HIV counseling and testing (HCT) is central to HIV/AIDS control efforts in many developing countries.1,2 Through HCT, individuals, couples, and families can learn their HIV status and receive personalized risk reduction counseling to help prevent acquisition or further transmission of HIV. HIV-positive individuals can be enrolled in rapidly expanding services for support, care, and treatment, and efforts to reduce onward transmission through “prevention with positives” programs. Furthermore, HCT can help communities address issues of prevention, denial, stigma, and discrimination associated with HIV/AIDS and can mobilize support for HIV/AIDS control efforts.1,3
In Kenya, the number of registered HCT sites (also known as Voluntary Counseling and Testing centers) increased from 3 in 2000 to more than 900 by the end of 2007 (Kenya National AIDS and STI Control Program, personal communication VCT coordinator, Carol Ngare, 2008). Most registered HCT sites are integrated into existing health facilities (79% in 2007); however, to date, these facilities have had few staff dedicated specifically to HCT, and health care workers are occupied with other clinical tasks. As a result, many people receiving HCT in Kenya were tested at fixed locations in nonmedical settings “stand-alone HCT” during the study period.4,5
Despite the increasing availability of stand-alone HCT centers in Kenya, services are concentrated in urban areas, and knowledge of HIV status remains low. In 2003, only 14% of adults 15-49 years reported having been tested for HIV and receiving their results,6 rising to 36% in 2007.7 To extend HCT coverage to population groups with limited access to existing services, new models for delivering HCT have been developed beyond the familiar “stand-alone” method.3,8-10 These new models focus on bringing services to rural and hard-to-reach populations through a variety of “mobile HCT” approaches using trucks, vans, tents, and existing community facilities such as empty school rooms and churches. Some sites also report low utilization of their stand-alone HCT centers, and mobile HCT can be a welcome addition for increasing numbers of people reached and utilizing staff time. Although Kenya's HCT Quality Assurance Strategy11 suggest counselors are limited to serving no more than 10 HCT clients per day, many counselors report only serving an average of 2-4 clients at stand-alone sites, indicating there is capacity to serve more.
Although reports from service providers suggest that uptake and acceptance of mobile HCT is high,3,10 (J. Moore, unpublished report May 2002), it is unclear whether it is as effective as stand-alone HCT at meeting overall HIV prevention, care, and treatment objectives when all costs and outcomes are considered. Such objectives include: (a) increasing knowledge of HIV status, especially among persons previously untested; (b) identifying HIV discordant couples, for whom the HIV prevention benefits of HCT have been demonstrated12-18; and (c) newly identifying HIV-infected persons eligible for care and treatment services. Prior studies evaluating stand-alone HCT have reported cost per client tested ranging from $13 to $36,19-23 but little evaluation has been conducted on the cost or effectiveness of HCT outside stand-alone sites. One recent study in Uganda evaluating nontraditional HCT approaches compared stand-alone with hospital-based and home-based HCT approaches, finding that these alternative methods fared favorably in terms of cost per client and were effective at reaching priority groups including previously untested clients and HIV-infected individuals.23
The purpose of our study was to describe and compare stand-alone HCT with 3 approaches to providing mobile HCT that were being implemented in Kenya. Comparisons were made between stand-alone HCT and each of the 3 mobile HCT approaches and secondarily, between stand-alone HCT and mobile HCT overall. In comparing client characteristics, HIV testing results, and costs of the 4 HCT approaches, we are able to draw conclusions regarding the effectiveness of these approaches in contributing to the above HIV control objectives.
