Preliminary Version The Impact of Social Interactions on the Spread of HIV Infection among Injection Drug Users: A Cellular Automaton Model

Injection drug users (IDU) who share needles are at high risk for contracting human immunodeficiency virus (HIV) infection. Social and behavioral influences that promote needle sharing can, therefore, impact HIV transmission. HIV spreads rapidly in IDU communities and interventions that target needle-sharing have had variable results. We constructed a cellular automaton model to study the dynamics of the HIV epidemic in an IDU community, in the presence of influences that promote or discourage sharing of used needles. Peer influences are tracked by a counter associated with each individual, who begin or stop sharing needles once a threshold level of influences from neighbours is reached. The simulated epidemic exhibited a strong nonlinear response to social influence on needle-sharing behaviour. An epidemic phase diagram for the parameter space of social influences revealed two states for HIV prevalence. The endemic state above the phase transition curve is characterised by stable HIV prevalence of approximately 35%. Parameter values below the phase transition curve lead to the extinct state. This is similar to a herd immunity effect, as the epidemic in this region of the parameter space is eventually driven to extinction. The behaviour of the system implies that public health interventions aimed at reducing needle sharing may have little effect if coverage is limited. If coverage exceeds the phase transition threshold, interventions are expected to be highly effective in stemming HIV epidemics in IDU communities. We used the computer algebra system Maple for all simulations in this project.