McCreesh, Nicky MSc; White, Richard G. PhD
Department of Infectious Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK School of Medicine, Pharmacy and Health Durham University, Durham, UK
Department of Infectious Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine London, UK
We very much welcome the comments from Eaton and colleagues1 on our work. We agree with them entirely when they say that the results of our study2 should not be used “…to draw conclusions about the impact of a successful ‘network-based’ behavior change communication intervention.”1 These interventions target, and hopefully affect, a range of risk behaviors. Instead, as we state in the article, our results should be interpreted as providing an indication of the contribution that concurrency reduction (alone) may have in these campaigns.
Eaton and colleagues refer to the Ugandan “Zero Grazing” campaign3 that targeted both multiple partnerships and concurrency in the late 1980s and early 1990s and coincided with reductions in HIV transmission. We do not argue that this campaign and others like it were unsuccessful in reducing the incidence of HIV. Nor do we argue that similar campaigns that are currently being planned or implemented will not reduce the incidence of HIV if they successfully reduce the prevalence of a range of risk behaviors. However, our results do suggest that in contemporary rural Uganda and, perhaps, in similar populations, the contribution of reductions in the prevalence of concurrency may be modest.
In our modeling study, to isolate the effect of concurrency, we kept the overall number of sexual partnerships constant as we reduced the prevalence of concurrency. Eaton and colleagues state, “However, it is highly unlikely to reflect the impact of a successful real world intervention because there is no reason to believe that such compensatory behavior would occur in real life. Why would people with no partners take on new partnerships in response to such a campaign?” We would argue from the opposite perspective that it is implausible that all women (and men) dissuaded from becoming a concurrent partner by a successful intervention would remain single. It is perhaps more reasonable to think that they would attempt to form a partnership with someone else. This may particularly be the case in populations in which women rely on sexual partners for financial support.4 If so, and the concurrency intervention had the intended effect of strengthening social norms against forming partnerships with individuals with current partner(s), then individuals without current partner(s) would become relatively more desirable, thus increasing the probability that people with no partners would indeed take on new partnerships.
Overall, though, as we said in our discussion, if a concurrency intervention did also lead to a reduction in total number of partners, then this may have led us to underestimate impact. However, as we also stated in our article, there are reasons to think that we may have overestimated the impact of a concurrency intervention. We did not simulate higher condom use between concurrent partners or coital dilution,5 both of which may reduce the impact of an intervention against concurrency. The use of more complex models and data collection could help inform this debate further. So far though, we do not believe that our conclusions are too far off the mark.
It goes without saying that we also wholeheartedly support the continued funding of rigorous evaluations of behavior change campaigns using biological end points. This is essential for identifying the most efficient prevention programs. However, as was clear when we set out to do this work, it will not be possible in current evaluations to separate out the effects on HIV transmission of any reductions in the number of concurrent partnerships in the population from other aspects of risk behavior. Regrettably, this is probably truer now than it was then, given the funding cuts that Eaton and colleagues highlight. As such, our modeling study2 provides valuable information on the potential role of concurrency reduction in any observed impact on HIV, which will not be obtainable from these empirical studies of real-life campaigns.
Nicky McCreesh, MSc
Department of Infectious Disease
Epidemiology, Faculty of Epidemiology
and Population Health, London School of
Hygiene and Tropical Medicine, London, UK
School of Medicine, Pharmacy and Health
Durham University, Durham, UK
Richard G. White, PhD
Department of Infectious Disease
Epidemiology, Faculty of Epidemiology and
Population Health, London School
of Hygiene and Tropical Medicine
1. Eaton JW, Hallett TB, Epstein H. What might be the impact of sexual partnership “concurrency” behavior change communication campaigns? Sex Transm Dis 2012; 39: 899.
2. McCreesh N, O’Brien K, Nsubuga RN, et al. Exploring the potential impact of a reduction in partnership concurrency on hiv incidence in rural uganda: A modeling study. Sex Trans Dis 2012; 39:407–413. 410.1097/OLQ.1090b1013e318254c318284a.
3. Kirby D. Changes in sexual behaviour leading to the decline in the prevalence of HIV in Uganda: Confirmation from multiple sources of evidence. Sex Transm Infect 2008; 84 (suppl 2): ii35–ii41.
4. Luke N. Age and economic asymmetries in the sexual relationships of adolescent girls in Sub-Saharan Africa. Stud Family Plann 2003; 34: 67–86.
5. Sawers L, Isaac A, Stillwaggon E. HIV and concurrent sexual partnerships: modelling the role of coital dilution. J Int AIDS Soc 2011; 14: 1–9.