Population and Setting
Kenya has a generalized HIV epidemic primarily driven by heterosexual sex. National HIV prevalence in 2007 was estimated at 9.2% and 5.8% in 15 year-old to 49-year-old women and men, respectively, and among married persons living with HIV in Kenya, 45% had a partner who was not also infected with HIV.7 We collected data from 6 stand-alone HCT centers in Kenya, all of which had been operating for 2-4 years before data collection. Each stand-alone HCT centre also offered 1 of the 3 mobile HCT approaches. At the time of the study, these were the only programs regularly and consistently offering these types of mobile HCT approaches. The stand-alone HCT centers were based in different geographic areas, with HIV prevalence in each of the 6 study areas ranging from 5% to 15%. For programmatic reasons, the types of mobile services were not evenly spread across these areas. For example, semimobile services focused on Central, Western, Rift Valley, and Eastern Provinces, where HIV prevalence was approximately 5%, and the fully-mobile truck operated only in the higher prevalence area of Nyanza province, estimated at 15%.7
Stand-alone and mobile HCT approaches followed standardized procedures for service delivery.5,24 HCT was free, voluntary, and confidential and was delivered by trained counselors after a serial algorithm of rapid HIV antibody tests with finger-prick blood sample collection. Counselors delivered pretest counseling to individuals, couples, families, or groups to discuss basic HIV/AIDS information; explain the HIV testing process; and to discuss client risk behavior. Immediately after rapid testing, post-test counseling was conducted with all clients to explain test results, develop personalized prevention strategies, discuss partner testing and disclosure, and offer appropriate referrals. Additional programmatic characteristics of the 4 HCT approaches are presented in Table 1 and are described briefly below.
Stand-alone HCT was developed in the mid-1980s and has been considered the standard model for HCT service delivery in Kenya.2,8 Using this approach, HCT was provided at free-standing, fixed centers, which typically included rooms for at least 2 counselors and 1 reception area. These centers were not attached to a medical or health facility and were generally located in urban or semi-urban areas. Most stand-alone HCT centers relied upon Kenya's national mass media campaign for raising awareness about HCT services.25 Some local community mobilization was also undertaken to encourage uptake of services.
All mobile HCT approaches assessed here operated with stand-alone HCT centers as their base, and counselors from the stand-alone center traveled to nearby rural sites to deliver the mobile services. On average, mobile HCT sites were within 50 km of the stand-alone HCT center, with semimobile container and fully mobile truck HCT reaching greater distances than community-site mobile HCT. In addition to the national mass media campaign, mobile HCT approaches included advance mobilization with community leaders to generate support for HCT in host communities. Typically, 3-5 counselors provide services at each mobile HCT site. Three mobile HCT approaches were assessed in this study.
Community-Site Mobile HCT
Community-site mobile HCT, which utilized existing community sites such as churches, empty school rooms, clinics, and other community buildings to provide temporary HCT, was introduced in Kenya around 2002. This low-technology approach relied on locally available methods of transportation (eg, walking, riding a bicycle or motorbike, renting a vehicle) to reach nearby rural communities and hard-to-reach populations. Testing supplies were carried in bags and boxes, and hand-made posters advertising the dates and location of HCT were posted throughout the host community in advance.
Semi-Mobile Container HCT
In 2005, the US Centers for Disease Control and Prevention (CDC) introduced this mobile HCT approach as a means of increasing HCT access to more distant rural communities. A converted shipping container was hauled to each site by a detachable 4-wheel drive truck and was placed in a community for 7-10 days before being towed to its next prescheduled location. The container had 2 counseling rooms with built-in laboratory tables and space for carrying testing supplies. Counselors traveled to the testing site by local transportation or hired vehicle, and tents were pitched alongside the container for additional counseling rooms.
Fully Mobile Truck HCT
Fully-mobile truck HCT (also introduced by CDC in 2005) is very similar to semimobile container HCT, in that a large truck moves from site to site to provide mobile HCT services in communities. The truck contains 2 fully equipped HCT counseling rooms, and tents may be pitched to provide additional rooms. The primary difference is that fully mobile truck HCT is provided via a 4-wheel drive truck that is entirely self contained, and the truck can move to a new site when client flow becomes low, typically after 2 or 3 days. The truck has seats to transport up to 5 counselors.
Data Collection and Analysis
A national data form was completed for all clients receiving stand-alone and mobile HCT in Kenya. Data for this study were anonymous unlinked data collected routinely by project monitoring systems. Each service user was issued with a unique client code, and no personal identifiers were collected. Ethical approval for this study was received from the US CDC. Study projects were assessed for a 12-month period from May 2005 to April 2006. Data elements including total client volume, demographics, testing history, individual or couple attendance, and HIV diagnosis were extracted retrospectively for all clients receiving HCT at one of the study sites.
