Political-economic transitions in the Soviet Union, Indonesia, and China, but not the Philippines, were followed by HIV epidemics among drug users. Wars also may sometimes increase HIV risk. Based on similarities in some of the causal pathways through which wars and transitions can affect HIV risk, we use the term "Big Events" to include both. We first critique several prior epidemiological models of Big Events as inadequately incorporating social agency and as somewhat imprecise and over-generalizing in their sociology. We then suggest a model using the following concepts: first, event-specific HIV transmission probabilities are functions of(a) the probability that partners are infection-discordant; (b) the infection-susceptibility of the uninfected partner; (c) the infectivity of the infected-as well as (d) the behaviours engaged in. These probabilities depend on the distributions of HIV and other variables in populations. Sexual or injection events incorporate risk behaviours and are embedded in sexual and injection partnership patterns and community networks, which in turn are shaped by the content of normative regulation in communities. Wars and transitions can change socio-economic variables that can sometimes precipitate increases in the numbers of people who engage in high-risk drug and sexual networks and behaviours and in the riskiness of what they do. These variables that Big Events affect may include population displacement; economic difficulties and policies; police corruption, repressiveness, and failure to preserve order; health services; migration: social movements; gender roles; and inter-communal violence-which, in turn, affect normative regulation, Youth alienation, networks and behaviours. As part of these pathways, autonomous action by neighbourhood residents, teenagers, drug users and sex workers to maintain their economic welfare, health or happiness may affect many of these variables or otherwise mediate whether HIV epidemics follow transitions. We thus posit that research on whether and how these interacting causal pathways and autonomous actions are followed by drug-related harm and/or HIV or other epidemics can help LIS understand how to intervene to prevent or mitigate such harms. (C) 2008 Elsevier B.V. All rights reserved.
