HIV/AIDS Skepticism

Pointing to evidence that HIV is not the necessary and sufficient cause of AIDS

Posts Tagged ‘cost-benefit analysis of antiretroviral therapy’

HIV/AIDS SCAM: Have antiretroviral drugs saved 3 million life-years?

Posted by Henry Bauer on 2008/07/06

In the previous post [Antiretroviral therapy has SAVED 3 MILLION life-years, 1 July 2008], I showed that the impression conveyed by “millions” is misleading. The claim actually amounts to an estimate that HAART has saved, at an annual cost of about $20 billion, only about 13% of AIDS victims, in other words about 1 in 8, which is hardly what’s implied by the commonly used description of HAART as “lifesaving”.

I referred also in that earlier post to “dishonesty” in the Walensky et al. article. That charge reflects the fact that only a by-the-way sentence on the fourth page of the article modifies drastically the claim, made in the Abstract, of “at least 3.0 million” life-years saved: “Of these, 1,184,851 years have already been realized, and 1,629,041 years are being accrued by current patients”. The claimed 3 million turns out to be less than 1.2 million! Yet that is once again fudged or masked by the last sentence of the article: “Ten years after the introduction of potent combination ART, at least 3 million years of life have been saved in the United States” [emphasis added]. Counting projected future savings as already in hand might not survive an independent audit.

Repeating the calculations in the earlier post with the lower figure of 1.2 million of actually realized savings, we find that there were saved by 2003 not 13% of patients but only 6%, at expenditure of more than $180,000 per saved life-year, or $12.5 million per life; and our productivity in GDP terms then represents a measly return of 0.36% on this human capital. Such are the numbers that Fauci apparently believes to justify current expenditure on HIV/AIDS. One can be sure, moreover, that the computer model was designed and the calculations made with a view to presenting as rosy a picture as possible. If this is the best they can come up with, then it’s time to stop talking about HIV/AIDS as a manageable, chronic but not fatal disease.

Not only is the claimed benefit of treatment much less than impressive, the claim actually lacks any solid foundation whatsoever. It relies on a computer model that makes a number of unjustifiable assumptions, and it ignores such central issues as the acknowledged toxicity of the antiretroviral treatment as well as how the definition of AIDS has changed, and thereby the health-state of people being treated.

Here is the essence of the Walensky article: “The Cost-Effectiveness of Preventing AIDS Complications (CEPAC) model was used to estimate per-person survival benefits. CEPAC is a widely published computer-based state-transition simulation model of HIV disease that incorporates CD4 cell count; HIV RNA level; ART efficacy; OI incidence, treatment, and prophylaxis; and other important clinical information [16–18, 21]. “State transition” means that the model characterizes the progression of disease in an individual patient as a sequence of transitions from one ‘health state’ to another. . . . In the model, the level of HIV RNA determines the rate of CD4 cell count decline, and the absolute CD4 cell count governs the monthly risk of OIs and death”.
One hardly needs to read any further, given that Rodriguez et al. (JAMA 296 [2006], 1498-1506) found a lack of correlation between “HIV RNA level”—otherwise known as viral load—and the rate of CD4 decline. This fact alone would be enough to vitiate the model; but there are also no valid studies of ART efficacy using untreated controls. As to prophylaxis of opportunistic infections, more is said below.

Walensky et al. considered “6 distinct eras of HIV/AIDS treatment from 1989 to 2003”. But about 50,000 AIDS deaths had already been reported up to 1989 (CDC 1990), about 35,000 of those during the AZT monotherapy years of 1987 and 1988. Some (or most or perhaps even all) of those deaths were caused or hastened by the AZT, and those lost life-years should surely be subtracted from the savings calculated from 1989 on, since HAART typically incorporates AZT or an analog of it, albeit at much lower doses than in the monotherapy era. Instead, Walensky et al. apparently seek to hide AZT toxicity by saying that “we excluded the early benefits of antiretroviral mono- and dual-drug therapy when survival benefits were more limited”, a fine illustration of double-speak: “more limited benefits” here stands for “no benefit, just caused harm”.