Costs and Cost-Effectiveness
Costs were assessed from a programmatic perspective. The costs of HCT services for the 12-month study period were collected from accounting reports, receipts, and interviews with project personnel. Cost estimates were based on actual expenses for staff time, vehicles, equipment, supplies, per diems, overheads, and building rental. Shared costs, including building rental costs, were divided between program components, including mobile and stand-alone activities, based on direct allocation.26 Input prices for items such as vehicles, equipment, and testing supplies were standardized across projects, and capital costs were annuitized over a 5-year useful life with a discount rate of 3%.27 Vehicle costs specific to mobile HCT were allocated fully to these approaches. All costs were converted to 2007 US dollars (1 USD = 74.03 KSh at the time of this study). Client time and travel costs were not assessed. Potential costs or cost savings incurred by the health system through HIV treatment for individuals testing positive, or averted HIV treatment from HIV prevention, were also not included.
Descriptive analyses compared HCT utilization for different age groups, gender, and key target groups. Economic analyses calculated the incremental cost-effectiveness of adding mobile HCT to stand-alone HCT for a range of outcomes: all clients, new clients, HIV-positive clients identified, and new HIV-positive clients identified. Aggregate results were calculated using weighted average costs and outcomes across all 6 sites. Different mobile HCT approaches were compared in terms of the cost per client and cost per previously untested client. Analyses were conducted using SAS software, version 9.1 [Copyright (c) 2002-2003 by SAS Institute Inc, Cary, NC].
Client Population Characteristics
Data were assessed for a total of 62,173 HCT clients. Population characteristics of clients attending stand-alone and mobile HCT are presented in Table 2. A total of 14,634 clients accessed stand-alone HCT and 47,539 accessed mobile HCT during the study period May 2005 to April 2006. The greatest number of clients was seen via community-site mobile HCT compared with the other approaches, and the smallest number was seen via the fully mobile truck. However, there was only one fully mobile truck in use during the 1-year study period. Although gender and age distribution were similar for both stand-alone and mobile HCT, mobile HCT attracted a greater proportion of new testers compared with stand-alone [88.0% (95% CI = 87.7 to 88.3) vs. 57.5% (56.7 to 58.3) P < 0.0001]. The breakdown by mobile HCT approach showed that semimobile container HCT served a greater proportion of women and young clients (15-24 years) than the other 2 mobile approaches and more couples (8%). Stand-alone HCT served a greater percentage of couples compared with mobile HCT [17.9% (17.3 to 18.5) vs. 1.9% (1.8 to 2.0), P < 0.0001].
Client Testing Results
Table 3 shows HIV prevalence in clients receiving HCT at stand-alone and mobile HCT sites during the study period. Overall, HIV prevalence was higher among all clients attending stand-alone than mobile HCT [14.1% (13.5 to 14.7) and 9.0% (8.7 to 9.2) respectively, P < 0.0001]. HIV prevalence was also higher among females, youth 15-24 years, and new testers at stand-alone HCT sites compared with mobile HCT (P < 0.0001). HIV prevalence was highest for clients tested at fully mobile truck HCT [16.9% (16.1 to 17.6)], and lowest for clients tested at semimobile container HCT [5.4% (5.0 to 5.7)]. Despite the small proportion of couples tested at fully mobile truck HCT, a substantially high proportion of these couples was discordant [19.7% (13.2 to 26.2)] compared with 11.6% (10.4 to 12.8) at stand-alone HCT, 4.8% (2.2 to 7.4) at community-site mobile HCT, and 2.8% (2.0 to 3.7) at semimobile container HCT.
The incremental cost-effectiveness of adding mobile HCT to stand-alone services is presented in Table 4. Overall, the incremental cost per outcome achieved was lower when mobile HCT was added to stand-alone services. The cost per HIV-positive client identified and new HIV-positive client identified was lower for mobile HCT than stand-alone HCT. However, these outcomes could not be assessed for each mobile HCT approach because there are possible differences in background HIV prevalence for each of the mobile approaches. Yet a comparison of the 3 mobile approaches does reveal that the incremental cost per HCT client tested and per new HCT client tested was lower for each of the 3 mobile approaches compared with stand-alone HCT, and community-site mobile HCT was the lowest cost approach for each of these outcomes.
