Share this article on:

Characteristics of HIV Voluntary Counseling and Testing Clients Before and During Care and Treatment Scale-Up in Moshi, Tanzania

Shorter, Meghan M MD*; Ostermann, Jan PhD†‡; Crump, John A MBChB*‡§‖; Tribble, Alison C MD*; Itemba, Dafrosa K BA; Mgonja, Anna; Mtalo, Antipas; Bartlett, John A MD*‡§‖; Shao, John F MD, PhD§‖; Schimana, Werner MD#; Thielman, Nathan M MD, MPH*‡

JAIDS Journal of Acquired Immune Deficiency Syndromes: December 2009 - Volume 52 - Issue 5 - p 648-654
doi: 10.1097/QAI.0b013e3181b31a6a
Epidemiology and Social Science

Objectives: We evaluated changes in characteristics of clients presenting for voluntary counseling and testing (VCT) before and during care and treatment center (CTC) scale-up activities in Moshi, Tanzania, between November 2003 and December 2007.

Methods: Consecutive clients were surveyed after pretest counseling, and rapid HIV antibody testing was performed. Trend tests were used to assess changes in seroprevalence and client characteristics over time. Multivariable logistic regression models were used to estimate the contribution of changes in sociodemographic and behavioral risk characteristics, and symptoms, to changes in seroprevalence before and during CTC scale-up.

Results: Data from 4391 first-time VCT clients were analyzed. HIV seroprevalence decreased from 26.2% to 18.9% after the availability of free antiretroviral therapy and expansion of CTCs beyond regional and referral hospitals. Seroprevalence decreased by 27 % for females (P = 0.0002) and 34% for males (P = 0.0125). Declines in seropositivity coincided with decreases in symptoms among males and females (P < 0.0001) and a more favorable distribution of sociodemographic risks among females (P = 0.002). No changes in behavioral risk characteristics were observed.

Conclusions: Concurrent with the scale-up of CTCs, HIV seroprevalence and rates of symptoms declined sharply at an established freestanding VCT site in Moshi, Tanzania. If more HIV-infected persons access VCT at sites where antiretrovirals are offered, freestanding VCT sites may become a less cost-effective means for HIV case finding.

From the *Department of Medicine, Duke University Medical Center, Durham, NC; †Center for Health Policy, Duke University, Durham, NC; ‡Duke Global Health Institute, Duke University, Durham, NC; §Kilimanjaro Christian Medical Centre, Moshi, Tanzania; ‖Kilimanjaro Christian Medical College, Tumaini University, Moshi, Tanzania; ¶KIWAKKUKI (Women against AIDS in Kilimanjaro), Moshi, Tanzania; and #Elizabeth Glaser Pediatric AIDS Foundation, Moshi, Tanzania.

Received for publication May 16, 2008; accepted May 6, 2009.

Supported by Roche Laboratories; National Institutes of Health: AIDS Clinical Trials Group (5U01 AI069484-02), International Studies of AIDS associated coinfections (U01 AI062563-04), (Drs. Bartlett, Crump, and Thielman), Mid-career Investigator (K24 AI-0744-01), and Center for AIDS Research (P30AI64518) (Dr. Bartlett); US Department of State Fulbright Program (Dr. Thielman); Hubert-Yeargan Center for Global Health, Duke University Medical Center (Drs. Tribble and Ms Mayhood).

Presented at 4th International AIDS Society (IAS) Conference on HIV Pathogenesis, Treatment, and Prevention, July 22-25, 2007, Sydney, Australia.

The authors report no conflict of interest.

Correspondence to: Nathan M. Thielman, MD, MPH, Box 90519, Room 109 Trent, Division of Infectious Disease and International Health, Duke Global Health Institute, Duke University Medical Center, Durham, NC 27710 (e-mail:

Back to Top | Article Outline


Despite substantial advances in HIV diagnostics and therapeutics over the past decade, approximately 2.5 million people were newly infected in the past year, bringing the global count of HIV-infected people to 33.2 million.1 Although many resources have been allocated to address the epidemic in sub-Saharan Africa, this region remains disproportionately affected, accounting for 68% of HIV-infected adults and 76% of adult and child deaths from AIDS worldwide.1 Continuing efforts to focus on prevention and treatment remains critical to containing the HIV/AIDS epidemic.2,3

Voluntary counseling and testing (VCT) has long been seen as an important intervention in sub-Saharan Africa, offering an individualized, client-centered approach that addresses prevention of transmission between partners and between mother and child.4 VCT also provides opportunities for early identification of infection,5 allowing for more effective treatment of HIV/AIDS and its coinfections,6,7 especially as access to antiretroviral therapy (ART) continues to expand.8

In Tanzania, free access to ART began in September of 2004 under the HIV/AIDS Treatment and Care Plan 2003-2008.9 With a goal of treating 400,000 HIV-infected Tanzanian residents by the year 2008, this plan also emphasized the need to expand VCT services as the primary entry point for care and treatment.

