RESUMEN
BACKGROUND: Medical male circumcision is a promising HIV prevention tool in countries with generalized HIV epidemics, but demand creation interventions are needed to support scale-up. We piloted a peer referral intervention in which circumcision clients were offered incentives for referring their peers for circumcision. METHODS: The intervention was implemented between June 2014 and February 2015 in 6 randomly selected health facilities in Southern Province, Zambia. For the first 5 months, circumcision clients ≥18 years of age were given referral vouchers that allowed them to refer up to 5 peers for circumcision within a 3-month period. An incentive of US$2 was offered for each referral. The primary outcome was the number of circumcisions performed per month in each facility. To assess the effect of the intervention, a difference-in-difference analysis was performed using longitudinal data from the intervention facilities and 22 nonintervention facilities. A questionnaire was also implemented to understand men's perceptions of the intervention. RESULTS: During the 8-month intervention period, 1222 men over 18 years of age were circumcised in intervention facilities. In the first 5 months, 699 circumcision clients were enrolled and 385 clients brought a referral voucher given to them by an enrolled client. Difference-in-difference analyses did not show a significant increase in circumcisions performed in intervention facilities. However, circumcision clients reported that the referral incentive motivated them to encourage their friends to seek male circumcision. Peer referrals were also reported to be an important factor in men's decisions because 78% of clients who were referred reported that talking with a circumcised friend was important for their decision to get circumcised. CONCLUSIONS: The peer referral incentive intervention for male circumcision was feasible and acceptable. However, the intervention did not have a significant effect on demand for male circumcision. Barriers to circumcision and features of the intervention may have limited the effect of the intervention. Further efforts regarding encouraging male-to-male communication and evaluations with larger sample sizes are needed.