Safety of adult medical male circumcision performed by non-physician clinicians in Kenya: a prospective cohort study

Trained, experienced nurses and clinical officers provided safe voluntary medical male circumcision (VMMC) in public health facilities in Nyanza Province, Kenya, as evidenced by the low 2% adverse event rate (most commonly, excess swelling). Task shifting for male circumcision can improve access to quality VMMC services.


INTRODUCTION
I n 2007, based on results from 3 randomized controlled studies, 1-3 the World Health Organization (WHO) and the Joint United Nations Programme on HIV/AIDS (UNAIDS) determined there was compelling evidence that male circumcision reduces the risk of heterosexually acquired HIV infection in men by approximately 60%. 4,5 Later studies found that this protective effect against HIV is sustained over extended follow up (up to 6 years in the studies). 6,7 As such, in countries with high HIV prevalence, a generalized heterosexual HIV epidemic, and low rates of male circumcision, WHO and UNAIDS recommend safe, voluntary circumcision for adult men as one important component of a comprehensive strategy to prevent HIV. 4 Nyanza Province in Kenya meets these criteria for voluntary medical male circumcision (VMMC) services. In 2008, the overall HIV prevalence rate in Kenya was 5%, with approximately 83% of men circumcised, while in Nyanza Province, HIV prevalence was 12%, with only 46% of men circumcised. 8 Luos, an ethnic group of Kenya concentrated in Nyanza Province, 9 had the highest rate of HIV (20%), with only 22% of Luo men circumcised. 8 In 2009, the Kenyan Ministry of Health rolled out a policy for adult male circumcision in response to the WHO/UNAIDS recommendation. The goal was to provide a framework to ensure safe, accessible, and sustainable male circumcision services. 10, 11 Although Kenya is committed to rolling out VMMC services, availability of health care providers, particularly physicians, is limited. [10][11][12] Clinical officers were initially trained as the primary provider of male circumcision services. These are mid-level, non-physician clinicians, similar to physician assistants, whose preservice training focuses on clinical skills, often including minor surgeries. In 2009, the government approved nurses to also perform the procedure. 13 Task shifting to include nurses as circumcision providers may be one short-term strategy for addressing the gap in human resources, especially in rural communities where there tend to be fewer clinical officers and physicians. Indeed, by the end of 2011, 10 of the 14 sub-Saharan African countries where scale up of VMMC services was deemed to be a priority had implemented task shifting, and 4 of these countries had explicit taskshifting policies. 14 Ensuring safety of the male circumcision procedure is of critical importance, particularly when implemented as part of mass campaigns or when conducted by lower-level cadres. Adverse event (AE) rates in controlled trials have ranged from 1.5% in Kisumu, Kenya, 2 to 3.6% in Rakai, Uganda, 3 and 3.8% in Orange Farm, South Africa. 1 Adverse events reported in these studies included pain, 1 bleeding, 1,2 swelling, 1,2 and infection, 1-3 among others. A recent systematic review and meta-analysis of circumcisions conducted in the context of task shifting or task sharing found AE rates ranging from 0.70% (95% confidence interval [CI]50.44%-1.02%) to 37.36% (95% CI527.54%-47.72%), with an overall pooled proportion of 2.31% (95% CI51.46%-3.16%). 15 This article reports findings from a study conducted by EngenderHealth, under the umbrella of the Male Circumcision Consortium, to assess the safety of circumcisions performed by nonphysician clinicians (clinical officers and nurses) as part of routine service provision, to better inform the Government of Kenya and the Ministry of Health regarding the feasibility of rolling out circumcision services by non-physicians. The Male Circumcision Consortium worked closely with the Kenyan Ministry of Health to help inform the provision and scale up of safe VMMC services.
The objectives of the study were: Informed consent was obtained from all participants in their local language prior to participation in the study. Minors (those under age 18) were asked to sign an assent form with consent provided by a parent.

Study Procedures
Circumcision procedures and all associated VMMC services were carried out in accordance with the ''Policy on Male Circumcision in Task   Providers were asked to complete a questionnaire about their demographic, educational, and professional profiles. Using standardized case report forms, trained research assistants and male circumcision providers collected data on clients' sociodemographic background, details of the preoperative clinical examination, procedures used during the circumcision, and surgical complications/AEs experienced immediately after the circumcision, as well as 7 days and 60 days following the procedure.
All clients were interviewed about satisfaction with their circumcision 7 days and 60 days following the procedure. A 7-day follow-up visit is standard in the Kenya VMMC program. A follow-up visit at 60 days was included in the study to evaluate AEs that may not have been apparent until healing was advanced or complete (such as torsion, excessive skin removal) and to gather information on client satisfaction after healing was complete.
Participants were given a modest travel allowance (the equivalent of approximately US$3.00) at their 7-day and 60-day follow-up visits to enhance retention throughout the study.
The study was approved by the Institutional Review Board at the Kenya Medical Research Institute as well as the Protection of Human Subjects Committee at FHI 360.

