The 2016 UNAIDS Report – Get on the Fast Track: a Life-Cycle Approach to STI Prevention/STI/blogs – underlines the particular vulnerability to infection of women at a relatively early phase of the life-cycle, especially in limited resource settings such as sub-Saharan Africa as a result of structural factors .  These can seem intractable, but the authors of the report propose a number of practical measures for this group that include sexuality education in schools, social transfers and PrEP.  So the recent publication of a Cochrane Review of studies assessing the effectiveness of two of these interventions – sexuality education in schools and social transfers in the form of material incentives for girls to remain in school – is very timely – especially as five of the eight studies included in the analysis are based in the limited resource settings of sub-Saharan Africa that are the focus of the UNAIDS report.

A number of both European and non-European countries have incorporated some form of sexuality education into their school syllabuses, and there have been attempts to investigate their effectiveness in a number of places, including the UK (Stephenson & Johnson/STIs; Stephenson & Oakley; Henderson & Hart).  Unfortunately many of these trials rely on self-reported data, a tendency that has been shown to be problematic in this area (Langhaug & Cowan/STIs; Plummer & Hayes/STIs). The importance of the recent Cochrane Review is that it focuses largely on biological outcomes: namely, rates of STIs and pregnancy at follow-up.  The studies of sexuality education that are based on these outcomes have been undertaken in sub-Saharan Africa: Ross & Hayes (R&H,Tanzania); Cowan & Pascoe (C&P, Zimbabwe); Duflo & Kremer (D&K, Kenya). The same applies to studies of social transfers based on biological outcomes (Baird & Oezler (B&O) and Duflo & Kremer (D&K)).  (The non-African studies in the analysis use pregnancy prevalence as their preferred outcome (Cabezon & Garcia (C&G, Chile); Stephenson & Oakley (S&O, England); Henderson & Hart) (H&H, Scotland)).)

So what biological evidence do we find of the effectiveness of school-based education interventions? Practically none, say the reviewers.  No difference was reported between intervention and control groups for HIV or for other STIs – except in the case of R&H for syphilis prevalence at the end of follow up (RR 0.81: CI 0.47-1.39).  Even the statistically significant aggregate outcome for long-term pregnancy prevalence (0.55: CI 0.34-0.91) (C&G, S&O, H&H, D&K) seems largely accounted for by the results of C&G, which were at particularly high risk of bias.  When the latter were excluded, differences in pregnancy prevalence dropped to 0.93.

When it came to the other element of this survey, social transfers (B&O and K&K), only B&O reported a significant effect for HIV prevalence and HSV-2 prevalence.  This evidence was considered weak because, in the former case, it concerned data for school ‘drop-outs’ which, say the reviewers, the B&O study was not powered to detect, and, in the latter, because there was no measurement of prevalence at baseline.  As for pregnancy outcomes, both studies reported a reduction in short-term prevalence (0.76).

The trials of educational interventions may simply have been underpowered to detect small, but clinically important effects (especially with HIV).  On the other hand, the authors of the review also point to a growing consensus among experts that the determinants of sexual health outcomes and sexual risk-taking are wider structural factors such as poverty and cultural gender norms that lie beyond the capacity of school-based education programmes to influence.  The evidence for the effectiveness of incentives to stay at school, though as yet very inadequate, seems more encouraging.  This is evidently a field that requires further research.