A recent article has highlighted how Aids statistics can be manipulated by dissidents to prove their point of view Belinda Beresford There are damned lies. There are statistics. And then there is noseweek. That venerable organ of expos’s and investigative journalism has itself been used in a malicious distortion of facts to advance the arguments of Aids dissidents. The magazine has used inaccurate information and statistical analysis in an attempt to disprove one of the major arguments that there is an Aids epidemic in South Africa – that large numbers of people, in particular young men and women, are dying. The editorial in noseweek (issue No. 30) ridicules research on mortality done by the Medical Research Council (MRC), and presented to the presidential Aids advisory panel by MRC head Professor Malegapuru (William) Makgoba in July. The study found profound changes in the patterns of death in South Africa, with mortality among the younger age groups considerably higher in the late 1990s compared to the early part of the decade. The information was obtained by the Sunday Times, which ran a front-page story highlighting aspects on the day the international Aids 2000 conference opened in Durban. However, according to noseweek, the Sunday Times had “been used to publish a grotesque piece of propaganda that was calculated to mislead the public and embarrass the president”.
The MRC had also created “panic through the selective and incorrect use of statistics. Makgoba’s timing suggests he also hoped to embarrass and discredit President Thabo Mbeki.”
The magazine bases its allegations on an analysis of the Sunday Times’s report which was done by Statistics SA at the request of the Department of Home Affairs shortly after the article appeared. Unfortunately for noseweek, as Statistics SA readily confirmed in response to the Mail & Guardian’s questions, that initial assessment was flawed.
One criticism in that report, and a fundamental part of noseweek’s attack, was that the MRC figures for 1990 excluded the “bantu homelands”. In noseweek’s words this meant the figures described the “largely white, middle class, urban – and older – population”.
Introducing the black population of the “homelands” would lead to a much higher number of deaths, the magazine says. In addition the pattern of deaths would change, with more people dying at a younger age, reflecting the fertility and life patterns of black South Africans. What noseweek failed to point out, despite its being in the original Statistics SA document, is that it was only the populations of the “independent homelands” of Bophuthatswana, Transkei, Ciskei and Venda (TBVC states) that were excluded from the 1990 figures. These represented less than 20% of the total population of South Africa. The other six homelands were included – including the biggest, KwaZulu-Natal, which is most grievously affected by HIV/Aids, as well as the country’s black townships. And while the inclusion of the TBVC states would undoubtedly change both the number and pattern of deaths to an extent, it alone would not cause the radical changes in mortality over the past few years. But, more importantly, the MRC was well aware of this deficiency in the data collected by home affairs, and made adjustments for it. This is confirmed by a letter from Statistics SA to the M&G, in which acting head Ros Hirschowitz said that deaths in these former TBVC states “were indeed excluded from the 1990 database, but the MRC included them in their estimates, but it did so from other sources. We were not aware of this at the time when we wrote the letter to home affairs.”
Hirschowitz says that although the effect of including people from former TBVC states would lead to “younger distribution” of deaths, since blacks tend to die earlier than whites, “this does not necessarily mean that the impact of Aids is not present”. She also confirmed that the home affairs database included deaths among all population groups, including people living in the townships such as Soweto and Alexandra. There is some accuracy in noseweek’s allegations. For example, Hirschowitz says that in 1990 deaths among poor and rural people may have been excluded from official statistics. However, she points out that the same could apply in 1999, since “deaths in rural areas tend to be under-reported in South Africa”. In any case, given the inclusion of the former TBVC states, it is not the number of deaths in South Africa which is so telling of an epidemic, but the changing pattern of mortality. The graphs (below) come from a presentation by David Bourne, a researcher at the University of Cape Town, and his colleagues, based on data from death registration. The 1990 graph demonstrates what are considered normal distribution curves for deaths in South Africa, with a peak for very young children, followed by a convex curve as deaths decline against increasing age. The number of deaths then peaks again among the old, representing deaths related to ageing. However, the later graph shows the pattern changing, with a rapid increase in the proportion of deaths among young adults, particularly women. Part of the increase can be ascribed to better reporting of deaths but the researchers indicate that the shift in age structure can only be explained in terms of additional deaths due to Aids. Noseweek triumphantly quotes the original Statistics SA report as saying that the 1999 pattern of deaths presented by Makgoba “is, in fact, ‘not a new profile'”. Contradicting the authors of the noseweek editorial, Hirschowitz told the M&G: “Having later seen the more detailed presentation of the MRC, Statistics SA agrees that the mortality profile, especially that of females, has changed over the past decade.” Hirschowitz says: “While it is hard to separate out deaths due to HIV/Aids from those due to other causes, the mortality profiles portrayed suggest that the deaths in the younger ages could be possibly explained in part by the impact of the Aids epidemic.” Even more recent research from the MRC finds that there has been an increase in child and young adult morality since the 1990s. “This is particularly marked in recent years for women aged 20 to 45,” according to an MRC report.
Statistics SA confirms this in its letter, saying: “In summary, the HIV/Aids epidemic is probably having an impact on the age profiles of the South African population, particularly women. But a great deal of additional analysis and research is needed before the impact can be fully understood. This research is ongoing.” The MRC will shortly be releasing a detailed report addressing the complexities of interpreting the changing death patterns.
Noseweek, after calling for Makgoba’s dismissal from his post, announced that his presentation and the research behind it “have simply been absorbed into the myth that passes as ‘Aids science’ in South Africa today”.
But it is the noseweek article, and the erroneous arguments on which it was based, that have become the real myth. The article is being circulated on the Internet by Aids dissidents as an example of the “debunking” of HIV/Aids orthodox myths. After all, if people are not dying, then there is no epidemic, and no Aids. The managing editor of noseweek and co-author of the editorial, Marten du Plessis, is an office bearer of the dissident grouping Forum for Debating Aids in South Africa, with former journalist Anita Allen and controversial author Anthony Brink. Despite its name, and the fact that it is a member of the International Coalition to End Aids Censorship, Du Plessis and his associates have attempted to muzzle the M&G for publishing orthodox Aids views. The incorrect arguments behind the editorial were recently repeated by well-known free marketeer Leon Louw at a “trial” of HIV held at Rand Afrikaans University at the end of August. According to the organisers’ press statement, “Louw told a stunned audience that MRC president William Makgoba largely fabricated the much-publicised statistics he presented to the presidential Aids panel.” This argument was a significant factor in converting members of the audience to the dissident point of view. And so it seems that bad reporting risks reaching the status of a received truth, fuelled by the apparently desperate dissident desire for ammunition against overwhelming scientific – and anecdotal – evidence that thousands of South Africans are sick and dying as a result of HIV/Aids.