HIV continues to be one of the greatest health challenges in the world with an estimated 34 million people living with HIV infection globally by the end of 2011.1 Sub-Saharan Africa carries the heaviest burden of HIV infection, accounting for 69% of the total number of people living with HIV and 71% of all new HIV infections.
Improvements in the coverage of antiretroviral therapy (ART) have resulted in broad-based health gains in the sub-Saharan Africa region, saving an estimated 9 million life years.1 Costs to the health care system for sustaining treatment for the number of persons living with HIV are high and underscore the importance of reducing transmission and new infections if an AIDS-free generation is to be realized.
HIV testing and counseling (HTC) is an essential step towards receiving HIV care and treatment among HIV-infected persons and plays an important role in HIV prevention for both HIV-infected and uninfected persons. Among people living with HIV, knowledge of HIV status has been associated with more than 60% reduction in HIV transmission through improved risk-reduction behavior.2 ART itself is now recognized to confer substantial HIV prevention benefit, with a demonstrated 96% reduction in transmission within serodiscordant couples, further emphasizing the importance of HIV diagnosis.3
In the quest for universal access to HIV testing, a national target for HTC and knowledge of HIV serostatus has been set at 80% of the adult population in Kenya.4 Although considerable strides have been made toward reaching this target, Kenya faces barriers to successful HIV prevention and treatment because of high rates of undiagnosed HIV infection. In 2007, only 36% of Kenyans had ever been tested for HIV, citing limited access to services and distance to testing sites as major barriers to HIV testing.5 Furthermore, among an estimated 1.4 million Kenyans who were HIV infected at that time, 84% were unaware of their HIV infection.6
In 2008, Kenya implemented a national HTC strategy to achieve universal testing and knowledge of HIV status by 2013.7 The strategy recommended a broader package of HTC modalities beyond client-initiated voluntary counseling and testing (VCT) with greater focus on services provided directly to individuals, including provider-initiated testing and counseling in health care settings and home-based testing and counseling (HBTC).8,9 Among 6.8 million HIV tests conducted in 2012, provider-initiated testing and counseling in outpatient settings and inpatient wards was the most common testing strategy, representing 42.7% of all tests conducted; traditional VCT accounted for 38.2% of tests conducted that year.10
HBTC, where HTC services are offered to families in the privacy of their home, has been successfully implemented in sub-Saharan Africa since the early 2000s.11 Before the availability of HBTC in Kenya, 83% of Kenyans in 2007 said they would be willing to accept HIV testing in the home should it be offered to them in the future. Since then, HBTC has been implemented in high HIV prevalence regions in the country and has resulted in high testing rates among those offered the service.12 Self-testing as part of HBTC could further increase HTC coverage while addressing the issues of confidentiality for the client13 but has had limited use to date.
In this article, we assess progress towards achievement of universal access to HTC in Kenya using the results from the second Kenya AIDS Indicator Survey (KAIS 2012).
Study Setting and Design
KAIS 2012 was a population-based household survey of persons aged 18 months to 64 years covering 9 of 10 programmatic regions in Kenya. North Eastern region was not included in the survey because of regional insecurity at the time of the study. A 2-stage stratified sampling design was used to produce national and regional estimates of HIV indicators. A fuller description of the survey methods is described in detail elsewhere.14 This analysis was restricted to persons aged 15 to 64 years.
Participants were administered a questionnaire that collected information on demographics, sexual behavior, and HIV testing. Questions on HIV testing included whether participants had ever been tested for HIV, type of HIV testing received, location and frequency of HIV testing, and self-reported HIV status based on the last HIV test conducted.