Study Measures
We defined AEs based on standardized definitions outlined in the WHO male circumcision manual 16 and listed on Kenya's VMMC Adverse Events reporting form, 17 as well as based on definitions used by other researchers. 2,3 Each post-discharge AE was graded on level of severity (mild, moderate, or severe) by the male circumcision provider examining the study participant. We then convened a panel of research, HIV, and clinical experts to review the AEs in order to avoid inclusion of events that were not considered causally associated with the circumcision procedure or were within normal limits. We report only moderate and severe AE rates because most mild AEs are generally considered within normal limits; this has become the standard for reporting clinical outcomes of male circumcision studies. [18][19][20][21][22] We examined both provider-and clientrelated factors as possible risk factors associated with AEs. Provider factors included:

Statistical Analysis
We calculated the proportion of men who experienced an AE associated with the circumcision prior to discharge and at 7 days and 60 days post-circumcision. We compared provider and client characteristics of those who presented with an AE at 7 days post-circumcision with those who did not by calculating odds ratios (ORs) and corresponding 95% confidence intervals. Odds ratios were derived using generalized estimating equations (GEE) to account for clustering of client outcomes within facilities.
Variables eligible for inclusion in multivariate analysis were conceptually associated with risk of AEs and statistically associated (P , .20) with the procedure in bivariate analysis. Adjusted ORs were also generated using GEE, accounting for clustering by facility. No AEs were reported when men were discharged following their circumcision, and few AEs were reported at 60 days post-circumcision, which precluded us from evaluating risk factors for AE at those points in time.

Provider Characteristics
A total of 26 trained circumcision providers (15 nurses and 11 clinical officers) performed VMMC on study participants. The number of procedures performed by nurses ranged from 5 to 206, and by clinical officers, from 2 to 259. All the clinical officers were male; 10 of the nurses were male and 5 female. On average, nurses were approximately 5 years older than clinical officers (average age of nurses, 39.3 years; clinical officers, 34.6 years).
Nurses worked, on average, about 4 more years in their profession than clinical officers (nurses, 13.7 years; clinical officers, 10.1 years). Somewhat before the study began, nurses' official scope of work did not include minor surgeries including male circumcision. However, they reported, on average, roughly the same number of years' experience performing minor surgeries as clinical officers (nurses, 10.1 years; clinical officers, 9.9 years) and male circumcisions (nurses, 4.3 years; clinical officers, 5.3 years).
In this study, clinical officers performed 818 (36%) circumcisions and the nurses, 1,426 (64%). On average, nurses conducted the surgery in slightly less time than clinical officers (mean, 22.4 minutes versus 24.5 minutes, respectively).

Participant Characteristics
The study recruited 2,244 men and boys. Most study participants were ethnically Luo (98%) and were Christian (96%) ( Table 1). Mean age was 22.7 years. The majority (72%) were single (not married, divorced, or widowed, or not living with a partner). Most men had at least some primary education (98%), and over 70% had at least some secondary education. Most of the participants were either students (42%) or employed (39%). In general, the study participants lived relatively close to the health care facility where they had their circumcision performed, traveling on average about 5.5 kilometers, and 58% walked to the clinic. About 6% reported they had HIV infection, and about 3% reported they had a sexual partner (wife, primary partner, or any other partner) who had HIV infection. According to the participantreported data, among the men who had HIV infection, one-third had a discordant partner. Of the 2,244 men who were circumcised during the study, 2,192 (98%) returned for the 7-day followup visit, and 1,845 (82%) returned for the 60-day follow-up visit.

Participants Experiencing an Adverse Event
No study participant experienced surgical complications or immediate postoperative AEs.
At 7 days post-circumcision, 2.0% of the men experienced a moderate or severe AE; the rates were similar between nurses and clinical officers (2.1% vs. 1.9%, respectively) ( Table 2). The most common AE 7 days post-circumcision was excess swelling (about 1% among both nurses and clinical officers). There were no severe AEs, other than 2 cases of severe pain, as self-reported by men. Men 18 years and older were significantly more likely than boys under 18 years old to experience an AE at 7 days post-circumcision. Specifically, compared with younger participants, men 18 years and older were significantly more likely to experience bleeding (0.5% versus 0%) and swelling (1.3% versus 0.2%) at the 7-day follow-up visit.
At 60 days post-circumcision very few men (0.2%) experienced moderate AEs, and no men had a severe AE (Table 2). There was no difference in AEs at 60 days between circumcisions performed by nurses and clinical officers.
At 7 days following their circumcision, 89% of the men reported they had resumed work, and 76% had resumed their leisure activities. Among those who had not resumed work activities by this time, 28% said they had too much pain or discomfort to do so. Among the men who had not resumed leisure activities by 7 days, 10% reported they had not done so due to pain and discomfort.

