Turned up to eleven: Fair and Balanced

Monday, April 15, 2002

Several people (Charles Murtaugh and J Bowen) have mentioned Michael Fumento's The Myth of Heterosexual AIDS on their websites in posts I have responded to, and after the suggestion by "No Watermelons" that I check it out, I did. Here is my gut reaction, from browsing through a couple of his articles, and consulting CDC for additional information. Many of Fumento's assertions are true, especially those that revolve principally around media coverage, rather than actual scientific data. Like all controversial stories, there have been outlandish claims, ludicrously bad predictions, and bitter political and scientific battles over description, prevention, and control of the AIDS epidemic. Mr. Fumento makes a very controversial claim, however, which is that heterosexual AIDS does not exist!! (actually, he is claiming that it is very rare, and that the epidemic is over).

The biggest problem with Mr. Fumento's work, however, is that he continually conflates and allows the reader to confuse HIV infection with AIDS. This is a vexing problem for a science writer, to be sure. HIV, the virus that almost everyone (finally!) agrees causes AIDS, has a highly variable latency period. As an aside, there is research suggesting that this is not latency in the traditional sense at all, but merely the result of a long and difficult fight by the immune system against a relatively slow growing but persistent virus. That is a debate for another time and forum. What is actually happening right now in the HIV/AIDS epidemic, however, is much more complex than Mr. Fumento or his detractors would like to believe.
Before I continue, let me address the issue that originally brought this to the fore, which was my AIDS in Africa post. In the writings I have read so far on Mr. Fumento's site, he does not address the issues I mentioned at all, except to say that HIV/AIDS advocates add in numbers from the African epidemic to make it seem like a bigger problem. Here is the only reference I have found (I've read 4 of his articles so far);

Another trick we’re increasingly seeing is a blending of the U.S. AIDS epidemic and the worldwide pandemic.

So, in summary, I don't think Mr. Fumento is much of a source on AIDS in Africa
Mr. Fumento has made a career out of this one subject, and his basic thesis is this; number of heterosexual cases is high because people lie about their homosexual experiences. I won't dispute that there may be discrepancies between what people tell the epidemiological surveyor, and the realities, but it seems to me that Mr. Fumento is stretching when he suggests that this accounts for most of the heterosexual cases in the U.S. I will quote at length from Mr. Fumento's article in 1999 from American Spectator, the infamous, now defunct, conservative magazine, known for its campaign against Bill Clinton (just had to get that in there!)

Sex, Lies, and Florida’s Faux Heterosexuals

Yet then, as now, while even the CDC’s statistics didn’t show the epidemic the alarmists repeatedly claimed they did, the data nevertheless suffered a huge flaw which the agency has clearly acknowledged – and even proved – internally but never acknowledged outwardly. Its definition of "heterosexual transmission" actually means nothing more than that’s how the person claimed to have gotten the disease. If a man diagnosed with AIDS has had anal sex with 1,000 men and shared needles on 500 occasions, and he tells his local health department he had no risk factors other than sex with a woman, he would go straight into the heterosexual transmission category. Again, it hardly takes the proverbial rocket scientist to know that a lot of men are going to be embarrassed to admit to homosexual activity or worried about admitting to illegal drug usage and would simply deny these risk factors.

This is an accurate statement, as far as I know, but accompanied by rather silly hyperbole. Is there any evidence of such a person making that claim? It seems to me highly unlikely that vast numbers of gay men and I.V. drug users are lying. These populations are in fact quite small compared to the heterosexual population.
Let's do a little thought experiment. Suppose that approximately 10% of the gay population of the United States were to lie about it when asked (this is a guess). Lets set the gay population at 5% of the US population (low end estimate is 3%, high end 10%, but I think that is a myth, and the real population is 3-5%). For the purposes of our thought experiment, we will classify anyone who has had intercourse with someone of the same sex as gay. Only MSM (men having sex with men) are counted in these surveys. I don't know the percentages in the gay population, but lets just assume that it reflects the population at large (~50% men). Therefore 0.25% of the U.S. population is composed of gay men who are lying about it. This is about 700,000 people (very rough calculation, based on 280 million population). The low-ball estimate of people infected with HIV is 800,000 (CDC), and 42% are MSM (336,000), or ~10% of the "out" gay male population. If we assume that IV drug users wouldn't lie about their sexual preference, then we are left with the 33% who claim to be straight, non-IV drug users (264,000 cases). The overall gender percentages make sense if they split 50/50 in the non-MSM category, which leaves us with 132,000 men, claiming to be straight, but really gay men who are HIV-infected. The percentage of "closeted" gay men infected with HIV is therefore maximally ~1/6, or 18%. Not outlandish, but divergent from the "out" population. We can play with these numbers a bit, and make it less or more plausible, but for now, I think this possibility can be accepted.

As an aside, no scientist disputes that female-to-male HIV transmission is much harder than male-to-female or male-to-male, but the question has always been whether or not this "barrier" would be sufficient to slow disease spread. It appears, in the US, that HIV spread in the population is fairly well contained, in a population that contains a plurality, if not a majority, gay men. This does not mean, however, that heterosexual AIDS is a "myth." If it is, then where do all those HIV + women come from? Are they all IV drug users lying too?

