HIV infection represents a major health and development problem in Kenya. By December 2003, an estimated 1.4 million Kenyans were living with HIV or AIDS, with most unaware of their infection status.1 Many HIV prevention efforts have focused on behavior change communications that involve some or all of the so-called “ABCs” of HIV prevention: abstinence, being faithful to one's partner or fidelity (which includes a reduction in one's number of partners), and condom use. Promotion of condoms has continued as a harm reduction approach to HIV prevention throughout the epidemic. Substantial data from East Africa and southern Africa as well as from Asia support the effectiveness of condoms in reducing acquisition of HIV and other sexually transmitted infections (STIs), some of which are strong cofactors in HIV acquisition and transmission.2 Analyses from Uganda, Thailand, and elsewhere have demonstrated an increase in abstinence (delay of sexual debut) and fidelity (reduction in partner acquisition) in association with declining HIV acquisition among adolescents and young adults.3,4
Recent debate has emerged over the relative importance and emphasis of the 3 components of the ABCs over others in HIV prevention messages. The President's Emergency Plan for AIDS Relief (PEPFAR) espouses a “targeted approach” whereby different components of the ABCs are given different levels of emphasis for subpopulations with different life circumstances. The cornerstone of this approach involves the promotion of abstinence among youth who have not yet engaged in sex, elimination of casual partnerships and improving skills for the maintenance of committed relationships among those who are engaged in sexual activity, and targeted promotion of condoms for high-risk subpopulations (eg, female sex workers [FSWs] and their clients).5 Critics of this approach contend that targeted approaches oversimplify the complexity of individual behaviors and community realities that structure individual sexual activity.6 Furthermore, they assert that a cohesive preventive strategy equipping individuals with a wide repertoire of skills early in life is likely to be most useful throughout the life course.
We surveyed knowledge, attitudes, beliefs, and practices, using specific questions about beliefs in the ABCs of prevention, in a sample of adult men in Nairobi, Kenya. We present the demographic correlates of HIV prevention beliefs and associations between these beliefs and self-reported HIV risk and prevention behaviors, with a focus on the ABCs of HIV prevention.
We conducted this study during 2002 at the Special Treatment Center (STC) (the primary STI referral clinic in the city) and at the Tumaini Health Clinic in the Kibera slum area of Nairobi, Kenya. The Tumaini Foundation, a Christian relief agency with a large presence in resource-poor areas of Kenya, sponsors the Tumaini Health Clinic. Using respondent-driven sampling techniques beginning with men attending STI clinics, we enrolled a sample of 500 sexually active men older than 18 years of age.7 Same-sex interviewers performed all interviews in Kiswahili or English.
HIV Prevention Beliefs: ABCs of HIV Prevention
With an open-ended question, we asked men to name “the 3 best ways you can protect yourself from HIV infection.” Four dichotomous variables were created from responses to this question based on whether men specifically cited abstinence, fidelity (being faithful to 1 partner and avoiding multiple partners at the same time), or condom use as the “best” way to prevent HIV and whether they cited at least 2 of these as best methods (a general measure of HIV prevention beliefs).
The interview assessed several demographic variables, some previously associated with HIV knowledge and risk behavior: age (a median split variable at <27.7 years and at greater than and including 27.7 years), educational attainment (secondary school level or greater vs. less than secondary school), income in the last month (a median split variable at less than and including 3200 Kenya Shillings (∼$40 US) and at greater than that value), ethnic/tribal background (5 variables, including the 4 largest tribes and all other tribes combined into “other”), religious affiliation (Catholic, Protestant, Muslim, and “other” less populous Christian faiths), years of residence in Nairobi (<10 vs. ≥10 years), current marital status, number of children (any vs. none), number of acquaintances who had died of HIV/AIDS (a variable split at less than the median of 4 and at 4 and greater), and their perception of risk of becoming infected with HIV (a variable dichotomously coded as somewhat or extremely likely or somewhat or extremely unlikely).