Factors Associated With Adverse Events
In bivariate analyses, client age of 18 years and older was predictive of experiencing a moderate or severe AE, while students were less likely to experience a moderate or severe AE. Provider professional experience 6 years or higher was also associated with a lower likelihood of a participant experiencing a moderate or severe AE (P , .05) ( Table 3).

Client Satisfaction With the Circumcision
At 60 days post-circumcision, over 99% of study participants said they were satisfied; 99% reported their partner was satisfied; and over 99% said

DISCUSSION
The intent of this study was to evaluate the safety of trained non-physician cliniciansnurses and clinical officers-performing adult male circumcisions. These clinicians already had experience conducting minor surgeries and were trained on male circumcision techniques until they achieved competency. Men in this study experienced no immediate adverse outcomes associated with the circumcision or with postoperative procedures. Some did present with AEs at 7 days post-circumcision; fewer presented with AEs at 60 days post-circumcision. There was no difference in the rate of moderate and severe adverse events between participants who were circumcised by a nurse or a clinical officer.
Our findings of AE rates of 2.1% and 1.9%, for nurses and clinical officers, respectively, are lower than moderate/severe AE rates (3.8%) reported by Gray et al. in the clinical trial setting. 3 AE rates in such controlled trials may be higher than those found in observational studies, due to heightened vigilance. However, our findings are also similar to those found in a more typical service-delivery setting with circumcisions performed by trained non-physicians; Herman-Roloff and colleagues reported an AE rate (moderate and severe) of 2.1% in the context of passive surveillance (routine clinical data reported by providers) and 7.5% in the context of active surveillance (outreach by research staff to identify and report events) in Nyanza Province, Kenya. 19 While our detection of AEs was passive (that is, among participants returning to the clinic), 98% of participants returned for the 7-day follow-up visit. AE rates in this study were thus within an acceptable level compared with data from both the clinical-trial and servicedelivery settings.
Our findings are supported by similar studies evaluating the provision of circumcision by nonphysician providers. Buwembo and colleagues found no significant differences in moderate or severe AE rates between men whose circumcision was performed by physicians versus clinical officers (1.5% and 0.68%, respectively), after adjusting for other factors. 20 Herman-Roloff and colleagues found no difference in the odds of experiencing an AE when providers of any type had reported previously performing at least 100 circumcisions. 19 Others have also found that AE rates decreased as experience increased Adverse event rates for non-physicians in this study were similar to rates reported in typical service-delivery settings. (based on number of male circumcisions performed). [23][24][25] In our study, having conducted at least 100 circumcisions during the study period did not predict moderate or severe AEs; although participants circumcised by providers with less professional work experience (, 6 years) had increased risk of experiencing moderate or severe AEs.
We also investigated client-related risk factors for AEs, as some of these may be predictors or proxies for clients' behavior and risk postsurgery. Adult men, ages 18-54 years, were at higher risk of experiencing moderate or severe AEs than youth, ages 13-17 years. Adult men were more likely to experience bleeding than young men, which may not be surprising given differences in vascularization in pre-adolescents compared with adults. Adult men were also more likely to experience swelling (but not hematoma). From our data we cannot say why this occurred, and further investigation is warranted. Being employed was a risk factor for experiencing an AE. It is possible that employed men returned to work soon after the procedure, and in so doing disrupted the healing process.

Limitations and Strengths of the Study
Some limitations should be kept in mind. We did not have data on the number of circumcisions performed by each provider prior to participating in the study. Thus, we cannot evaluate provider proficiency prior to starting the study, although we know that providers were deemed competent at completion of their training. In addition, although we report AEs that we deemed to be causally associated with the circumcision, the rates we found may either overestimate or underestimate truly causal AE rates. We may have been underpowered to detect small significant differences, and results of our multivariate analysis are limited by wide confidence intervals and potentially unstable estimates. Finally, findings with regard to provider skill and experience must be interpreted with caution given the small number of providers who conducted circumcisions in the study. Our findings may not be generalizable to other service-delivery settings.
This study also has several strengths. Our retention rate at the 7-day follow-up visit was exceptionally high, and considerably higher than 7-day follow-up rates in the routine servicedelivery setting, which averaged 27.5% from 2008 to 2011, 26 decreasing the possibility that men returning for follow up were more likely to be those experiencing AEs. Further, as male circumcision is a key preventive measure for HIV, it is imperative to better understand risk factors associated with AEs after the procedure in order to make it as safe as possible. Information on such risk factors is currently limited in the literature. 19,20 Our study contributes to a better understanding of these risk factors, and as such provides information that can strengthen male circumcision services. Our study also adds to the current body of evidence regarding task shifting for male circumcision in the routine servicedelivery setting, further suggesting that trained nurses and clinical officers can provide safe medical male circumcision, and thereby facilitate increased availability and access to circumcision services.