One fact, two interpretations (Science Studies 104: “Theory-laden facts”)
Posted by Henry Bauer on 2009/04/26
When I changed my academic field of interest to Science Studies after a quarter century of research in electrochemistry, I had considerable difficulty in understanding, let alone appreciating, axioms and approaches taken for granted in the humanities and in the social sciences. The experience of putting together a multidisciplinary endeavor, a Center for the Study of Science in Society, taught me much — for example, how very difficult truly interdisciplinary efforts are, in part because academic disciplines are more like cultures than like abstract intellectual exercises. I gained further insights into those cultural differences through observing the criteria on which people in the disparate fields base their professional evaluations of one another (chapters 15 & 16, “Evaluations” & “Tribal Stereotypes”, in To Rise Above Principle: The Memoirs of an Unreconstructed Dean ).
The mantra, “facts are theory-laden”, which philosophers and sociologists in particular were wont to reiterate, gave me particular trouble for quite a long time. I’d spent a couple of decades doing or supervising experiments; reading numbers off meters or digital displays had never seemed to me problematic (apart from the inevitable range of stochastic uncertainty). The facts we observed in chemistry might not be perfectly precise numerically, but they were surely accurate, I thought, in the sense of truthfully representing the objects being studied. Yet philosophers had been arguing for many decades about this sort of thing, with something like a consensus having emerged that pure empiricism is impossible: You may “see” something as clearly and accurately as ever could be, and yet that doesn’t convey what the seen object actually “is”. A popular illustration of this is by means of ambiguous drawings:
Is it a duck looking to the left or a rabbit looking to the right? We “see” the object — black lines on white background — quite faithfully, but we aren’t sure what it is.
Applied to matters of science, consider the question of classifying things. Science really began as natural history, observing and classifying natural objects. Before modern chemistry and an understanding of the composition of materials was available, it was natural enough to classify by shape; but in that way 15th-century compendia of knowledge grouped together conical objects that, nowadays, we recognize as having no meaningful relation to one another, since they included belemnites (minerals), fossilized shark’s teeth (bits of once-living things), and axe- or arrow-heads (human artefacts). *
In other words, one cannot “objectively” recognize the significance of a given “fact”, or piece of evidence.
Facts do not speak for themselves.
An important corollary, directly pertinent to HIV/AIDS, is that controversies can persist over extended periods of time when a certain set of phenomena lends itself to opposing interpretations; see, for example, in my book about Loch Ness how every piece of evidence can be quite plausibly incorporated into the view that “The monster is a myth” (chapter 1) and equally also into the view that “The monster exists” (chapter 2).
When I was doing research in chemistry, it had seemed to me quite unproblematic, how to explain the data I gathered: either the explanation fitted with our prior understanding, or it pointed somewhere else; but I didn’t question that the facts spoke unequivocally. Well, I’ve learned otherwise. But many scientists who have not benefited from the wisdom of philosophy of science, or sociology of science, or science studies, continue to see nothing problematic in how they interpret data. That’s one of the huge problems with HIV/AIDS theory: its adherents and its groupies and its vigilantes see everything only from their own viewpoint and regard every other interpretation as simply wrong. How could “the facts” in “thousands of research papers” be “wrong”? Don’t those facts “speak for themselves”?
Well, of course, it’s not that the facts are necessarily wrong, it’s that they are equally able to “speak for themselves” a quite different explanation, namely, that “HIV” tests don’t detect a pathogen. They detect a wide range of antibodies (and in newer tests nucleic-acid bits) that correlate partly with heredity and partly with life experiences, particularly health challenges. Higher levels of “HIV” sometimes indicate poorer health, but not as a result of an infection; and in any case “HIV” is not the cause, it’s a reflection. In Dr. Christian Fiala’s nice analogy, testing “HIV-positive” is rather like running a fever.
Here’s one example where CDC researchers jumped to a conclusion on the basis of ambiguously interpretable facts. They had observed that in a group of young African men, mortality among the “HIV-positive” was higher than among the others, and they claimed this as proof that HIV causes death [Dondero and Curran, Lancet 343 (1994) 989-90]. They apparently forgot that association doesn’t prove causation; and they may well have lapsed into their ignorant supposition because they didn’t recognize that there’s an alternative explanation, namely, that those who are more ill in the first place are more likely to test “HIV-positive”, just as those who are ill are more likely to run a fever than those who are not ill.
