In 2007, the first Kenya AIDS Indicator Survey (KAIS 2007) found that the national prevalence of HIV in Kenya was 7.1% among persons aged 15–64 years.1 KAIS 2007 also found that 85.0% of men in Kenya had been circumcised, and that the national prevalence of HIV infection in circumcised men was 3.9% [95% confidence interval (CI): 3.3 to 4.5] compared with 13.2% (95% CI: 10.8 to 15.7) in uncircumcised men.1 The highest prevalence of HIV (14.9%) and the lowest prevalence of male circumcision (46.4%) were both in Nyanza region.1
In November 2008, Kenya's Ministry of Health began the implementation of an ambitious voluntary medical male circumcision (VMMC) program.2,3 The program was initiated after the results of 3 randomized controlled trials showed that male circumcision could reduce the risk of acquiring HIV infection among men by approximately 60%.4–6 The World Health Organization and Joint United Nations Programme on HIV/AIDS subsequently issued recommendations encouraging the use of VMMC as an effective HIV prevention intervention.7 Additionally, 1 modeling study showed that 5 HIV infections could be averted for every 36 men circumcised, saving up to 6200 US dollars per every HIV infection averted before 2025 in Kenya's Nyanza region.8 Planning for the VMMC program in Kenya relied heavily on data from the 1999 census, the 2003 Kenya Demographic and Health Survey, and KAIS 2007, which showed that 15% of Kenyan men aged 15–64 years were uncircumcised.1,9,10 Circumcision prevalence varied by region, with Nyanza region having the lowest prevalence (46.4%) followed by Nairobi, Rift Valley, and Western regions.1 The number of HIV-uninfected men who were uncircumcised at the time of KAIS 2007 was estimated to be 1.2 million, and the Nyanza and Rift Valley regions were estimated to have the largest number of uncircumcised HIV-uninfected men (601,709 and 235,688, respectively). Together with Nairobi and Western regions, these 4 regions accounted for nearly 90% of all uncircumcised HIV-uninfected men in the country.1
Kenya set a strategic target to increase the percentage of circumcised men nationally from 85% to 94% in 3–5 years by performing 860,000 VMMCs (80% of estimated need) in men aged 15–49 years by 2013. The brief duration was designed to maximize the public health impact11 and was planned to result in the reduction of the percentage of uncircumcised men from 15% to 9% nationally. Programmatic data indicated that 560,000 VMMCs had been completed in Kenya by the end of December 2012, bringing the national VMMC program achievement to within 65% of its strategic target. Moreover, 80% of these VMMCs were in the Nyanza region alone (Personal Communication, Kenya Ministry of Health, June 26, 2013).
Using data from the second Kenya AIDS Indicator Survey (KAIS 2012), we describe the progress that has been made in the Kenya VMMC program from 2007 to 2012. We further describe the current unmet need, including comparing the characteristics of the uncircumcised HIV-uninfected and the recently circumcised men in Kenya.
KAIS 2012 was a national population-based cross-sectional study, which sampled households across 9 programmatic regions of Kenya between October 2012 and February 2013. Details of its methods have been described elsewhere.12 Because of regional insecurity, the North Eastern region was not included in the survey. A 2-stage cluster sampling design was used to derive estimates of the prevalence and sociodemographic, behavioral, and biological correlates of HIV infection that were representative at the regional level for the study's adult and adolescent population (aged 15–64 years).
All individuals aged 15–64 years, who were either usual residents or had stayed overnight in the household on the night before the survey, were eligible to participate. For this analysis, we have restricted our study population to men aged 15–64 years.
Trained interviewers administered structured questionnaires that included sociodemographic characteristics and sexual behaviors. The questionnaire administered to men included items on knowledge of the benefits of male circumcision, intention to be circumcised, and for circumcisions that had occurred in last 3 years, where, when, and by whom the circumcision had been performed. To eliminate data entry time lag and ensure high standards of data quality, the survey used low-cost portable netbooks for data capture in the field (Mirus Innovations, Mississauga, Ontario, Canada). The netbooks allowed on-screen validations and data management quality checks, minimizing data transcription errors while simultaneously increasing data security and confidentiality.13 In addition, a blood sample was taken for biologic testing at the National HIV Reference Laboratory in Nairobi.