The first era commences with “prophylaxis for Pneumocystis jiroveci pneumonia (PCP) starting in 1989”, ignoring that Michael Callen and Josef Sonnabend had pioneered prophylaxis against PCP in AIDS victims years earlier. Callen cites data from the Centers for Disease Control and Prevention that more than 30,000 people had died of PCP by February 1989 even though the possibility of prophylaxis had been known since 1977. Callen himself had urged Fauci in May 1987 to recommend prophylaxis, but Fauci refused; nearly 17,000 PCP deaths occurred between May 1987 and February 1989 (pp. 30-31 in Michael Callen, Surviving AIDS, HarperCollins 1990). That is in direct contradiction to the claim that “88% of eligible patients in the pre-ART era were receiving OI [opportunistic infection, includes PCP] prophylaxis” (Walensky et al., p.12). How many of those 30,000 or 17,000 PCP deaths should be subtracted from Walensky’s 3 million — or actually 1.2 million — saved years of life? 30,000 lives lost to PCP, after all, already represent more than 2 million life-years, and even 17,000 lives amount to over a million-and-a-quarter life-years, either of which would wipe out entirely all the life-years claimed to have been saved between 1989 and 2003.

Walensky et al. cite an estimate that only 57% of known “HIV-positive” people are receiving treatment, and they assert that additional life-year-savings would result if more were being treated. But how many of the non-treated are avoiding antiretroviral drugs by choice? Certainly among gay men, knowledge of the fearsome “side”-effects of antiretroviral drugs has been widespread for two decades. Moreover, any reader of the official Treatment Guidelines learns that “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 . . . is greater than the risk for AIDS”. How then could Walensky et al. legitimately ignore these toxicities, as they do: “The analysis did not account for later ART-related toxicities that may result in, for example, cardiac disease or diabetes”. They wave aside the iatrogenic harm from ART even further by opining that “hyperlipidemia reduces overall life expectancy by ~1 month”. When your doctor tells you that your cholesterol is too high and that you should begin a lifelong course of expensive statin drugs — whose deleterious “side” effects also call for regular doses of Coenzyme Q10, which few doctors will tell you, however —, try responding that you have it on good authority that the potential benefit of conquering hyperlipidemia is only about 1 month of extra life. Let me know what your doctor says to that.


Noted by Walensky et al. is that “after 1992, ~70% of new AIDS diagnoses were made according to a CD4 cell count criterion of <200 cells/mm3 alone”. This criterion for an AIDS diagnosis is unique to the United States, and patients thus diagnosed may display no symptoms of illness. Thus up to 70% of “AIDS” patients receiving antiretroviral drugs in the United States since 1993 have been clinically healthy when they begin “treatment”. It would then be hardly surprising that survival rates increased from the years before 1993 when this CD4-count criterion was introduced, for initially healthy people will surely survive toxic drugs longer than people who are already ill: “projected per-person survival after an AIDS diagnosis increased from 19 months (1.6 years) in the absence of treatment to 179 months (14.9 years) by 2003, a gain of 160 months (13.3 years)” [emphasis added]. The all-knowing computer model can apparently be sure already in 2003 that patients will survive on average into 2018. But even this projection hardly justifies the assertion that AIDS is now “a highly treatable chronic condition”, given that even by 2004 — 8 years into the “lifesaving” HAART era — most deaths from “HIV disease” were still occurring among people around 40 years of age, just as two decades earlier [Table 42, p. 236, in “National Center for Health Statistics: Health, United States, 2007 with Chartbook on Trends in the Health of Americans”, Hyattsville, MD, 2007; see “HIV DISEASE” IS NOT AN ILLNESS, 19 March 2008].

Puzzling is the statement that “Mean per-child survival gains for the averted infections ranged from 60.5 years if the child was born before 1996 (before combination ART) to 45.8 years during 1996-1999, when combination ART was available”. If ART is better, why is the survival gain from it only ¾ of the earlier survival gain from pre-ART prophylaxis of opportunistic infections?


It also remains for me a continuing mystery that so many AIDS researchers, reviewers of HIV/AIDS manuscripts, and editors of journals that publish this material are so lacking in elementary numeracy as to pepper their articles with numbers like “832,179 years in ART 3”, “2,813,892” years saved, and so on. Numbers no less than words should convey meaning. The only thing conveyed clearly by “2,813,892” is that the writers take computer outputs as sacrosanct and don’t think about what the numbers mean. Those extra digits are not only meaningless, they positively distract the reader, making necessary a mental rounding-off to recognize that the substantive claim is “about 2.8 million”; not many people, after all, are used to digesting 7-digit numbers and savoring their significance. In the Walensky et al. article, the mystery of this abusive mathematical incompetence is only deepened by the fact that the authors’ affiliations include departments of epidemiology and biostatistics.


Walensky et al. “employed a model-based approach, conducting repeated analyses to explore the clinical consequences of alternative patient-care-innovation pathways”. The whole article deals not with actual patients but with “hypothetical patients”. This fulfills the suggestion, reported in a previous post, that one no longer needs human beings for clinical studies, computers can conveniently substitute [VIRTUAL HIV/AIDS RESEARCH AND TREATMENT, 17 June 2008]. The most convenient thing about this, of course, is that it’s much easier to get the results you want from a computer model you have yourself designed than from observations of real people.

