It's not Condoms that are Failing to Protect Against HIV, it's UNAIDS

[Cross posted from the Don't Get Stuck With HIV blog.]

At the beginning of this month, David Gisselquist took a careful look at UNAIDS' 'Modes of Transmission' model and found it seriously lacking, grossly overestimating HIV transmission among couples in long term relationships in Malawi. As a result of this flaw, the model gives results which appear to support the extremely racist view that most Africans in high HIV prevalence countries, male and female, engage in a lot of unsafe sex, and mainly sex with people other than their partners.

David shows how the Modes of Transmission model currently estimates that 81% of Malawi's 95,000 new HIV infections were accounted for by spousal transmission. If you remove the flaw, the percentage goes down to 20%, leaving 60% of all infections unaccounted for by the model (non-sexual transmissions from mother to child make up much of the remainder). How were all those other people infected, including the women who are said to have infected their babies?

It is very likely that a substantial number of HIV infections in Malawi and other high prevalence countries are a result of non-sexual transmission, such as through unsafe healthcare, cosmetic procedures and traditional practices. The much lauded 'ABC' (Abstain, Be faithful, use Condoms) approach to HIV prevention does not work, not because many Africans actually live up to the stereotypical 'all men are bastards, all women are hapless victims', but because HIV is not always transmitted through heterosexual sex.

Consider condoms, which are a great technology for reducing unplanned pregnancies, many sexually transmitted infections and sexually transmitted HIV, through anal and vaginal intercourse. But a number of surveys have found that HIV prevalence is very high among those who use condoms. Indeed, prevalence is often higher among those who at least sometimes use condoms than among those who never use them. The following chart is from the relevant Demographic and Health Survey for four countries.

In some cases, HIV prevalence is 50% higher among those who sometimes use condoms than among those who never use them, sometimes 100%. Shocking? Only if you think HIV transmission in high prevalence African countries is all about sex. Consider another set of figures, this time for condom use at last sexual intercourse in past 12 months. The figures for those who have not had sex in the past 12 months also raise questions (data from DHS surveys). You could suggest that people are not honest, or that people who are infected are 'abstaining', but it is far more vital to figure out exactly how people are being infected in order to prevent further infections.

Why are HIV prevalence figures so much higher among people who say they sometimes use condoms? I can only tell you what I think; condom use is completely irrelevant to non-sexually transmitted HIV. That sounds obvious, but UNAIDS insist that almost all transmission is through heterosexual sex, yet they stand by figures like these. It is not possible for HIV prevalence to be so much lower among those who never use condoms if almost all HIV transmission is sexual. But there may be an explanation for why those who sometimes use condoms seem so much more likely to be infected.

HIV prevalence is often highest among wealthier, urban dwelling, employed, female, better educated people who live in wealthier countries that have reasonable access to reproductive healthcare services, a relatively low population density and sometimes a higher urban population (but not always). People who answer that description, people who can tick at least some of those boxes (some of the factors are interdependent), it seems, are also more likely to use condoms.

So it is not a case of people with the above characteristics using condoms, yet still being more likely to be infected with HIV, but rather a case of those same people being more likely to be infected with HIV through unsafe healthcare or some other non-sexual route. Once you challenge the sexual behavior paradigm the rest is clear: condoms are irrelevant to non-sexual HIV transmission. It only sounds unintuitive if you keep clinging to the sexist, racist and extremely dangerous reflex about sexual behavior, so beloved by UNAIDS, WHO, CDC, PEPFAR, the Gates Foundation and various universities that have been prominent in the HIV industry.

Given what we so often hear about HIV being inextricably linked with poverty, unemployment, lack of education, isolation, poor access to health services, etc, it is worth emphasizing that the virus may often be more closely linked to the opposite of these factors. Of course, all of these factors are abhorrent and it should be the aim of every wealthy country to ensure that such conditions are alleviated. But if HIV is being transmitted through unsafe healthcare and other routes, all healthcare development must be SAFE healthcare, all HIV education must include information about non-sexual transmission, all employment 

and environments must exclude risks of bloodborne transmission of HIV, as much as possible.