That type of one-sided, blind misunderstanding is rampant throughout the literature of HIV/AIDS. Here’s another instance:
With the exception of conditions included in the 1987 AIDS surveillance case definition, pneumonia, with or without a bacteriologic diagnosis, is the leading cause of HIV-related morbidity and death (55, 56). In addition, several studies have shown that persons with HIV-related immunosuppression are at an increased risk of bacterial pneumonia (57-59). For example, one study found that the yearly incidence rate of bacterial pneumonia among HIV-infected IDUs without AIDS was five times that found in non-HIV-infected IDUs (58). Recurrent episodes of pneumonia (two or more episodes within a 1-year period) are required for AIDS case reporting because pneumonia is a relatively common diagnosis and multiple episodes of pneumonia are more strongly associated with immunosuppression than are single episodes. For example, data from the ASD Project indicate that the risk of an HIV-infected person having had one episode of pneumonia in a 12-month period is approximately five times higher among infected persons with CD4+ T-lymphocyte counts of less than 200/uL (320/2,411) than among those with higher CD4+ T-lymphocyte counts (90/2,792). In contrast, data from the same study indicate that the risk for multiple episodes of pneumonia in a 12-month period is approximately 20 times higher among HIV-infected persons with CD4+ T-lymphocyte counts of less than 200/uL (67/2,411) than among those with higher CD4+ T-cell counts (4/2,792) (CDC, unpublished observations)” [emphases added].
I’ve highlighted the “facts” that are clearly open to the alternative explanation that people with bacterial pneumonia, and people with multiple episodes of pneumonia, are more likely to test “HIV-positive” and to have low CD4 counts than are people without those experiences. Pneumonia is the cause; “HIV-positive” status and low CD4 counts are effects, not causes.
Increasingly insistent in mainstream discourse is the drive for wider testing, even universal testing. Since it’s now believed that HAART reduces enormously the risk of passing on “HIV”, if all the “HIV-positive” people in the world could be identified and put on HAART, “HIV” could be wiped out! [Clare Wilson, “Are we about to eliminate AIDS?”, New Scientist, 19 February 2009, 38-41]
One problem, of course, is that HAART has dangerous “side”-effects, opening the ethical question of “treating” perfectly healthy people: “Persuading everyone with HIV to start therapy purely for public health reasons could be ethically dubious” [to put it mildly, not to say euphemistically]. “Perhaps the most medically contentious part of the elimination plan, in any country, is that all those diagnosed positive would begin antiretroviral treatment immediately. At present there is no firm evidence that HIV does any damage to an individual as long as their CD4 count is above 350. ‘There are great big ethical problems about recommending treatment to someone when it’s not clinically beneficial to that person’”.
Yes indeed. The analogy with eugenics springs to mind, sterilizing people with “poor genes” for the benefit of future generations. Here is one of the benefits of an education in Science Studies: the analogy with eugenics springs to mind. For those technically (in both senses) expert only in HIV/AIDS, it doesn’t spring to mind.
But I want here to look at the interpretation of facts, not at ethical questions:
“We know now that starting treatment earlier than at a CD4 count of 200 brings health benefits. As well as reducing the risk of opportunistic infections, a large study showed last year that people who began treatment with a CD4 count above 350 are less likely to develop conditions usually seen as unrelated to HIV, such as heart or kidney disease (The Journal of Infectious Diseases, vol 197, p 1133). Researchers now suspect that long-term HIV infection causes a low-level activation of the immune system that can damage the heart, kidneys and liver. For these reasons, the treatment threshold in wealthy nations is now 350.”
The “fact” is that “HIV-positive” people who begin HAART with CD4 > 350 develop organ failure at a lower rate than those whose CD4 is ≤200. The official explanation is that it’s better to attack “HIV” earlier, before health has deteriorated from CD4 > 350 to CD4 ≥ 200. But there’s a perfectly plausible alternative explanation:
IF HAART has “side” effects that include organ failure, and IF higher CD4 counts bespeak a healthier immune system, plausibly bespeaking better health in general, then higher CD4 counts protect against the “side” effects of HAART. When you treat healthier people with organ-failure-causing drugs, they’ll survive longer than if you administer organ-failure-causing drugs to people who are already ill.
