Mwangi, Mary PhD*; Bunnell, Rebecca ScD, MEd*; Nyoka, Raymond MSc†; Gichangi, Anthony PhD, MSc*; Makokha, Ernest PhD, MSc*; Kim, Andrea PhD, MPH‡; Kichamu, George MSc§; Marum, Lawrence MD, MPH‖; Ichwara, Jared MA¶; Mermin, Jonathan MD, MPH*; 2007 KAIS Study Group
Increased access to antiretroviral treatment (ART) and decreased mortality coupled with persistent HIV incidence in sub-Saharan Africa has resulted in increasing numbers of people living with HIV and renewed prevention efforts targeting the general population and persons living with HIV/AIDS.1 Several studies have documented high-risk sexual behaviors among HIV-infected persons in Africa2 and other areas,3 highlighting the need to scale-up positive prevention interventions.4,5
Meta-analytic reviews of studies of prevention interventions among HIV-infected persons provide evidence of the efficacy of behavioral and biomedical prevention strategies in reducing HIV transmission among heterosexual discordant couples6-8 and men who have sex with men.9,10 In Uganda, ART, prevention counseling and partner voluntary counseling and testing (VCT) for HIV-infected adults over a 6-month period resulted in a 70% reduction in unprotected sex with partners of unknown or negative HIV status.11 Awareness of HIV status is associated with substantial reduction of high-risk sexual behaviors.12,13 In Rwanda, interventions promoting VCT and condom use resulted in a large increase in condom use among HIV discordant couples.14 Consistent condom use is estimated to reduce transmission within HIV discordant couples by 80%,15 whereas ART and prevention interventions among HIV-infected persons can result in substantial reductions in viral load and HIV transmission.16,17
Many countries with generalized epidemics, including Kenya, have embarked on integrating and scaling-up positive prevention interventions in facility and community-based HIV programs. The World Health Organization guidelines define essential prevention and care interventions for adults and adolescents living with HIV in resource-limited settings.18 The successful design and implementation of these interventions at the country level depends on a better understanding of the sociodemographic, sexual, and other characteristics of HIV-infected persons in specific geographical and sociocultural contexts.2
Many resource-limited countries lack national-level data to guide HIV prevention, care, and treatment programs although they have a growing population of HIV-infected individuals. Population-based analyses of transmission risk among HIV-infected adults in Uganda, Kenya, and Malawi have relied on limited definitions of knowledge of HIV status and sexual transmission risk2,19 attributable to lack of data on participants' knowledge of their own HIV status and that of recent sexual partners. In previous population-based analysis of sexual behaviors of HIV-infected adults in East Africa,12,19 respondents who had ever been tested and received their HIV test results were regarded as knowing their HIV status. Further, high-risk transmission sex was defined as unprotected last sex regardless of partner HIV status because earlier surveys did not collect data on the HIV status of respondents' most recent sexual partners. The inherent limitations in these definitions mean that previous analysis may have either over or underestimated sexual transmission risk behaviors among HIV-infected adults in East Africa.
To inform the design of HIV prevention interventions for HIV-infected persons in Kenya, we integrated detailed questions on HIV status and sexual behavior in a nationally representative survey to collect information on awareness of self and partner HIV status and partner-specific sexual behaviors among HIV-infected persons. We examined factors independently associated with knowledge of HIV infection, being sexually active in the year before the survey and having unprotected sex with partners of unknown or known HIV-negative status (unsafe sex) among HIV-infected adults.
The 2007 Kenya AIDS Indicator Survey (KAIS) was a cross-sectional, nationally representative population-based survey of men and women aged 15-64 years. The survey was a stratified 2-stage cluster design. In the first stage, 415 clusters were selected from the cluster sampling frame. In the second stage, 10,375 households were selected from these clusters with equal selection probability.20 Data were collected on sociodemographic characteristics, HIV/AIDS related indicators such as knowledge, sexual behavior, HIV-testing, and health service utilization. Venous blood samples were collected from consenting adults and tested for HIV-1/2, Herpes Simplex Virus-2, and syphilis. Respondents who could not provide venous blood had the option of a finger prick to prepare dried blood spots. CD4 cell enumeration was performed for all HIV-reactive samples. Study participants collected test results approximately 6 weeks after the interview. The protocol was approved by the Institutional Review Boards of the Kenya Medical Research Institute and the US Centers for Disease Control and Prevention.