Here’s the point to bear in mind whenever the gurus parade the outputs of their computer models:
A computer model can be guaranteed to mimic reality faithfully only if everything about that reality is already known in every detail. But if that is so, then one doesn’t need a computer model. Computer models are experiments carried out on surrogates of reality, surrogates that are unavoidably simplified and based on assumptions about reality. In the Walensky et al. case, the model incorporates assumptions about what happens to a person with a given viral load and CD4 count under no treatment, and what happens to individuals with given viral loads and CD4 counts under a variety of treatment regimens: all of which are based on guesses, because clinical trials with proper controls have never been carried out to determine properly the parameters needed for such a model. Moreover, as earlier mentioned, the article by Rodriguez et al. found no correlation between “viral load” and subsequent decline in CD4 counts. Further, the article ignores the well established phenomenon of “long-term non-progressors” or “elite controllers”, individuals who demonstrate that being “HIV-positive” does not necessarily lead to destruction of the immune system, illness, and death. How could the fates of non-treated “patients” be modeled when this phenomenon is ignored? When it is not even known what proportion of people are potentially elite controllers?

The outputs of this model deserve no credence whatsoever. The claim of more than 3 million saved life-years is utterly bogus. Even were it not bogus, it would reveal the claimed benefits of antiretroviral therapy to be at best marginal and procured at egregiously excessive cost.

Posted in antiretroviral drugs, clinical trials, experts, Funds for HIV/AIDS, HIV absurdities, HIV/AIDS numbers | Tagged: , , , , , , , , , , , , , | 4 Comments »

Antiretroviral therapy has SAVED 3 MILLION life-years

Posted by Henry Bauer on 2008/07/01

It used to be understood, at least by statistics gurus and the few statistically literate observers, that there are “Lies, damned lies, and statistics”, because the data can be arranged so easily to serve any desired mis-interpretation; just as survey researchers and social scientists know that they can get any desired result from a poll by phrasing suitably the questions they ask. Nowadays, however, the preferred way to fool the public is by means of computer models [HIV NONSENSE: TODAY AND EVERY DAY, 22 November; HIV/AIDS: NUMBERS THAT DON’T ADD UP, 29 November 2007].

A brother-in-law of mine years ago had been one of several vice-presidents of a major corporation. One of the VPs always succeeded in getting what he asked for in the way of annual budget, but none of the others did. Eventually those others asked the successful one for his secret. He explained that he always presented his request to the Budget Committee of the Board in the form of a computer print-out. Who would have the temerity to argue with that? Computers don’t lie, do they?

Far too many people are far too easily misled by computer outputs. Computers don’t think. They are more gullible than any human being. They believe whatever you tell them, no matter how absurd or how obviously wrong. Hence

GIGO: Garbage In, Garbage Out.

But only computer gurus and the few computer-literate users know that. Most people dissolve in obeisance when presented with something spewed out by a computer.

More illustrations pop up all the time of commentariat, media, and policy makers accepting conclusions based on computer models instead of on happenings in the real world. Yet those computer outputs are no better than the assumptions and uncertainties fed into the computer in the first place. Computers can only say, “IF this is so, then that follows”—PROVIDED WE UNDERSTAND EVERYTHING ABOUT ALL THE PROCESSES INVOLVED.
If even one assumption in the model is wrong, then the model’s output is worth not even as much as an informed guess; it will likely be totally misleading.

HIV/AIDS has offered innumerable illustrations of Garbage Out, in the form of computer-model estimates of HIV and AIDS numbers issued for two decades by UNAIDS, the World Health Organization, the Centers for Disease Control and Prevention, and in peer-reviewed articles. Several insiders have recently been forced to acknowledge that those estimates were wildly off the mark—so wildly that predicted epidemics never happened, and rates of “HIV-positive” have had to be revised drastically downwards across the globe [WHO SAYS that WE’VE BEEN VERY WRONG about HIV and AIDS? (Clue: WHO = World Health Organization), 10 June 2008].

Recently I came across an outstanding specific example of computer-model nonsense, the assertion that through antiretroviral therapy, “at least three million years of life have been saved in the United States alone” (Anthony S. Fauci, “Twenty-five years of HIV/AIDS” (Op-Ed), Manila Times, 29 May 2007) . Op-Ed pieces don’t bother giving citations, but Googling for that assertion immediately revealed its source: “The survival benefits of AIDS treatment in the United States”, Walensky et al., Journal of Infectious Diseases 194 [2006] 11-19; abstracted, unsurprisingly enough, among the News items on the website of Fauci’s Institute.

