The Epidemiology of Fear

by John Lauritsen

Psychological warfare is being waged against gay men in the United States. For the past month or so the media have been disseminating hostile propaganda, with the message that we will all die, that we must die. These death threats do not issue from the usual bigots -- not from Roman Catholic agitators, or menopausal beauty queens, or fundamentalist TV hustlers, or quack psychiatrists, or Hasidic zealots. We are not being drummed to death by voodoo witch doctors, or anathematized by prurient priests. We are being cursed in the name of science, and the imprecations directed against us have the imprimatur of the Public Health Service (PHS). The prognosis of doom is emanating from that peculiar form of medical survey research known as "epidemiology".

HIV Antibodies = Death?

Michael Specter, writing in the Washington Post, was one of the first to propound the death message, in his June 3, 1988 article, "AIDS Virus Likely Fatal To All Infected". Specter writes:

The AIDS virus [sic] will almost certainly kill everyone it infects unless effective drugs are developed to treat it, federal researchers have predicted for the first time....

After studying a group of gay men from San Francisco for the past decade, however, researchers have produced a statistical model that predicts 99 percent of those infected will eventually develop acquired immune deficiency syndrome "if they do not die from other causes."

Because no one has ever been cured of AIDS, a 99 percent AIDS rate means that virtually all would die unless a treatment is developed.

These grim statements are allegedly based on epidemiological research conducted in San Francisco, as discussed in a report that appears in the June 3, 1988 issue of Science, "A Model-Based Estimate of the Mean Incubation Period for AIDS in Homosexual Men". The authors are Kung-Jong Lui, a mathematician with the Centers for Disease Control (CDC); William W. Darrow, of the CDC's AIDS program; and George W. Rutherford, III, of the AIDS Office in the San Francisco Department of Public Health.

The headline on the second page of Specter's article is even more emphatic, "AIDS Infection Proving Fatal in All Cases". After inaccurately describing the San Francisco study, and repeating the latest doomsday estimates from the PHS (300,000 AIDS cases in the U.S. by the end of 1992), Specter lays out the ramifications of the "finding" that everyone with HIV antibodies will develop AIDS:

"Public health service officials...hope the new study will encourage those at highest risk to be tested so that they will seek medical attention if needed...."

Many physicians are prescribing AZT for their patients who are infected but have not developed AIDS, although the drug has not yet been proven effective for those patients. Public health officials say that this study is likely to encourage other doctors to prescribe it to patients infected with HIV.

Now, let's step back for a moment and observe what's happening here. First, a number of crucial semantic distinctions are being obliterated. "AIDS", a condition or disease that is said to be invariably fatal, is now being conflated with "HIV infection", which in turn is a misnomer for testing positive for antibodies to a retrovirus that has not yet been shown to be harmful. (Readers of the Native are aware that Peter Duesberg, a molecular biologist at Berkeley, has provided a powerful, and so far unanswered, critique of the hypothesis that HIV is the cause of AIDS.) The concept of AIDS is expanding to encompass not only AIDS-Related Complex (ARC), but also so-called "HIV infection", and even membership in a "high risk group". To be a gay man is becoming more and more equivalent to being a person with AIDS (PWA).

Second, AZT is being promoted as the appropriate treatment for "HIV infection". Persons who test positive for HIV antibodies will now find themselves between the Scylla of AIDS and the Charybdis of AZT poisoning, with the long-term prognosis of the latter being worse than that of the former. This amounts to a reinstatement of the Judeo-Christian death penalty for sodomy. Lovers of other men must die. ("They shall surely be put to death; their blood shall be upon them." "The wages of sin is death.") (I have written two major exposes on AZT for the "Native", the upshot being that the FDA- conducted trials of AZT, on which approval of the drug was based, were invalid and indeed fraudulent; that there is no scientifically credible evidence AZT has benefits of any kind; and that AZT is a highly toxic drug, incompatible with life. ["AZT on Trial", Issue 235, and "AZT: Iatrogenic Genocide", Issue 258.])

Specter was not alone in putting forth this interpretation of the San Francisco study. On June 3, 1988, Paul Reger, a science writer for the Associated Press, wrote: "AIDS eventually will kill 99 percent of the people infected with the virus, according to a new study that says it takes an average of 7.8 years for the disease itself to show up." And a New York Times article by Bruce Lambert, "New York Called Unprepared on AIDS " (July 14, 1988), contained a header, "Almost all carriers of the virus are expected to become ill", and quoted Dr. James O. Mason, director of the CDC, as saying, "We have to assume that everyone infected will ultimately become symptomatic."