Adult men and women aged 15-64 years who were usual residents or had spent the previous night in selected households were eligible to participate after providing informed consent. For minors aged 15-17 years, parental consent and minor assent were obtained before participation.
Laboratory and Clinical Data
HIV-1/2 testing was conducted in a serial algorithm [Vironostika HIV-1/2 combi-antigen/antibody enzyme immunoassay (EIA) for screening, Murex 1.2.0-antibody EIA for confirmation, and discrepancies were resolved by polymerase chain reaction (Roche Amplicor HIV DNA PCR v1.5)]. CD4 cell count was performed using automated flow cytometer (FACSCalibur, BD Biosciences, San Jose, CA) with acquisition and analytic capabilities [Multi-Test IMK with TrueCount tubes]. Syphilis was screened using indirect particle agglutination assay (Serodia TPPA, Fujirebio Inc, Japan) and confirmed with RPR (Macro-Vue, BD). Antibodies to Herpes Simplex Virus-2 were detected using antibody EIA (Kalon Herpes Simplex Virus-2). All tests were performed according to manufacturer's instructions and laboratory quality assurance procedures.
Participants who had ever been tested and were willing to disclose their HIV status were asked to disclose their HIV status. Laboratory-confirmed HIV-infected participants who self-reported HIV positive were considered to have accurate knowledge of their HIV status. Laboratory-confirmed HIV-infected participants who had never been tested and those who had been tested and self-reported HIV negative based on their last HIV test were considered as not knowing their HIV status.
Participants reporting a sexual partner in the year before the survey were regarded as sexually active. Partner HIV status and partner-specific condom use was derived from respondents' reports on their last 3 sexual partners in the past year. Respondents who self-reported inconsistent condom use with a sexual partner of unknown or known HIV-negative status in the past year were regarded as having engaged in unsafe sex.
We examined differences in sociodemographic, behavioral, and serological characteristics of HIV-infected men and women. Three separate multivariate analyses were conducted to determine factors independently associated with correct knowledge of HIV status, being sexually active and engaging in unsafe sex. In all multivariate models, we included variables that were significantly associated with the specific outcome variable (P < 0.05) in bivariate analysis and other demographic variables that were not significant but shown to be associated with the outcomes in other studies, for example, sex, marital status, and residence to adjust for potential confounding. Variables were also assessed for colinearity. We conducted further analysis to explore for potential effect modification in all models. Odds ratios and associated 95% confidence intervals (CIs) are presented. The analysis was conducted using survey procedures in SAS version 9.1 (SAS, Cary, NC). Analyses took into account stratification and clustering in the survey design, and estimates were weighted to account for sampling probability.
Characteristics of HIV-Infected Adults in Kenya
Overall, 9,691 (97%) of 10,025 eligible and occupied households completed the household interview. Of 19,840 eligible adults, 17,940 (90%) completed interviews (87% of men and 93% of women). Among those who completed interviews, 15, 853 (88%) consented to blood draw (88% of men and 88% of women). Of 15,853 adults who consented for blood draw, 1104 (6.9%) were HIV infected, which after weighting gave a national HIV prevalence estimate of 7.1%. The median age of HIV-infected adults was 33 years (31 years for women, interquartile range (IQR) 25-39, and 37 years for men, IQR: 29-44), Wilcoxon P = 0.002. Of all HIV-infected adults, 67.6% were female and half resided in Nyanza (32.1%) and Rift Valley (19.6%) provinces (Table 1). A large majority (62.7%) were currently married or cohabiting, 24.7% were previously married (14.3% widowed, 10.4% separated or divorced), and 12.7% had never married. Among those married or cohabiting, 15.9% were in a polygamous union. A large majority resided in rural areas (72.5%) and reported primary or no formal education (72.7%).
