BACKGROUND AND OBJECTIVES: Three randomized trials demonstrated male circumcision decreases female-to-male HIV incidence by 60%. Male circumcision research in sub-Saharan Africa has focused on adolescents and adults. Modeling suggests infant male circumcision (IMC) will be cost saving for HIV prevention in high to moderate seroprevalent regions. This study examined parental decision-making and differences in characteristics of parents accepting and declining IMC services in western Kenya.
METHODS: This case-control study was conducted in 2010 at 5 government hospitals in Nyanza Province, Kenya. Cases were mothers and fathers accepting circumcision for their son. Controls were parents who declined IMC services. A questionnaire comprising 41 questions was administered.
RESULTS: A total of 627 mothers and 493 fathers enrolled. In multivariable logistic regression modeling, factors associated with accepting IMC among mothers were the following: father circumcised (odds ratio [OR] = 2.30, P < .001) and agreeing with the father about the IMC decision (OR = 4.38, P < .001). Among fathers, factors associated with accepting IMC were the following: being circumcised (OR = 1.77, P = .016) and agreeing with the mother about IMC (OR = 11.0, P < .001). Fathers were the primary decision makers in most instances (66%). Few parents (3%) reported they would prefer a future son to remain uncircumcised.
CONCLUSIONS: Fathers are important in the IMC decision-making process. Fathers, as well as mothers, should be targeted for optimal scale-up of IMC services. Circumcision programs should offer services for males of all ages, as male circumcision at some age is highly acceptable to both men and women.
- sub-Saharan Africa
- male circumcision
- HIV prevention and control
- health knowledge
- health attitudes
- health practice
- acceptance of health care
- CI —
- confidence interval
- IMC —
- infant male circumcision
- MC —
- male circumcision
- OR —
- odds ratio
What’s Known on This Subject:
Male circumcision reduces risk of HIV acquisition in men by 60% and is associated with other health benefits. Compared with adult circumcision, infant male circumcision is safer, less expensive, and represents a cost-saving intervention for HIV prevention in many settings.
What This Study Adds:
IMC is little known in East Africa and is not routinely practiced. This is the first study to assess acceptability and uptake of IMC in East Africa among parents who were actually offered the procedure.
Three randomized controlled trials and numerous observational studies have demonstrated that male circumcision (MC) significantly decreases risk of HIV acquisition in men.1–5 World Health Organization guidelines recommend MC services be provided as a component of a comprehensive HIV prevention package.6 Given the limited armamentarium of proven HIV-prevention techniques, MC could play an important role in the containment of the epidemic in high-prevalence areas where the disease is primarily transmitted heterosexually and where circumcision rates are low, as is the case in many regions of sub-Saharan Africa.
To date, research on the acceptability and provision of circumcision services in sub-Saharan Africa, as well as rollout of services, have focused on adolescent and adult males.7–9 Compared with adolescent and adult MC, the circumcision of an infant is safer, less technically challenging, faster, easier to care for postoperatively, and likely to reduce chances of risk compensation.10–14 Infant male circumcision (IMC) is more easily integrated into existing medical infrastructure than adolescent/adult MC because routine antenatal, delivery, and maternal child health care is widely available. Further, a recent study in Rwanda estimated the cost per IMC procedure is $15, in comparison with $59 for an adolescent or adult procedure.13 Benefits to boys circumcised in infancy include reduction in urinary tract infections in early life and avoidance of phimosis.15,16 The same health benefits afforded to circumcised men later in life accrue to those circumcised in infancy. These include reduction in ulcerative sexually transmitted infections and penile cancer.17–20 Benefits to female sexual partners of circumcised men include reduced risk of bacterial vaginosis, trichomonas, and cervical cancer.20–22 Despite the many advantages of circumcising infants, a drawback is the lengthy interval between the intervention and impact on the HIV epidemic, even if a recent analysis has shown that infant circumcision is cost-saving for HIV prevention under conditions that prevail in many African countries.13
IMC is practiced in Ghana and other parts of West Africa but is little known in East and southern Africa.23 Studies from areas in East and southern Africa where MC (at any age) is not traditionally practiced, report levels of acceptability for MC of ∼75% under the conditions that MC is protective against HIV acquisition, and that it is offered safely and affordably.8 Most research has found a greater proportion of men and women prefer adolescent circumcision to infant circumcision, with the exception of Botswana, where 1 study showed 55% to 63% of adults preferred circumcision to be performed on infants or young children and another study showed 81% of postpartum mothers of male infants felt the best time to circumcise was within the first year of life.24,25 In regions where acceptability studies have been conducted, those who favor infant over adolescent or adult circumcision cite reduced pain, fast healing, the ability to maintain a controlled environment, and having the boy habituated to his circumcision status before he becomes an adult.8,25–30 The reasons given for opposing infant circumcision include the fragility of babies, traditional beliefs (eg, that a mother should not see her son’s circumcised penis), fear of bleeding, and allowing the boy to consent to the procedure himself.8,25–30
This study assessed parental decision-making, barriers and facilitators to IMC uptake, and differences between parents who accept and decline IMC services in Nyanza Province, Kenya. The dominant ethnic group in Nyanza is the Luo, a Nilotic people who do not traditionally practice MC. Since 2008, MC services have been scaled up in the area, and prevalence of MC in adults has increased from ∼25% to 50%.31 In 2011, HIV prevalence among pregnant women in Nyanza was 12% (Dr Charles Okal, Provincial AIDS Control Officer Nyanza Province, personal communication, 2012). This study provides insight into why parents actually choose or decline IMC for a son when the service is offered.