The interaction between prevention efforts and treatment rollout has been highlighted as an important strategy for effectively fighting the HIV epidemic.10,11 In developing countries where access to ART may be limited, it has been postulated that synergistic effects between the two could lead to 29 million fewer infections and 10 million deaths averted, whereas unsuccessful integration3 or sole focus on treatment would be insufficient to control the epidemic.12

Several studies have evaluated interactions between ART access and VCT, finding more favorable attitudes toward testing,13 expansion of education activities, and changing community attitudes toward HIV and testing14 after the availability of drugs. In Haiti,15 South Africa,16 and Botswana,17 VCT uptake increased after ART became accessible. However, it has yet to be determined how VCT client characteristics, including risk behaviors, may change after the regional scale-up of access to free care and treatment. Greater access to care and treatment centers (CTCs) might influence the characteristics of populations accessing freestanding VCT centers. Understanding such changes may help shape policy and practice around the optimal means for HIV case finding and care delivery. Using data from a cohort of clients presenting at an established stand-alone VCT Centre in Moshi, Tanzania, we examined changes in seropositivity, sociodemographic characteristics, and HIV risk behaviors and symptoms of clients from 2003 through 2007, a period intersected by the introduction of free ART in September 2004 and the rapid scale-up of multiple CTCs.

Back to Top | Article Outline


Location and Context

Subjects were recruited from the AIDS Information Centre in downtown Moshi, Tanzania, an established stand-alone VCT site operated by Kikundi cha Wanawake Kilimanjaro Kupambana na UKIMWI (KIWAKKUKI; Women Against AIDS in Kilimanjaro). VCT services have been offered since March of 2003, in conjunction with other services that have been incorporated since the founding of KIWAKKUKI in 1990 such as home-based care, education and outreach, and orphan care and assistance services. Initially, each test cost 1000 Tanzanian shillings (US $0.95 at 2003 exchange rates), but testing for clients became free in May 2004. In September 2004, the Tanzanian government began offering access to free ART; hence therapy became available locally at Kilimanjaro Christian Medical Centre (KCMC), though initially in limited supply. Scale-up of regional HIV CTCs through which patients could receive free ART progressed rapidly. Data collection for this report began November 21, 2003, and ended December 31, 2007.

Back to Top | Article Outline

Voluntary HIV Counseling and Testing Procedures

All counseling and testing was performed in accordance with guidelines provided by the Tanzanian Ministry of Health National AIDS Control Programme.18 Clients presenting for testing received confidential pretest counseling, including risk assessment and risk reduction planning with a trained counselor, after which informed consent was obtained for clients aged 18 and above. Clients younger than 18 were tested with a parent's permission but not enrolled in this study. Counselors then administered a structured questionnaire designed to obtain information on sociodemographic characteristics, reasons for testing, past and current sexual behavior, partner relationship status, HIV testing history, potential HIV-related symptoms (fever, cough, bloody cough, diarrhea, rash, night sweats, genital ulcers, and weight loss), ART knowledge, and planned changes in behavior after testing.19 For marital status specifically, clients were given the option to choose cohabitating, divorced/separated, widowed, married (polygamous or monogamous), or single (including never married, have girlfriend or boyfriend, or engaged). The provision of VCT services was not contingent upon participation in the study.

