Dynamic models of the AIDS epidemic

This paper describes HIV transmission models built using STELLA, a new language for the Apple Macintosh. These include a basic model, or template, of sexual transmission of HIV; a model of heterosexual transmission constructed from the template; and a model which simulates transmission caused by needle sharing. The paper also describes a computer program, based on the template rate equations, which optimizes key epidemiological parameters associated with the early epidemic in San Francisco. With this program the authors have shown that the per contact transmission probability (infectivity) of the antigen-positive first stage of infection may be significantly higher than the infectivity of the antibody-positive subsequent stage. This finding stresses the importance of screening for viral antigens to detect first stage infection in high risk groups.