Objective: Due to high rates of undiagnosed and untreated HIV infection in Africa, we compared HIV counseling and testing (VCT) uptake among household members of patients receiving antiretroviral therapy.
Methods: HIV-infected persons attending an AIDS clinic were randomized to a home-based or clinic-based antiretroviral therapy program including VCT for household members. Clinic arm participants were given free VCT vouchers and encouraged to invite their household members to the clinic for VCT. Home arm participants were visited, and their household members offered VCT using a 3-test rapid finger-stick testing algorithm. VCT uptake and HIV prevalence were compared.
Findings: Of 7184 household members, 3974 (55.3%) were female and 4798 (66.8%) were in the home arm. Home arm household members were more likely to receive VCT than those from the clinic arm (55.8% vs. 10.9%, odds ratio: 10.41, 95% confidence interval: 7.89 to 13.73; P < 0.001), although the proportion of HIV-infected household members was higher in the clinic arm (17.3% vs. 7.1%, odds ratio: 2.76, 95% confidence interval: 1.97 to 3.86, P < 0.001). HIV prevalence among all household members tested in the home arm was 56% compared with 27% in the clinic arm. Of 148 spouses of HIV-infected patients, 69 (46.6%) were uninfected. Persons aged 15-24 were less likely to test than other age groups, and in the home arm, women were more likely to test than men.
Conclusions: Home-based VCT for household members of HIV-infected persons was feasible, associated with lower prevalence, higher uptake, and increased identification of HIV-infected persons than clinic-based provision.
From the *Centers for Disease Control and Prevention-Uganda, Global AIDS Program, National Center for HIV, STD and TB Prevention, Centers for Disease Control and Prevention, Entebbe, Uganda; †Medical Research Council Program on AIDS, Entebbe, Uganda; and ‡The AIDS Support Organization, Kampala, Uganda.
Received for publication January 11, 2010; accepted May 25, 2010.
The study was supported by Cooperative Agreement Number IU01- PS000065-01 from the US Centers for Disease Control and Prevention and the United Kingdom Medical Research Council.
The contents of this publication are solely the responsibility of the authors and do not reflect the official views of the Centers for Disease Control and Prevention or the United Kingdom Medical Research Council.
Correspondence to: Dr. Eric Lugada, MD, PhD, Centers for Disease Control and Prevention-Uganda, Uganda Virus Research Institute, PO Box 49, Entebbe, Uganda (e-mail: firstname.lastname@example.org).
Although HIV counseling and testing (VCT) is the entry point for HIV prevention, care, and treatment services,1 throughout Africa knowledge of HIV status among those currently HIV-infected remains low.2,3 For example, in Kenya's most recent national serosurvey, 84% of HIV-infected persons were unaware of their status.4 Despite this low coverage, unmet desire for testing is high4suggesting that access barriers exist.
In sub-Saharan Africa, 45%-50% of cohabitating partners and spouses of HIV-infected adults are HIV uninfected, although the vast majority of these couples are unaware of their HIV discordance.5,6 Increasing VCT for sexual partners of those infected facilitates access to prevention, care, and treatment services.3,7 Knowledge of HIV status may be associated with reduced risk behavior8,9, and providing VCT together with prevention services including condoms to HIV-discordant couples decreases their HIV transmission risk by as much as 80%.10 Additionally, up to 19% of children younger than 5 found in households of HIV-infected adults in Uganda had undiagnosed HIV infection.11,12 Providing VCT to family members of persons taking antiretroviral therapy (ART) has the potential additional advantage of reducing risk of primary infection with ART-resistant HIV through adherence support and treatment access for eligible household members.
Despite its potential benefits,11 implementation of family VCT is not yet a standard component of HIV/AIDS care and treatment services. Evidence on the best programatic approaches to implement family VCT is still lacking. Facility-based prevention-of-mother-to-child transmission programs have attempted to promote partner testing, but male partner uptake rates were below 5% in the Democratic Republic of Congo,13 Rwanda,14 and Zimbabwe.15 In contrast, home-based VCT has been successfully implemented in Uganda for more than 500,000 people in several settings16,17 among sexual partners and other family members of HIV-infected adults11,16,18 but has never been compared directly with clinic-based VCT.