HIV Risk and HIV Prevention Behavioral Outcomes
We also asked men about a series of specific behaviors potentially related to increased or decreased risk of HIV acquisition and transmission. Responses were dichotomously coded as follows: (1) use of condoms ever versus never, (2) sex with an FSW ever versus never, (3) concurrent sexual partnerships during the l ast 3 months (sexual partnerships that overlapped in time during the last 3 months) versus no concurrent partnerships, (4) 100% versus <100% condom use with lower risk partners (wives and girlfriends), and (5) 100% versus <100% condom use with higher risk partners (casual partners and FSWs).
We first evaluated the association between demographic variables and belief in 1 or more of the ABCs as best HIV prevention strategies as outcomes in a set of bivariate and multivariate models using logistic regression. In multivariate models, all demographic variables were entered and then eliminated using a backward stepwise process until the final model contained only variables associated (at the P ≤ 0.1 level) with the outcome of interest.
We then measured associations between beliefs in 1 or more of the ABCs as best prevention strategies, adjusted for demographic variables, and HIV risk and prevention behaviors as outcomes in a set of bivariate and multivariate models similar to those used in the first step of the data analysis. The demographic and HIV prevention belief variables were entered into the multivariate models and removed using the same backward stepwise process mentioned previously.
The participants had a wide age distribution, were evenly split between those who were married and unmarried and between and those who had and had not fathered children, and represented some of the major ethnic and religious groups in Kenya (Table 1). More than 50% of men earned at least $40 US per month, and most had lived in Nairobi for at least 10 years and had attained at least an intermediate level of secondary school education. Most men knew several individuals who had died of HIV. Approximately 45% perceived their risk of HIV acquisition to be “somewhat or extremely likely.”
HIV Prevention Beliefs
Prevention beliefs that included the ABCs were associated at the bivariate or multivariate level with several demographic variables, including age, marital status, education, number of children, ethnicity, religious affiliation, years of residence in Nairobi, and HIV risk perception (Tables 2, 3). Men with at least a secondary school education cited abstinence and condom use individually as best prevention methods and cited ≥2 of the ABCs significantly more often than those with less education. Married men cited abstinence, condom use, and ≥2 of the ABCs less frequently and fidelity more frequently than their unmarried counterparts. Luhyas (a tribe from western Kenya) and Kambas (an eastern Kenya province tribe) cited fidelity more frequently than Luos (a Nyanza province tribe historically known for polygynous unions). Muslims and members of “other” less populous Christian faiths cited abstinence significantly less frequently than Catholics. Men living in Nairobi for at least 10 years cited abstinence less frequently than more recent migrants to the city. Men who perceived their risk of acquiring HIV to be somewhat or extremely high cited ≥2 of the ABCs more frequently than those who believed their risk was lower.
HIV Risk and Preventive Behaviors
In bivariate analyses, all the HIV prevention beliefs that included the ABCs individually and in aggregate and all demographic factors, except for the number of acquaintances dead from HIV, were associated with 1 or more HIV risk or protective behaviors (Table 4). In multivariate analyses, use of condoms ever and use of condoms consistently with wives or girlfriends were significantly more common among men who cited ≥2 of the ABCs than among men who cited <2 of the ABCs as best prevention methods (Table 5). Those citing fidelity as a best prevention method had ever used condoms less often than others who did not cite this prevention method. Men who mentioned abstinence or fidelity as a best prevention method less often ever had sex with an FSW and engaged in concurrent sexual partnerships during the last 3 months less often than men who did not cite abstinence or fidelity as a best prevention method. Additionally, those who recognized condom use as a best prevention method engaged in greater numbers of concurrent partnerships.
As shown in Table 5, reported HIV risk and prevention behaviors were also associated with demographic characteristics (eg, education, ethnicity, residence in Nairobi, income), factors that are proxies for stages of the life cycle (eg, age, marital status, parenthood), and perceptions of HIV risk and exposure to HIV mortality in the community.