So first we need to recognize that HIV is not solely transmitted through 'unsafe' sex and that it can be transmitted, perhaps far more easily, through unsafe healthcare, cosmetic procedures and traditional practices. ABC 'strategies' do not work because HIV transmission is not all about sex, not because Africans are too careless, promiscuous or ignorant (or even 'disempowered') to follow its patronizing advice. Safe sex has its place, but safe healthcare is a far more urgent issue in high HIV prevalence African countries right now. It's not condoms that are failing to protect people against HIV, but the intransigence of UNAIDS and the rest of the HIV industry.

Misinformation from UNAIDS' flawed Modes of Transmission model

[By Dr David Gissselquist - Cross-posted from the 'Don't Get Stuck With HIV' website and blog]

To defeat HIV/AIDS in Africa, UNAIDS recommends: “Know your epidemic.” The best way to do so is to investigate to trace the source of infections – especially in children with HIV-negative mothers, virgins, and married people with HIV-negative spouses and no outside partners.

But that’s not what UNAIDS urges African governments to do. Instead, UNAIDS urges governments to use its Modes of Transmission (MOT) model to estimate numbers of infections from various risks.

But the MOT model contains a glaring error. Because of this error, whoever uses the model ends up estimating far too many infections coming from spouse-to-spouse transmission.

In Uganda, for example, the MOT model estimates that 60,948 married adults got HIV from their spouses during 2008. This is two-thirds of the model’s estimated total new infections from all risks in Uganda in 2008.

The MOT model got this number by supposing that 5.9% of married adults (421,000 adults) were HIV-negative with HIV-positive spouses, and that 14.5% of these spouses at risk got HIV from husbands or wives in 2008 (60,948 = 14.5% x 421,000).

But the number of spouses at risk is far, far less. Uganda’s 2004/5 HIV/AIDS Sero-behavioral Survey reports that 6.2% of husbands and 5.2% of wives were HIV-positive.  But – and this is the important fact the MOT model ignored – most HIV-positive husbands and wives were married to each other. Only 2.8% of wives and 1.8% of husbands were HIV-negative with HIV-positive spouses.

Overall only about 2.3% of married adults (averaging 2.8% of wives and 1.8% of husbands) were HIV-negative with HIV-positive spouses – only 222,000 vs. the 421,000 estimated in the MOT model. If 14.5% of these 222,000 adults got HIV from their spouses in a year, that would account for 32,100 new infections (14.5% x 222,000), far less than the 60,948 estimated in the MOT model.

Why is this important? Because if fewer infections are coming from spouses, how did so many Ugandans get HIV in 2008? In other words, the MOT not only over-estimates HIV from spouses, but also underestimates infections from other risks.

What risks are underestimated? Hold on now! Don’t run away with sexual fantasies about young people and some married adults having too much fun with non-spousal partners. Indulging in racist and stigmatizing sexual fantasies is something too many official AIDS experts like to do. But the evidence does not support such fantasies. The best information on sexual behavior does not come close to explaining Uganda’s epidemic.

Setting aside sexual fantasies, the underestimated risks are more likely to be those that UNAIDS’ staff and other health professionals want to ignore – skin-piercing procedures with unsterile instruments, such as injections, dental care, manicures, etc. This is true not only in Uganda but also in more than 15 other African countries that have used the MOT model to get ridiculous figures on numbers of HIV infections from spouses.

Remember how we began: The best way to “know your epidemic” is to trace infections. Let’s challenge HIV/AIDS researchers -- finally -- to do their job. Although it’s decades too late, tracing is still needed to find all the important risks and to stop Africa’s generalized HIV/AIDS epidemics.

[Note: This blog summarizes evidence and arguments in: Gisselquist D. UNAIDS' Modes of Transmission model misinforms HIV prevention efforts in Africa’s generalized epidemics, available at:  http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2315554.]

