, because virtually
all of the men in the trials eventually came down with AIDS.

The more important question is an analysis of the *earliest*
AIDS cases.  In these first known cases of AIDS, was there
a statistically significant, disproportionate representation
of vaccine trial participants?

The prevalence of vaccine trial participants among the first
AIDS cases has sometimes been dismissed on the grounds
that promiscuous men were chosen for the vaccine trials,
and therefore, supposedly, it was nothing suspicious that
they should be the first to get AIDS.

Relative to a supposedly non-promiscuous straight population,
that might have been true.  However, the relevant comparison
is not against the straight population.  The promiscuous
vaccine trial participants, and their rate of AIDS, needs to
be compared against that of other promiscuous gay males,
in the same cities, during the same time period.

Dr. Paul O'Malley, the health investigator who headed up
the Merck/CDC hepatitis B study, reported that of the first
24 AIDS cases in San Francisco, 11 were from the vaccine trial [1].

Similarly, in New York, of the first 41 AIDS cases, some
25-50% were reported to be participants of the vaccine
trials [2].  For our purposes, we can take a middle figure of
37.5%, or 15 men.

In San Francisco, some 6700 men participated in the
hepatitis vaccine studies.  In New York, it had been some

Not all the 6700 participating in the SF study actually
received the vaccine.  I am attempting to verify the
actual number.  However, because the "11 of 24"
figure is known, it actually biases the analysis in
favor of "natural occurrence" if we treat the full
6700 as trial members.

The only remaining information needed to make a
ballpark estimate of the statistical odds are figures
for the numbers of promiscuous gay men that existed
in these cities, in the late 1970s.

It is difficult to get accurate information, since there is
no formal gay census.  However, you can make some
reasonable, conservative estimates.

The population of the city of San Franciso is about
700,000, and the metropolitan area is about 1.7
million.  Most gays would estimate 10% of the population
as being gay, while I tend to use a more conservative 5%
figure.   New York and San Francisco should have higher
percentages than the national average, but to be conservative,
you can stick to the overall average.

Another crude yardstick is the Pride Parade
attendance, about 200,000 in San Francisco.

Estimates about venereal disease, and surveys about
unsafe sex practices can give an estimate of what percent
of the gay population might be termed "promiscuous".

Another possible measurement is difference in the
number of children per household, which is lower in SF
than for the national average, for obvious reasons.  An
estimate based on this yielded roughly the similar
results (details available for anyone who is curious).

>From such sources, I'm estimating about 100,000 sexually
active gay males in metropolitan San Francisco in the late 1970s,
and another 100,000 in New York.  We are safe in our conclusions
if our estimates are more likely conservative than inflated.

These estimations are crude, but could be refined.  For a
first-cut, ball-park estimate, they ought to be adequate.
When a ball-park estimate yields results that are sufficiently
overwhelming, the margin of error is sometimes wide enough,
that major adjustments in the starting figures will do nothing
to alter the bottom-line result.

This becomes a relatively straightforward problem
in binomial distributions.  If you want to calculate the
probability of getting r of one kind of thing in a sample
of n things, the probability is:

      (n! / (n-r)!r!)  x  (q ^ (n-r) x  p^r)

where "!" means "factorial, "^" means "raised to the power".

p is the probability of getting the item of interest, and q is
the probability of getting the other item.

For SF, p is (6700/100000) = .067.   q is (93300/100000)
= .933.   n is 24 and r is 11 (as discussed above).

The above formula gives the probability of getting 11
vaccine participants out of a sample of 24.  What we want
is the probability of getting 11 OR MORE, so I wrote
a program to sum repetitively for the chances of getting
11, 12, 13 . 24.

For San Francisco, the odds against getting such a
high percentage of vaccine participants in the first
24 AIDS cases are about 1 in 7.7 million.  In New
York the chance of getting 15 vaccine participants
in the first 41 AIDS cases were even more dramatic,
about 1 in 2.2 x (10 ^ 19).

For the overall probability in NY and SF, you can
take the product of these two figures, giving about
1.7 in 10 ^ 26   (this being a figure of 1 followed by
a string of 27 zeroes).

There is another major reason why these figures are
likely to be, if anything, extreme underestimates of
the statistical correlation.

An important issue is the true size of the "pool" from
which you draw.  It can be argued that, of the first AIDS
cases, the true pool size for the "general population" of
promiscuous gays is not merely the numbers in San
Francisco or New York.  It is not surprising that AIDS
should have broke out in these cities.  However, there is
nothing to confine the chances for the AIDS first cases to
have been in these cities.  There are also many mobile,
promiscuous gays in Boston, Chicago, Florida, etc- virtually
any of them had equal chance of being among the first

In truth, the size of the "general population" is all the
promiscuous gays, all over the world.  In view of this,
quibbling over questions such as how many promiscuous
gays existed in San Francisco becomes clearly

There is a clear conclusion, here:  the connection between
the vaccines and the outbreak of AIDS is not random, and is
not explained by the promiscuity of the men who took the

What, then does this mean?  Are we looking at a tragic
accident, something gone wrong in the manufacture of
the vaccines?  Might the vaccines have been maliciously
laced with HIV, maybe something secretly discovered in
remote villages in Africa?  What of similar, alleged connection
to polio and smallpox vaccine programs in Africa?

Future installments will examine these questions.

Tom Keske
Boston, Mass.

[1] "Emerging Viruses, AIDS & Ebola", Leonard Horowitz
[2] "Designer Diseases',

More information about the Microbio mailing list