Modeling HIV risk

Objective: To provide HIV risk estimates for specific local population subg roups using an HIV risk index combining HIV risk behaviors, prevalence rate s, and transmission probabilities. Methods: A sample of 270 individuals was studied. Respondents described sex ual and injection risk behaviors and partners in the previous 30 days. An H IV risk index was computed for each individual, combining reported sexual a nd injecting risk behaviors. HIV prevalence estimates for partners and HIV transmission probabilities for each of the risk behaviors. Partner HIV prev alence rates were estimated from a national sampler and HIV transmission pr obabilities were obtained from previously published studies. Results: Projected risk estimates were found to differ a great deal within major demographic categories. Highest 10-year risk was found among African- American male gay injectors (72%) and the lowest among white male heterosex ual noninjectors (<.01%). Conclusions: The method used here for estimating HIV risk can be calculated for specific at-risk population subgroups of any size and composition, inc luding individuals. By understanding which specific subpopulations are at r isk, specific interventions and public health campaigns may be better targe ted.