Bulletin of the World Health Organization
Print version ISSN 0042-9686
HERMAN-ROLOFF, Amy; BAILEY, Robert C and AGOT, Kawango. Factors associated with the safety of voluntary medical male circumcision in Nyanza province, Kenya. Bull World Health Organ [online]. 2012, vol.90, n.10, pp. 773-781. ISSN 0042-9686. http://dx.doi.org/10.2471/BLT.12.106112.
OBJECTIVE: To determine factors associated with the incidence of adverse events associated with voluntary medical male circumcision (VMMC) for the prevention of HIV infection in Nyanza province, Kenya. METHODS: Males aged 12 years or older who underwent VMMC between November 2008 and March 2010 in 16 clinics in three districts were followed through passive surveillance to monitor the incidence of adverse events during and after surgery. A subset of clinic participants was randomly selected for active surveillance post-operatively and was monitored for adverse events through a home-based, in-depth interview and a genital exam 28 to 45 days after surgery. Performance indicators were assessed for 167 VMMC providers. FINDINGS: The adverse event rate was 0.1% intra-operatively and 2.1% post-operatively among clinic system participants (n = 3705), and 7.5% post-operatively among participants under active surveillance (n = 1449). Agreement between systems was moderate (κ: 0.20; 95% confidence interval, CI: 0.09-0.32). Providers who performed more than 100 procedures achieved an adverse event rate of 0.7% and 4.3% in the clinic and active surveillance systems, respectively, and had decreased odds of performing a procedure resulting in an adverse event. With provider experience, the mean duration of the procedure also dropped from 24.0 to 15.5 minutes. Among providers who had performed at least 100 procedures, nurses and clinicians provided equivalent services. CONCLUSION: To reduce the adverse event rate, one must ensure that providers achieve a desired level of experience before they perform unsupervised procedures. Adverse events observed by the provider as well as those perceived by the client should both be monitored.