Stephen Lewis: UN "Sloppy" on AIDS Numbers

Stephen Lewis says the United Nations, where he used to work, "was stubborn and it was sloppy" in its compilation of global AIDS statistics. The result was last week's embarrassing need to restate the number of people living with HIV from 40 million to 33 million.

The tone is so blistering that one wonders if Peter Piot, head of UNAIDS, will take it as a personal attack from Lewis, the former UN Special Envoy for HIV/AIDS in Africa?

I'll have more to say in future posts, but wanted to post what Lewis, who is now co-director of something called AIDS-Free World, said as soon as possible. The speech itself has been making the e-mail rounds, but I have also received it from Lewis's executive assistant.

Here's what Lewis told the World Health Editors Network last Friday (Nov. 23) at the Imperial College in London:


"It’s fair to say that any speech designed to address leadership and AIDS must start with the events of this week: the publication of the UNAIDS Epidemic Update, 2007.

Let me get two matters out of the way at the outset. I think the new set of numbers is much closer to the truth of the pandemic, although I’m inclined to believe that they’re still too high and that another awkward revision lies ahead. Second, I don’t believe for a moment that UNAIDS inflated the figures for the purpose of extracting money. That seems to me too conspiratorial by half. If by some bizarre happenstance, Machiavelli’s apprentices were involved, they make lousy fund-raisers: we’re billions of dollars short of where we should be, old estimates or new.

I take issue with the report on different grounds. For years, knowledgeable epidemiologists have been telling the UN that the figures were too high. They didn’t whisper their criticisms: they wrote books and articles. They lobbied behind the scenes. No one paid them heed. It doesn’t take a Nobel prize statistician to guess that prevalence rates based on urban antenatal clinics should not be extrapolated to the entire country and presented as holy writ. That became compellingly clear when the spate of population-based household surveys, country after country, invariably showed lower prevalence.

But the UN chose a course of delay and dithering. It can never admit that it’s wrong. So finally, and predictably, came the moment of truth: the result is an overall prevalence rate that is lower by almost seven million than last year’s estimate.

Sure, it can be rationalized by arguing that it’s just a methodological adjustment, rooted in superior statistical-gathering techniques. And that might even be persuasive if there hadn’t been academicians and epidemiologists clamouring for revision for years. The UNAIDS explanation sounds good; the fine print has the ring of arcane scientific authority. But down here, in the mortal universe, where people aren’t easily taken in, it doesn’t wash.

The UN was stubborn and it was sloppy. In the process, it undermined public confidence in the reliability of the figures, introducing completely unnecessary levels of doubt, contention and confusion.
Where HIV/AIDS is concerned, there is no room for the jolting of confidence. The new estimates confirm a continuing apocalypse for sub-Saharan Africa: 22.5 million infections, 61% of them women, 68% of world-wide infections, 76% of all deaths, 11.4 million orphans … this is where the focus must be, this is where it should always have been; not a report cluttered by mathematical adjustments so that virtually every story that’s written begins with the news of a statistical volte-face. If the recording of data had been more scrupulous all along, we could have welcomed this report as the latest instalment in a record of declining numbers, showing some strong hints of progress, and plausibly leading to universal access for treatment, prevention and care by 2010.

Instead, all of us have to run to the trenches to remind the world that more money is still desperately needed and that the situation, in many places, remains grim, bleak, funereal.

As a matter of fact, at the risk of knocking the nail through the wall, allow me to add that I earlier used the word ‘sloppy’ advisedly. For a number of years in the 1990’s, I was the Deputy at UNICEF and oversaw a raft of publications. I would never have permitted this Update to go out as is.

Let me explain by way of example.

In one of the dense explanatory notes, there’s a statistical adjustment described that attracted little notice, but seems to me to be of enormous import. If I may attempt a straightforward simplification, the report seems to be saying that, on average, the population-based random household surveys in countries with generalized HIV epidemics produced estimates that were 20% lower than the estimates produced by antenatal surveys. Therefore, says the report, for all those countries with generalized epidemics that have not yet done random surveys, we’re applying a reduction factor of .8 … that is to say, the new figures will be 80 % of the old figures.

That’s quite a reduction! Don’t you think it might have been useful to produce a table (there were, after all, tables galore) to show what it might mean, in practice, country by affected country? No such luck.