This study shows that demographic factors are associated with differing beliefs in abstinence, fidelity, and condom use as best methods for HIV prevention and that these beliefs, in turn, are associated with corresponding behaviors important to the transmission or prevention of HIV in sub-Saharan Africa. Moreover, these beliefs seem to be complimentary in that they are applied in different contexts involving HIV risk.
Associations between demographic factors and beliefs in abstinence, fidelity, and condom use as best methods to prevent HIV were plausible and relevant to the context specificity of uptake of the ABCs. For example, married men more frequently reported fidelity and less frequently reported abstinence as a best prevention method when compared with their unmarried counterparts.
Further, those whose prevention beliefs included abstinence or fidelity reported less frequent concurrent partnerships as might be expected but also reported less sex ever with FSWs. These relationships demonstrate that emphasis on the ABCs in aggregate and individually are associated with behaviors that one would intuitively expect but also with other risk behaviors that are more indirectly associated with these beliefs. For example, men who believed in ≥2 of the ABCs as best prevention methods, regardless of which 2, reported using condoms ever and more consistently with wives or girlfriends. This suggests the power of a comprehensive repertoire of prevention beliefs, which can be applied in different HIV risk contexts.
These findings do fit generally with other studies of HIV behavior change interventions that employed components of the ABCs. Kamali et al3 found a significant increase over time in Uganda in respondents' report of ever using a condom and also in the age of sexual debut concurrently with a decline in HIV incidence. Furthermore, evidence from national survey data in Uganda showed positive changes in all 3 of the ABCs occurring in tandem with a reduction in new HIV infections.8 Additionally, ecologic and longitudinal data show a concomitant reduction in the number of partners or sex with FSWs and, in some cases, increases in uptake of condoms and reductions in HIV incidence among subpopulations living in Cambodia, Zambia, Ethiopia, and the Dominican Republic.9-14 Our data, on a cross-sectional level, also show strong associations between men applying HIV prevention skills employing all the ABCs individually and in aggregate and experimentation with and general maintenance of new preventive behaviors.
This study has several limitations. First, the cross-sectional design of the study does not establish causality between prevention beliefs and behaviors. A prospective design with a baseline assessment of demographics and prevention beliefs followed by a randomization of subpopulations to an intervention incorporating promotion of all of the ABCs and subsequent monitoring of HIV risk behaviors, beliefs, and practices would best define how promotion influences beliefs in and adoption of new behaviors. Second, we did not ascertain reasons for consistent condom use. The association of general HIV prevention beliefs with consistent condom use among lower risk partners from more stable relationships may sometimes represent attempts by men to regulate fertility. Family size is decreasing in Kenya, and men may be using birth control to reduce the likelihood of impregnating their partners.1,15 Third, we did not ascertain how normative and societal factors, including male community norms for sexual behaviors, economic and cultural circumstances that encourage the comodification of sexual activity among women, and decision-making power differentials between men and women in sexual scenarios, may influence the extent to which men take preventive action.16-19 Systematic attention to these factors may inform the development of future successful interventions.
In conclusion, the ABCs of HIV prevention were plausibly associated with demographic characteristics and with risk and preventive behavioral outcomes in a complimentary and context-specific manner. The findings raise the question as to whether the A, B, or C of HIV prevention should be promoted in a context-specific way (as with current US government policies) or in aggregate, beginning at an early age, to be used throughout different life stages. The fact that these prevention beliefs and practices were related to varying risk contexts and to important demographic factors (eg, age, marital status) suggests that promotion of a broader range of preventive beliefs early in life may be the most sensible approach rather than reserving specific types of prevention beliefs, such as condom promotion, only for certain subpopulations (eg, migrants, FSWs). This is particularly true, because (1) it was young, unmarried, and less educated men who were least likely to believe in fidelity and most likely to believe in condoms; (2) it was older married men who were most likely to believe in fidelity; and (3) it is difficult to envision how promotion of condom use only in men who were young, single, or married and who had casual or commercial sex would be feasible or would spare other married men from the dangers of education about condoms.