Overall, 92.5% of HIV-infected adults had heard about drugs that help HIV-infected people live longer and 36.6% knew about antiretroviral (ARV) drugs. Overall, 12.1% of HIV-infected adults reported they were taking daily cotrimoxazole. Of the 1018 HIV-infected adults with CD4 count results, 18.8% had a cell count of 250 cells per microliter or less and were eligible for ARV, however, only 17 (11.4%) reported they were receiving ARV treatment. Of all HIV-infected adults, about half (54.1%) knew about vertical transmission of HIV and 76.9% had heard about drugs that reduce mother-to-child transmission of HIV. Less than half (38%) of all HIV-infected adults (40.8% of men and 36.7% of women, P = 0.241) knew that HIV discordance within a couple was possible. Overall, 52 (7.1%) of the 676 HIV-infected women aged 15-49 years were pregnant and 13.5% of HIV-infected men aged 15-64 years had a pregnant spouse or partner at the time of the survey.
Overall, 9.6% of HIV-infected adults reported a past diagnosis of tuberculosis (TB), with 2.7% reporting a TB diagnosis in the year before the survey. Of 869 HIV-infected adults who had heard about sexually transmitted infections (STIs) and were sexually active in the past year, 52 (5.3%) self-reported having had STI symptoms in the past year.
Overall, 80.7% of HIV-infected adults were serologically positive for Herpes Simplex Virus-2 infection (73.7% of men and 84% of women, P < 0.001) and 4.2% for syphilis (6.4% of men and 3.2% of women, P = 0.079). Among all HIV-infected men aged 15-64 years, 62.9% were circumcised. Of the 856 HIV-infected adults with household data, 7.7% reported chronic illness in the past 3 months.
Knowledge of HIV Status
Overall, 87.8% of HIV-infected persons had heard of VCT, and 85% knew where one could get tested for HIV. However, less than half (44.5%) had been tested for HIV (35.1% of men and 49% of women, P < 0.001). Of those ever tested, 48.3% had their last test within less than 12 months before the survey, 23.7% within the past 12-23 months and 28.0% 2 or more years before the survey. Of 497 HIV-infected persons who had been tested for HIV and received test results, 98.2% were willing to disclose their HIV status to interviewers. Of the 363 HIV-infected adults who had been tested for HIV and were sexually active in the past year, 88.9% reported they had disclosed their HIV status to all their sexual partners in the past year.
A large majority of those who had never been tested (88.3%) expressed willingness to test at home with a trained VCT counselor. Among the 607 HIV-infected adults who had never been tested, 87.8% had heard of VCT and 85.0% knew where people can get tested. One-third (33.3%) of HIV-infected adults who had never been tested for HIV did not test because they perceived low risk of HIV infection. Other reasons for not having ever tested included fear others would know about their test results (7.1%) and long distances to the nearest VCT center (5.8%).
Overall, 916 (83.8%) of HIV-infected adults did not know their HIV status (86.0% of men and 82.8% of women); 56.0% had never been tested for HIV; and 27.8% self-reported HIV negative based on their last HIV test. Among 692 currently married or cohabiting HIV-infected adults, 379 (55.5%) had never been tested for HIV and 202 (29.1%) self-reported they were uninfected based on their last HIV test. In multivariate analyses (Table 2), correct knowledge of HIV infection was associated with being 30 or older (30-39 years adjusted odds ratio (AOR): 5.8, CI: 2.7 to 12.3; 40-49 years AOR: 5.2, CI: 2.0 to 14.0; 50-64 years AOR: 4.8, CI: 1.6 to 14.5). Further analysis (data not shown) revealed that the association between knowledge of HIV status and age differed by sex; knowledge of HIV status increased with age for women and not for men. Compared with women aged 15-24 years old, women in the 30-39 age group had the highest odds ratio (AOR: 6.9, CI: 3.1 to 15.3). Also, correct knowledge of HIV status was associated with having completed primary education (AOR: 6.8, CI: 1.8 to 24.9), having recently tested for HIV (HIV test less than 12 months before survey AOR: 4.1, CI: 2.2 to 7.9), self-reported previous diagnosis of TB (AOR: 14.1, CI: 6.3 to 31.6), and not being sexually active in the past year (AOR: 5.0, CI: 2.2 to 11.2).