The CDC knows this from common sense and from history. At the turn of the century, it was common to blame contracting syphilis on touching doorknobs, using public toilets, or drinking from cups left at public fountains. The CDC also knows it from the New York City experience back when its health department still conducted individual interviews with men, and it knows it from its own research.

Early 20th century history notwithstanding, this passage is irrelevant. Of course some people lie; everyone knows that. The question is, how many, and how do you adjust for them?

It was the CDC itself that sent epidemiologists to Florida to find out why that state contributed so many cases to the heterosexual transmission category. Their study, published in the American Journal of Public Health in 1993, described how they analyzed the cases categorized as "heterosexual transmission" in two southeastern Florida counties and found, just by going through the patient files, that about a fifth had been misclassified. For example, a diagnosis of anal gonorrhea in a man is a pretty good presumptive indicator of homosexual activity, and many of the male "heterosexual transmission" cases had medical records indicating treatment for this. Although some were unavailable for interview (presumably having died), most were and some of these were also reclassified based on their own admissions. Ultimately, of the non-Haitians, slightly over half of the presumed heterosexual men were reclassified, as were over ten percent of the women. An additional third of the men re-interviewed but not reclassified had evidence of anal disease that may or may not have been caused by sex with another man.
(emphasis added)
Here we have some concrete evidence, according to this, in 1993 ~1/2 of the men were reclassified as MSM, so in our above sample, that implies 66,000 straight men with HIV. Not exactly a "myth". To contrast, there were 364,000 cases of gonorrhea, 12% in the MSM population, or 43,680. Is that a "myth"? 66,000 straight men with HIV is 1.5 per 100,000. Auto accident deaths in the total population (not fair, but gives perspective), are .0001785 per 100,000. In any event, I think the term "myth", which was clearly meant originally to counter the idea that HIV can spread in the heterosexual population as easily as in the homosexual population, is now being abused.

There is no reason to believe that Florida’s recording system is any poorer than that of most states, therefore no reason to believe that if the CDC looked closely at cases categorized as heterosexual transmission in the other 49 it wouldn’t come up with similar findings. This would be more consistent with the evidence from Nancy Padian’s study. There ought to be a big fat asterisk next to the "heterosexual transmission" category in the CDC’s semi-annual HIV/AIDS reports, with the footnote stating: "This is based strictly on the assertions of the diagnosed person and has not been confirmed."

Oh, really? Based on Florida's recordkeeping in recent history, this statement is deeply suspect! (OK, that's a cheap shot, but it's fun). Seriously, though, is this statement backed up by any data at all? Don't you think that there are lots of states that may or may not have better public health systems than Florida? Perhaps some actual research would help...Based on my observations, CDC is highly dependent on academic laboratories for help in surveillance for infectious disease, so the quality of information depends largely on 1) the state public health system, and 2) the academic research going on in a particular disease in that state.

I will summarize with what I think is a fair assessment of Mr. Fumento's work. I think he should be given credit for exposing the hyperbolic rhetoric of some within the Public Health community. In fairness, I think there was a justified fear that the picture of HIV/AIDS as a "gay" disease would lead to apathy towards this infectious disease within the community (scientific and general) which would deter research. Clearly, heterosexual men and women are at less risk than homosexual men in particular. It seems to me, actually, that being a lesbian would be your safest course (of course, this isn't an option for me!). Nevertheless, as I stated before, this is irrelevant to the situation in Africa, which is what I was interested in to begin with. In Africa, they have different environmental (other endemic diseases, poverty, poor drinking water) and epidemiological (an additional widespread strain of HIV, HIV-2, which has a different disease course) factors which alter drastically the impact of the disease, without even addressing the difficulty of getting treatment to people there. Mr. Fumento is correct to say that the expansion phase of HIV is over, and the disease has leveled off in the population, in terms of new infections. Interestingly, this means the number of people with HIV will rise, because AIDS death rates have fallen. Public health measures to treat pregnant women with HIV have reduced vertical transmission (mother to child), and better blood screening methods have reduced the rates in blood transfusion to almost zero. What is interesting to me, however, is that Mr. Fumento's credit-worthy work in exposing the overwrought rhetoric of what I am sure he would characterize as "the left" is matched by the overheated rhetoric of his supporters. HIV/AIDS is a serious disease, and a worldwide epidemic issue. Ignoring it will not make it go away. The U.S. has been fortunate. We have a very good public health system, and that system, among other things, has allowed us to contain this disease. Other people around the world are not so fortunate. They don't care what percentage of American gay men with HIV lie about it. They just want to live.
UPDATE: Here is a site that sums up Fumento's book more succinctly than I did.
UPDATE II: Jay Manifold writes to correct my bad early morning math. The comment and correction are in the comments section, but the actual rates per 100,000 for HIV infection in straight males is 49.6, and the actual rate of auto accident deaths is 17.85. In retrospect, I am not sure that comparison was valuable, but these #'s should be correct, in any case.