Male circumcision, sexual risk behaviour and HIV infection in Uganda. A mixed methods study among men age 15-59 years

Introduction: HIV/AIDS is a significant public health threat, especially in the sub Saharan African region. Enormous global efforts to control and prevent new infections are still needed on top of managing the huge number of over 36 million people living with the virus. Transmission through heterosexual intercourse remains the main contributor to the HIV epidemic in sub Saharan Africa. The WHO in 2007 recommended adoption of safe male circumcision (SMC) as part of the comprehensive HIV prevention interventions. However, like other risk reduction interventions, SMC is prone to challenges when implemented at such large-scale population levels. Behavioural risk compensation and, sociocultural beliefs and misconceptions in the post-circumcision period may affect successful implementation. Therefore, it is prudent to explore the existence of risk compensation and the beliefs that may shape sexual behaviour of men both before and after circumcision, so as to contribute to the success of the SMC programme scale-up. General objective: To estimate the associations of male circumcision with sexual risk behaviour and HIV prevalence before and just after the national scale up of the safe male circumcision (SMC) programme in Uganda, and to understand the sexual behaviour choices among men circumcised under the SMC programme in Wakiso district, Uganda. Methods: A mixed method study was conducted in Uganda that included an analysis of the Uganda AIDS Indicator surveys of 2004 and 2011 (papers I and II), and a qualitative sub study in 2015 and 2016. The two surveys had nationally representative samples and employed a two-stage stratified cluster sampling design. The analysis in paper I is based on data from 14,875 (6,906 in 2004 and 7,969 in 2011) circumcised and uncircumcised men, while paper II includes 5,776 uncircumcised men from the 2011 survey alone. Both samples are of men who reported to ever have had sex and were aged 15-59 years. In paper I, the dependent variables were HIV sero-status obtained from blood sample tests in both surveys, and sexual risk behaviours, while the main independent variable was self-reported circumcision status. In paper II, willingness to be circumcised was the dependent variable while the sexual risk behaviours were the independent variables. In the qualitative sub study, SMC clients were enrolled and followed up after receiving services at health facilities in Wakiso district located in central Uganda (papers III and IV). In 2015 twenty-five participants were purposively selected from health facilities where they reported for voluntarily receive male circumcision. They were interviewed twice, just after circumcision and six months later. Baseline indepth interview topics included discussions of motives for circumcision, influences on the decision, sexual behaviours, experiences with health education received and any known beliefs. Follow-up interview topics included experiences with healing, beliefs and post circumcision sexual behaviours. Results: Circumcised men reported higher prevalence of all sexual risk behaviours than uncircumcised men, except for transactional sex. Use of condoms with the last non-marital sexual partner among circumcised men was lower in the 2011 survey (PR 0.88; 95% CI: 0.79-0.99) compared with the 2004 survey (PR 1.07; 95% CI: 0.98- 1.18), but there were no other statistically significant changes in sexual risk behaviours between the two surveys. Circumcised men were less likely to test HIV positive than the uncircumcised in both surveys, (PR 0.63; 95% CI: 0.48-0.82) in 2004 and (PR 0.62; 95% CI: 0.49-0.80) in 2011 (paper I). Willingness to be circumcised was higher in uncircumcised men reporting multiple partners (PR 1.19; 95% CI: 1.11-1.29), non-marital sex with (PR 1.71; 95% CI: 1.59-1.85) and without a condom (PR 1.47; 95% CI: 1.35-1.59), or transactional sex (PR 1.61; 95% CI: 1.39- 1.87) in the 2011 survey (paper II) compared to those who did not report these risk behaviours. Findings from the qualitative study showed that the important factors influencing men to opt for circumcision were female sexual partners and a perceived need to reduce HIV transmission risk. According to participants’ reports, emphasis was mainly put on the immediate problems of wound care and the surgical procedure during health education for circumcision clients at health facilities, and less on post circumcision sexual behaviour. All the men, however, were aware that circumcision only offers partial risk reduction for HIV infection (paper III). In the baseline interviews, the men talked about beliefs that could influence their sexual behaviour, while in the follow up interviews they discussed how these had indeed shaped their behaviour after circumcision. All men had heard that it was important that the initial sexual intercourse post circumcision was with someone else other than the stable partner. Some of the men strongly believed in this and consequently ended up engaging in one-off sex without using condoms. There was also some misunderstanding of what comprised complete wound healing, and indeed a few men had sex before the completion of the recommended abstinence period. Men also correctly believed that the risk of acquiring HIV remained even after SMC and as a result the majority continued to practice safe sexual behaviour such as maintaining one sexual partner or using condoms with extra marital partners (paper IV). Conclusion: The higher level of willingness to be circumcised among men reporting sexual risk behaviour may suggest that the early adopters of SMC were likely to be those in particular need of this additional HIV protective measure. There was no clear evidence of behavioural risk compensation after circumcision in the 2011 UAIS although sexual risk behaviours were more common among circumcised than uncircumcised men. The qualitative study indicated gaps in health education for clients at health facilities, with no attention being given by health care providers to detrimental beliefs influencing sexual risk behaviour decisions, yet these beliefs were widespread among men who were interviewed