But does HAART cause organ failure? The NIH Treatment Guidelines say so quite plainly (January 2008, p. 13):
“In the era of combination antiretroviral therapy, several large observational studies have indicated that the risk of several non-AIDS-defining conditions, including cardiovascular diseases, liver-related events, renal disease, and certain non-AIDS malignancies [97-102] is greater than the risk for AIDS in persons with CD4 T-cell counts >200 cells/mm3; the risk for these events increases progressively as the CD4 T-cell count decreases from 350 to 200 cells/mm3.”
This is perfectly clear, isn’t it? Conditions that had never before been considered AIDS-defining, nor associated with “HIV”, threaten people on HAART more seriously than does being “HIV-positive”, the condition for which they are supposedly being “treated”. The higher the CD4 count when they start on these “treatments”, the less likely that the “side” effects will damage them. That’s the same “fact” as cited above from the Journal of Infectious Diseases.
But the notion that HAART rather than “HIV” is the major source of death nowadays is absolutely unpalatable, unthinkable for mainstream researchers and their hangers-on. There must be some other way of interpreting these “facts”. And sure enough there is: “Researchers now suspect that long-term HIV infection causes a low-level activation of the immune system that can damage the heart, kidneys and liver”.
Now, there’s absolutely no independent evidence for this “suspicion”. It’s precisely the sort of “ad hoc” hypothesis that is traditionally used to avoid admitting that a theory has been falsified. Since the mainstream has no doubt that “HIV” kills, and that HAART saves, there just has to be some way, like this one, to fit every fact to their theory.
But such “suspicions” should be tested. And there exists an available group of subjects on whom it can be tested almost immediately: the “long-term non-progressors” or “elite controllers” known to the mainstream, as well as the many “HIV-positive” people who belong to dissident organizations and who have remained healthy while avoiding antiretroviral drugs for two decades or more. A prospective study should compare the rate of death from organ failure among untreated “HIV-positives” with the rate of death from organ failure among people on HAART. One might also compare those rates with the rate of death from such organ failures in the population at large.
In medicine, which interpretation to choose among those available can become a matter of life and death. The belief that organ failure results from “HIV” rather than from HAART has led to further studies:
“Still, no one really knows what the effects of starting treatment earlier are. This question should be answered by a large international trial called START, organised by the US National Institutes of Health, to compare the health of people who start therapy at 350 with that of people who start at over 500.”
This is being done despite the fact cited in the NIH Treatment Guidelines, that HAART patients experience more organ failure than they do AIDS illnesses. One can reasonably expect that people with CD4 > 500 will “do better” than those with lower CD4 counts, since they are presumptively healthier to begin with and will resist iatrogenic poisoning better and longer. The “success” of the study will then be cited as support for testing everyone and treating every “HIV-positive” person, no matter how young or healthy, with antiretroviral drugs — and not for some short time, all the way to death.
Is there any way to avoid the dilemmas that stem from the theory-ladenness of facts, the possibility of interpreting any given piece of evidence in different, even opposing ways?
Of course there is. One doesn’t engage in argument over the ambiguous evidence, one looks for evidence that simply cannot be fitted into one or the other viewpoint. With respect to HIV/AIDS, such evidence is quite plentiful. For instance, the fact that “HIV” isn’t infectious, as shown by the epidemiology of positive “HIV”-tests in the United States; and the fact that mortality among “people living with AIDS” varies little with age, nothing like the exponential increase with age that is seen with actual illnesses and diseases; and the fact that “HIV” numbers and “AIDS” numbers don’t correlate chronologically, geographically, nor in their relative impacts by sex or by race.
* FOOTNOTE: William B. Ashworth Jr., “Sharks-teeth, ax-heads, and belemnites: Problems of nomenclature in 17th-century paleontology”, National Meeting, History of Science Society, Toronto, 19 October 1980.
This entry was posted on 2009/04/26 at 1:23 pm and is filed under antiretroviral drugs, clinical trials, experts, HIV as stress, HIV does not cause AIDS, HIV risk groups, HIV skepticism, HIV tests, HIV/AIDS numbers. Tagged: "HIV-positive" analogous to fever, ambiguous evidence, “HIV-positive” as marker, iatrogenic HAART, Science Studies 104, theory-laden facts, William B. Ashworth. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.