After pretest counseling and administration of the questionnaire, a 2 mL blood sample was drawn and tested using both Capillus (Trinity Biotech PLC, Bray, County Wicklow, Ireland) and Determine (Abbott Laboratories, Abbott Park, IL) rapid HIV1/2 antibody tests. If the two test results did not match, the blood samples were sent to the referral laboratory at KCMC for confirmatory testing using the enzyme-linked immunosorbent assay method. For quality control purposes, every 10th sample also underwent repeat testing using the enzyme-linked immunosorbent assay method at the KCMC laboratory. Clients who tested positive were referred to the local CTCs, offered to join peer support groups and the home-based care program and encouraged to have sexual partners and children tested. For clients who tested negative, post-test counseling focused on prevention of infection with HIV, and clients were encouraged to return for repeat testing at the time interval recommended by the Ministry of Health. Regular testing of the sexual partner was also emphasized. Clients who screened positive for possible sexually transmitted diseases, domestic violence, and/or tuberculosis were referred to the appropriate care centers.

Back to Top | Article Outline

Ethical Considerations

Ethical approval for this study was obtained from the KCMC Research Ethics Committee, the Duke University Health System Institutional Review Board, and the Tanzanian National Institute of Medical Research National Medical Research Coordinating Committee. All study subjects were provided with Kiswahili versions of the written consent, and only consenting adults were enrolled in the study.

Back to Top | Article Outline


Data were entered using Epi Info 2002 or Epi Info 3.3 software (Centers for Disease Control and Prevention, Atlanta, GA) or Teleform 9.0 (Cardiff, Visa, CA). Data were validated by randomly sampling 10% of the questionnaires, with an acceptable error rate of <1 error per 5 forms. Data were compiled into one database using JMP 6.0 (SAS Institute, Inc, Cary, NC), and all analyses were performed using STATA 10 (StataCorp, College Station, TX).

Differences in client characteristics by gender were assessed using χ2 and Student t tests. Nonparametric trend tests were used to descriptively assess changes in the characteristics of the client population over time. For analyses of changes in seropositivity and correlates of seropositivity, the study period was divided into 3 time intervals: pre-CTC period (before October 1, 2004); a transitional early CTC period (October 1, 2004 to June 30, 2006); and a CTC scale-up period (July 1, 2006 or later). The early CTC period was defined by the time of free-ART introduction at only the zonal and regional referral hospitals until decentralization to district hospitals and health centers.

Generalized Hausman tests were used to evaluate whether correlates of seropositivity, identified using gender-specific multivariable logistic regression models, differed significantly between the pre-CTC and the scale-up CTC periods. All models included the following variable domains as covariates: (1) sociodemographic characteristics (age, education, rural versus urban residence, and marital status), (2) behavioral risk (number of lifetime partners, 3 or more lifetime partners), and (3) presence of any HIV-related symptoms and the number of symptoms. All analyses were stratified by gender.

Combined models, stratified by gender and covering the entire study period, were estimated to identify the effects of changes in the client population on rates of seropositivity and to identify the relative effects of changes in the distribution of each of the 3 explanatory domains: sociodemographic characteristics, behavioral risk characteristics, and symptoms. Parameter estimates from the combined models were used to predict each client's probability of seropositivity. To determine domain-specific contributions to changes in seropositivity, we iteratively analyzed changes in the predicted probabilities over time, holding all characteristics in the respective other domains constant at the sample mean. Time effects were analyzed using linear regression models with predicted probabilities as dependent variables and time as the sole explanatory variable. The time variable was specified in 3 segments: the pre-CTC period was used as the reference group, the transitional (early CTC) period was described by a spline ranging from October 1, 2004 to June 30, 2006), and the CTC scale-up period was specified as a binary indicator variable. Significance was assessed by linear combinations of the spline and the CTC scale-up parameters. The process was repeated for each domain, separately for males and females.

Back to Top | Article Outline


Between November 21, 2003, and December 31, 2007, 8647 consecutive clients presented for VCT services at KIWAKKUKI, and, of these, 4391 (50.8%) were first-time testers. Seropositivity rates among first-time testers declined significantly in the CTC scale-up period (Fig. 1), with a decrease from 14.9% to 9.9% (P = 0.0124) in males and from 35.4% to 25.9% (P = 0.0002) in females, compared with the pre-CTC period. For comparison, also shown in Figure 1 are available Kilimanjaro region-specific antenatal clinic and demographic surveillance data from this period.20-22 Sociodemographic characteristics, behavioral risk characteristics, and client-reported symptoms in the pre-CTC period and the CTC scale-up period, along with trends across the entire study period, are summarized in Table 1. Age distribution did not change significantly across time. The percent of female clients reporting a primary school education or less decreased significantly (P < 0.0001) as did the proportion residing in rural areas for both males and females (P < 0.0001). Significantly, more men and women reported that they were divorced/separated during the later period (P = 0.0001 and P = 0.0058, respectively), and fewer women reported being widowed (P = 0.0090). The proportion of clients reporting 3 or more lifetime sexual partners increased over time from 55.6% to 66.7% for men (P = 0.0172) and from 26.4% to 31.4% for women (P = 0.0951). The number of lifetime sexual partners increased significantly for males (P = 0.0199) but not for females. Over time, the proportion of clients reporting any of the 8 listed HIV-related symptoms decreased by more than half for men and by more than a third for women (P < 0.0001); the mean number of symptoms did not change.