Central Laboratory Testing
Following the questionnaire, participants provided blood samples for centralized HIV testing at the National HIV Reference Laboratory in Nairobi. Specimens were screened for HIV antibodies with Vironostika HIV Uni-Form 2012 Plus O Enzyme Immunoassay (bioMérieux, Marcy d'Etoile, France) and confirmed using the Murex HIV HIV.1.2.O HIV Enzyme Immunoassay (DiaSorin, SpA, Saluggia, Italy). Specimens testing negative on the screening assay were classified as HIV-negative, whereas specimens testing positive on the screening and confirmatory assays were classified as HIV-positive. Specimens with discordant results after confirmatory testing were re-tested using the same testing algorithm. Specimens with discrepant results after re-testing were tested with polymerase chain reaction (Cobas Amplicor HIV-1 Monitor Test, version 1.5, Roche Molecular Diagnostics, Pleasanton, CA) for final HIV results. Central laboratory results were not linked to individuals or households and therefore, the results of these tests were not returned to participants.
Survey participants were offered HBTC to learn their HIV status in their home using the national HIV testing algorithm for rapid HIV testing.7 After providing informed consent, participants were administered pretest counseling by trained HBTC service providers. Rapid testing was conducted using a sample of venous blood collected for central laboratory HIV testing, unless a venous sample was not provided, in which case a capillary sample was collected. An individual was diagnosed as HIV-positive in HBTC after receiving a reactive test result on the screening test [Determine HIV-1/2 (Inverness Medical, Waltham, MA)] and a subsequent reactive test result on the confirmatory test [Unigold (Trinity Biotech PLC, Bray, Ireland)]. Individuals were determined to be HIV-negative after receiving a nonreactive result on the screening test. If test results were discrepant between the screening and confirmatory test, individuals were referred to a nearby health facility for re-testing. Individuals who tested HIV-positive through HBTC were offered point-of-care CD4 testing using the PIMA CD4 Analyzer (Alere, Walltham, MA). HBTC service providers delivered test results back to participants with post-test counseling and provided referrals for follow-up services where needed.
Our two outcomes of interest were ever been tested for HIV and being unaware of HIV infection. Persons who had at least 1 HIV test before the survey were defined as persons who had ever been tested for HIV. Persons who tested HIV-positive in the central laboratory but self-reported that their last HIV test result was HIV-negative, indeterminate, or unknown; had not received the test result back; or had never been tested for HIV were defined as being unaware of their HIV infection. We examined these outcomes across select sociodemographic, behavioral, and clinical variables. These variables included sex, age, education, marital status, wealth, residence, region, sexual behavior, history of sexually transmitted infection (STI), and pregnancy status for women.
We conducted univariate analysis to describe HIV testing coverage, HIV testing behavior, and unawareness of HIV infection by select variables. Bivariate analysis was conducted to compare persons who accepted and did not accept HBTC. Data were weighted to account for sampling probability, adjusted for survey nonresponse, and presented as weighted proportions and 95% confidence intervals. A χ2 test for homogeneity was conducted to test for equality of proportions at any specified level of categorical variables. A χ2 test for independence was used to test for equality of proportions between populations. P values of less than 0.05 were considered statistically significant. Data were analyzed using SAS version 9.3 (SAS Institute, Cary, NC).
The survey protocol was approved by the Ethical Review Committee of the Kenya Medical Research Institute, the Committee on Human Research of the University of California, San Francisco, and the Institutional Review Board of the US Centers for Disease Control and Prevention. Verbal informed consent was obtained from survey participants aged 18 years and older and emancipated minors aged 15–17 years (the latter defined as persons below 18 years who were married, pregnant, or a parent). Consent was provided by parents or guardians for minors aged 15–17 years, and assent was provided by the minor participant. HBTC was conducted in a private setting around the home, with strict adherence to confidentiality. HBTC test results were not shared with any of the survey staff.
Out of 16,383 eligible persons aged 15–64 years, 13,720 (83.7%) agreed to be interviewed. Of these, 7954 (58.0%) were women and 5766 (42.0%) were men. Of 13,655 participants who provided information on HIV testing history, 71.6% had ever been tested for HIV (Table 1). Women were more likely to have ever been tested compared with men (80.4% vs. 62.5%, P < 0.001).