Home-based health strategies, including provision of ART,19 malaria treatment,20 and neonatal care,21 are effective in improving coverage and health. To understand the effectiveness of home-based VCT as a service delivery strategy in the context of home-based ART, we compared VCT uptake and testing results for household members of HIV-infected individuals participating in a randomized trial comparing home versus clinic-based ART in Jinja, Uganda.22
This study was conducted in the districts of Jinja, Kamuli, Iganga, Mayuge, and Mukono in southeastern Uganda, with an estimated HIV prevalence of 5.6% among 15- year to 49-year olds compared with 6.4% national prevalence.3,23 Implementation was in collaboration with The AIDS Support Organization (TASO) clinic, a local nongovernmental organization that has provided counseling, support services, and medical care to more than 12,600 HIV-infected clients in Jinja since 1990, including 4000 active clients in 2004. Most clients registered at TASO lived within a 75-km radius from the clinic. Most were poor, had low levels of education, and limited access to medical services other than at TASO. In September 2004, TASO began providing a comprehensive package of free ART services to registered clients, including family VCT. In addition, VCT services in TASO Jinja catchment area were provided in several hospitals and at a large counseling and testing center in Jinja.
This was a substudy of a cluster-randomized trial comparing home and clinic-based antiretroviral treatment in Uganda. The main study used a cluster randomization approach with 44 clusters, stratified by distance to the TASO Jinja clinic and estimated number of HIV infected clients, randomized in equal numbers to either the home or clinic arm, as detailed elsewhere.24 Consenting patients meeting the inclusion criteria were randomized by cluster to receive home-based or clinic-based ART delivery. VCT was offered to all household members in support of national policy to increase HIV testing among all people at risk of infection, to ensure equity of access to services, and to maximize efficiency in HIV testing it was not part of the main study procedures. Patients were encouraged to disclose their HIV status to household members and were required to designate a “drug companion” to support adherence. HIV-infected family members were encouraged to attend the TASO clinic for free HIV care services, ART clinical eligibility screening, and, if indicated, free ART.
Household members were defined as persons who shared a hearth and slept in the same house or cluster of houses as the index patient for at least 5 days a week for 3 months before recruitment into the main study. All household members, regardless of age or relationship, were eligible for VCT. Both home and clinic arm index patients completed a census form listing all their household members. Participants in the main study's clinic arm were given free VCT vouchers for each household member and encouraged to invite all their household members to the clinic for VCT by a trained counselor.
Household members in the home arm were visited by appointment between 8.30 am and 5.00 pm by trained lay field officers riding motorcycles. Household members received basic HIV and antiretroviral education, and consenting individuals were offered VCT using a 3 rapid test sequential algorithm in a private place at or near their home, in accordance with Uganda national counseling and testing policy. Routine TASO VCT cards were used to record household member HIV testing data and results. VCT cards were completed for all household members in the home arm present during home visits and all household members in the clinic arm who presented to the clinic with the VCT voucher. In both arms, household members did not necessarily know the HIV status of index clients.
The tests used were Abbot Determine HIV-1/HIV-2 (Abbott laboratories, Japan) as a screening test, Uni-Gold HIV-1/HIV-2 (Trinity Biotech plc, Bray, Ireland) as a confirmatory test, and Stat-Pak (Chembio Diagnostic systems, Inc. USA) as a “tie-breaker”. A positive specimen was defined as one that was positive on at least 2 assays. HIV testing was performed by trained lay field officers in the home arm and by laboratory technicians in the clinic arm. Household members were encouraged to disclose their HIV status to other household members. Participants in either arm experiencing disclosure challenges were offered counselor-assisted disclosure.
From February 2005 through February 2007, VCT was offered to all household members free of charge. For those aged 10-17 years, counseling included the child and a parent or guardian. For those aged 0-9 years, only the parent or guardian was counseled and asked to consent for the child's testing. For children younger than 24 months, a finger prick sample was collected on filter paper and sent to the laboratory for DNA polymerase chain reaction testing.