As a service to study participants, respondents were given the opportunity to learn their HIV status within their homes through home-based testing and counseling using venous blood samples collected for the survey or a separate capillary sample. Methods for home-based testing and counseling were consistent with national guidelines for HIV testing.14 All specimens for centralized testing were transported from the field to the National HIV Reference Laboratory, where HIV tests were performed. Specimens were screened with Vironostika HIV-1/2 UNIF II Plus O Enzyme Immunoassay (BioMérieux, Marcy l'Etoile, France). Specimens testing negative by the screening assay were reported as a final result negative. Specimens testing positive by the screening assay were confirmed using the Murex HIV.1.2.O HIV Enzyme Immunoassay (DiaSorin SpA, Saluggia, Italy). Samples showing discordant results after confirmatory testing were tested again with the 2 assays. Polymerase chain reaction (Cobas AmplicorHIV-1 Monitor Test, version 1.5; Roche Molecular Diagnostics, Pleasanton, CA) was performed to resolve specimens with twice-discordant results. All positive specimens and 5% of negative specimens were retested for quality assurance purposes using the same testing algorithm at the Kenya Medical Research Institute Laboratory.
We compared the proportion of circumcised men between KAIS 2007 and KAIS 2012 to assess the progress of Kenya's VMMC program. We also calculated the number of HIV-uninfected men who had not been circumcised from estimates of the population and of the prevalence of circumcision and HIV infection in each region. National and regional numbers of circumcised, uncircumcised, and HIV-uninfected uncircumcised men were estimated based on the observed prevalence of circumcision and 2012 projected population estimates from the 2009 Kenyan Population and Household Census.15 Uncertainty bounds were calculated for the population estimates based on 95% confidence intervals (CI). Final population estimates presented were rounded to the nearest 1000 persons. We also conducted descriptive analyses to examine the demographic characteristics and risk and protective behaviors of recently circumcised and uncircumcised HIV-uninfected men. Finally, using multivariable logistic regression analysis, we calculated adjusted odds ratios (aOR) and 95% CI to identify variables independently associated with being HIV-uninfected and uncircumcised in the VMMC priority region of Nyanza. We selected variables with a P value of <0.1 in bivariate analyses for final multivariable models and used backwards elimination, if they did not remain significant at a P value of <0.05. We performed all analyses using the procedures for surveys in SAS software (version 9.3; SAS Institute Inc., Cary, NC). Analyses accounted for the stratified cluster design of the survey. Each response was weighted to account for its sampling probability and adjusted for survey nonresponse. As estimates based on small sample sizes are unreliable and are not likely to be nationally or regionally representative, we suppressed estimates based on denominators less than 25 observations.
We obtained verbal informed consent from all participants for both the interview and blood sample. The study was approved by the Kenya Medical Research Institute's Ethical Review Committee, the Institutional Review Board of the US Centers for Disease Control and Prevention, and the Committee on Human Research of the University of California, San Francisco.
We identified 16,383 eligible adults in 9189 eligible households. Of these, 7452 (45.4%) were men; 5766 (77.3%) were interviewed, and 4836 (64.8%) provided blood specimens.
Overall, 91.2% of men (95% CI: 89.7 to 92.7; n = 5229) reported being circumcised, 8.8% (95% CI: 7.3 to 10.3; n = 509) reported being uncircumcised, 0.1% (n = 8) did not report their circumcision status, and 0.3% (n = 20) had missing data. The overall prevalence of HIV infection in uncircumcised men was 16.9% (95% CI: 13.1 to 20.7) compared with 3.1% (95% CI: 2.5 to 3.8; P < 0.001) in circumcised men.