TABLE 1-a HIV Testin...Image Tools
HIV testing rates were highest among persons aged 25–34 years (84.6%) and lowest among persons aged 55–64 years (50.1%) (P < 0.001). Persons who were divorced or separated had the highest testing rates (78.9%) followed by persons who were married or cohabiting (77.6%) (P < 0.001). Testing rates increased with increasing education level (56.3% for those reporting no primary education compared with 74.3% for those reporting secondary education or higher, P < 0.001) and wealth index (61.8% for those in the poorest wealth category compared with 81.6% among those in the richest wealth category, P < 0.001). Urban residents were more likely to have been tested compared with rural residents (79.5% vs. 67.0%, P < 0.001). Nairobi region (84.2%) followed by Nyanza region (80.0%) had the highest levels of testing, whereas Rift Valley region had the lowest level (66.3%) (P < 0.001).
TABLE 1-b HIV Testin...Image Tools
Those who perceived themselves to be at great risk for HIV had high testing rates (83.2%) compared with those who perceived themselves to be at no risk for HIV (64.1%) (P < 0.001). Among women, those who had ever been pregnant had higher levels of testing (88.9%) compared with those who had never been pregnant (59.7%) (P < 0.001). Nearly all (95.1%) women who were currently pregnant had ever been tested for HIV. Among women who had given birth in the past 5 years, 93.0% reported that they received an HIV test during their last pregnancy (data not shown).15
Among testers, 56.1% reported being tested for HIV in the past year (Table 2). The most common testing settings were outpatient clinics (44.0%) followed by VCT settings (30.2%). Men were more likely to have been tested at VCT settings (41.3% vs. 22.0%, P < 0.001) but equally likely to have tested at outpatient clinics compared with women (43.5% vs. 44.4%). In total, 69.4% of participants had been tested for HIV more than once in their lifetime. The median number of HIV tests per person was 3 (interquartile range, 2–4) (data not shown).
The majority of testers had been tested on their own. However, over one-third (37.2%) reported that they were tested with a sexual partner in the past. Among those, men were more likely to have been tested with a partner (42.0%) compared with women (33.8%) (P < 0.001). A small proportion (3.5%) of testers reported that they had ever used an HIV self-test kit, with significantly more men citing use of such kits compared with women (5.2% vs. 2.3%, P < 0.001). Among all testers, 3.7% self-reported that they were HIV-positive based on the results of their last HIV test.
A total of 648 (5.6%) persons were HIV infected based on central laboratory testing. Among those, 53.1% were unaware of their HIV infection (Table 3). Unawareness of infection was high among men (62.0%), women who had never been pregnant (82.5%), young persons aged 15–24 years (82.0%), those reporting no primary education (61.1%), individuals who had never been married or cohabited (77.7%) or had been divorced or separated (64.6%), those in the richest wealth quintile (65.2%), and persons who did not use a condom with their most recent sexual partner in the past 12 months (67.7%). Among HIV-infected persons unaware of their infection, 76.4% had been tested for HIV in the past. Of these, 90.7% reported that they had tested HIV-negative on their last HIV test, and of these, over half (55.9%) reported having been tested for HIV in the past year (data not shown).
TABLE 3-a Unawarenes...Image Tools
In total, 9874 (72.0%) of 13,720 survey participants accepted HBTC and received their HIV results in the home (Table 4). Compared with persons who accepted HBTC, persons who refused were more likely to be in the richest wealth quintile (25.1% vs. 19.0%, P < 0.001), live in urban residences (42.0% vs. 35.1%, P = 0.003) and Coast region (11.3% vs. 8.4%, P < 0.001), report condom use with the most recent sexual partner in the past 12 months (18.8% vs. 15.6%, P = 0.002), have been tested for HIV in the past 3 months (26.1% vs. 17.7%, P < 0.001), and self-report HIV-positive status (4.6% vs. 1.9%, P = 0.002).
