Abstract:
Introduction:
HIV is one of the worst epidemics ever, with about 22.5 million people in sub-Saharan Africa living with the virus. The epidemic is extremely heterogenic and dynamic. HIV prevention depends upon a good understanding of population-specific transmission determinants. Therefore, this study of the magnitude and determinants of the recently observed decrease in HIV infection in Tanzania will help define the future direction of HIV interventions in the country.
Methods:
Data from the 2003/04 and 2007/08 AIDS indicator surveys from Mainland Tanzania were used in this analysis. In these surveys a multi-stage cluster sampling method utilizing two-stage sample design was employed to recruit participants age 15-49 from throughout the country. Socio-demographic, behavioral and biological determinants for HIV transmission were collected, and HIV testing was done to all consenting participants. In this analysis, data were restricted to Mainland Tanzania, people who reported to be sexually active, and people who had a valid HIV test result. A total sample of 3,944 men and 5,213 women from the 2003/04 survey and 3,708 men and 5,311 women from the 2007/08 survey were included in this analysis.
Results:
Apart from a significant increase between 2003/04 and 2007/08 in the proportion of people reporting having ever had an HIV test, other socio-demographic, behavioral and biological variables remained similar across the two surveys. Overall, HIV infection decreased by 14% (from 7.9% to 6.8%) over the four years between surveys, but this decrease was only significant among urban males. In both surveys among both women and men, formerly married individuals were at higher HIV risk, In the 2007/08 survey, however, single women were 60% less likely than married women to be infected. Risk of infection increased with wealth for both sexes and in both surveys. Risk was two to three times higher among rural residents compared with urban residents. In the 2007/08 survey educational attainment was associated with HIV infection in rural areas but not in urban areas. Having many lifetime sexual partners was a risk factor among both men and women, in both rural and urban areas, and in both surveys. Having a sexually transmitted infection (STI) was an important biological determinant for HIV infection in the 2007/08 survey, with twice the likelihood of infection for infected men and women compared with those uninfected. Male circumcision was associated with 60-80% lower odds of HIV infection.
Conclusion and recommendation:
HIV prevalence is decreasing among men in urban areas of Tanzania. Overall, men and women have similar determinants of HIV transmission, as do rural and urban populations. However, the study identified a number of transmission determinants that varied by sex and by urban-rural residence across the two surveys. Hence, interventions must address specific determinants in order to be effective. Similarly, interventions must address urban and rural populations separately.
HIV Prevalence and Associated Risk Factors: Analysis of Change over Time in Mainland Tanzania (PDF, 546K)