Overall, 1080 (97.4%) of 1104 HIV-infected adults had ever had sex in their lifetime. Of these, 14.4% reported having one lifetime sexual partner (3% of men and 19.8% of women, P = 0.070), 83.1% had 2 or more lifetime partners (94.3% of men and 77.7% of women, P < 0.001), and 911 (82.6%) were sexually active in the past year (94.2% of men and 77.0% of women, P < 0.001). Among the 911 sexually active HIV-infected adults, 102 (11.4%) had 2 or more sexual partners in the past year (22.7% of men and 4.7% of women, P < 0.001).
In bivariate analysis (Table 3), HIV-infected adults who did not know their HIV status were less likely to have ever used a condom (39.6% vs. 56.3%, P = 0.005) and more likely to be sexually active (86% vs. 65.5%, P < 0.001) and to engage in unsafe sex (85.2% vs. 24.0%, P < 0.001) compared with HIV-infected adults who knew their HIV status. In multivariate analysis, adults aged 25-64 years were less likely to be sexually active compared with those aged 15-24 years (Table 4). Sexual activity was also associated with incorrect knowledge of HIV status (never tested for HIV, do not know HIV status AOR: 5.5, CI: 2.8 to 10.7; ever tested for HIV, do not know HIV status AOR: 6.5, CI: 2.1 to 19.6) and having multiple lifetime sexual partners (2 partners AOR: 4.1, CI: 1.5 to 10.9; 3 partners AOR: 8.5, CI: 3.4 to 21.6; 4 or more partners AOR: 18.7, CI: 8.0 to 43.8) and urban residence (AOR: 2.4, CI: 1.0 to 5.4).
Overall, 79.3% of HIV-infected adults believed condoms protect against diseases. Men were significantly more likely than women to believe that condoms diminish a man's pleasure (49.6% vs. 26.5%, P = 0.006). One in 5 (22.9%) HIV-infected adults felt that buying condoms was embarrassing. Less than half (41.9%) of adults who had ever had sex had ever used a condom (50.9% of men and 38.5% of women, P = 0.003). In bivariate analysis, HIV-infected adults who knew their HIV status were more likely to have ever used a condom than adults who did not know they were infected (Table 3). Overall, 72.7% of HIV-infected adults knew where to buy male condoms (86.8% of men and 66% of women, P < 0.001) and 18.8% knew where to buy female condoms (15.1% of men and 20.6% of women, P = 0.065). Less than half (36.7%) of the 735 HIV-infected women said they would ask a husband or partner to use a condom if they wanted him to do so.
Of the 243 HIV-infected adults who were sexually active and reported the HIV status of all their sexual partners, 75.3% reported having an HIV-negative partner or a partner of unknown HIV status. Of the 861 sexually active HIV-infected adults who provided information on the HIV status of at least 1 sexual partner and condom use in the past year, 675 (76.9%) had unprotected sex with a partner of unknown or HIV-negative status (male 74.2%, female 78.5%). In multivariate analyses (Table 5), having unsafe sex was highly associated with incorrect knowledge of HIV status with the odds of engaging in unsafe sex trending higher among HIV-infected adults who had never been tested for HIV (never tested AOR: 51.7, CI: 27.3 to 97.6; ever tested, incorrect knowledge of HIV status AOR: 18.6, CI: 8.6 to 40.5). Women, adults aged 50-64 years and those currently or previously in a marital or cohabiting relationship were more likely to engage in unsafe sex compared with men (women AOR: 1.9, 1.1 to 3.2), adults aged 15-24 years (50-64 years AOR: 4.9, CI: 1.1 to 22.0), and those never married (married/cohabiting AOR: 11.1, CI: 5.2 to 23.7; widowed/divorced/separated AOR: 3.3, CI: 1.3 to 8.4).