New York City Health Commissioner, Dr. Stephen C. Joseph, was quoted as saying:

"I don't know anybody in the field who does not agree that eventually the overwhelming percentage of infected people will have serious if not severe symptomology, in the high 80's, 90's -- as close to universal as you get in medicine."

Before analyzing the San Francisco study, which does not support the statements made by Specter, Reger, Lambert, Mason, and Joseph, a basic point needs to be emphasized. Although there is undeniably a correlation between HIV antibodies and the development of AIDS, the correlation is far from perfect, and it is only a hypothesis that the relationship is causal. Duesberg has persuasively argued that, even in patients who are dying from AIDS, HIV remains biochemically inactive, or latent; and that a virus, like anything else, has to do something to get something done. It has yet to be proven, in even a single case, that HIV has played any role at all in causing AIDS. It remains to be demonstrated whether, and if so how, HIV is capable of causing illness of any kind.

The San Francisco Study

The Science article, "A Model-Based Estimate of the Mean Incubation Period for AIDS in Homosexual Men", has the typical shortcomings of reports written by public health officials. In particular, the report contains an inadequate description of methodology, which does not even appear in one place; part of the methodology appears on the first page, and then more methodology appears, incongruously, on the second page. So far as I can tell, this is what was done:

A number of epidemiological studies have utilized a cohort of 6709 homosexual and bisexual men who enrolled at San Francisco City Clinic between 1978 and 1980, in order to participate in various studies of hepatitis B. Investigators Lui, Darrow and Rutherford obtained a subsample of 84 of these men, for whom the approximate date of seroconversion could be estimated -- that is to say, men who had a positive HIV-1 antibody test within 12 months of a negative antibody test. The authors offer the following description: "The 84 men include 83 men who were selected at random or returned for hepatitis B vaccine follow-up, could be located and gave written consent for their stored sera to be tested for HIV-1 antibody, and one man who died from AIDS in 1982."

In the time period involved, from 1978 to the present, 21 of the men (25% of the total) developed AIDS. On the average, for these 21 men, the time between seroconversion and a diagnosis of AIDS (which the authors refer to as "the incubation period") was 4.8 years.

Using these data, Lui developed an arcane mathematical model, whose projections were intended to estimate two things:

1.the proportion of the total sample of "infected" men who would eventually develop AIDS, and

2.the "mean incubation period" for those who would develop AIDS. He estimated the latter at 7.8 years. With regard to the former, the following conclusion was reached:

Let p be the proportion of infected individuals who will eventually develop AIDS.... The maximum likelihood estimate of p is 0.99 with a 90% confidence interval (0.38, 1.00)....

Confronted with this statement, Specter, who is obviously unfamiliar with statistical language, simply latched on to the "maximum likelihood estimate" of 99%, and ignored what followed. And yet the statement, "with a 90% confidence interval (0.38, 1.00)" is crucial. Translated into plain English, the above statement reads as follows:

Let "p" be the proportion of individuals with HIV antibodies, who will eventually develop AIDS.... With about 90% certainty, p lies somewhere between 38% and 100%.

Note the difference. With only a 90% confidence interval, the estimate of "p" has a 62 percentage point spread, all the way from 38% to 100%. Statistically, this means that the estimate is wildly unstable. In fact, if someone asked me to analyze data with a confidence interval anywhere near this large, I'd simply tell him to go away, and to come back when he had data worth looking at. And a 90% confidence level is rather low. Normally in research one prefers at least a 95% confidence level, in which case, according to Lui, "p" would be somewhere between 27% and 100%! At any rate, these statistics are a far cry from Michael Specter's statement, "The AIDS virus will almost certainly kill everyone it infects."

To make sure that I had interpreted the key statement correctly, I called both Kung-Jong Lui and William Darrow, and to my near amazement, they both agreed with me on almost everything. Lui said that my rewording of the conclusion regarding "p" was correct, and that the statements made in the press had been inaccurate and misleading. He said that Specter's statements, which I read to him, were wrong, and that if Specter had called him, he would have told him so. Darrow also agreed that media coverage of their article had been far from satisfactory, and that existing data were not adequate to estimate, with any degree of precision, the proportion of all people with HIV antibodies who would eventually develop AIDS.