This paper is the product of work carried out by dedicated field staff based in Nairobi. The authors specifically acknowledge Charity Maingi, Charles Muga, Zachary Kwena, Norton Mutai, Calvin Obuya, and Edwin Waweru for their constant effort on this project. They also thank the clinic staff at the Special Treatment Center and at the Tumaini Health Clinic for their permission and support for this study.
1. Central Bureau of Statistics (CBS) [Kenya], Ministry of Health [Kenya], and ORC Macro. Kenya Demographic and Health Survey 2003
. Calverton, MD: CBS, MOH, and ORC Macro.
2. National Institute of Allergy and Infectious Diseases, National Institute of Health, Department of Health and Human Services. National Institute of Allergy and Infectious Diseases Workshop Summary Report: Scientific Evidence on Condom Effectiveness for Sexually Transmitted Disease (STD) Prevention, Herndon, Virginia, 2001. Available at: http://www3.niaid.nih.gov/NR/rdonlyres/84AF59B3-F28E-4971-8297-775D1C020FE4/0/condomreport.pdf
. Last accessed September 20, 2006.
3. Kamali A, Carpenter LM, Whitworth JA, et al. Seven-year trends in HIV-1 infection rates, and changes in sexual behaviour, among adults in rural Uganda. AIDS
4. Shelton JD, Halperin DT, Nantulya V, et al. Partner reduction is crucial for balanced “ABC” approach to HIV prevention. BMJ
5. Office of the United States Global AIDS Coordinator. President's Emergency Plan for AIDS Relief: Sexual Transmission of HIV and the ABC Approach to Prevention
. Washington, DC, December, 2005.
6. Barnett T, Parkhurst J. HIV/AIDS: sex, abstinence and behavior change. Lancet Infect Dis
7. Heckathorn D. Respondent-driven sampling: a new approach to the study of hidden populations. Soc Probl
8. Singh S, Darroch JE, Bankole A. A, B and C in Uganda: the roles of abstinence, monogamy and condom use in HIV decline. Reprod Health Matters
10. National Center for HIV/AIDS, Dermatology, and STDs. Behavioral Sentinel Survey (BSS). V: Sexual Behavior Among Urban Sentinel Groups, Cambodia, 2001
. Phnom Penh, Cambodia: National Center for HIV/AIDS, Dermatology and STD, 2003.
11. Fylkesnes K, Musonda RM, Sichone M, et al. Declining HIV risk behaviors in Zambia: evidence from surveillance and population-based surveys. AIDS
12. Agha S. Declines in casual sex in Lusaka, Zambia: 1996-1999. AIDS
13. Mekonnen Y, Sander E, Akilu M, et al. Evidence of changes in sexual behaviors among male factory workers in Ethiopia. AIDS
14. Green EC, Conde A. Sexual partner reduction and HIV infection. Sex Transm Infect
15. Dodoo FN, Ampofo A. AIDS-related knowledge and behavior among married Kenyan men: a behavioral paradox? J Health Hum Serv Adm
16. Aggleton P, O'Reilly K, Slutkin G, et al. Risking everything? Risk behavior, behavior change, and AIDS. Science
17. Meekers D, Calves AE. ‘Main’ girlfriends, girlfriends, marriage, and money: the social context of HIV risk behaviour in sub-Saharan Africa. Health Transit Rev
. 1997;7(Suppl 7):361-375.
18. Vanlandingham M, Knodel J, Saengtienchai C, et al. In the company of friends: peer influence on Thai male extramarital sex. Soc Sci Med
19. Wojcicki JM, Malala J. Condom use, power and HIV/AIDS risk: sex-workers bargain for survival in Hillbrow/Joubert Park/Berea, Johannesburg. Soc Sci Med
Keywords:© 2006 Lippincott Williams & Wilkins, Inc.
ABCs; AIDS; Kenya; men; prevention; sexual behavior