Lack of Logic in the Received View of the HIV Pandemic

Something I have always found mysterious about UNAIDS' view (it's something of a received view) of the course of the HIV epidemic is that they estimate that the number of new infections peaked in Sub-Saharan Africa (SSA) some time in the mid 1990s. And they reckon that the reason new infections began to drop from then on can be put down to the success of HIV prevention and education programmes in changing the sexual behaviour of people, especially men who have sex with men (MSM), commercial sex workers (CSW), intravenous drug users (IDU) and young women.

With few exceptions, most SSA countries were doing very little to treat people with HIV or to prevent the transmission of HIV in the 1990s. Treatment was in its infancy and was inaccessible to the majority of Africans. And where prevention programmes had been implemented, they consisted of little more than mass education campaigns. They had very little influence on people's behaviour in the 1990s. And why would they have much influence? They had only started and only in a few countries, Uganda being one of the countries that started HIV prevention early. But even the nature and effectiveness of Uganda's HIV prevention campaign is still being hotly debated. Prevalence there has changed little in years and sexual behaviour indicators have long been sliding in the wrong direction.

What bothers me is that even if widespread prevention activities started in the mid 1990s, it would take many years for them to have much effect. That's if they actually had any effect at all. Ok, I can't research every country in SSA, but in the case of Kenya, very little was being done in the 1990s. It was only in the early 2000s that some serious work started, say 2002 or 2003. And the Kenya Aids Indicator Survey (KAIS) makes it quite clear that HIV prevalence, which had been dropping before 2003, actually increased and is now higher, after half a decade of HIV prevention work.

What I'm getting at is this: if rates of HIV transmission peaked in the mid nineties, then it did so for some reason other than the fact that every country had implemented widespread prevention programmes. The reason I suggest this is because prevention just wasn't a big thing then, at least, not big enough to explain why the epidemic started to 'decline'. I'm not saying that rates of transmission didn't drop, just that they didn't drop because of prevention programmes.

Another reason for thinking that prevention programmes didn't have much influence on rates of HIV transmission is because even after they did start, there is little evidence that they could have been the cause of the drop. There is plenty of evidence that most current HIV prevention programmes have little or no effect. In Kenya's case, scaling up HIV prevention programmes seem to have resulted in an increase in prevalence, the total number of people living with HIV. This doesn't tell us if transmission rates have decreased, so what about transmission? Are there still lots of people becoming newly infected?

According to the KAIS, transmission patterns are changing. Numbers infected in urban areas have dropped but numbers infected in rural areas have increased, especially among men. The majority of Kenyans, 75% or more, live in rural areas. Poorer and less well educated people are now being infected in greater numbers. The majority of poor and less well educated people live in rural areas and most Kenyans are poor and badly educated. These trends all follow what KAIS refer to as a 'rapid scale up of HIV prevention, care and treatment services'.

A recent article in AllAfrica.com quotes UNAIDS as claiming that their successful prevention and education programmes have *finally* begun to change the behaviour of those who are most at risk. If this is only happening in recent times, how can they claim that it had anything to do with a decline in incidence that began in the mid 1990s. But Kenya, along with many other SSA countries, have explicitly not targeted some of the groups who are thought to be most at risk, MSM, CSWs, IDUs and young women. The well presented 'Modes of Transmission Survey' for Kenya makes it quite clear that these groups are still being ignored.

There may be isolated signs of people's behaviour changing in some ways. All sorts of movements may have achieved great things, especially relating to HIV treatment and increasing access to treatment. I certainly wouldn't claim that all the billions that have been poured into HIV for over two decades has been wasted. But I have yet to see clear evidence that HIV transmission has declined as a result of prevention efforts. I think the epidemic has its own dynamics, like any epidemic, but I am not convinced that the enormous Aids industry has had much influence on its course. I just hope I'm wrong.