Simple calculations of direct impact for the initial assessment of the value of primary HIV prevention interventions.

IntroductionOver the course of the HIV pandemic prevention and treatment interventions have reduced HIV incidence but there is still scope for new prevention tools to further control HIV. Studies of the transmission dynamics and cost effectiveness of HIV prevention tools are often done using detailed complex models but there is a role for simpler earlier analyses. MethodsEquations are defined to calculate the cost effectiveness, budget impact, and epidemiological impact of HIV prevention interventions including equations allowing for multiple interventions and heterogeneity in risk across populations. An efficiency ratio of primary HIV prevention and IV treatment as prevention is defined. ResultsAs HIV incidence declines the number needed to treat to prevent one HIV infection increases. The cost effectiveness of HIV is driven by incidence, along with efficacy, duration, and costs of the intervention. The budget impact is driven by cost, size of the population and coverage achieved, and impact is determined by the effective coverage of interventions. Heterogeneity in risk could in theory allow for targeting primary HIV prevention but current screening tools do not appear to sufficiently differentiate risk in populations where they have been applied. DiscussionSimple calculations provide a tool to readily assess the cost-effectiveness, impact, and budget impact of HIV prevention interventions and can include heterogeneities in risk of HIV acquisition. These calculations provide rough initial estimates that can be compared with more sophisticated transmission dynamic and health economic models. ConclusionHIV incidence is declining making primary prevention tools less cost effective. If we require prevention to be more cost effective either we need to target primary prevention tools or they need to be less expensive. Simple equations allow for an exploration of the cost effectiveness of HIV interventions but the sensitivity of results to assumptions needs to be tested by comparison with transmission dynamic models.