Strategies for the Prevention of HIV Infection and the Enhancement of Reproductive Health Among Adolescents in Rural Tanzania: MEMA Kwa Vijana Trial

Background: Adolescents are at high risk of adverse reproductive health outcomes, including HIV, other STIs and unwanted pregnancies. However, there is little empirical evidence to guide the choice and implementation of effective interventions. We assessed the impact of an intervention programme on the sexual health of adolescents in rural Tanzania.

Methods: Using data from a prior population-based survey of 9,445 15-19 year olds, 20 communities were stratified and randomly allocated to either receive the new interventions (Intervention Group) or standard interventions (Comparison Group) during Phase 1 (January 1999-December 2001) of the MEMA kwa Vijana Project. The new intervention programme had four major components: community activities; teacher-led, peer-assisted sex education in the last three years of primary school (Years 5-7); training and supervision of health workers to provide "youth-friendly" STD and family planning services; and peer condom social marketing for youth (from January 2000). The pre-defined primary outcomes were HIV incidence and Herpes simplex virus type 2 (HSV2) in a cohort of 9,645 adolescents, mean age 15.5 years, [95% range 14.1-18.3 years], who were recruited in late 1998 before entering Year 5, 6 or 7 of primary school. Secondary outcomes included six further biomedical, five behavioural, one attitudinal, and three knowledge outcomes.

Findings: At the follow-up survey in late 2001-early 2002, the intervention had had a statistically significant impact on all knowledge and attitudinal outcomes and also on reported condom use and reported STI symptoms in both males and females. Significantly fewer males in the intervention communities reported sexual debut during follow-up, or having multiple sexual partners during the past 12 months, but no difference was seen for these two outcomes among females. There were only five HIV seroconversions in males. Among females, the adjusted rate ratio for HIV incidence (intervention vs comparison communities) was 0.76 (95%CI: 0.35,1.65). Overall prevalences of HSV2 were 11.9% in males and 21.1% in females, with adjusted prevalence ratios (PRs) of 0.92 (95%CI:0.69,1.22) and 1.05 (95%CI:0.83,1.32) respectively. There was no consistent impact on the other biological outcomes, with adjusted PRs varying from 0.78 (95%CI:0.46,1.30) for syphilis in males to 1.94 (95%CI:1.01,3.72) for gonorrhoea in females. A non-significant trend towards greater beneficial impact among students enrolled in Year 4, who potentially received all three years of the in-school programme, was seen for most outcomes. The beneficial impact on knowledge and reported attitudes was confirmed in a cross-sectional survey in a different group of students who were in Year 7 of primary school in mid-2002.

Conclusions: The intervention substantially improved knowledge, reported attitudes and reported condom use in both sexes, and reported sexual behaviour in males, but had no consistent impact on biological outcomes within the three-year trial period. The data suggest a dose-related effect for several outcomes, with greater impact among those receiving two or three years of the in-school programme.