The statistical analysis accounted for the fact that study subjects were not independent; first, since randomization was by cluster; and second, because subjects in a given cluster were grouped into households, and there were 2 levels of clustering; subjects in households and households in communities. A robust data analysis approach that uses parallels between cluster randomized trials and multistage sample surveys was followed since it also accounts for clustering of subjects within households.25 Analyses were conducted using Stata Release 9 (Statacorp, College Station, TX). Variance estimates are based on calculations at the primary sampling-unit level (cluster). These variance estimates do not require information about the household level; the estimates of variance produced were approximately unbiased or if biased, they are biased toward more conservative estimates.26 Categorical variables and the proportions testing and receiving their results in each arm were compared by using the second-order correction to the Pearson χ2 statistic for 2-way tables from complex multistage surveys as per Rao and Scott.27
In addition, the odds of testing and receiving results were compared between arms, adjusting for age and gender. Because the age distributions differed slightly between arms, a logistic regression model with robust confidence intervals (CIs) was fitting to account for the clustering effects of the design. Further exploratory logistic regression models were fitted separately to each arm to investigate factors associated with testing and receiving results. Potential explanatory variables considered were age of client (grouped), gender of client, CD4 count of index client in the household (grouped as 0-49, 50-99, 100-149, 150-199, or 200+ cells/μL, respectively), client education (none, primary, secondary or higher/technical), client gender, and number of household members.
The proportion of total HIV-infected household members identified in each arm was estimated assuming that the true HIV prevalence was equal in the 2 arms, by applying the age and gender-specific HIV prevalence found among household members in the home arm.
The study was approved by the Science and Ethics Committees of the Uganda Virus Research Institute, the Uganda National Council of Science and Technology, and the Institutional Review Boards of the US Centers for Disease Control and Prevention and Liverpool School of Tropical Medicine.
By the end of December 2006, 1453 index clients had enrolled: 859 in the home arm and 594 in the clinic arm (Wilcoxon rank sum P = 0.035). A larger proportion of households in the home arm were enrolled [791 (92.1) % of all possible households] than in the clinic arm [455 (76.6%) of households]; and 4 of the 5 biggest clusters were in the home arm. Overall, there were 4798 (66.8%) household members in the home arm and 2386 (33.2%) in the clinic arm (Table 1). Of 7184 household members, 3974 (55.3%) were female. Median household size was 6, including the index client. Slightly over half (52%) of all household members were younger than 15 years of age. There were no significant differences in demographic characteristics between the home and clinic arms. VCT information was available for 1246 households; 791 (63.5%) were in the home arm.
Overall, 2938 (40.9%) household members completed VCT and received their results (Table 2). In the home arm, 2678 (55.8%) of 4798 received their results, compared with 260 (10.9%) of 2386 in the clinic arm. The design effect for VCT uptake was much higher in the home arm (9.0) than in the clinic arm (2.6). Acceptance of HIV testing was high (89%) among household members present at the time of the home visit. Analysis of repeat home visits revealed that the majority (over 93%) were likely to be tested at the first visit. VCT uptake was higher in the home arm for all age, gender, and index client characteristic categories. Adjusting for both age and gender, persons in the home arm were much more likely to undergo HIV testing and receive their results than persons in the clinic arm [adjusted odds ratio (aOR) 10.41; 95% CI: 7.89 to 13.73; P < 0.001] (Table 3).
Analysis of home arm household member data revealed that the index client having a CD4 count above 200 cells per microliter was associated with reduced odds of HIV testing uptake (aOR: 0.69; 95% CI: 0.52 to 0.91; P = 0.012) (Table 4). Higher education level of the index client was not associated with reduced testing uptake (aOR: 0.85; 95% CI: 0.71 to 1.02; P = 0.084). Persons aged 15-24 were least likely to receive an HIV test, with women more likely than men to get tested (aOR: 1.20; 95% CI: 1.07 to 1.35; P = 0.004).
In the clinic arm, household members of index clients with either the lowest (<50 cells/μL) or highest CD4 counts (>200 cells/μL) seemed less likely to receive an HIV test (Table 5). Persons aged 15-24 were least likely to get tested, with household members who lived in a household with a higher number of persons less likely to get tested (aOR: 0.90; 95% CI: 0.85 to 0.95; P < 0.001). Trends for gender of the household member and education level of the index client were similar to those of the home arm but were not statistically significant. Gender of the index client was not independently associated with household member testing in either arm.
