AVERT - A USER-FRIENDLY MODEL TO ESTIMATE THE IMPACT OF HIV SEXUALLY-TRANSMITTED DISEASE PREVENTION INTERVENTIONS ON HIV TRANSMISSION

Objective: To introduce the newly developed AVERT model by describing the purpose, logic, advantages and limitations of the model, to valida te the model's estimates against seroconversion data from a large rand omized controlled trial, and to provide practical examples of its appl ications. Design: Static, deterministic spreadsheet-type model based o n per sex act HIV-1 transmission probabilities. Methods: Data from a r ecently completed trial carried out in Cameroon were used to validate the estimated number of new HIV infections generated by the AVERT mode l. A relatively limited set of biological and behavioral parameters wa s used to estimate the impact of a targeted HIV/sexually transmitted d isease (STD) prevention intervention in a South African mining communi ty. Results: The comparison of AVERT estimates with actual seroinciden ce data from the Cameroon trial not only confirmed the validity of the model's outputs but also illustrated its potential to provide additio nal options in data analysis. Modeling the pre-and post-intervention s cenarios for the South African mining community with AVERT provided es timates of the number of HIV infections averted due to targeted period ic presumptive STD treatment and community-based peer education. Concl usions: With a small number of accessible input variables, AVERT can p rovide plausible and defendable impact estimates of intervention effec ts on the reduction of HIV transmission. The AVERT model may be a help ful tool for decision-makers and planners in setting appropriate progr am priorities and analysing the cost-effectiveness of different interv ention packages. (C) 1998 Lippincott Williams & Wilkins.