Seattle, Washington
March 4–7, 2019


Conference Dates and Location: 
February 13–16, 2017 | Seattle, Washington
Abstract Number: 



Katrina F. Ortblad1, Guy Harling2, Joshua A. Salomon3, Frank Tanser4, Deenan Pillay5, Tinofa Mutevedzi6, Till Baernighausen7

1Harvard T.H. Chan SPH, Boston, MA,2Harvard T. H. Chan SPH, Boston, MA,3Harvard SPH, Boston, MA,4Africa Centre for Pop Hlth, Mtubatuba, South Africa,5Africa Hlth Rsr Inst, Mtubatuba, South Africa,6Africa Hlth Rsr Insitute, Mtubatuba, South Africa,7Heidelberg Univ , Inst of Pub Hlth, Heidelberg, Germany

Abstract Body: 

Voluntary medical male circumcision (VMMC) has been proven in a number of randomized clinical trials (RCTs) to reduce HIV transmission by 60%. However, the benefits of circumcision might be negated by risk compensation, i.e., increases in risky sexual behaviors because of the biological HIV risk reduction following circumcision. To date, data on risk compensation in sub-Saharan Africa has been largely limited to RCTs. We test the risk compensation hypothesis for the first time using data from a population-based cohort study in sub-Saharan Africa.

A population-based cohort in KwaZulu-Natal, South Africa was followed longitudinally from 2003 to 2014. Self-reported circumcision status and sexual behavior was collected for all individuals annually, 2009-2014. Four variables were used to measure sexual behavior: (1) condom use at last sex, (2) regular condom used, (3) number of partners in the last 12 months, and (4) number of concurrent partners. Multivariable models with individual fixed effects were used to determine the impact of circumcision uptake on the self-reported sexual behavior variables.

From 2009 to 2014 14,997 unique men reported their circumcision status (median age 25 years, IQR: 19-41 years). During this time circumcision prevalence rose dramatically (2% in 2009 to 12% in 2014) and 954 individuals partook in circumcision interventions (as indicated by changes in their circumcision status over time). No significant changes in sexual behaviors were observed before and after circumcision uptake. The odds of condom use at last sex were 1.1 (95% CI: 0.4 – 3.0) for individuals post-circumcision compared to pre-circumcision and individuals post-circumcision had 0.9 times (95% CI: 0.7 – 1.2) the reported number of sexual partners in the past 12 months compared to number of partners reported pre-circumcision.

We find no evidence for risk compensation following circumcision in a community in rural KwaZulu-Natal. The often-hypothesized risk compensation phenomenon is unlikely to reduce the impact of VMMC campaigns on population HIV incidence in this and similar real-world settings. Circumcision should continue to be vigorously scale-up as a key HIV prevention strategy and newly circumcised males should continue to be counseled on the importance of condom use post-circumcision.

Session Number: 
Session Title: 
Presenting Author: 
Till Baernighausen
Presenter Institution: 
Heidelberg University, Institute of Public Health