This case-control study was conducted between March and October 2010 at 5 government health facilities in 3 districts in western Kenya. Cases were mothers and fathers, aged 18 years or older, accepting circumcision for their son at a participating health facility. Controls were mothers and fathers who had been offered IMC for an eligible son and declined the service. All participants provided written informed consent.
Mothers delivering on the maternity ward or present at the maternal child health clinic (for antenatal care, vaccinations, well-baby visits) were given group health talks on the benefits and risks of IMC. At the 3 urban health facilities (a district hospital, a provincial hospital, and a health center), talks and IMC services were offered Monday through Friday. At the 2 smaller, peri-urban facilities (both district hospitals), talks and IMC services were offered twice per week. Information provided in the group talks included the following: that MC protects a man from heterosexual HIV acquisition by 60%; that urinary tract infections are less common in circumcised infants; and that complications associated with the procedure are rare but could include pain, bleeding, and infection. Women were approached individually after the group talks. Because of movement of women through the health facilities, neither the number of women at the group health talks nor the number of women approached individually were recorded. Mothers who had a male child <2 months of age and who declined IMC were referred to a research assistant, enrolled as controls, and interviewed in a private location. Women who accepted circumcision for their son were referred to the IMC procedure room. All women who presented for IMC services and provided written consent for the medical procedure were offered participation in the study as a case. Trained nurses and clinical officers (similar to physician assistants) provided circumcision services in an IMC room on the maternity ward at each study facility. The study interview took place on the same day as the circumcision.
By design, we aimed to enroll 300 case mothers and 300 control mothers. Sample size calculation was based on comparison of demographic variables and circumcision beliefs between those choosing circumcision and those declining. Estimated proportions for sample size determination were based on previously published research on MC acceptability in Nyanza.32 To achieve equity in the number of cases and controls, case enrollment was reviewed weekly and additional controls were enrolled as needed. After a woman consented to participate, she was asked if the father of the boy could be contacted for a separate interview. If she agreed, study personnel recorded contact information for the father and attempted to trace and interview him. Research assistants fluent in English, Kiswahili, and DhoLuo conducted face-to-face interviews (separately for mothers and fathers) lasting ∼30 minutes using a questionnaire consisting of 41 closed-ended questions (types of questions included yes/no and multiple response), many of which allowed the interviewer to choose “other” and write a unique response. Data were entered into Microsoft Access 2007 (Microsoft Corporation, Seattle, WA) and imported into SAS software version 9.2 (SAS Institute Inc, Cary, NC) for analysis. The Kenyatta National Hospital Ethics and Research Committee and the University of Illinois at Chicago Institutional Review Board provided ethical approval for this study.
Differences between cases and controls in demographic variables, beliefs, and attitudes about MC, and decision-making surrounding IMC were computed by using odds ratios (ORs) or Pearson’s χ2 test for independence. A predefined α level of 0.05 was used to assess significance. Multivariate logistic regression models were used to identify predictors of preference for IMC and potential confounders. Because mothers and fathers are not independent, we built 2 separate models. Demographic variables and variables significant at the P < .10 level in bivariate analysis were entered into exploratory logistic models. Variables with considerable correlation (Pearson’s r > 0.60) were not entered into the same model. Backward elimination of nonstatistically significant variables (P > .05) was performed 1 variable at a time and β coefficients, likelihood ratio statistics, and Hosmer and Lemeshow goodness-of-fit statistics were compared between full and reduced models after each elimination. Because Muslim and Nomiya participants (n = 54 mothers and 40 fathers) traditionally practice IMC, they were considered noninformational and were therefore excluded from multivariable models.