A large majority of HIV-infected adults in Kenya were unaware of their infection. Most were sexually active and engaged in unprotected sex, placing their uninfected partners at risk of HIV infection. Adults who knew about their infection were less likely to be sexually active and to engage in unsafe sex compared with those who did not know their status. This study also showed a trend toward higher levels of unsafe sex among HIV-infected adults who had never been tested compared with those ever tested but unaware of their HIV infection, a finding that suggests that having tested in the past could be beneficial. Unsafe sex was more common among women, older adults aged 50 or older, and those currently and previously married or cohabiting.
These findings provide additional population-level evidence of an association between knowledge of HIV status and safer sexual behaviors among HIV-infected persons and reduction in sexual transmission risk after HIV counseling and testing.8,13 The association between unsafe sex and lack of awareness of HIV status among HIV-infected adults in this survey underscores the importance of accelerating HIV testing in the general population and the potential benefit of expanding positive prevention interventions among persons living with HIV and AIDS. Kenya recorded phenomenal growth in VCT sites and clients in the past decade.21 However, although the proportion of adults aged 15-49 years ever tested for HIV grew, it remained low between 2003 (15.2%) and 2007 (36.6%).20 Low awareness of HIV status in the general population and among HIV-infected adults in this analysis support the need for innovative and targeted HIV testing and prevention approaches. Kenya has adopted a more focused approach that involves door-to-door testing in high-prevalence regions and expansion of provider-initiated testing in health care settings.22 The increased knowledge of HIV status among women aged between 30-39 years and participants who reported a previous diagnosis of TB could be attributed to increased uptake of routine HIV testing in antenatal20 and TB clinics in Kenya. Additionally, the high acceptance of home-based HIV testing in the general population19 and among HIV-infected adults who had never been tested for HIV in this analysis present an opportunity to achieve universal access to HIV testing in high-prevalence regions.
The association between having recently tested for HIV and correct knowledge of HIV status in this survey underscores the need to redefine universal HIV testing goals to include and promote repeat HIV testing among sexually active adults. Defining how often sexually active persons should test for HIV remains a challenge.
Our findings also highlight the related and urgent need to expand HIV prevention interventions that target persons living with HIV/AIDS and their partners. Kenya has integrated a comprehensive HIV prevention component into its care and treatment package for HIV-infected persons and supported the design and implementation of a national curriculum for health providers.23 The higher levels of unsafe sex among HIV-infected women, older adults, and those currently or previously married or cohabiting in this study emphasizes the need for targeted HIV testing and positive prevention interventions for women, couples, and older adults. Positive prevention interventions will need to address perceptions and feelings about condoms that could negate safer sex among persons living with HIV/AIDS.
This analysis used respondents' self-reports of partner HIV status to determine if sexually active participants had engaged in unsafe sex. The low awareness of HIV status among Kenyans20 and among HIV-infected persons in this analysis suggests there could have been under recognition and therefore underreporting of partner HIV infection in this survey. Self-report of partner HIV status is the only realistic measure of unsafe sex in a large population-based survey. Further, individuals' knowledge and perceptions about the HIV status of their sexual partners and associated sexual behaviors, whether correct or incorrect, provide a reasonable measure of sexual risk taking among HIV-infected persons in the general population.
Our analysis is based on a subset of the survey sample. Due to few observations in some strata, some of the variables assessed in our models and some of the results of effect modification analysis (data not shown) had very wide CIs, which limit the precision of some of our estimates. Despite these limitations, the low awareness of HIV status in the general population19 and among HIV-infected adults in our analysis suggests high levels of HIV transmission risk. As access to HIV care and treatment expands, Kenya and other resource-limited countries need to intensify HIV prevention interventions among HIV-infected persons to achieve a substantial reduction in HIV incidence.
We asked participants sensitive questions about their HIV status and that of sexual partners in a nationally representative HIV survey that included blood sample collection. These questions were acceptable and provided information that will give directions for programatic HIV prevention activities and benchmarks for monitoring progress of HIV prevention efforts.
1. Joint United Nations Programme on HIV/AIDS (UNAIDS). Report on the Global HIV/AIDS Epidemic 2008
. Geneva, Switzerland: UNAIDS/08.25E / JC1510E; 2009.