Summarized in Table 2 are client reasons for testing during the pre-CTC and CTC scale-up periods with trend tests for all clients across the study period, by gender. Over time, a greater percentage of men and women cited having multiple sexual partners (P < 0.0001 for each), suspecting unfaithfulness in their partner (P < 0.0001 for each), and premarital planning (P = 0.0167 for males; P = 0.0002 for females). Women citing preconception planning as a reason for testing increased significantly (P = 0.0068). Multivariate analysis of correlates of seropositivity from 2003 to 2007 revealed no significant changes over time (Table 3). Among males, receiving less education [odds ratio (OR): 1.45, 95% confidence interval (CI): 1.07 to 2.07, P < 0.05], residing in a rural area (OR: 1.49, CI: 1.07 to 2.09, P < 0.05), cohabitating (OR: 5.76, CI: 3.49 to 9.51, P < 0.0001), being divorced/separated (OR: 2.99, CI: 1.67 to 5.34, P < 0.0001), being widowed (OR: 6.43, CI: 3.11 to 13.31, P < 0.0001), being married (OR: 1.80, CI: 1.11 to 2.90, P < 0.05), having more than 3 lifetime sexual partners (OR: 1.53, CI: 1.04 to 2.25, P < 0.05), and having a higher number of symptoms (OR: 1.81, CI: 1.50 to 2.19, P < 0.0001) were associated with increased risk. Among females, cohabitating (OR: 2.73, CI: 1.94 to 3.86, P < 0.0001), being divorced/separated (OR: 2.13, CI: 1.50 to 3.03, P < 0.0001), being widowed (OR: 4.57, CI: 3.21 to 6.49, P < 0.0001), having more than 3 lifetime sexual partners (OR: 1.97, CI: 1.55 to 2.51, P < 0.0001), having any symptoms (OR: 2.07, CI: 1.44 to 2.97, P < 0.0001), and having a higher number of symptoms (OR: 1.76, CI: 1.52 to 2.04, P < 0.0001) were associated with increased risk.





The differential effects of the 3 explanatory domains (sociodemographic characteristics, behavioral risk characteristics, and symptoms) on seroprevalence are summarized in Table 4. For females, changes in sociodemographic characteristics and symptoms contributed to decreased seroprevalence (P = 0.002 and P < 0.0001, respectively). For males, symptom-related decreases in seropositivity (P < 0.0001) were partially offset by small increases in behavioral risk characteristics (P = 0.0590).



Back to Top | Article Outline


In this analysis of over 4300 consecutive clients presenting for first-time VCT over 50 months and including intervals before and after CTC expansion in Moshi, Tanzania, we describe a significant decrease in HIV seroprevalence for both women and men. Reductions in symptoms for both men and women and a more favorable distribution of sociodemographic risk factors among women contributed to this decrease.

Although causality between the scale-up of CTCs and changes in VCT client characteristics cannot be assessed from this study, it is evident that the proportion of clients found to be HIV seropositive decreased significantly in this region with the addition of more CTCs. Several reasons could underlie the observed results. Persons who suspect that they are HIV infected (either because they have a family member being cared for at a CTC or because they are symptomatic) may prefer to be tested at sites where treatment is known to be available. It is also possible that because treatment availability may reduce community stigma, community members may be more inclined to access testing even when well.23 It has been shown in South Africa that VCT coupled with ART availability changes community perceptions favorably.24 Similarly, educational campaigns concurrent with CTC scale-up may have resulted in a broader population accessing VCT services. This is supported by the changing sociodemographic characteristics of women. Another possibility is that seroprevalence is decreasing in the general population. Nationally, Tanzania has seen a decreasing trend over the past few years, with seropositivity rates dropping from 9% in 20031 to 7% in 200425 to 6.5% in 2005 to 6% in 2007-2008.1 As shown in Figure 1, population-based survey estimates of HIV seroprevalence in the Kilimanjaro Region were 7.3% in 2003-2004 and 1.9% in 2007-2008. Even though prevalence decreases of such magnitude seem unlikely, it is quite possible that seroprevalence is actually decreasing in this region;22 additional data are needed to corroborate these observations. Our data and those from other similar cohorts suggest that all of the explanations may have contributed to these changes.