Among those who accepted HBTC, 48.7% were men, one-third (33.8%) were aged 15–24 years, 29.7% had never been married or cohabited, 52.0% reported secondary education or higher, and 64.9% resided in the rural areas. Most (70.7%) had been tested for HIV in the past, with 54.4% reporting that their last HIV test was conducted in the past 12 months. Over 85% of respondents who accepted HBTC felt they were at no (42.3%) to low (43.1%) risk for HIV, while only 4.7% felt they were at great risk for HIV. Among participants who accepted HBTC, 1.9% reported a previous HIV-positive diagnosis. The proportion of HIV-infected persons who were unaware of their HIV infection and accepted HBCT was 68.6% (data not shown).
TABLE 4-b Select Cha...Image Tools
Overall, 4.1% participants tested HIV-positive in HBTC. Of those, 26.6% (n = 78) were on ART, and 73.4% (n = 258) were not on ART (data not shown). Among those who accepted point-of-care CD4 testing and were not on ART (n = 244), 42.5% had CD4 cell counts ≤350 cells per microliter, and 18.4% had CD4 cell counts ≤200 cells per microliter.
In 2012, over 70% of Kenyans aged 15–64 years had ever had an HIV test, double the rate observed in 2007 when only 36.6% of adults reported ever having been tested for HIV.6 Among women, testing rates increased from 40.7% in 2007 to 80.4% in 2012, achieving the universal access target for testing. Among men, a substantial increase in HIV testing was also observed from 24.9% in 2007 to 62.5% in 2012, but this still remained below the universal access target.
Routine HIV testing for pregnant women attending antenatal clinics (ANC) for prenatal care has played a major role in increasing HIV testing rates among women. In 2012, over 90% of women who had given birth in the past 5 years received an HIV test at an ANC visit for their last pregnancy.15 Still, approximately 5% of women do not attend ANC for prenatal care and require alternative approaches for accessing HTC, linkages to ART, and prevention of mother-to-child transmission of HIV services.
We found that men who had ever been tested were more likely to have been tested in community-based settings (eg, VCT facility and mobile VCT) compared with other venues. Community testing programs, therefore, could play an important role in expanding testing to underserved men such as those who are less educated and economically disadvantaged. Furthermore, innovative testing strategies for men in clinical settings may be needed, irrespective of whether they are at the clinic for their own care or accompanying a family member, including in antenatal care settings.
Only one-third of testers had ever tested together with a sexual partner. Approximately 4 in 10 new infections in the country occur within steady heterosexual partnerships,16 and the majority of persons in discordant relationships are unaware of their HIV status. Expansion of couples-centered HTC can offer major prevention benefit by identifying discordant couples and presenting options for reducing risk and transmission within these relationships.17,18 The World Health Organization recommends that couple testing be expanded in settings where routine HIV testing is offered, with support for mutual disclosure to empower couples to make informed decisions about HIV prevention and family planning.19 As Kenya begins to implement these recommendations, there will be need to assess their feasibility, acceptability, and behavioral and clinical impact among couples.
More than half of HIV-infected Kenyans were unaware of their infection, presenting a major barrier to HIV prevention and treatment in the country. Over 90% of persons who were unaware of their infection had reported that they were HIV-negative based on their last HIV test result; over half of these had been tested in the past year. A number of reasons may explain the discrepancies between self-reported HIV-negative status and laboratory-confirmed HIV infection, including recently acquired HIV infection since the last HIV test, reluctance in reporting HIV-positive status in a survey setting, false-negative results on previous HIV tests, or lack of understanding of previous positive test results. HIV-infected persons who were unaware of their HIV infection perceived themselves to be at no or low risk for HIV infection despite recent testing behavior. The national HTC guidelines recommend annual testing for persons with ongoing risk and more frequent testing after known occurences of HIV exposure.7 Counseling messages during HTC should include comprehensive risk assessment measures and provide recommendations on when to re-test for HIV infection.
Through HBTC, 72.0% of survey participants were tested and received their HIV test results within the privacy of their homes. Furthermore, 69% of HIV-infected persons who were unaware of their infection learned their HIV status through HBTC. In KAIS 2007 where HIV testing was conducted at a central laboratory, test results were made available at a nearby health facility approximately 6 weeks after survey teams had visited the households. As a result, only 49% of participants who agreed to HIV testing in the survey accessed their test results.5 The successful implementation of HBTC in KAIS 2012 demonstrates the feasibility of its inclusion in surveys with important benefits to survey participants, including immediate knowledge of HIV status, counseling, and early linkage to HIV care.
