Background: Large differences in the spread of HIV have been observed within sub-Saharan Africa.
Goal: The goal was to identify factors that could explain differences in the spread of HIV within sub-Saharan African populations.
Study design: Ecologic comparison of data from population-based surveys in high and relatively low HIV prevalence rural areas in Zimbabwe, Manicaland, and Tanzania, Kisesa.
Results: HIV prevalence in Manicaland and Kisesa was 15.4% and 5.3% in men aged 17-44 years and 21.1% and 8.0% in women aged 15-44 years (odds ratios, 3.3 and 3.1, respectively). Marriage is later, spatial mobility more common, cohabitation with marital partners less frequent, education levels are higher, and male circumcision is less common in Manicaland. However, adjustment for differences in these factors increased the odds ratios for HIV infection in Manicaland versus Kisesa to 6.9 and 4.8 for men and women, respectively. Sexually transmitted infection levels were similar, but syphilis was only common in Kisesa. Respondents in Kisesa started sex earlier and reported more sexual partners. Age differences between partners were similar in the 2 locations.
Conclusion: Substantial differences exist between the contemporary sociodemographic profiles of rural Manicaland and Kisesa. However, these differences did not translate into measurable differences in the biologic or behavioral factors for which data were available and did not explain the much higher HIV prevalence found in Manicaland. These findings might reflect more extensive AIDS-selective mortality and behavior change or greater bias in reporting of sexual behavior in Zimbabwe.