We approached 629 eligible women for participation in this study; of these, we enrolled 312 mothers who declined IMC services, 315 mothers who accepted IMC services, and 2 mothers (0.63%) declined participation. Of the 312 control mothers, 28 (9%) did not give consent to contact the father of the baby, 31 fathers (10%) could not be traced or were not available to be interviewed, and 253 fathers (82%) enrolled. Of the 315 case mothers, 32 mothers (10%) did not provide consent to contact the father, 43 fathers (14%) could not be traced or were not available, and 240 fathers (76%) enrolled. No fathers refused participation outright. A total of 1120 individuals (627 mothers and 493 fathers) are included in the analyses. In comparison with mothers whose partners were not enrolled in the study, mothers whose partners did enroll were more likely to live with their spouse/partner (P < .001), to have consulted the father about the IMC decision (P < .001), and to report that circumcised men experience greater sexual pleasure (P = .02).
Characteristics of Mothers and Fathers
The median age of mothers was 25 years (interquartile range 21–30) and the median age of fathers was 32 years (interquartile range 28–37). Most participants (79% of women and 83% of men) were of Luo ethnicity and the remaining were from 18 different ethnic groups. Close to half of fathers were circumcised (45% by mother’s report and 43% by father’s self-report). Ninety percent of parents (570 women, 440 men) were Christians, 6% (40 women, 31 men) were Muslim, 2% (14 women and 9 men) were Nomiya (a Kenyan Christian sect traditionally practicing IMC on the eighth day of life), and few participants (<1% of mothers and 2% of fathers) reported not belonging to any religion. Most women (73%) were unemployed and 68% reported having earned no income in the previous month. A smaller proportion of men reported being unemployed and earning no income in the previous month (23% and 13%, respectively). Other demographic characteristics are listed in Table 1.
Among mothers, cases and controls were similar in terms of ethnic origin and current employment status (Table 2). Women who adopted circumcision for their infant (cases) were more likely to know the circumcision status of the father and report the father was circumcised compared with women who declined circumcision for their son (controls). Case mothers were more likely to have no formal education or to have finished postsecondary education, to be Muslim or Nomiya, and to report some earnings in the previous month. Case mothers were less likely to be living with their husband/partner. Finally, in questions that ascertain knowledge, beliefs, and attitudes about MC, cases were more likely to believe circumcised men enjoy sex more, that women enjoy sex more with circumcised men, and that a circumcised penis looks better.
Among fathers, cases and controls were similar in terms ethnic origin, current marital status, and any income earned in the past month. Fathers who had their son circumcised were more likely to be circumcised themselves, to have completed no education or postsecondary education, and to be Muslim or Nomiya. Fathers who did not have their son circumcised were more likely to be employed.
Separate multivariate logistic regression models for mothers and fathers were built to predict acceptance of IMC in non-Muslim, non-Nomiya participants (Table 3). Among the 573 women and 453 men, 6 (<1%) were excluded because of missing data.
In the final model selected for mothers, the following characteristics were all associated with a preference for IMC: circumcised father, not residing with the husband/partner, agreeing with the husband/partner about the IMC decision, believing a circumcised man enjoys sex more, and having either no education or having finished postsecondary education (versus primary and secondary school finishers). The variable that explained the most variance in accepting IMC among mothers in multivariate analysis was belief that circumcised men experience greater sexual pleasure (OR = 3.77; 95% confidence interval [CI] 2.51, 5.67; Wald χ2 40.7), followed by agreeing with the father about circumcision (OR = 4.38; 95% CI 2.63, 7.32; Wald χ2 31.9).
In the multivariate logistic regression model for fathers, the following variables were associated with accepting IMC: agreeing with the mother about IMC (OR = 11.0; 95% CI 4.78, 25.2; Wald χ2 32.0), believing women enjoy sex more with circumcised men (OR = 3.57; 95% CI 2.27, 5.63; Wald χ2 30.1) and being circumcised (OR = 1.77; 95% CI 1.11, 2.81; Wald χ2 5.81).