Instead, the report notes that “Of the total difference in the estimates published in 2006 and 2007, 70% are due to changes in six countries: Angola, India, Kenya, Mozambique, Nigeria and Zimbabwe.” I was stumped by Angola because its prevalence rate is so low, it’s hard to imagine that it could make a great difference to the numbers. India everyone knows about. Kenya and Zimbabwe have oft been talked about (although the alleged decline in Zimbabwe seems to me to be complicated by the numbers of deaths, the out-migration, and the difficulty of reliable surveys in a sociological wasteland). Nigeria was clearly possible. But Mozambique was a mystery because I remembered from my last visit that the most recent data had shown an increase in prevalence.

This is no small matter. This is a report that is telling the world that the dramatic revision in the figures, and the figures themselves, are now the best that science has to offer.
Naturally, therefore, I looked at all the entries in the report dealing with Mozambique. On page 12, there is a section on “Recent HIV and sexual behaviour trends among young people” that chronicles reductions in HIV prevalence in various countries from Kenya to Malawi to Zimbabwe to rural Botswana. However the paragraph ends with these words: “There was no evidence of a decrease in HIV infection levels among young people in Mozambique …” (emphasis mine).

On page 13, there’s an elaborate table titled, in part, “2006/2007 analysis of trends among 15 to 24-year-olds in high prevalence countries: HIV prevalence among pregnant women, 2000-2006 in sentinel surveillance systems …” Mozambique is in the table. Below the list of countries, there’s an asterisk which reads: “Analyses of countries with more than three years of data based in the following number of consistent urban and rural sites …” For Mozambique there are 20 sites listed, South, North and Central. Beside Mozambique is another legend symbol which reads: “No evidence of decrease” (emphasis mine).

On page 15 we encounter the critical “Regional Overview”, leading with the section on “Southern Africa”. It indicates that this sub-region includes the eight countries with prevalence rates over 15%: Botswana, Lesotho, Mozambique, Namibia, South Africa, Swaziland, Zambia and Zimbabwe. The narrative then points out, with some gratification, that in Zimbabwe there is an actual decline in national HIV prevalence, and that “the epidemics in most of the rest of the sub-region have either reached or are approaching a plateau” The next sentence reads: “Only in Mozambique latest HIV data (in 2005) have shown an increase in prevalence over the previous surveillance period.” (emphasis mine).

Finally, on page 17, Mozambique is dealt with separately, as are all the countries of the sub-region. The opening paragraph reads “In the other lusophone country of this sub-region, Mozambique, the epidemic has again started to increase in all three zones after appearing to have stabilized in the early 2000s.” (emphasis mine … the three zones are south, central and north, and the other lusophone country is of course Angola.)

If I was in the Ministry of Health in Mozambique, I would want to know why all of the narrative evidence of the report seems to contradict the assertion that Mozambique is one of the six countries in the world that has most significantly contributed to the reduced numbers, 2007 over 2006. And if the explanation is --- as it must be --- the application of the new 80% rule applied to countries, ie Mozambique, that haven’t yet done population-based surveys, then those figures should be set out conclusively. Otherwise this report does what so many other reports have done: it slides casually over critical information. It confuses where it ought to be clarifying. It’s blurred where it should be coherent.

For me, the Epidemic Update, 2007, is simply a symbol, a symbol of insufficient leadership, within the United Nations, against the pandemic of AIDS. The time has come --- we’re just a week away from another World AIDS Day --- for the new Secretary-General to throw the full weight of his office behind a campaign to subdue the pandemic, with a particular consuming focus on Africa.
The agenda lies in the UNAIDS report.
For example, one of the most startling statistics is the revelation that women now constitute 61% of the infections in Africa … close to 14 million women infected. There are no words. It’s a catastrophe rooted in gender inequality, and everyone in the highest citadels of the United Nations knows it, but virtually nothing changes.

We have a report from a High-Level Panel on UN Reform, pointing out the lamentable UN record on women, and recommending the creation of a new international agency for women. The proposal lies dormant on the order paper of the General Assembly, crying out for leadership from the Secretary General. The Deputy Secretary General has spoken, strongly and bravely although, given the inevitable and nasty internal rivalries, her words are too often given to rhetorical sleight-of-tongue.

Where is the UN Secretary General when the AIDS pandemic rages, and the women of Africa need him most? No one pretends that the women’s agency is the sole answer, but you can bet that things would not be so excruciatingly horrendous if women had an international vehicle to draw upon, with resources and voice.