The design effect for HIV infection was similar in the 2 arms, with an overall design effect of 1.9, indicating that the effect of clustering on HIV status was low. HIV prevalence amongst those tested was higher in the clinic arm compared with the home arm (17.3% vs. 7.1%; odds ratio: 2.76; 95% CI: 1.97 to 3.86; P < 0.001). Prevalence was highest among young children and persons aged 25-44 in both arms (Table 2). Household members of index clients with higher levels of education did seem to have an increased HIV prevalence.
Of 148 spouses of index clients tested, 69 were HIV negative, giving an HIV serodiscordance rate of 46.6%. The serodiscordance rate was similar in the home arm (47.9%) and the clinic arm (41.4%) (Rao-Scott F = 0.44; P = 0.51). The serodiscordance rates were also very similar for male (47.4%) and female index clients (45.1%). The serodiscordance rate was highest in couples in which the nonindex partner had lower primary education or less (62.5% in such couples compared with 30.2% in couples in which the nonindex client had higher primary education (P5-P7) and 38.7% in couples in which the nonindex client had secondary education or more; Rao-Scott F = 6.24; P = 0.005). There was however no association between the education level of the index partner and serodiscordance (Rao-Scott F = 1.03; P = 0.38). Of 69 serodiscordant married couples, the husband was the HIV-infected member in 46 (66.7%).
An approximate calculation was carried out to compare the proportions of all HIV-infected household members identified in the 2 arms. Because in the clinic arm it is possible that individuals who suspected that they may be HIV infected were more likely to go for VCT, whereas the individuals who were tested in the home arm were more representative of the general study population, we assumed that the age and gender-specific HIV prevalence in the home arm applied to both arms. Assuming the age and gender-specific home are HIV prevalence applied to both arms, 346 HIV-infected individuals would be expected in the home arm and 165 in the clinic arm. Thus, an estimated 54.6% (189 of 346) of all HIV-infected household members in the home arm were identified compared with 27.3% (45 of 165) in the clinic arm.
HIV testing offered at home to household members of HIV-infected adults initiating ART was associated with a 10-fold increased likelihood of testing compared with HIV testing offered in the clinic. Acceptance of home-based HIV testing was very high (89%) among household members present at the time of the home visit. These striking results are similar to studies previously conducted in Southwestern and Eastern Uganda11,16-18,28 that indicated a large proportion of persons accept VCT delivery at home, with the highest rate, 98%, among household members of HIV-infected individuals. In one of these studies, when the option of home or community receipt of results was offered,16 few participants presented at the community-based counseling office.
These findings are also consistent with studies elsewhere in Africa. In a population-based study conducted in an urban setting in Zambia,29 participants were randomly assigned to receive VCT either at a clinic or a setting of their choice, which included home counseling; only 12% opted for the clinic. The highest risk population for HIV infection, that of biological family members of HIV-infected individuals, is more likely to access HIV testing and subsequent care and treatment services through home testing.11,18 In this setting, HIV stigma was low, and HIV status disclosure to household members common, and previous studies have not documented increases in gender-based violence with provision of home-based testing.30 In settings where stigma is higher, specific efforts to reduce stigma, protect against gender-based violence, and support disclosure may be necessary for successful home-based programs.
Low VCT uptake by family members at the clinic may be due to unaffordable transportation costs, difficulty of the index client disclosing HIV status to household members, index clients not referring household members for VCT, time constraints among household members,17,24,31,32 or limited use of free VCT vouchers.33 Transportation cost and time constraints may partially explain the finding that household members from homes with higher numbers of persons were less likely to come to the clinic for HIV testing. Strategies to decrease barriers to clinic-based VCT, such as facilitating transportation costs, actively supporting disclosure, or providing incentives to index clients to successfully refer household members, should be explored.
