AiS Challenge Team Abstract Team Number: 088 School Name: Desert Academy Area of Science: Epidemiology Project Title: The Law of Unintened Consequence: Treating AIDS

Definition of the problem:

The treatment of AIDS is a humanitarian concern of the highest order. A number of methods of diagnosis and treatments are now available as a result of this humanitarian concern. However no method of diagnosis is 100 percent accurate in ruling in or ruling out if a person indeed has AIDS. Similarly no treatment is capable of a 100 percent “cure”. In addition, AIDS is a medical problem of a relatively small percentage of the population even though it significant in terms of the growing and absolute number of people infected with it. These three factors: (1) diagnosis, (2) treatment and (3) level of population infection present a classic situation in which Bayesian probability can be used to assess the costs and benefits of treating or not treating AIDS. Purpose of the project and Anticipated Results The core problem to be explored is how to apply Baysian probability to a complicated situation in which there are several models which attempt to explain how AIDS infects a population initially and then spreads. The purpose is to develop a more comprehensive model in which both classic statistical probability modeling is combined with complexity modeling. The dynamic changes that can be visualized with StarLogo are to be combined with the static probabilities created by Bayesian statistics. One anticipated result is that the diagnosis and treatment of AIDS is actually worse for the population as a whole. Benign neglect may result in an overall healthier population as well as AIDS patients themselves. This result is based in part on what happened when patients with syphilis were not treated in the 1930/40s. Plan of action Several simulation models will be developed. The first part of a model will involve creating a Baysian distribution, building upon how AIDS emerges, is diagnosed and treated in a particular population. The second part is integrating this statistical model into a dynamic model using StarLogo. The final part of the simulations is to compare each of the models to one another.