So, too, women in conflict zones. My colleagues and I carefully watched the Security Council debate on Resolution 1325 just a couple of weeks ago, and have read carefully the proceedings of the Security Council debate this week on the protection of civilians in armed conflict. The speeches are getting better: more feeling, more informed, more urgent (that is particularly true of the Undersecretary General for Humanitarian Affairs; on the other hand, the Secretary General’s speech last Tuesday on the protection of civilians in armed conflict was so pro forma as to make one weep. He managed to mention Sudan, Somalia, Afghanistan, Iraq and a passing aside to Darfur. Not a syllable on the Democratic Republic of the Congo). When you have a savage war on women, as in the DRC, with huge implications for transmission of the AIDS virus, speeches are the road to hell. When will the United Nations actually take hold? There are suggestions cited of the Secretary General leading a new campaign to eliminate violence against women. I recently saw an early draft of this potential Secretary General initiative, and I can categorically say that in more than twenty years of association with multilateralism, I’ve rarely seen anything more vapid, fatuous and insubstantial. It was as if the illusion of progress, dressed up in the Byzantine underworld of United Nations processes was sufficient unto itself. Public relations for inconsolable grief.

The Epidemic Update report acknowledges that only in a few instances has behaviour change played a serious role in the reduction of the numbers. In fact it says that the differences in estimates “result largely from refinements in methodology rather than in trends in the pandemic itself.” There is no question that a small proportion of the decline stems from changes in behaviour, but there is equally no doubt that the vast majority of the difference lies in the methodological adjustment.

Now, then, is the perfect time to rally the UN community to the side of prevention, with much greater focus on high-risk groups; on the increasingly persuasive thesis of concurrent partners; on the roll-out of male circumcision (which should have been pursued vigorously much more quickly, but like so much else in the response to the pandemic got caught up in the overweaning proclivity for excessive scientific inquiry when the case is already clear, not to mention the faint hearts in the face of potential controversy); on harm reduction around which there has been inexcusable ambivalence; on the Prevention of Mother to Child Transmission, possibly the easiest preventive intervention of all, and therefore the most grievous emblem of multilateral negligence; on the absolute, resolute need to pursue, with unflinching tenacity the continued quest for a microbicide and a vaccine, regardless the setbacks. This is the time for a crescendo of United Nations voices.

The Epidemic Update is also shocking in the information that 76% of the overall deaths lie in Africa: one million, six hundred thousand adults and children. It tells you everything you didn’t want to know about the painfully slow rollout of treatment, and how vital it is that keeping people alive not get lost in some senseless, artificial tension between treatment and prevention. Nor should we ever forget where so many of the deaths come from, and how silent were the voices of the United Nations leadership while President Mbeki pursued his fatal denialism. South Africa has far and away the highest number of HIV infections in the world: five and a half million. There are eight hundred to a thousand deaths a day. No one has ever been held to account. Talk about impunity.

And as always, there’s the question of resources. It’s hard to know how the donors will react to the new numbers published by UNAIDS. They may simply yawn, and give whatever they intended. They may feel bamboozled, and want to cut back. They may see possibilities in the reduced numbers, possibilities of greater progress because fewer people are involved. Whatever the reaction, we must somehow persuade the world that we’re way behind, billions and billions of dollars behind, when it comes to funding all the components of the pandemic, from orphans to second-line drugs.

All of these elements, so many of them flowing generically from the Epidemic Update, are the stuff of the United Nations. UNAIDS, the Joint UN Programme on HIV/AIDS, oversees or coordinates ten co-sponsors: UNHCR, UNDP, UNICEF, WFP, UNFPA, UNODC, ILO, UNESCO, WHO and the World Bank. That’s a compendium of multilateral aristocracy and power rivaled in influence only by the G8 and the International Financial Institutions. And yet, members of the UN family, in the face of the AIDS pandemic, have sometimes acted with a kind of catatonic passivity. There is no excuse for it.

More than twenty-five years into the pandemic, we have an epidemic update that is --- let’s face it --- horrifying in its implications. Whether it’s 40 million or 33 million, this plague continues to ravage humankind. I simply do not believe that the United Nations has done everything it can possibly do to turn the tide. And I don’t mean just the member states, I mean the agencies and the secretariat.

I’m a multilateralist to the core of my being. But that doesn’t mean that the UN is above criticism. And it most emphatically doesn’t mean acting as an apologist for those who are chosen to lead and who have failed to lead."