A high proportion of HIV infections in sub-Saharan Africa occur within long-standing couples,34,35 and their children.11 Discordant couples accounted for almost half (47%) of all couples tested in our study, similar findings have been reported from Cote d'Ivoire, Uganda, and Kenya.4,7,11,12 In this study, 29% of HIV infected yet undiagnosed family members were children aged younger than 14 years, emphasizing the need for proactive public health interventions that offer routine access to household members of persons infected with HIV. Partner testing improves provision of care to those infected and helps protect those who are not from potential infections.36 VCT may be associated with reduced risk behavior1,34,37,38 and lower rates of HIV transmission among HIV-discordant couples.34
HIV prevalence among those tested in the clinic arm was higher compared with the home arm. As the cluster randomization should have ensured similar HIV prevalence in the 2 arms, this suggests that clinic arm household members who test may have done so because of illness or of suspecting that they are HIV infected.
In contrast, a higher percentage of the estimated total HIV-infected household members were identified in the home arm compared with the clinic arm. Increased knowledge of HIV infection allows people to benefit from ART,39-41 tuberculosis screening, treatment, and preventive therapy, cotrimoxazole prophylaxis,42,43 and potential prevention of other opportunistic infections associated with HIV infection. HIV-infected children of index clients were more than twice as likely to be identified in the home arm as in the clinic arm. HIV-infected children have a high mortality and are especially difficult to access. Home-based testing of children of index clients may be an especially efficient strategy for identifying this population and linking them to care and treatment.
In both study arms, household members with high CD4 counts were less likely to get tested, and in the clinic arm, household members with very low CD4 counts were also less likely to get tested. Household members with higher levels of education also seemed to be less likely to get tested. This could be either due to higher levels of stigma among more educated household members or a lack of empowerment to decline VCT among less educated household members. The reasons for these findings should be investigated further and may inform strategies to improve VCT access and services.
The striking difference in the design effect for VCT between study arms also suggests that in the home arm, although consent for HIV testing was approached on an individual basis, VCT was effectively a household level decision as opposed to an individual level decision. This may be due to the high proportion of children among home arm household members. Although 89% of those found at home received VCT, only half of all home arm family members were tested. Substantial numbers of school-aged children and men were not at home, resulting in a higher proportion of adult females testing at home. Repeat home visits revealed that the majority (more than 93%) were likely to be tested at the first visit reflecting a rather low utility in conducting repeat home visits for HIV testing. Repeat home visits for repeat HIV testing of HIV-uninfected household members may be useful; however, this was not investigated in this study. There were low levels of VCT uptake among persons 15-24 mainly young adult males who are least likely to be at home, innovative ways of reaching them should be investigated. Providing home-based VCT in the evening or weekend hours may improve access to this population.
The home arm had a substantially larger number of participants than the clinic arm. This was likely due to inexact estimates of the size of HIV-infected populations when initially determining the size of clusters to randomize. As this ratio was consistent throughout the enrollment period and the number of refusals was low (24), randomization seems to have been successful.
Although VCT has been shown to be cost effective in reducing HIV infection,44 it is currently not available to the majority of Africans who are poor, live in rural areas, and have little or no access to transportation. In Malawi, persons living in lower income households were less likely to have access to facility-based VCT more likely to accept home-based testing.45 Evidence from the current study and others46 indicates that VCT uptake would be improved if it were standard practice for home-based VCT to be provided to household members of persons infected with HIV. It is an efficient strategy to both identify undiagnosed HIV infections among adults and children and also identify HIV-discordant couples, thus facilitating provision of effective interventions to reduce HIV transmission. Home-based visits may also provide opportunities to identify non-HIV-related health issues and allow linkage with other health services.
Implementing home-based VCT presented logistical challenges; cost of hiring a huge pool of field officers, fuelling and maintaining motorcycles, and mobilization of family members at their homes were effects of stigma against HIV/AIDS still persist. However, they were not insurmountable and need to be carefully considered since home-based VCT has been used in many other situations in Uganda17,18 and elsewhere in Africa.6,45
Throughout the continent, home-based VCT models are being recommended,47 including by the new Health Minister in South Africa,48 and cost analyses of home-based approaches have been documented.49,50 Home-based HIV testing should be more widely available in Africa and be considered in national policies and as a standard practice by health facilities serving populations affected by HIV.
We thank all TASO clients and their household members who participated in this study. All TASO Jinja and Centers for Disease Control and Prevention Uganda staff dedicated to providing care and treatment to those in need in addition to generating a public health evidence base for better and improved services.