This study assessed client characteristics, HIV testing history and results, and costs of 4 different approaches to providing HCT in Kenya. Although multiple approaches to HCT service delivery are used to provide access to different population groups and to meet HIV prevention, care, and treatment goals,3,28,29 past studies have primarily focused on describing the characteristics of traditional stand-alone HCT. This study compared stand-alone HCT with 3 alternative mobile HCT approaches: community-site mobile HCT, semimobile container HCT, and fully mobile truck HCT. The findings demonstrate that mobile HCT approaches compare favorably with stand-alone HCT in terms of cost per client and effectiveness for reaching key target population groups. This is consistent with the findings of another recent study, which showed that other nontraditional HCT approaches were less expensive than stand-alone HCT and were successful in reaching different populations.23
Our data show great utilization of HCT using mobile HCT approaches, with more than 3 times as many clients accessing HCT via mobile approaches than at stand-alone HCT for the same period. Although, overall fewer women accessed HCT services than men, our data show comparatively higher utilization of women at community-site and semimobile container HCT than stand-alone. This may be related to the barriers women face in getting time away from domestic responsibilities and cost of transport to attend distant stand-alone HCT sites. Mobile HCT approaches were also more successful than stand-alone HCT at reaching individuals who were learning their HIV status for the first time, with new clients representing more than 85% of clients at all mobile HCT approaches, compared with 58% at stand-alone HCT. Mobile HCT was introduced to improve access to HCT, to increase knowledge of HIV status in underserved communities,3,8-10 and to identify new HIV-positive individuals so that they may be referred for treatment and care services. These data indicate that mobile HCT achieved these objectives.
Our data indicate low utilization by couples at mobile HCT sites compared with stand-alone. Couples are a key target group for HCT, and couples HCT has been well documented as an effective HIV prevention intervention, especially for discordant couples.12-18 In countries such as Kenya, rates of discordance among married couples can be as high as 45%.7,30 Targeting couples with HCT is a high priority in these countries, and more effort is needed to attract couples to receive HCT together, especially at mobile sites. The high rate of discordance among couples at fully mobile truck HCT (20%) is likely explained by the high HIV prevalence in the geographic area in which this method was implemented, suggesting that HCT may be most effective at reaching discordant couples in areas known to have high HIV prevalence.
Other differences in utilization among couples may be explained by demographic differences and inherent differences in the various HCT approaches. Attending HCT as a couple requires planning, discussion, and preparation by partners. Stand-alone sites at permanent locations operate with fixed hours, allowing couples to plan for HCT services at their convenience, and they actively promote couples HCT through advertising and encouraging clients to return with their partners. Mobile services may be introduced to communities on short notices, giving partners little time to prepare for receiving couples HCT together. One exception is home-based mobile HCT, which has reported high utilization by couples and is also cost-effective compared with stand-alone HCT.23 This approach is now being scaled up rapidly in Kenya.
HCT contributes to HIV care, treatment, and prevention goals by identifying HIV-infected persons who are eligible for care and treatment and who can prevent onward transmission to others.31 The HCT strategies studied here showed great variation in their effectiveness at serving HIV-infected clients. It was no surprise that mobile HCT approaches implemented in areas with higher HIV prevalence reached higher numbers of HIV-infected individuals. Similarly, low reported HIV prevalence at semimobile HCT is likely due to the lower background prevalence in communities where this approach was used (5%-7%). Because 1 HIV-control objective is to identify HIV-infected persons, mobile HCT approaches will be more effective at achieving this objective if they are implemented in areas with high HIV prevalence.
Overall, the addition of community-site mobile HCT services is more cost effective than stand-alone services in this study, in terms of incremental cost per client tested and per new HCT client. This is not surprising, as transport and other costs are minimal with this “low-tech” approach. HCT providers typically live in the same community where stand-alone HCT operates, and from there, they may walk, bicycle, or use local transportation to get to each community-site mobile HCT. However, counselors report the burden of carrying bags and boxes with testing supplies and utilizing often unreliable means of transportation such as public transportation or faulty rental cars. Due to the transportation difficulties of this method, HCT services are often provided close to main roads or nearby the central location of the stand-alone HCT center, and counselors also report challenges related to locating temporary facilities in communities that are accessible and visible to community members but also confidential, private, and clean. Still, the 2 programs using this approach saw high numbers of clients, reflecting the fact that lower costs of mobile HCT are also linked to higher numbers of clients seen using these approaches. For example, 1 implementing partner providing community outreach served 10 times as many clients using this approach than at the stand-alone HCT, with comparable service times available. Furthermore, stand-alone HCT centers reported operating at low volume during conventional business hours, whereas the mobile HCT approaches assessed here reported a constant flow of clients throughout the day and typically extended their hours into the evening to serve more clients.
