I attended an interesting satellite session on Monday on the decline in HIV prevalence in Zimbabwe. The speakers attempted to address whether the decline in prevalence was due to natural dynamics (e.g. mortality or migration) and if not, was there evidence available to suggest that changes in risk behavior affected the epidemic. Unfortunately, the available data would not allow the speakers to demonstrate a causal relationship, so they:
- used mathematical modeling to demonstrate that mortality and migration would not have been sufficient to cause the magnitude and timing of the decline;
- stated that if behavior change impacted HIV prevalence starting in 2004 then behaviors would have had to change at some point during 1999-2003;
- analyzed quantitative behavior to identify behavioral changes during 1999-2003 and used qualitative research tools to assess what was happening programmatically in the early 1990s compared to 1999 compared to today.
They found that a reduction in numbers of partners, changes in social norms and the economic situation, which drove some people into sex work, but for most men made sex work too expensive may have contributed to the decline. An increase in condom use with non-regular partners was also observed during the key time period. Furthermore, although no single program stood out, a combination of awareness programs, more widespread condom distribution, and the increased availability of HIV services may have changed people's attitudes about HIV. In addition, personal experience with mortaility and a fear of HIV also may have played a role. Again, though they can't attribute causality, the evidence indicates that the decline in HIV prevalence in Zimbabwe was due to a reduction in risky sexual behavior.