1. 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
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. Uganda Ministry of Health. Uganda HIV/AIDS Sero-Behavioral Survey 2004-2005
. March 2006.
5. Guthrie BL, de Bruyn G, Farquhar C. HIV-1-discordant couples in sub-Saharan Africa: explanations and implications for high rates of discordancy. Curr HIV Res
6. Kimaiyo S, Wools-Kaloustian K, Boit L, et al. Effectiveness and outcomes of door-to-door HIV testing in a rural district of Western Kenya. Presented at: XVII International AIDS Conference, August 5, 2008; Mexico City, Mexico. Abstract: TUPE0389.
7. N'Gbichi JM, De Cock KM, Batter V, et al. HIV status of female sex partners of men reactive to HIV-1, HIV-2 or both viruses in Abidjan, Cote d'Ivoire. AIDS
8. Turner AN, Miller WC, Padian NS, et al. Unprotected sex following HIV testing among women in Uganda and Zimbabwe: short- and long-term comparisons with pre-test behaviour. Int J Epidemiol
9. 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
10. Hugonnet S, Mosha F, Todd J, et al. Incidence of HIV infection in stable sexual partnerships: a retrospective cohort study of 1802 couples in Mwanza Region, Tanzania. J Acquir Immune Defic Syndr
11. Were WA, Mermin JH, Wamai N, et al. Undiagnosed HIV infection and couple HIV discordance among household members of HIV-infected people receiving antiretroviral therapy in Uganda. J Acquir Immune Defic Syndr
12. Kabatesi D, Ransom R, Lule JR, et al. HIV prevalence among household members of persons living with HIV in rural Uganda. Presented at: XIV International AIDS Conference; July 2002; Barcelona, Spain. Abstract: TuPeD4910.
13. Heyward WL, Batter VL, Malulu M, et al. Impact of HIV counseling and testing among child-bearing women in Kinshasa, Zaire. AIDS
14. Ladner J, Leroy V, Msellati P, et al. A cohort study of factors associated with failure to return for HIV post-test counselling in pregnant women: Kigali, Rwanda, 1992-1993. AIDS
15. Dube SMR, McFerland W. HIV voluntary counselling and testing of couples in Harare: problems and prospects. Presented at: XIIWorld AIDS Conference; July 9-15, 2000; Durban, South Africa. Abstract TuPeD3773.
16. Matovu JK, Kigozi G, Nalugoda F, et al. The Rakai Project counselling programme experience. Trop Med Int Health
17. Wolff B, Nyanzi B, Katongole G, et al. Evaluation of a home-based voluntary counselling and testing intervention in rural Uganda. Health Policy Plan
18. Were W, Mermin J, Bunnell R, et al. Home-based model for HIV voluntary counselling and testing. Lancet
19. Mermin J, Were W, Ekwaru JP, et al. Mortality in HIV-infected Ugandan adults receiving antiretroviral treatment and survival of their HIV-uninfected children: a prospective cohort study. Lancet
20. Kidane G, Morrow RH. Teaching mothers to provide home treatment of malaria in Tigray, Ethiopia: a randomised trial. Lancet
21. Bang AT, Bang RA, Stoll BJ, et al. Is home-based diagnosis and treatment of neonatal sepsis feasible and effective? Seven years of intervention in the Gadchiroli field trial (1996-2003). J Perinatol
. 2005;25(Suppl 1):S62-S71.
22. Jaffar S, Amuron B, Birungi J, et al. Integrating research into routine service delivery in an antiretroviral treatment programme: lessons learnt from a cluster randomized trial comparing strategies of HIV care in Jinja, Uganda. Trop Med Int Health
23. Uganda Bureau of Statistics: 2002 Uganda Population and Housing Census
24. Amuron B, Coutinho A, Grosskurth H, et al. A cluster-randomised trial to compare home-based with health facility-based antiretroviral treatment in Uganda: study design and baseline findings. Open AIDS J
25. Donner A, Klar N. Design and Analysis of Cluster Randomization Trials in Health Research
. London, United Kingdom: Arnold; 2000.