The raison d’être for the Walensky effort was that, since the US Government was allocating $21 billion to HIV/AIDS activities in 2006, questions might well be raised about what benefit the United States was getting in return.

Here’s some back-of-the envelope arithmetic. The Walensky calculation is for life-years saved through 2003. Funding began at zero in 1981 and rose steadily. If the increase was linear, then by 2003 roughly $225 billion would have been spent to save 3 million life-years, in other words, about $75,000 per life-year. Given an average lifespan of 75 years, that works out at well over $5 million per human life.

I personally don’t much care for such calculations. I think comparing lives and dollars is rather like comparing dogs and rocks. But I recognize that insurance companies, lawyers, and similar realities of modern life do need to make such calculations, and for their purposes, $5 million per life might not seem out of whack. Unless, of course, one thinks not of the rare court cases that make headlines but of all citizens. In terms of pay, benefits, and compensation for injuries, we certainly don’t compensate most members of our armed forces at $75,000 per life-year or $5 million per life. And if 300 million Americans are each worth $5 million, then our population’s total value is more than a thousand trillion dollars — 1015 —, which seems a bit high. After all, our annual Gross Domestic Product per capita is about $45,000, which on a human-capital investment of $5 million represents an unattractively low yield of less than 1%.

So if you must value people in dollar terms, the Walensky calculation suggests that we are spending significantly too much on HIV/AIDS. Of course neither the Walensky article nor the Fauci commentary made this sort of calculation. They left it at trumpeting millions of life-years saved, which the unwary might easily confuse with millions of actual LIVES saved. But at a lifespan of 75 years, 3 million life-years represents 40,000 lives, not quite so impressive a number. After all, by 2003 there had been more than 520,000 AIDS deaths in the United States, according to the Centers for Disease Control and Prevention (HIV/AIDS Surveillance Report for 2003, vol. 15). The Walensky calculation therefore amounts to a claimed saving of only about 7% of AIDS victims (40,000 out of [520,000 + 40,000] patients; 40/560 = .071) .

Again, I don’t myself care to calculate like this, because I think every life is literally invaluable. But if one insists on cost-benefit calculations to justify expenditures, as Fauci and Walensky et al. do, this 7% success-rate is somewhat less than impressive—especially for treatment that they like to describe as lifesaving, as having converted a fatal illness into a chronic, manageable one.

But perhaps one should consider only the HAART era, beginning in 1995, with the really successful treatment by “cocktail” or “combination” antiretroviral therapy?

From 1996 to 2003, Walensky et al. calculate 2.35 million life-years saved. At 75 years per life, that yields about 31,000 lives, compared to about 200,000 reported AIDS deaths during that period. So “lifesaving” HAART actually saved the lives of 13% of AIDS patients. That’s much better than 7%, of course, but not exactly what most people would understand by “lifesaving”.

Consider again costs and benefits: To save the lives of 13% of “AIDS” victims, we apparently need to spend $21 billion annually. Therefore to save all the victims, we would presumably need to spend about $160 billion a year.


I do apologize if some readers find this rather flippant, tongue-in-cheek commentary offensive. For my part, I find offensive this propaganda about millions of life-years saved when the reality is so much less impressive. The article begins by citing expenditures as the reason for calculating benefits, but then fails to compare costs and savings. As sketched out above, such a comparison indicates that the savings have been bought at a very steep price; one that society would probably be reluctant to pay were it given the choice—especially if the information were also provided about how disproportionate are these expenditures on HIV/AIDS in comparison to funds aimed at the major causes of death, cancer and heart disease [STOPPING THE HIV/AIDS BANDWAGON—Part II, 1 February 2008]. Moreover, under the Ryan White Care Act, the government acts as payer of last resort and “provides some level of care for around 500,000 people a year and, in 2004, provided funds to 2,567 organizations. The Ryan White program also funds and provides technical assistance to local and state primary medical care providers, support services, healthcare provider and training programs” (, where the pertinent official source documents are cited). HIV/AIDS is a unique instance of government-funded socialized medicine in the United States.

The Walensky et al. article is also technically incompetent and dishonest in several respects, as will be discussed in a later post.

Posted in antiretroviral drugs, experts, Funds for HIV/AIDS, HIV/AIDS numbers | Tagged: , , , , , , , , , , | 10 Comments »

%d bloggers like this: