Abstract

Predictions of future HIV and AIDS incidence are derived from mathematical models based on the intrinsic and extrinsic epidemiological factors of HIV. The two most important intrinsic factors are HIV infectiousness, and HIV incubation period. The most important extrinsic factor is risky behaviors leading to HIV infection. The major risk groups include homosexual/bisexual men, intravenous drug users, promiscuous heterosexuals and recipients of blood products.

The three different types of models used to make HIV and AIDS incidence predictions are direct extrapolation models, backcalculation models, and dynamic transmission models. Direct extrapolation models use a function to describe the future trend of AIDS, back-calculation models use estimates of incubation periods to predict future HIV and AIDS incidence, and dynamic transmission models use systems of differential equations that describe the overall process of the epidemic.

The model constructed in this thesis is a dynamic transmission model. Its efficiency is based on the accuracy of short-term and long-term predictions. The short-term predictions used data from 1981-1986 to determine if the model can make accurate predictions for HIV and AIDS cases through the year 1992. The model proved to be inaccurate for short-term predictions; the probable reason being that not enough parameters were accounted for in the model. The results of long-term predictions were made through the year 2050. The results predicted that HIV and AlDS would cause up to a 25% reduction in the population. These results demonstrate that the dynamic transmission likely exaggerates the effects of HIV on the entire population possibly because HIV is only transmissible through risky behaviors involving the exchange of blood and bodily fluids.