The proportion of participants who tested HIV-positive in HBTC was 4.1%, lower than the prevalence of HIV infection reported in the survey (5.6%).20 Around 28% of survey participants did not accept HBTC and of these, 14.2% were HIV-positive based on central laboratory HIV testing. These data suggest that some individuals with previously known HIV infection declined to participate in HBTC. Although HBTC provides an essential service to survey participants, use of HBTC data alone for estimating HIV prevalence in the broader population may be subject to bias.
Through point-of-care CD4 testing, we were able to determine that over 40% of HIV-positive persons identified through HBTC were in need of immediate treatment based on the current immunologic criterion for HIV treatment (CD4 ≤350 cells per microliter)21; 17% had CD4 ≤200 cells per microliter, indicative of advanced HIV disease. Sub-Saharan Africa is particularly challenged by high rates of late HIV diagnosis of persons in advanced stage of HIV disease, late enrollment into HIV care, and delayed initiation of ART leading to poor clinical outcomes and higher mortality.22,23 Continued expansion of HTC toward universal access, routine re-testing of persons at high-risk for HIV exposure, and point-of-care CD4 testing at the point of HIV diagnosis can help to increase identification of HIV-infected persons and improve early linkages to HIV care for better health outcomes.
This analysis had a few limitations. Our main outcome variables relied on self-report of HIV testing behavior and HIV status. It is possible that participants answered according to what they perceived to be socially desirable, resulting in a dilution of observed findings. Additionally, we relied on historical data around the respondent's last HIV test, including when and where the last HIV test was conducted, allowing for the potential for recall bias. Third, because of regional insecurity at the time of the survey, North Eastern region was not included in the KAIS 2012 sample, and therefore, results are not generalizable to the country as a whole. However, North Eastern region is sparsely populated and the least affected by HIV of all regions of Kenya, with an estimated prevalence of HIV infection of less than 1%.5 It is thus unlikely that exclusion of North Eastern region biased our results substantially. Finally, HBTC and point-of-care CD4 testing were provided to participants who wanted to learn their HIV status on the day of the survey. As a result, the estimates presented on HBTC are specific for a limited sample and not generalizable to the broader Kenyan adult population.
Despite these limitations, the results presented in this analysis provide an acceptably representative picture of the status of HIV testing in Kenya. KAIS 2012 found high testing rates among adult and adolescent Kenyans, nearly achieving the national testing goal for the country. Yet, the majority of HIV infection remains unidentified, contributing to ongoing HIV transmission and disease progression in the country. We recommend that the national HTC program expands all testing modalities to rapidly identify HIV-infected persons so that they can access the care they need immediately. Greater emphasis is required on increasing uptake of HTC among couples, promoting retesting among those at high risk for HIV infection, and increasing HIV testing in men. The inclusion of HBTC and point-of-care CD4 testing in this national survey was demonstrated to be feasible, improved knowledge of HIV status among participants, and identified a group in need of treatment that would have otherwise not sought care. These data will be essential as new strategies for HTC are implemented, moving Kenya closer to achieving its universal access targets of prevention, care, and treatment.
The authors thank the fieldworkers and supervisors for their work during KAIS data collection. They also acknowledge all the individuals that participated in this survey. The authors thank Kevin DeCock, George Rutherford, Mike Grasso, and Joy Mirjahangir for reviewing and providing input on the article; 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, Maya 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, Bharat Parekh, George Rutherford, Sandra Schwarcz, Shanaaz Sharrif, Victor Ssempijja, Lydia Tabuke, Yuko Takanaka, Mamo Umuro, Brian Eugene Wakhutu, Celia Wandera, John Wanyungu, Wanjiru Waruiru, Anthony Waruru, Paul Waweru, Larry Westerman, and Kelly Winter.