The Ugandan Ministry of Health and Columbia University are collaborating in a longitudinal population-based HIV cohort study in rural Rakai District, southwestern Uganda. Preliminary behavioral and HIV-1 serological data from the cohort suggest that geographic location, mobility and related variables influence sexual behavior and the structure of sexual networks in the study population, contributing in turn to the rapid spread of HIV infection. In 1990, adult HIV-1 prevalence in the cohort was 20.1%; extrapolated district level HIV prevalence was 13.9% (95%CI = 9.7% - 17.7%, design effect 3.5). Between 1989-1990, one year incidence among all cohort adults was 2.6%; in men 20 to 24, it was 5.9%, and in women in the same age group, 6.4%. Substantial gradations in HIV prevalence and in reported behavioral variables (mobility, travel to urban areas, numbers of partners) were noted between Rakai's three geographic strata: main road trading centers, small trading villages on secondary roads and rural agricultural villages. We propose to collect detailed qualitative and quantitative data in Rakai District to define sexual networks and behaviors critical to the transmission of HIV infection, and to elucidate their determinants, including place of residence, geographic mobility and social/ethnic status. A specific objective will be to determine whether there are potential behavioral barriers to HIV spread from high to low prevalence areas, and whether certain communities act as bridging populations facilitating HIV transmission into lower prevalence villages. Additional data will be collected in neighborhoods known to be hubs of travel and migration in the regional capital of Masaka and the national capital of Kampala. Movement of persons between all study sites, and its effects on sexual behavior, will be examined. In Rakai, subjects for the behavioral research will be drawn in part from the cohort study, which is collecting serological health status and knowledge, attitudes and behavior data on a random sample of the population. For those subjects from Rakai who are not enrolled in the existing cohort, as well as for respondents in Masaka and Kampala, we will collect filter paper serological samples for ELISA and Western blot testing, in order to match serostatus with reported behaviors. The behavioral study will provide information critical to the development of more effective preventive strategies for Rakai and Uganda as a whole. Improved prevention programs are needed since HIV transmission remains elevated despite high levels of knowledge regarding AMS, the existence of educational programs, and condom distribution. The data will also permit more exact modelling of the HIV epidemic.