26. Statcorp Stata Statistical Software: Release 5.0: User's Guide (U36: Overview of Survey Commands)
. College Station, TX: Stata Corporation; 1997.
27. Rao J, Scott A. On chi-squared tests for multiway contigency tables with cell proportions estimated from survey data. Ann Stat
28. Mermin J, Lule J, Ekwaru JP, et al. Cotrimoxazole prophylaxis by HIV-infected persons in Uganda reduces morbidity and mortality among HIV-uninfected family members. AIDS
29. Fylkesnes K, Siziya S. A randomized trial on acceptability of voluntary HIV counselling and testing. Trop Med Int Health
30. Apondi R, Bunnell R, Awor A, et al. Home-based antiretroviral care is associated with positive social outcomes in a prospective cohort in Uganda. J Acquir Immune Defic Syndr
31. Laver SM. Voluntary testing and counselling for HIV. “Are adults in rural communities ready to test?” A descriptive survey. Cent Afr J Med
32. 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
33. Corbett EL, Dauya E, Matambo R, et al. Uptake of workplace HIV counselling and testing: a cluster-randomised trial in Zimbabwe. PLoS Med
34. Allen S, Tice J, Van de Perre P, et al. Effect of serotesting with counselling on condom use and seroconversion among HIV discordant couples in Africa. BMJ
35. Painter TM. Voluntary counseling and testing for couples: a high-leverage intervention for HIV/AIDS prevention in sub-Saharan Africa. Soc Sci Med
36. De Cock KM, Bunnell R, Mermin J. Unfinished business-expanding HIV testing in developing countries. N Engl J Med
37. Marks G, Crepaz N. HIV-positive men's sexual practices in the context of self-disclosure of HIV status. J Acquir Immune Defic Syndr
38. Roth DL, Stewart KE, Clay OJ, et al. Sexual practices of HIV discordant and concordant couples in Rwanda: effects of a testing and counselling programme for men. Int J STD AIDS
39. Ho DD. Time to hit HIV, early and hard. N Engl J Med
40. Hogg RS, Yip B, Chan KJ, et al. Rates of disease progression by baseline CD4 cell count and viral load after initiating triple-drug therapy. JAMA
41. Wong KH, Lee SS, Low KH, et al. Temporal trend and factors associated with late HIV diagnosis in Hong Kong, a low HIV prevalence locality. AIDS Patient Care STDS
42. Anglaret X, Chene G, Attia A, et al. Early chemoprophylaxis with trimethoprim-sulphamethoxazole for HIV-1-infected adults in Abidjan, Cote d'Ivoire: a randomised trial. Cotrimo-CI Study Group. Lancet
43. Mermin J, Lule J, Ekwaru JP, et al. Effect of co-trimoxazole prophylaxis on morbidity, mortality, CD4-cell count, and viral load in HIV infection in rural Uganda. Lancet
44. 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. Lancet
45. Helleringer S, Kohler HP, Frimpong JA, et al. Increasing uptake of HIV testing and counseling among the poorest in sub-Saharan countries through home-based service provision. J Acquir Immune Defic Syndr
46. McKenna SL, Muyinda GK, Roth D, et al. Rapid HIV testing and counseling for voluntary testing centers in Africa. AIDS
. 1997;11(Suppl 1):S103-S110.
47. Matovu JK, Makumbi FE. Expanding access to voluntary HIV counselling and testing in sub-Saharan Africa: alternative approaches for improving uptake, 2001-2007. Trop Med Int Health
48. Leon NH, Colvin CJ, Lewin S, et al. Provider-initiated testing and counselling for HIV - from debate to implementation. S Afr Med J
49. Menzies N, Abang B, Wanyenze R, et al. The costs and effectiveness of four HIV counseling and testing strategies in Uganda. AIDS
50. Terris-Prestholt F, Kumaranayake L, Ginwalla R, et al. Integrating tuberculosis and HIV services for people living with HIV: costs of the Zambian ProTEST Initiative. Cost Eff Resour Alloc
Keywords:© 2010 Lippincott Williams & Wilkins, Inc.
Africa; antiretroviral therapy; clinic-based; HIV; home-based; randomized trial; voluntary counseling and testing; Uganda