Although client numbers were large and statistical tests reached significance, we were only able to evaluate a small number of projects in provinces with widely varying HIV prevalence rates. Only 1 fully mobile HCT truck was functional at the time of the study, and this was based in Nyanza province in western Kenya with the highest HIV prevalence in the country (15%). Consequently, results related to HIV prevalence should be interpreted with caution. Cost per HIV-positive client and cost per new HIV-positive client were not reported for each mobile approach individually, given the possible bias from differences in background prevalence. In addition, although the prevention impact of HCT may be higher where services are targeted to HIV-positive individuals and discordant couples,12-17 the extent of behavioral risk reduction after HCT is unclear, particularly for new service delivery approaches such as mobile HCT. For this reason, it was not possible to calculate outcomes estimating the number of HIV infections prevented by HCT, and more research in this area is needed.
This study demonstrates that the cost of providing HCT through the addition of mobile services is generally lower than the cost of stand-alone HCT across 4 different outcomes. Mobile HCT may be an effective strategy for improving HCT access in rural and underserved urban populations and is a cost effective approach for expanding access to HCT services. Women and youth readily access mobile HCT, though based on our data more efforts are needed to attract couples to mobile HCT, such as advertising well in advance to give couples time to prepare for testing together. Although stand-alone HCT was more expensive than mobile HCT in this study, it is necessary to support the daily activities of mobile HCT, as the mobile HCT approaches studied here relied upon stand-alone HCT sites for operational support. Policymakers need to support a variety of HCT modalities, including stand-alone HCT and mobile HCT, to expand the reach of HCT to a diverse range of populations and meet HIV care, treatment, and prevention goals.
The authors wish to acknowledge John Blandford, CDC Atlanta, GA, for his technical guidance with the study. The following individuals and programs are also recognized for their assistance in conducting field work: Leonard Nambubi, Edwin Ochieng, Carol Ojuok and Carol Owiti with CDC in Nairobi, Kenya; Kennedy Ogweno, Jacob Okal, and Anthony Waruru with CDC in Kisumu, Kenya; Carolyne Muthoni and Nduku Kilonzo with Liverpool VCT, Treatment and Care in Nairobi, Kenya; Manase Amollo and Pamela Menya with Liverpool VCT, Treatment and Care in Kisumu, Kenya (formerly Tivoli AIDS Resource Centre); Andrew Ndunda and Rayola Osana-Nyyeneque with Hope Worldwide Kenya, Makindu Centre of Hope VCT and Youth Centre in Makindu, Kenya (formerly Chagua Maisha VCT and Youth Centre); Shem Andege and Carren Charus with Olemila VCT in Kapsabet, Kenya, Martin Simuyu and Sylvia Nabwire with Soy AIDS Resource Centre in Soy, Kenya; and Anastasia Juma and Paul Ochieng with Our Lady of Perpetual Support in Kisumu, Kenya.
1. World Health Organization (WHO). The Right to Know: New Approaches to HIV Testing and Counseling
. Geneva, Switzerland: WHO; 2003.
2. WHO. Increasing Access to Knowledge of HIV Status: Conclusions of a WHO Consultation, 3-4 December 2001
. Geneva, Switzerland: WHO; 2001.
3. Khumalo-Sakutukwa G, Morin SF, Fritz K, et al. Project accept (HPTN 043): a community-based intervention to reduce HIV
incidence in populations at risk for HIV
in sub-Saharan Africa and Thailand. J Acquir Immune Defic Syndr
4. National AIDS and STI Control Programme (NASCOP), Ministry of Health, Kenya
. AIDS in Kenya: Trends, Interventions, and Impact
. 7th ed. Nairobi, Kenya
: NASCOP; 2005.
5. Marum E, Taegtmeyer M, Chebet K. Scale-up of Voluntary HIV
Counseling and Testing in Kenya
6. Central Bureau of Statistics (CBS) [Kenya
], Ministry of Health (MOH) [Kenya
], and ORC Macro. Kenya Demographic and Health Survey 2003
. Calverton, MD: CBS, MOH, and ORC Macro; 2004. Available at: http://www.measuredhs.com
. Accessed online March 3, 2007.
7. National AIDS and STI Control Programme (NASCOP), Ministry of Health, Kenya
. Kenya AIDS Indicator Survey 2007: Preliminary Report
. Nairobi, Kenya
: NASCOP; 2008.
