Response to Pettifor et al. ‘Young people's sexual health in South Africa’: HIV prevalence and sexual behaviors from a nationally representative household survey

Abstract

The attempt to determine the relationship between an HIV/AIDS campaign and HIV prevalence by Pettifor et al. [1] is novel. However, the finding that participation in ‘at least one loveLife program’ produces an adjusted odds ratio of 0.6 in relation to HIV prevalence raises a number of questions. The concept of causal impact by a national HIV prevention intervention requires considerable explanation. In South Africa there are a wide range of national-level HIV/AIDS campaigns and programmes. Many, including loveLife, have an explicit youth and HIV prevention orientation in their activities. Over and above these national-level interventions, there are many other influences on the internalization of knowledge about HIV/AIDS and also the response to HIV prevention by young people. These include provincial and local HIV/AIDS campaigns and programmes; news items, documentaries, features and talk shows about HIV/AIDS; knowledge of HIV/AIDS statistics; engagement with local HIV/AIDS organizations; discussions as part of faith-based activity; interpersonal communication with family, friends, health service providers; knowing individuals who are living with HIV or who have died of AIDS, and so on [2]. This complex of sources and contexts of HIV/AIDS discourse are described in Table 1.Table 1: Sources and contexts related to HIV/AIDS for youth in South Africa.HIV risk and the ability to prevent HIV infection is also relative to a range of factors, including access to resources such as condoms as well as power differentials related to culture, sex and economic power [3]. There is a great deal of overlap within this milieu, including between HIV/AIDS programmes and campaigns as a product of their explicit focus on youth in South Africa. The authors were unable to take into account ‘participation’ in other campaigns (as this was not measured), whereas the wider complex of factors that might well influence HIV infection was not presented. What has occurred as a consequence is the representation of loveLife as a mutually exclusive intervention directly influencing HIV prevalence. This is extremely misleading. The authors recognized that the use of HIV prevalence as a determining variable is problematical, given that the timing of HIV infection cannot be inferred from prevalence data, and it can therefore not be reliably correlated with participation. Notwithstanding this concern, important variations in HIV prevalence in South Africa do not appear to have been taken into account in their analysis using this variable. HIV prevalence varies widely between the provinces. In the original study report [1], this ranged from 4.8% in Limpopo Province to 14.8% in KwaZulu-Natal Province, and from 8.7% in rural informal areas to 17.4% in urban informal areas. Urban HIV prevalence also varies between informal urban settlements (17.4%) and formal urban settlements (9.5%) [1]. These variations do not appear to have been controlled for. Overall, considerably more explanatory information would have been expected in the article in relation to the factors outlined above. The adjusted odds ratio of 0.6 for participation in loveLife programmes against HIV prevalence would have alerted the authors to the need to follow more careful and detailed analysis, given that what is suggested by the data is that participation in the loveLife programme is nothing short of a ‘magic bullet’ for HIV prevention. Findings in the initial study report conducted by the RHRU/loveLife consortium, of which the authors are part, should have underscored the need to be cautious. For example, respondents were asked to indicate, in an unprompted question, what they had done as a result of ‘what they saw or heard about loveLife’. The following was noted: ‘While being aware of loveLife and participating in its programmes is an essential element, it is hoped that through this interaction youth will change their behaviour or act in some positive way as a result. Among all youth, 24% reported that they had done something as a result of what they saw or heard about loveLife. Fifteen percent of all youth report having talked to someone about loveLife as a result of what they saw or heard. Fewer reported looking for more information on sex, sexuality and relationships (4%), looking for more information on loveLife (3%), or calling thethaJunction (1%). Sixty-one percent did nothing as a result, and 16% had not heard of loveLife.’ [4]. In effect, knowing about or participating in loveLife, does not appear to contribute strongly to HIV prevention behaviour. This would also suggest more caution in relation to the conclusion by the authors that their ‘present analysis is consistent with the hypothesis that the [loveLife] program is having an effect on HIV risk among young people.’ [1].