The proportion of men who reported being circumcised increased by 6.2% (95% CI: 5.1 to 7.3; P < 0.001) from 85.0% (95% CI: 83.2 to 86.8) in KAIS 2007 to 91.2% (95% CI: 89.7 to 92.7) in KAIS 2012 (Table 1). The proportion of circumcised men increased in all age groups, with the largest increase of 9.4% (95% CI: 7.2 to 11.6; P < 0.001) in men aged 15–24 years and lowest increase of 1.9% (95% CI: −1.2 to 5.0; P = 0.541) in men aged 55–64 years. Proportions of circumcised men increased from 2007 to 2012 in all regions, with highest increases of 18.1% (95% CI: 13.6 to 22.6; P < 0.001), 9.0% (95% CI: 5.9 to 12.1; P < 0.001), 4.1% (95% CI: 1.7 to 6.5; P = 0.001), and 5.1% (95% CI: 2.3 to 7.9; P < 0.001) in VMMC priority regions of Nyanza (from 48.2% to 66.3%), Nairobi (from 83.2% to 92.2%), Rift Valley (88.7%–92.8%), and Western regions (87.8%–92.9%), respectively. Increases in circumcision prevalence by age group were highest in the Nyanza region, where the proportion increased by 25.2% (95% CI: 18.2 to 32.2; P < 0.001) and 21.8% (95% CI: 12.4 to 31.2; P < 0.001) in men aged 15–24 years and aged 25–34 years, respectively (Table 1). The biggest percentage increases in circumcision prevalence by ethnic groups were among men belonging to the Luo tribe (30.6%, 95% CI: 26.0 to 35.2; P < 0.001) and among men belonging to smaller tribes, for which individual tribal affiliation was not collected and were categorized as “Other” (25.9%, 95% CI: 20.5 to 31.3; P < 0.001). The Other category included Teso and the traditionally noncircumcising tribes who reside in the Turkana region.
Table 2 shows the estimated number of men aged 15–64 years in the 7 former provinces of Kenya by circumcision and HIV status. The percentage of HIV-uninfected men who were uncircumcised at the time of the survey was estimated to be 7.3% of the total population of men aged 15–64 years, which corresponded to 756,000 men. The priority regions of Nyanza, Rift Valley, and Nairobi continued to have the highest HIV prevalence among uncircumcised men aged 15–64 years [25.9% (95% CI: 20.5 to 31.3), 10.2% (95% CI: 2.7 to 17.7), and 9.5% (95% CI: 2.3 to 16.7), respectively] and large estimated numbers of uncircumcised HIV-uninfected men (345,000, 180,000, and 85,000, respectively) (Table 2). These 3 regions account for 80.7% of all uncircumcised HIV-uninfected men in the country. The priority region of Western, however, had the lowest HIV prevalence at 4.2% (95% CI: 0.0 to 10.7) and the lowest estimated number of uncircumcised HIV-uninfected men at 73,000.
The majority (84.9%, 95% CI: 80.8 to 88.9) of HIV-uninfected men who were circumcised in the last 3 years were aged 15–24 years, and 28.0% (95% CI: 21.1 to 35.0) were from the Luo ethnic tribe (Table 3). More than half lived in Rift Valley (34.8%, 95% CI: 27.1 to 42.5) or Nyanza region (22.5%, 95% CI: 16.2 to 28.9), 68.2% (95% CI: 62.0 to 74.5) lived in rural areas, and 43.0% (95% CI: 35.8 to 50.1) had reported secondary education or higher education. The majority (86.8%, 95% CI: 82.9 to 90.7) had never married or cohabited, and 52.2% (95% CI: 46.1 to 58.3) were not yet sexually active. Most (57.6%, 95% CI: 50.7 to 64.4) reported ever having been tested for HIV, and over half (52.6%, 95% CI: 43.6 to 61.5) of those sexually active in the past 12 months used a condom with the last sexual partner.
Among HIV-uninfected men who were uncircumcised, 51.4% (95% CI: 46.0 to 56.8) were aged 15–24 years and 52.5% (95% CI: 47.0% t 58.0%) had never married or cohabited (Table 3). The majority (61.2%, 95% CI: 53.9 to 68.5) were from the Luo ethnic tribe, and nearly half (46.7%, 95% CI: 39.2 to 54.3) lived in the Nyanza region. The majority of sexually active HIV-uninfected and uncircumcised men (84.6%, 95% CI: 79.8 to 89.5) were not using condoms at all times with their last sexual partner in the past 12 months, and 40.4% (95% CI: 34.2 to 46.7) reported never having been tested for HIV. The majority (72.1%, 95% CI: 66.1 to 78.1), however, knew circumcision protected men somewhat from HIV, and over half (53.0%, 95% CI: 46.7 to 59.2) reported that they intended to be circumcised in the future. Intention to be circumcised was highest among young men aged 15–24 years (69.6%, 95% CI: 61.7 to 77.4) and men who had never married or cohabited (73.6%, 95% CI: 66.1 to 81.1) (data not shown).