We also note persisting differences among VCT clients by gender. Disparity of power within relationships and in society are prominent within sub-Saharan culture and are often cited as contributing factors to higher HIV risk among women in this region.26 It is possible that greater access to care and treatment may attenuate gender disparities by providing more personal control of health to men and women equally. Despite this hypothesis, our results show a continuing gender differential in HIV seroprevalence, suggesting that, for females, testing may be viewed less as a means of prevention and more as an entry point to care, plausibly leading to gender-specific responses to the increased access to CTCs.

We note several limitations to this study. All results were self-reported by clients; underestimation of current sexual behaviors may have occurred due to social desirability bias.27 Similarly, clients are a self-selected group who attend counseling and testing services voluntarily and cannot be considered a sample of the general population. Also, our pre-CTC period estimates are limited to a relatively short period, and thus we cannot fully distinguish what occurred during this time frame; there may have been changes in anticipation of the provision of free ART that we could not detect. Finally, other correlates of seropositivity, not included in our analyses, may have contributed to the observed changes in seropositivity over time. It is not possible to ascertain the contribution of other characteristics from our data.

Collectively, these findings hold several important policy-level and programmatic implications for VCT programs. First, as CTC sites increase in number and become more decentralized, more symptomatic clients may access HIV testing services at sites where antiretrovirals are offered. Freestanding VCT sites may thus become a less cost-effective means for HIV case finding. Given the rationale and mandate for universal testing in this region and anticipated worsening fiscal constraints, alternative methods for identifying HIV-infected persons, such as mobile VCT, provider-initiated testing and counseling, home-based testing, and CTC-based testing campaigns and their relative cost-effectiveness should be explored. Second, because the declines in seroprevalence among first-time testers at this VCT site seem to mirror those of population-based serosurveys for the Kilimanjaro region, seroprevalence data from stable VCT sites, such as this one, may be useful for ascertaining trends reflective of the larger population, despite the inherent bias of this self-selected, self-referred sample. Third, we observed no appreciable changes in behavior risk characteristics over time; there remains a pressing need to sharpen and focus HIV risk behavior reduction messaging.

In conclusion, we identified a decrease in seroprevalence in both genders. Declines in symptoms contributed to this overall finding in both men and women, as did a more favorable distribution of sociodemographic risks among women. Although this analysis cannot establish a causal relationship between these events and the scale-up of CTCs within this region, it is clear that the client population presenting for VCT has changed. It will be important for VCT programs to incorporate this new knowledge into their prevention programs if they are to appropriately target the changing needs of their client population.

Back to Top | Article Outline


We would like to thank the study participants and the staff, volunteers, and counselors of KIWAKKUKI.

