Open letter: To HIV epidemiologists and surveillance experts

What is causing the HIV prevalence decline in Zimbabwe?

Recent reports state that ‘official’ HIV prevalence estimates in Zimbabwe have continued the decline seen in recent years: A decline was first reported in 2004, and again in 2006, and new figures indicate the downward trend has continued, with rates falling by 10 percent over the past 5 years.

What does this HIV prevalence decline really mean?

Are HIV prevention programmes working? That would be great news, of course. But with the current economic and health service situation in Zimbabwe, some people are sceptical and looking for another explanation of the drop in HIV. Are they underestimating the resilience and AIDS competence of the people of Zimbabwe?

Or is it time for us to start thinking that maybe some of the factors that influence HIV prevalence (good and bad) are not always obvious or intuitive, no matter how familiar the term now seems? I’ll give a few examples of the discussions currently heating up the HIV listservs, but really hope that some of you experts will weigh in on this to put me, and others straight.

I vaguely recall similar arguments being made by two researchers (Stoneburner and Low-Beer) in the late 90s, when they looked closely at apparent HIV prevalence declines in Uganda. So tell us, were they wrong, or just ahead of their time?

What drives HIV prevalence up?

Instinctively, we would probably say right away that HIV prevalence goes up when more people get infected with HIV (bad).

But presumably another perfectly reasonable explanation for rising HIV prevalence is better access to care and anti-HIV drugs. Think about it: Anti-retroviral treatments make people with HIV live longer, so it follows that significantly increased access to ARVs and other care services (good) should be expected to increase HIV prevalence.

What’s more, common sense tells us to anticipate this effect being most evident during the roll-out years of initial ARV programmes in particular (i.e. now in many places). Right?

What drives HIV prevalence down?

Again, instinctively you probably would say that HIV awareness and prevention programmes are working (good), people are changing their sexual behaviours (good), and as a result HIV prevalence is falling.

Another feasible explanation, presumably, is that more people with HIV are dying (bad).

If there is poor access to health care, treatments, and/or poor nutritional status among people with HIV, the delay between HIV infection and disease, and ultimately to death, is thought to be shorter (bad). If the number of people dying goes up relative to the number of new HIV infections, then HIV prevalence should be expected to down.

These factors are not mentioned much in news about of (falling) HIV prevalence, including in the recent coverage of Zimbabwe’s HIV ‘decline’. The headlines: "Youth getting the message" or "Awareness changes behaviour" leave no room for anything other than the belief that HIV prevention is working.

There is also a synergistic effect to consider that combines three factors to influence HIV prevalence – possibly in a compounded way:

1. Personal witness:

As more people with HIV find out about their HIV status, eventually get sick and pass away, more and more people are subject to the indirect impacts of the epidemic, and even more are personal witness to its ravages. They see their friends, family members, neighbours and lovers succumb to disease and die, they see children left without parents or teachers, and they see family assets stripped, food insecurity get worse and other impacts. This personal witnessing of the epidemic has to have a huge impact on people and their perceptions of risk.

Let’s just say for argument that some people are significantly affected by ‘seeing’ the HIV epidemic for themselves. This means that a proportion of people are going to voluntarily change their sexual behaviour as a result, and take themselves out of HIV’s way.

2. HIV has already ‘removed’ a proportion of people:

Remember too that in a country with sustained and high HIV prevalence, combined with less than ideal health services, a significant proportion of sexually-active people are already taken out of the HIV vulnerability ‘equation’ by sickness and death itself.

3. Out-migration

Poor income prospects and sky-rocketing inflation has also made labour out-migration from Zimbabwe for work even more frenetic. Who typically leaves for work? Young, sexually-active people.

Under the conditions currently seen in Zimbabwe (a relatively small country of 11.6 million people after all), should we expect to reach a point where the pool of uninfected, vulnerable and sexually-active people is going to get smaller and smaller?

Just like the labour shortages seen in rapidly-expanding economies, isn’t it possible that HIV just can’t ‘find’ enough young people to maintain previously high HIV prevalences?

Zimbabwe’s AIDS programme (and its international partners, incidentally) is openly celebrating the declining HIV prevalence as evidence that its HIV prevention programmes are working, but that could easily be just a part of the full picture.

There is a chance that one day soon we shall, in fact, look back and lament that ‘some’ HIV prevalence declines were evidence that specific nations or communities were witnessing the unnecessary destruction of entire generations, that people are not living with HIV as long as they could or should be, and that the socioeconomic conditions are driving young people elsewhere and to some exent at least out of HIV’s way.

Is this really that difficult a situation for experts to understand and explain? The AIDS intelligentsia must be worried how the “ARVs drives up HIV prevalence” story will be interpreted, but that’s no excuse to sit on it.

There must have been similar deliberations around how to make sure that the “circumcision prevents HIV transmission” data was not misinterpreted. Like then, Zimbabweans in particular have a right to know what declining HIV prevalence in their country really means.

With rumours flying around that the new UNAIDS Epi report (to be published today) will include ‘news’ of general declines in HIV prevalence in some countries (because of changes in the way estimates are made), there may have never been a more important time for epidemiologists to come clean and explain what they think declining HIV prevalence means under various HIV epidemic stages.

We are all ears.