The father was the primary IMC decision maker in most instances, according to interviews with mothers and fathers (60% and 72%, respectively). Mothers who had their sons circumcised were less likely to report the father was the primary decision maker in comparison with mothers who declined IMC services (51% vs 70%).
When asked all the reasons the primary decision maker chose IMC, 315 mothers gave 20 unique, unprompted reasons; the 4 most frequent reasons were protection against HIV (78%), protection against sexually transmitted infection (61%), penile hygiene (56%), and religious reasons (10%). Fathers reported similar reasons (see Table 2), although a greater proportion of fathers than mothers reported hygiene was a reason for choosing IMC (68% vs 56%).
Controls were asked the reasons that the primary decision maker had declined IMC, and 312 mothers gave 19 unique responses. The most frequently cited reasons among mothers were pain (58%), risk (45%), desire to defer circumcision to an older age (35%), the partner being against the circumcision (21%), and going against cultural tradition (14%). The variable was scored as “risk” if the mother cited bleeding, infection, swelling, injury, damage to the penis, lidocaine toxicity, or death as the reason for declining IMC. Responses from fathers were similar (see Table 2), although fathers were more likely than mothers to cite going against cultural tradition as a reason for not circumcising.
Nearly all cases (98% of mothers and 97% of fathers) said that they would prefer to have a future son circumcised in infancy. Although only 16% of control parents preferred a future son to be circumcised during early infancy, 76% of mothers and fathers declining infant circumcision reported a preference for a future son to be circumcised at a later age.
To our knowledge, this is the first published study of factors associated with uptake and decision-making surrounding IMC in sub-Saharan Africa among parents who are actually offered the service. Comparison of parents who accept and those who decline IMC and their reasons for doing so were achieved through our case-control study design. Adjusting for confounders, we found that the father being circumcised was associated with increased likelihood of accepting IMC. Although the primary stated motivation for having a son circumcised may be health or hygiene related, social acceptability of IMC among circumcised fathers appears to play a large role in the acceptability and uptake of IMC, consistent with findings from the United States.33 In Nyanza Province, where this study was conducted, ∼235 000 adolescent and adult circumcisions have been achieved in the past 3 years with a goal of an additional 200 000 to be performed by the end of 2013.34 As adult MC becomes more prevalent, demand for IMC is likely to increase. Future programming should design messages specifically for adult men who become circumcised to educate them about the availability and benefits of IMC.
Agreement between the mother and father about the IMC decision was clearly important in the decision-making process. Where disagreement about IMC existed, fathers were more likely than mothers to oppose IMC, indicating that IMC might be more acceptable among mothers. This finding is consistent with our previous studies of adult MC and those of others from sub-Saharan Africa,8,25–30 and suggests that fathers have more decision-making power over IMC than do mothers. It is notable that when parents disagreed about IMC, however, the decision not to circumcise tended to predominate, regardless of whether the mother or father was the one to decline. For example, when the father was against the procedure and the mother for it, only 2 (5%) of 41 infants were circumcised. Similarly, when the mother was against the procedure but the father for it, only 1 (5%) of 23 infants was circumcised.
Our results indicate cases and controls agree that protection against diseases and improved penile hygiene are the main reasons to choose IMC. However, all mothers received a health talk that included information about the benefits and risks of IMC before enrolling in the study, which likely influenced responses. Remarkably, although controls declined circumcision for their infants, nearly all (92%) expressed the desire for their son's circumcision at some age.
Pain and perceived health risks to the infant, including bleeding, swelling, infection, and penile damage, are the major barriers reported by those declining IMC services. Even among those accepting IMC, 25% report pain as the primary reason not to circumcise an infant boy. Educational campaigns and counseling about pre- and postoperative pain control and the low risk of complications will be needed in IMC programs.
Limitations of our study include the potential bias associated with convenience sampling and the inability to record the number of mothers screened for participation. Additionally, because eligibility for the study required that the parent had made a decision about IMC, those parents who were undecided were unlikely to be screened and asked to participate again after they had made an initial decision. Such parents might have been different from those enrolled into the study. Our results might not be generalizable to noncircumcising communities in Kenya, or to the general population in Nyanza Province, as recruitment took place from government facilities and particularly from maternity wards. Most women (56%) in Nyanza Province do not deliver in a heath care facility35; however, as 81% of women in Nyanza receive antenatal care from a provider in the government sector and 93% of infants receive the bacillus Calmette-Guérin vaccine,35 promoting IMC at government facilities among peri-natal women may be a feasible approach to scale-up of services.