Back to Top | Article Outline


1. UNAIDS, WHO. AIDS Epidemic Update: December 2007. Geneva, Switzerland: UNAIDS; 2007.
2. UNAIDS. Voluntary Counselling and Testing (VCT): UNAIDS Technical Update. Geneva, Switzerland: UNAIDS; 2000.
3. Salomon JA, Hogan DR, Stover J, et al. Integrating HIV prevention and treatment: from slogans to impact. PLos Med. 2005;2(1; e16):0050-0056.
4. Luo C. Achievable standard of care in low-resource settings. Ann N Y Acad Sci. 2000;918:179-187.
5. The Voluntary HIV-1 Counseling and Testing Efficacy Study Group. Efficacy of voluntary HIV-1 counselling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: a randomised trial. Lancet. 2000;356:103-112.
6. Van de Perre P. HIV voluntary counselling and testing in community health services. Lancet. 2000;356:86-87.
7. Godfrey-Faussett P, Maher D, Mukadi YD, et al. How human immunodeficiency virus voluntary testing can contribute to tuberculosis control. Bull World Health Organ. 2002;80:939-945.
8. Landman K, Thielman N, Mgonja A, et al. Risk factors for low HIV treatment literacy among clients presenting for voluntary counseling and testing in Moshi, Tanzania. Presented at: 3rd IAS Conference on HIV Pathogenesis and Treatment; July 25, 2005; Vol Rio de Janeiro, Brazil.
9. United Republic of Tanzania. HIV/AIDS Care and Treatment Plan 2003-2008. Dar es Salaam, Tanzania: Ministry of Health and Social Welfare; 2003.
10. Lamptey P, Wilson D. Scaling up AIDS treatment: what is the potential impact and what are the risks? PLos Med. 2005;2(2):102-104.
11. Stover J, Bertozzi S, Gutierrez JP, et al. The global impact of scaling up HIV/AIDS prevention programs in low- and middle-income countries. Science. 2006;311:1474-1476.
12. Baggaley RF, Garnett GP, Ferguson NM. Modeling the impact of antiretroviral use in resource-poor settings. PLos Med. 2006;3(4):493-504.
13. Nuwaha F, Kabatesi D, Muganwa M, et al. Factors influencing acceptability of voluntary counselling and testing for HIV in Bushenyi District of Uganda. East Afr Med J. 2002;79:626-632.
14. 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. 2006;20:85-92.
15. Mukherjee J, Farmer J, Leandre F, et al. Access to Antiretroviral Treatment and Care: the Experience of the HIV Equity Initiative, Cange, Haiti. Geneva, Switzerland: World Health Organization; 2003.
16. Medecins sans Frontieres South Africa, Department of Public Health at the University of Cape Town, Provincial Administration of the Western Cape SA. Antiretroviral Therapy in Primary Health Care: Experience of the Khayelitsha Programme in South Africa. Geneva, Switzerland: World Health Organization; 2003.
17. Warwick Z. The influence of antiretroviral therapy on the uptake of HIV testing in Tutume, Botswana. Int J STD AIDS. 2006;17:479-481.
18. The United Republic of Tanzania Ministry of Health National AIDS Control Programme. National Guidelines for Voluntary Counselling and Testing, 2005. Dar es Salaam, Tanzania: Ministry of Health and Social Welfare; 2005.
19. Chu HY, Crump JA, Ostermann J, et al. Sociodemographic and clinical characteristics of clients presenting for HIV voluntary counseling and testing in Moshi, Tanzania. Int J STD AIDS. 2005;16:691-696.
20. Ministry of Health. HIV/AIDS/STI Surveillance Report: January-December 2005, Report Number 20. Dar es Salaam, Tanzania: Ministry of Health and Social Welfare; 2007.
21. Ministry of Health. Tanzania HIV/AIDS Indicator Survey 2003-2004. Dar es Salaam, Tanzania: Ministry of Health and Social Welfare; 2005.
22. Tanzania Commission for HIV/AIDS. Tanzania HIV/AIDS and Malaria Indicator Survey 2007-08. Dar es Salaam, Tanzania: Ministry of Health and Social Welfare; 2008.
23. Glick P. Scaling Up HIV Voluntary Counseling and Testing in Africa: What Can Evaluation Studies Tell Us About Potential Prevention Impacts? Strategies and Analysis for Growth and Access; Ithaca, NY; 2005.
24. Levy NC, Miksad RA, Fein OT. From treatment to prevention: the interplay between HIV/AIDS treatment availability and HIV/AIDS prevention programming in Khayelitsha, South Africa. J Urban Health. 2005;82:498-509.
25. Mishra V, Vaessan M, Boerma JT, et al. HIV testing in national population-based surveys: experience from the Demographic and Health Surveys. Bull World Health Organ. 2006;84:537-545.
26. Interagency Coalition on AIDS and Development. HIV/AIDS and Gender Issues. Ottawa, Ontario, Canada: Interagency Coalition on AIDS and Development; 2006.
27. Gregson S, Mushati P, White PJ, et al. Informal confidential voting interview methods and temporal changes in reported sexual risk behavior for HIV transmission in sub-Saharan Africa. Sex Transm Infect. 2004;80:36-42.

antiretroviral therapy; care and treatment; HIV seroprevalence; risk factors; sub-Saharan Africa

© 2009 Lippincott Williams & Wilkins, Inc.