Dynamics of spread of HIV-I infection in a rural district of Uganda.

OBJECTIVE--To define the geographical distribution of HIV infection and the community characteristics associated with HIV prevalence in a rural population of Uganda. DESIGN--Seroprevalence survey and interviews of the population aged 13 years and older in 21 randomly selected clusters. SETTING--Rural population of Rakai district, south west Uganda. SUBJECTS--1292 adults, of whom 594 men and 698 women gave a blood sample and answered the questionnaire. MAIN OUTCOME MEASURES--HIV status determined by ELISA and western blotting in relation to community characteristics. RESULTS--The weighted seroprevalence of HIV for the district was 12.6% with prevalence by cluster varying from 1.2% to 52.8%. Seroprevalence was highest in main road trading centres (men 26%, women 47%), intermediate in rural trading villages on secondary roads (men 22%, women 29%), and lowest in rural agricultural villages (men 8%, women 9%). For both men and women, multiple regression showed a strong negative association between cluster seroprevalence and the proportion of the population employed in agriculture (beta = -0.677 for men, -0.807 for women). Among women, cluster seroprevalence increased with a higher proportion of the population reporting multiple sex partners (beta = 0.814), external travel (beta = 0.579), and injections (beta = 0.483). CONCLUSIONS--Community characteristics, particularly the proportion of the population in agriculture, are associated with HIV prevalence and can be used for targeting interventions. The seroprevalences of HIV suggest spread of infection from main road trading centres, through intermediate trading villages, to rural agricultural villages.