Objectives: To determine baseline HIV prevalence in a trial of improved sexually transmitted disease (STD) treatment, and to investigate risk factors for HIV. To assess comparability of intervention and comparison communities with respect to HIV/STD prevalence and risk factors. To assess adequacy of sample size.
Setting: Twelve communities in Mwanza Region, Tanzania: one matched pair of roadside communities, four pairs of rural communities, and one pair of island communities. One community from each pair was randomly allocated to receive the STD intervention following the baseline survey.
Methods: Approximately 1000 adults aged 15-54 years were randomly sampled from each community. Subjects were interviewed, and HIV and syphilis serology performed. Men with a positive leucocyte esterase dipstick test on urine, or reporting a current STD, were tested for urethral infections.
Results: A total of 12,534 adults were enrolled. Baseline HIV prevalences were 7.7% (roadside), 3.8% (rural) and 1.8% (islands). Associations were observed with marital status, injections, education, travel, history of STD and syphilis serology. Prevalence was higher in circumcised men, but not significantly after adjusting for confounders. Intervention and comparison communities were similar in the prevalence of HIV (3.8 versus 4.4%), active syphilis (8.7 versus 8.2%), and most recorded risk factors. Within-pair variability in HIV prevalence was close to the value assumed for sample size calculations.
Conclusions: The trial cohort was successfully established. Comparability of intervention and comparison communities at baseline was confirmed for most factors. Matching appears to have achieved a trial of adequate sample size. The apparent lack of a protective effect of male circumcision contrasts with other studies in Africa.
PIP: To measure the impact of a sexually transmitted disease (STD) treatment program on the incidence of human immunodeficiency virus (HIV) in Zimbabwe's Mwanza Region, a pre-intervention baseline survey was conducted. Included in the survey were approximately 1000 randomly selected adults from each of the six intervention communities (defined as the population served by a health center and its satellite dispensaries) and six matched comparison communities. Overall HIV seroprevalence was 4.1% (3.7% in men and 4.4% in women), with a range of 1.6-8.6% and no significant differences between intervention and control communities. Peak prevalences for both sexes were found in the 25-34 year age groups and in roadside communities. The following factors were associated with an increased likelihood of HIV infection: separation, divorce, or widowhood; multiple injections in the preceding year; educational achievement of at least Standard 4; travel out of the district in the prior year; history of genital ulcers or discharge; and past or present infection with syphilis. HIV prevalence was significantly higher in circumcised men, but not when adjustment was made for other risk factors. Syphilis prevalence ranged from a low of 4.2% in island communities to a high of 11.1% in roadside communities. The baseline survey indicates that intervention and control populations are generally comparable, and that the goal of locating a study area with a relatively low incidence of HIV and high rates of other STDs has been achieved.