In Nyanza region, of the 772 men we surveyed, 13.4% (95% CI: 8.7 to 18.2) (n = 104) were HIV-uninfected and recently circumcised (past 3 years), and 26.3% (95% CI: 18.9 to 33.7) (n = 198) were HIV-uninfected and uncircumcised (data not shown). Over half (51.8%, 95% CI: 41.2 to 62.3) of men aged 15–24 years compared with 93.9% (95% CI: 87.6 to 100.0) of men aged 45–64 years were HIV-uninfected and uncircumcised (Table 4). Seventy-five percent (75.8%, 95% CI: 67.1 to 84.4) of men with secondary or higher level of education were HIV-uninfected and uncircumcised compared with 59.1% (95% CI: 50.1 to 68.1) of men who had completed primary level education. A higher proportion of men who were married or cohabiting (80.5%, 95% CI: 72.3 to 88.8) were HIV-uninfected and uncircumcised compared with men who had never married or cohabited (46.6%, 95% CI: 36.0 to 57.2). Overall, 81.9% (95% CI: 74.1 to 90.0) of men who did not use a condom consistently with the last sex partner in the past 12 months compared with 50.5% (95% CI: 36.4 to 64.5) of men who did were HIV-uninfected and uncircumcised. Moreover, 74.3% of men who reported 4 or more lifetime number of sex partners compared with 42.0% (95% CI: 21.3 to 62.7) of men who reported no partners were HIV-uninfected and uncircumcised.
In multivariable logistic regression, being younger [aged 15–24 years (aOR: 0.12, 95% CI: 0.03 to 0.55) and 25–34 years (aOR: 0.13, 95% CI: 0.03 to 0.57)] was significantly associated with lower odds of being HIV-uninfected and uncircumcised compared with men aged 45–64 years. In addition, men who reported condom use with the last sexual partner in the past 12 months had significantly lower odds of being HIV-uninfected and uncircumcised (aOR: 0.31, 95% CI: 0.16 to 0.60).
We found a significant increase in the proportion of men aged 15–64 years who reported being circumcised from 2007 to 2012. The highest increase in VMMCs was observed in regions and subpopulations where Kenya's National VMMC program has targeted since 2008, including Nyanza region, men from the Luo tribe, and men aged 15–44 years. The gains made in VMMC coverage over the past 5 years can be attributable to adoption of focused strategies for scaling up VMMC.
Kenya's minimum VMMC package includes provider-recommended and offered HIV testing and counseling, risk-reduction counseling and behavior change communication, sexually transmitted infection screening and treatment, condom promotion and provision, and medical circumcision surgery as per World Health Organization guidelines. Circumcised men are also advised to abstain from sex for 6 weeks after circumcision to allow for healing and to use condoms when sexual behavior resumes.11 The vast majority of recently circumcised men reported ever having been tested for HIV. Almost half of the HIV-uninfected men who were recently circumcised reported consistently using condoms, and over half reported using a condom at last sex, suggesting that risk compensation is not as widespread as was originally feared, and that many of those who were recently circumcised are heeding the prevention messages included in the complete package of circumcision services.16
Conversely, HIV-uninfected uncircumcised men are still engaging in risky behaviors, with only 21% reporting that a condom was used with the last sexual partner in the past 12 months. However, the difference in condom use between those recently circumcised compared with uncircumcised HIV-uninfected men may also be because of the factors unrelated to exposure to prevention messages in the minimum package. Recently circumcised men may be more disposed to prevention approaches overall, leading to an increase in uptake of both circumcision and condom use. It is also possible that many of the uncircumcised men are older and in monogamous married relationships and therefore feel no need to adhere to either prevention intervention.
It is critical that Kenya's VMMC program remains robust and explores approaches to promote and encourage HIV-uninfected uncircumcised men to access VMMC, offered as part of a comprehensive HIV prevention package, which allows access to other prevention benefits, including risk-reduction counseling, HIV testing, and condom use. Targeted strategies are needed for those men in Nyanza region who are HIV-uninfected and uncircumcised, the majority of whom are aged 15–24 years, report first sex between the ages of 15–19 years, live in rural areas, have never married or cohabited, and have previously been tested for HIV infection. A positive finding is that more than half of all uncircumcised men stated an intention to be circumcised. This finding was strongest among young men aged 15–24 years and men who had never been married or cohabited.