A Representative Sample?

Even the grossly unstable estimate of "p" (38% to 100%, with 90% certainty) applies only to the sample studied: 84 homosexual/bisexual men, non-randomly selected from the San Francisco City Clinic Study. It would be wrong to assume that this sample was at all representative of the total universe of people with HIV antibodies. This is one of the most basic questions in survey research: How representative is a sample of a particular universe or population? To what extent is one justified in projecting findings from the sample to the target universe?

Michael Specter, in his article of June 3, says that "The researchers randomly selected 84 of the men for follow-up studies...." This is simply not true. (In research sampling, "random selection" has a precise meaning: namely, that every individual in the population being sampled has an equal and a known probability of being selected.) In fact, the investigators randomly selected 515 HIV-1 seropositive men from the total cohort of 6709, but were only able to determine the year of seroconversion for 84 (of whom one had been dead for 6 years). They settled for what they could get. Therefore, the 84 men may not even be representative of all seropositive men in the total cohort.

Normally reports on survey research contain a description of the sample. One wants to know the characteristics of the people studied, so one can have some idea how typical they are of the total population the sample is intended to represent. There is no such description in the Science report. However, William Darrow was also the principal author of another epidemiological report utilizing the San Francisco City Clinic cohort ("Risk Factors for Human Immunodeficiency Virus (HIV) Infections in Homosexual Men", American Journal of Public Health, April 1987). The AJPH report does cite some characteristics of a sample of 359 men drawn from the City Clinic cohort, who were seronegative when first tested (1978-1980). Darrow told me he saw no reason to assume the characteristics of this sample would differ greatly from those of the 84 men in the other study.

These 359 men were, putting it euphemistically, "living in the fast lane". They were indeed "burning the candle at both ends". With regard to recreational drug use, 84% were cocaine users, 64% used amphetamines, 51% used quaaludes, 41% used barbiturates, 20% used needle drugs, and 13% shared needles. The investigators asked about poppers ineptly, but it appears that the great majority of these men were into poppers as well. In the area of sex, 95% practised receptive anal intercourse with steady or nonsteady partners, 57% averaged more than four different sexual partners per month, 44% practised insertive or receptive fisting with nonsteady partners, and 18% shared douching equipment. In terms of medical history, 74% had been treated for gonorrhea, 73% had had hepatitis, 57% had experienced bleeding with intercourse, 30% had been treated for amebiasis, and 28% had been treated for syphilis.

I would like to make two points, as nonjudgmentally as possible. First, if the 84 men studied by Lui, Darrow, and Rutherford were at all similar to the 359 men in the AJPH study, then they can hardly be representative of the total universe of 1.5 to 3 million individuals in the U.S. estimated by the CDC to have HIV antibodies. Second, it would be surprising if people who lived like this did not become seriously sick; a lifestyle of heavy drug use, multiple venereal diseases with frequent antibiotic treatment, and unhealthy and dangerous sexual practices, may be quite sufficient to cause a condition of immune deficiency, with or without HIV or any other specific infectious agent.

Refutation: New York Blood Center Data

A basic principle of analysis is that data must make sense. This may seem too obvious to mention, but novice analysts often are slaves to the numbers they see in front of them, and will concoct bizarre explanations rather than come to grips with contradictions in the data. In actual practice, when data don't make sense, it is almost always because they are wrong. There are many ways that errors can occur in survey research -- from outright cheating, to errors in coding or study design or mathematics or sampling, to a finger slip on the part of the keyboard operator entering computer tabulation specifications. It is the task of a good analyst to spot and track down such errors.

In the case of epidemiological research, the data ought to make sense in the context of what is known about AIDS. If the findings from the Lui, Darrow and Rutherford study are to have predictive value beyond the 84 men studied, then they should bear comparison with other studies of seropositive individuals.

A study conducted at the New York Blood Center flatly contradicts the findings of the Lui study. According to a New York Times article by Lawrence K. Altman, "AIDS Mystery: Why Do Some Infected Men Stay Healthy?" (June 30, 1987):

In New York, at least 13 men who volunteered in 1978 for the hepatitis B vaccine trial were already infected with the AIDS virus [sic] and have lived for nine years without developing AIDS, according to Dr. Cladd E. Stevens, the head epidemiologist at the New York Blood Center.

An astonishing point is that the immune systems for all 13 of these men look 'perfectly normal,' Dr. Stevens said in an interview....