2. Bunnell R, Opio A, Musinguzi J, et al. HIV transmission risk behavior among HIV-infected adults in Uganda: Results of a nationally representative survey. AIDS
3. Tunthanathip P, Lolekha X, Bollen LJM, et al. Indicators for sexual HIV transmission risk among people in Thailand attending HIV care: the importance of positive prevention. Sex Transm Infect
4. Coates T, Richter L, Caceras C. Behavioral strategies to reduce HIV transmission: how to make them work better. Lancet
5. Kalichman SC. Positive prevention: reducing HIV transmission among people living with HIV/AIDS. N Engl J of Med
6. Crepaz N, Lyles CM, Wolitski RJ, et al. Do prevention strategies reduce HIV risk behaviors among people living with HIV? A meta-analytic review of controlled trials. AIDS
7. Johnson BT, Carey MP, Chaudoir SR, et al. Sexual risk reduction for persons living with HIV: research synthesis of randomized controlled trials, 1993 to 2004. J Acquir Immune Defic Syndr
8. Allen S. Effect of serotesting with counseling on condom use and seroconversion among HIV discordant couples in Africa. BMJ
9. Herbst JH, Sherba RT, Crepaz N, et al. A meta-analytic review of HIV behavioral interventions for reducing sexual risk behavior of men who have sex with men. J Acquir Immune Defic Syndr
10. Smoak, ND, Scott-Sheldon LA, Johnson, BT, et al, Sharp Research Team. Sexual risk reduction interventions do not inadvertently increase the overall frequency of sexual behavior: A meta-analysis of 174 studies with 116,735 participants. J Acquir Immune Defic Syndr
11. Bunnell R, Ekwaru JP, Solberg P, et al. Changes in sexual behavior and risk of HIV transmission after antiretroviral therapy and prevention interventions in rural Uganda. AIDS
12. Bunnell R, Opio A, Musinguzi J, et al. HIV transmission risk behavior among HIV-infected adults in Uganda: results of a nationally representative survey. AIDS
13. Marks G, Crepaz N, Senterfitt JW, et al. Meta-analysis of high-risk sexual behavior in persons aware and unaware they are infected with HIV in the United States: implications for HIV prevention programs. J Acquir Immune Defic Syndr
14. Allen S, Serufilira A, Bogaerts J, et al. Confidential HIV testing and condom promotion in Africa: impact on HIV and gonorrhea rates. JAMA
15. Weller, SC, Davis-Beaty K. Condom effectiveness in reducing heterosexual HIV transmission. Cochrane Database Syst Rev
. 2001;(1) Art: CD003255. DOI: 10.1002/14651858.CD003255.
16. Pearson CR, Kurth AE, Cassels S, et al. Modeling HIV transmission risk among Mozambicans prior to their initiating highly active antiretroviral therapy. AIDS Care
17. Bunnell R, Ekwaru JP, Solberg P, et al. Changes in sexual behavior and risk of HIV transmission after antiretroviral therapy and prevention interventions in rural Uganda. AIDS
18. World Health Organization (WHO). Essential Prevention and Care Interventions for Adults and Adolescents Living With HIV in Resource-Limited Settings/Coordinated by Kevin O'Reilly
. Geneva, Switzerland: WHO; 2008.
19. Aberle-Grasse JM, Diaz T, Anand A, et al. Knowledge of HIV status, sexual risk behaviors and contraceptive need among people living with HIV in Kenya and Malawi. AIDS
20. National AIDS and STI Control Programme (NASCOP), Kenya. 2007 Kenya AIDS Indicator Survey: Final Report
. Nairobi, NASCOP; September 2009.
21. Marum E, Taegmeyer M, Chebet K. Scale up of vouluntaty counseling and testing in Kenya. JAMA
22. Kenya National AIDS Control Council (NACC). Kenya National AIDS Strategic Plan: Delivering on Universal Access to services, 2009/2010-2012/2013
. Nairobi, Kenya: NACC; August 2009: 1-88.
23. Kenya National AIDS and STI Control Program (NASCOP). Prevention with Positives in Kenya: National Guidance for Policy Makers and Implementers (Draft)
. Nairobi, Kenya: NASCOP; October, 2009: 1-37.
© 2011 Lippincott Williams & Wilkins, Inc.