Our results are useful for identifying measures that will likely reduce barriers and increase access to IMC services. As MC programs are scaled up in sub-Saharan African countries, transitioning from adolescent and adult circumcision services to infant circumcision will be prudent for sustained, cost-efficient HIV prevention.
The authors thank Linda Rosul and Nelli Westercamp for their helpful comments regarding analyses. We recognize the contributions of the late Jeckoniah Ndinya-Achola. We acknowledge the funders of this study, FHI360, who provided funds to the Male Circumcision Consortium through a grant from the Bill & Melinda Gates Foundation. Finally, we thank the research participants, without whom this project would not have been possible.
- Accepted March 22, 2012.
- Address correspondence to Marisa Young, BA, SPHPI M/C 923, 1603 W. Taylor St, Chicago, IL 60622. E-mail:
Dr Bailey, Ms Young, Dr Odoyo-June, and Dr Agot conceived and designed the study; Ms Young and Dr Bailey analyzed the data and wrote the first draft; Ms Young, Dr Bailey, Dr Odoyo-June, Dr Nordstrom, Dr Irwin, Dr Ongong’a, Ms Ochomo, and Dr Agot wrote the manuscript; Ms Young, Dr Odoyo-June, Dr Nordstrom, Dr Irwin, Dr Ongong’a, and Ms Ochomo provided and coordinated training and supervision; Ms Ochomo collected data; and Ms Young, Dr Odoyo-June, Dr Nordstrom, Dr Irwin, Dr Ongong’a, Ms Ochomo, Dr Agot, and Dr Bailey read and met criteria for authorship and agreed with the manuscript’s results and conclusions.
The funders had no role in design of the study, data collection, data analysis, manuscript preparation, or the decision to publish.
FINANCIAL DISCLOSURE: Dr Bailey receives funding from the National Institutes of Health and FHI360 for research into male circumcision for HIV prevention; FHI360 reviewed and commented on the manuscript; the other authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This study was funded by the Bill & Melinda Gates Foundation through a grant to FHI360 (grant #47394; http://www.fhi.org/en/index.htm).
- ↵WHO, UNAIDS. New data on male circumcision and HIV prevention: policy and programme implications. Technical consultation on male circumcision and HIV prevention: research implications for policy and programming, Montreaux, Switzerland March 6-8, 2007. Available at: www.malecircumcision.org. Accessed April 13, 2012
- Weiss HA,
- Dickson KE,
- Agot K,
- Hankins CA
- Wiswell TE,
- Geschke DW
- Wiswell TE,
- Hachey WE
- Auvert B,
- Sobngwi-Tambekou J,
- Cutler E,
- et al
- Weiss HA,
- Thomas SL,
- Munabi SK,
- Hayes RJ
- Castellsagué X,
- Bosch FX,
- Muñoz N,
- et al.,
- International Agency for Research on Cancer Multicenter Cervical Cancer Study Group
- ↵Gray RH, Kigozi G, Serwadda D, et al. The effects of male circumcision on female partners' genital tract symptoms and vaginal infections in a randomized trial in Rakai, Uganda. Am J Obstet Gynecol. 2009;200(1):42.e1–7
- ↵Male circumcision: global trends and determinants of prevalence, safety and acceptability Geneva. World Health Organization, Joint United Nations Programme on HIV/AIDS; 2007. Available at: www.malecircumcision.org/. Accessed April 13, 2012
- Kebaabetswe P,
- Lockman S,
- Mogwe S,
- et al
- ↵Westercamp M, Agot K, Bailey RC. Risk factors for HIV infection in a general population sample of circumcised and uncircumcised men: Kisumu, Kenya [MOLBPE047]. In: 6th IAS Conference on HIV Pathogenesis, Treatment and Prevention; 2011; Rome, Italy July 17-20, 2011
- ↵Kenya National Bureau of Statistics (KNBS) and ICF Macro. 2010. Kenya Demographic and Health Survey 2008-09. Calverton, Maryland: KNBS and ICF Macro. Available at: www.measuredhs.com/pubs/pdf/FR229/FR229.pdf. Accessed April 13, 2012
- Copyright © 2012 by the American Academy of Pediatrics