We recognize the limitations of our study, including the fact that comparison of the KAIS 2007 and KAIS 2012 data must be done with caution because the sampling frameworks of the 2 surveys were slightly different. North Eastern region was not included in KAIS 2012 and had high rates of circumcision in KAIS 2007, so overall prevalence of circumcision in KAIS 2012 is potentially underestimated. Circumcision status, condom use, and sexual behavior were measured through participants' self-report, and, therefore, responses may have been biased toward socially desirable answers. Some noncircumcising tribes, for example, Turkana and Teso, were combined with other smaller tribes and could not be analyzed separately. Additionally, sample sizes were small for some subgroups and resulted in reduced precision in the estimates presented. Finally, because of the cross-sectional study design, the directionality of associations cannot be established.
Notwithstanding these limitations, our study presents nationally representative population-based data on VMMC that demonstrated substantial progress toward bringing an evidence-based biomedical intervention to scale within a short period. The success of the Kenyan VMMC program should encourage other countries that are at various stages of scaling up VMMC to continue implementing VMMC actively for HIV prevention. For example, the success of Kenya's 2009, 2010, and 2011 Rapid Results Initiative informed the Swaziland Male Circumcision Task Force's decisions to adopt surgical, nonsurgical, and human resource efficiencies in scaling up VMMC.17 Overall, KAIS 2012 data show increased uptake of VMMC across Kenya, demonstrating the success of the national VMMC program. Despite this accomplishment, the Nyanza region remains below the target to circumcise 80% of all eligible men aged 15–49 years between 2009 and 2013, with the aim of averting an estimated 900,000 HIV infections over 20 years.11 Moreover, KAIS 2012 found that men with the highest HIV prevalence were aged 45–49 years.18 Therefore, strategies are needed to continue targeted messaging about the benefits of VMMC is needed, especially in Nyanza region and among older men.
Furthermore, as new cohorts of young men enter into adulthood, they will also require consistent targeting. Finally, as part of the long-term strategy of the national VMMC program, implementation and scale-up of infant circumcision services, integrated into routine maternal and child health services, is expected to maximize the long-term public health impact of VMMC on the HIV epidemic.2 To ensure sustainability of the VMMC program, continued financial resources and coordinated planning are necessary.
The authors thank Kevin De Cock, Anthony Gichangi, Amanda Viitanen, and Rennatus Mdodo for reviewing and providing input in the development of the article. The authors also thank those individuals who participated in this survey, as well as the field workers and supervisors who collected the data; and the KAIS Study Group for their contribution to the design of the survey and collection of the data set: Willis Akhwale, Sehin Birhanu, John Bore, Angela Broad, Robert Buluma, Thomas Gachuki, Jennifer Galbraith, Anthony Gichangi, Beth Gikonyo, Margaret Gitau, Joshua Gitonga, Mike Grasso, Malayah Harper, Andrew Imbwaga, Muthoni Junghae, Mutua Kakinyi, Samuel Mwangi Kamiru, Nicholas Owenje Kandege, Lucy Kanyara, Yasuyo Kawamura, Timothy Kellogg, George Kichamu, Andrea Kim, Lucy Kimondo, Davies Kimanga, Elija Kinyanjui, Stephen Kipkerich, Danson Kimutai Koske, Boniface O. K'Oyugi, Veronica Lee, Serenita Lewis, William Maina, Ernest Makokha, Agneta Mbithi, Joy Mirjahangir, Ibrahim Mohamed, Rex Mpazanje, Silas Mulwa, Nicolas Muraguri, Patrick Murithi, Lilly Muthoni, James Muttunga, Jane Mwangi, Mary Mwangi, Sophie Mwanyumba, Francis Ndichu, Anne Ng'ang'a, James Ng'ang'a, John Gitahi Ng'ang'a, Lucy Ng'ang'a, Carol Ngare, Bernadette Ng'eno, Inviolata Njeri, David Njogu, Bernard Obasi, Macdonald Obudho, Edwin Ochieng, Linus Odawo, Jacob Odhiambo, Caleb Ogada, Samuel Ogola, David Ojakaa, James Kwach Ojwang, George Okumu, Patricia Oluoch, Tom Oluoch, Kenneth Ochieng Omondi, Osborn Otieno, Yakubu Owolabi, Boniface O. K'Oyugi, Bharat Parekh, George Rutherford, Sandra Schwarcz, Shahnaaz Sharrif, Victor Ssempijja, Lydia Tabuke, Yuko Takenaka, Mamo Umuro, Brian Eugene Wakhutu, Wanjiru Waruiru, Celia Wandera, John Wanyungu, Anthony Waruru, Paul Waweru, Larry Westerman, and Kelly Winter.