More astonishing, Dr. Stevens said, for unknown reasons only one of the 87 people in the New York Blood Center study who were found to have become infected with the AIDS virus [sic] since 1981 has developed AIDS.

So then, in New York only one out of 100 "infected" individuals (1%) developed AIDS, whereas in San Francisco 21 out of 84 (25%) developed AIDS. If HIV is the sole cause of AIDS, it is not possible for both sets of data to be correct, notwithstanding the possibility that the time periods may not be quite the same, or that the characteristics of the two samples may be different. The possibility that the difference (25% vs. 1%) could be due to chance is less than one in a million. If, on the other hand, AIDS is caused by toxins (like recreational drugs) and other lifestyle factors, then both sets of data might be correct -- it would mean that the San Francisco subjects pursued an AIDS lifestyle (or "deathstyle"), and the New York subjects didn't, and that in either case, HIV had little or nothing to do with the outcome.

Conclusions

Existing data do not support claims that all, or most, or even many individuals with HIV antibodies will develop AIDS. As usual, government "epidemiology" falls far short of the standards of professional survey research. However, in the present comedy of errors, the main culprits appear to be the media. Reporters like Michael Specter, lacking the necessary training, are not up to the task of interpreting AIDS epidemiology.

It is still nothing more than a shaky hypothesis that HIV has anything at all to do with causing AIDS. In a couple of weeks, an issue of Science is scheduled to run a forum or debate on the HIV hypothesis, with Peter Duesberg arguing that HIV does not cause AIDS; and Robert Gallo, William Blattner and H.M. Temin arguing that it does. It will be the first time that Gallo & Co. have been willing to defend their hypothesis in a civilized manner and in an appropriate publication, complete with references. I suspect that many readers of this debate will be shocked when they realize how skimpy, indeed pathetic, the arguments on behalf of the HIV hypothesis are. And of course, if HIV is not the cause of AIDS, what exactly is the point of attempting to estimate the proportion of HIV-infected individuals who will develop AIDS?

It is serious when death threats are directed against us. I sometimes think that too much attention and sympathy have been given to those who are sick and dying, and not enough to those of us who have healthy minds and healthy bodies. We, after all, are also targets of psychological warfare. We also are increasingly being portrayed as sources of pollution, as threats to the "innocent" heterosexual population. Perhaps we should form our own self-help organization, which might be called "mens sana in corpore sano" or "kalos kagathos" or "Gesundheit".

Our survival depends on not accepting the role of victim. If people direct death wishes at us, we should direct death wishes right back again at them. No one should be allowed to attack us with impunity. At the same time we need to retain a sense of cool: an appropriate balance of self-preservation, anger, and a sense of humor. Aside from the fact that our lives are at stake, current events really are pretty ridiculous, aren't they?

References

1.Michael Specter, "AIDS Virus Likely Fatal To All Infected", The Washington Post, June 3, 1988.

2.Kung-Jong Lui, William W. Darrow, and George W. Rutherford, III; "A Model-Based Estimate of the Mean Incubation Period for AIDS in Homosexual Men"; Science, June 3, 1988.

3.For Duesberg's ideas, see: Peter H. Duesberg, "Retroviruses as Carcinogens and Pathogens, Expectations and Reality," Cancer Research, March 1, 1987; "A Challenge to the AIDS Establishment", Bio/Technology, November 1987. Also, John Lauritsen, "Saying No to HIV: An Interview With Professor Peter Duesberg", New York Native Issue 220, July 6, 19877 and Christopher Street Issue 118; and "Kangaroo Court Etiology", New York Native, Issue 264, May 9, 1988.

4.John Lauritsen, "AZT on Trial", New York Native, Issue 235, 19 October 1987; and "AZT: Iatrogenic Genocide", New York Native, Issue 258, 28 March 1988.

5.Paul Reger, "AIDS Prognosis", Associated Press dispatch, June 3, 1988.

6.Bruce Lambert, "New York Called Unprepared on AIDS", New York Times, July 14, 1988.

7.Ibid.

8.William W. Darrow, Dean F. Echenberg, et al.; "Risk Factors for Human Immunodeficiency Virus (HIV) Infections in Homosexual Men"; American Journal of Public Health, April 1987.

9.Lawrence K. Altman, "AIDS Mystery: Why Do Some Infected Men Stay Healthy?", New York Times, June 30, 1987.

Copyright John Lauritsen 1996.