Does risk compensation behaviour neutralize the benefits of voluntary medical male circumcision?

The effectiveness and feasibility of voluntary medical male circumcision (VMMC) as a preventative intervention against HIV has been demonstrated in a variety of non-circumcising African communities.  The WHO has designated 14 countries in southern and eastern Africa as priority areas for VMMC scale-up.  Attempts to model the progress of the epidemic have long sought to factor in the potential contribution of VMMC (STIs/Hallett & Abu-Raddad).  However, the possibility of risk-compensation remains an ongoing concern (i.e. that the known preventative effects of VMMC will lead to increased sexual risk-taking).  Current evidence has been largely limited to behavioural evaluations and extended follow-up in populations where RCTs of VMMC were being conducted (e.g. Rakai, Uganda; Orange Farm, South Africa; Kisumu, Kenya).  The evidence has been reassuring, by and large.  Yet it is also inconclusive – for two reasons: first, on account of the rigorous risk reduction counselling invariably provided by these trials, which is far in excess of what would be offered in operational settings; second, due to the lack of certainty as to the preventative effectiveness of VMMC that prevailed at the time when these RCTs were being conducted.  These are precisely the issues that bedevil studies seeking to evaluate risk compensation in the context of pre-exposure prophylaxis (PrEP): see STIs/blog/Marcus & Grant; STIs/blogs/Mugwanya & Baeten.

A recent prospective, longitudinal study conducted in Nyanza Province, Kenya (Westercamp & Bailey) claims to mark an important step forward towards understanding the relation between VMMC and sexual risk-taking in everyday operational settings.  Participants in the study were not exposed to unrealistic levels of risk reduction counselling, and the preventative efficacy of VMMC had already been well established prior to commencement of the study.  Participants were followed up, at six monthly intervals over a two year period, having assigned themselves either to a circumcision (1,5888) or a control (1,599) group, using (as far as possible) audio computer-assisted self-interview questionnaires to investigate sexual behaviour.

The result?  No risk-compensation, apparently.  In fact, the findings show an increase of 30%/6% in condom use at last sex (for the circumcised/control group), and a broadly comparable decline for both groups across a range of indicators, including transactional sex in the last six months (26-12%), most recent sex with a casual partner (20-12%), and having multiple partners in the last six months (26-16%).  The only adverse finding was an increase in sexual activity for both groups among the younger participants – but this seems to be largely explained by transition to married status.

The decline might be due to some limited exposure to HIV education through participation, or reflect “cognitive dissonance” – i.e. the re-evaluation of behaviours in the light of the personal investment involved in getting circumcised.  But there is evidence it might be part and parcel of a decrease in sexual risk-taking in the community at large due to the implementation of the VMMC programme over the period of the study 2008-2011.  A curious finding was the greatly reduced perception of high HIV risk among the “cross-over group” of those 20% of the control group who subsequently chose to be circumcised, as against the perceptions of those who initially assigned themselves to the circumcision group.  This, according to the authors, suggests that men motivated to early adoption of VMMC may represent a higher risk group.  When this finding is taken together with recent evidence of high (90%) acceptance of VTC services among men undergoing circumcision, the case stacks up for ensuring the provision of high-quality counselling as a priority throughout the commencement and rapid initial sale up of VMMC services.