8. Matovu JK, Makumbi FE. Expanding access to voluntary HIV
counseling and testing in sub-Saharan Africa: alternative approaches for improving uptake, 2001-2007. Trop Med Int Health
9. Nahamya F, Wasswa L, Muyita A. Mobile voluntary counselling and testing services: an effective tool for access for all. Presented at: the President's Emergency Plan for AIDS Relief Annual Meeting: The 2006 HIV
/AIDS Implementers' Meeting; June 12-15, 2006; Durban, South Africa.
10. Morin SF, Khumalo-Sakutukwa G, Charlebois ED, et al. Removing barriers to knowing HIV
status: same-day mobile HIV
testing in Zimbabwe. J Acquir Immune Defic Syndr
11. NASCOP, Ministry of Health, Kenya
. National Quality Assurance Strategy for Voluntary Counseling and Testing
. Nairobi, Kenya
: NASCOP; 2005.
12. Allen S, Meinzen-Derr J, Kautzman M, et al. Sexual behavior of HIV
discordant couples after HIV
counseling and testing. AIDS
13. Allen S, Serufilira A, Bogaerts J, et al. Confidential HIV
testing and condom promotion in Africa: impact on HIV
and gonorrhea rates. JAMA
14. Allen S, Tice J, Van de Perre P, et al. Effect of serotesting with counseling on condom use and seroconversion among HIV
discordant couples in Africa. Br Med J
15. Kamenga M, Ryder RW, Jingu M, et al. Evidence of marked sexual behavior change associated with low HIV
-1 seroconversion in 149 married couples with discordant HIV
-1 serostatus: experience at an HIV
counselling center in Zaire. AIDS
16. Weinhardt LS, Carey MP, Johnson BT, et al. Effects of HIV
counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985-1997. Am J Public Health
17. Denison JA, O'Reilly KR, Schmid GP, et al. HIV
voluntary counselling and testing and behavioral risk reduction in developing countries: a meta-analysis, 1990-2005. AIDS Behav
18. Burton J, Darbes LA, Operario D. Couples-focused behavioral interventions for prevention of HIV
: systematic review of the state of evidence. AIDS Behav
19. Forsythe S, Arthur G, Ngatia G, et al. Assessing the cost
and willingness to pay for voluntary HIV
counseling and testing in Kenya
. Health Policy and Planning
20. Mota, R, Lara, A, Taegtmeyer M. A multilevel modeling of cost
of VCT services in Kenya
. 2003. [WedOa1054] Presented at: The XIIIth International Conference of AIDS and STDs in Africa; September 2003; Nairobi, Kenya
21. Sweat M, Gregorich S, Sangiwa G, et al. Cost-effectiveness
of voluntary HIV
-1 counselling and testing in reducing sexual transmission of HIV
-1 in Kenya
and Tanzania. The Lancet
22. Thielman NM, Chu HY, Ostermann J, et al. Cost-effectiveness
of free HIV voluntary counseling and testing
through a community-based AIDS service organization in Northern Tanzania. Am J Public Health
23. Menzies N, Abang B, Wanyenze R, et al. The costs and effectiveness of four HIV
counseling and testing strategies in Uganda. AIDS
24. Alwano-Edyegu M, Marum E. Knowledge is Power: Voluntary HIV Counseling and Testing in Uganda
. Geneva, Switzerland: UNAIDS; 1999.
25. Marum E, Morgan G, Hightower A, et al. Using mass media campaigns to promote voluntary counseling and HIV
-testing services in Kenya
26. Drummond MF, Sculpher MJ, Torrance GW, et al. Methods for the Economic Evaluation of Health Care Programmes
. 3rd ed. Oxford, United Kingdom: Oxford University Press; 2005.
27. Weinstein MC, Siegel JE, Gold MR, et al. Recommendations of the panel on cost-effectiveness
in health and medicine. JAMA
28. WHO. Towards universal access-Scaling up Priority HIV/AIDS Interventions in the Health Sector: Progress Report
. Geneva, Switzerland: WHO; 2007. Scaling up.
29. Coates TJ, Richter L, Caceres C. Behavioral strategies to reduce HIV
transmission: how to make them work better. Lancet
30. 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
31. King R, Lifshay J, Nakayiwa S, et al. The virus stops with me: HIV
-infected Ugandans' motivations in preventing HIV
transmission. Soc Sci Med