Background: Self-report outcome measures of sexual behaviour are the mainstay of most HIV/STI prevention trials amongst adolescents. All reviews to date have focussed on sexual risk behaviour as the primary outcome. This is partly due to the large sample sizes required to demonstrate a reduction in incidence of HIV or STIs. However, self-report measures are prone to recall and social desirability bias (Plummer 2004; Langhaug 2010) and the relationship between such outcomes and HIV or STIs is not known. Biological measures, such as incidence of HIV or STIs are the most convincing indicators of the success of prevention interventions.

Objective: To summarise the effect of school-based HIV prevention interventions on the incidence and prevalence of HIV and STIs and on sexual risk behaviour among adolescents.

Methods: We conducted a comprehensive search for randomised controlled trials, published between 1985 and 2010, evaluating school-based HIV infection prevention interventions in which the unit of randomisation was individual students, classrooms, schools or communities, and in which the outcome measures included HIV and/or other STI incidence or prevalence.

Results: Three recent large randomised controlled trials of HIV prevention interventions among adolescents and youth in sub-Saharan Africa have included biological measures of STIs as their primary outcomes in addition to self-reported sexual risk behaviour (Cowan 2010; Ross 2007; Jewkes 2008). These trials demonstrate that it is possible to include biological outcomes in adolescent HIV prevention trials.

Conclusions: Future systematic reviews and individual intervention studies should include biological outcome measures. The development and validation of acceptable and valid biological measures should be prioritized and included, where feasible, in future studies and trials of interventions to reduce sexual health risk.