The Effect of Partner Reduction on HIV Spread for Different Patterns of Sexual Behavior Evaluated With STDSIM

Background and Rationale: For the developing world, the prevention of HIV transmission through heterosexual contacts remains the primary mode of HIV control. Yet, little systematic research has been performed on the effects of different forms of sexual risk reduction. Objectives: To evaluate the effect on HIV spread of different forms of sexual risk reduction using the micro-simulation model STDSIM. Methods: STDSIM simulates life histories of hypothetical individuals including sexual contacts and acquisition and effects of HIV/STDs. Three risk profiles reflecting developing regions in Southeast Asia, urban Africa and South America, rural Africa, and one generic low risk profile, were used: (1) prostitution, no short relationships; (2) prostitution, concurrent short relationships; (3) no prostitution, concurrent short relationships; and (4) some prostitution, serial short relationships. Risk was reduced by a 2-year delay in age of sexual debut, a 50% reduction in relationship formation rates, or a 50% reduction in prostitution. The effectiveness of these measures was assessed by comparing HIV prevalence in the adult population. Results: In profiles 1 and 2, prostitution contacts caused a rapid initial spread. Concurrency caused a high final prevalence in profiles 2 and 3. In profile 1, only the reduction in prostitution had an effect. In profile 2, a reduction in partnership formation rate was most effective. In profile 3, HIV almost disappeared in 15 years because of the decrease in concurrency resulting from a reduction from 1.7 to 1.3 partners per year. In profile 4, all interventions were about equally effective. Conclusions: Projections of interventions showed that a decrease in average concurrency below a critical level can lead to a large reduction in HIV prevalence. Changes in size of the largest network component were concomitant with changes in concurrency.